Specificity and Properties of Anti-HLA Antibodies Associated with Renal Allograft Rejection

Transcription

1 Specificity and Properties of Anti-HLA Antibodies Associated with Renal Allograft Rejection Hooi Sian Eng Bachelor of Biomedical Science (Hon) Master of Medical Science A thesis submitted for the degree of PhD School of Medicine Faculty of Health Science University of Adelaide, South Australia February

2 i. Contents i. Contents... 2 ii. List of Tables... 6 iii. List of Figures... 8 iv. Summary... 9 v. Dedication vi. Acknowledgements vii. Declaration viii. Thesis Structure ix. Abbreviations x. Publications xi. Presentations xii. Awards Chapter 1 Literature Review Overview of Human Major Histocompatibility Complex (MHC) Gene Discovery Chronology of HLA Discovery Genetics of MHC Evolution of MHC Genes MHC Gene Polymorphism Human Leukocyte Antigen (HLA) System HLA Gene Inheritance and Frequency HLA Class I, II and III Regions Nomenclature Structure of HLA Molecules HLA Peptide-Binding Grooves HLA Epitopes HLA Expression Roles of HLA Molecules Antigen Presentation Endogenous Presentation Pathway Exogenous Presentation Pathway T-cell Receptor Repertoire Selection Regulation of NK Cell Activity Significance of HLA in Renal Transplantation Mechanism of Renal Allograft Rejection Isolated MHC Class I Disparity Isolated MHC class II Disparity MHC Class I and II Disparity Pre-transplantation HLA Antibody Detection and Crossmatching Negative T-cells / Positive B-cells Crossmatches (TN/BP) Crossmatch: Current Serum Negative / Historic Serum Positive (CN/HP) Crossmatch: DTT-Reducible Positivity Post-transplantation Monitoring Donor-Specific Antibodies MHC Class I Chain Related (MIC) Genes: MICA and MICB HLAMATCHMAKER Amino Acid Triplet Program Antibody Detection Techniques Complement-dependent Lymphocytotoxicity Assay Flow Cytometry Solid Phase ELISA Solid Phase Luminex Method Comparison Chapter 2 Materials & Methods HLA Antigen Typing Dynabead T-cell Isolation Dynabead B-cell Isolation Serology HLA Antigen Typing

3 2.1.4 DNA Quantitation Gel Electropheresis Visualization of PCR Product Sequence Base Typing Prepare PCR Mixture PCR Product Visualization AMPure PCR Purification ABI Big Dye Terminator Ready Reaction Kit CleanSEQ Dye Terminator Removal Luminex Single Stranded Oligonucleotide (SSO) HLA Typing Prepare PCR Mixture Visualization of PCR Product Hybridisation Crossmatch Direct Crossmatch DTT Crossmatch Anti-HLA Antibodies Screening SeraClean Luminex Screen and PRA Luminex Single Antigen Bead Luminex Single Antigen Beads Anti-HLA antibodies IgG subclass Database National Organ Matching System (NOMS) Statistics Reagent Preparation Equipments Kits Chemicals Consumables Chapter 3 Clinical Relevance of A positive B-cell Crossmatch in Renal transplantation Introduction Objectives Materials and Methods Study Cohort HLA Typing Screening for Autoantibodies Luminex Antibody Specificity Clinical Data Statistics Results Characteristics of Patients with T B+ crossmatch Luminex Specificity of Anti B-cell Antibodies Early Graft Rejection Correlation of T B+ and Early Graft Rejection Correlation of Early Graft Rejection and Antibody Specificity in T B Graft Function at 6, 12 and 36 month Post-transplantation Graft Survival Significance of T B+ in Graft Survival Correlation of Antibody Specificity in T B+ with Graft Survival Historic Positive B-cell Crossmatch BXM as A Predictor of Graft Failure in 5 years Post-transplantation Discussion Chapter 4 Anti-HLA Antibodies and Risk Factors Associated with Transplant Glomerulopathy Introduction Objectives Materials and Methods

4 4.3.1 Study Cohorts HLA Typing Detection of Anti-HLA Antibodies Clinical Data ANZDATA Analysis Statistics Results Univariate Risk Factors Demographic Comparison Immunosuppression Early Graft Rejection Significance of Anti-HLA Antibodies in Transplant Glomerulopathy CDC Class I PRA defined Presensitization Luminex defined Presensitization Pre-transplant DSA as A Risk Factor of TG De novo DSA Predict TG Multivariate Analysis DSA as A Predictor of TG Time to Biopsy Association of TG and Graft Survival Pre-transplant DSA Associated with Early TG Development Significance of DSA in TG Progression DSA are Predictive for Graft Loss in TG Discussion Chapter 5 A Retrospective Study of B-cell Crossmatch with Luminex Technology in Well-Matched Highly Sensitized Patients from The Australian National Renal Exchange Programme (ANREP) Introduction Objectives Materials and Methods Study Cohort Crossmatch HLA Typing Luminex Antibody Analysis Clinical Data Statistics Results Characteristics of Study Cohort Graft Outcomes of Highly Sensitized Patients Graft Rejection Graft Function Graft Survival Predictive Value of B-cell Crossmatch in Highly Sensitized Patients Prevalence of Positive B-cell Crossmatch Characteristics of Patients with Positive B-cell Crossmatch Correlation of B-cell Crossmatch and Graft Rejection Graft function B-cell Crossmatch and Graft Function Correlation of B-cell Crossmatch and Graft Survival Predictive Value of Luminex defined Donor-specific Antibodies Presence of Donor-specific Antibodies Specificity of DSA defined by Luminex Correlation of DSA and Graft Rejection Correlation of DSA and Graft Function Correlation of Luminex defined DSA and Graft Survival Luminex defined DSA in B-cell Crossmatches Correlation of Luminex defined DSA and BXM with Graft Rejection Multivariate analysis of BXM and Luminex defined DSA in predicting rejection Discussion

5 6 Chapter 6 Discussion Significance of Preformed Anti-HLA Antibodies Detected by B-cell Crossmatch Antibodies Causing Positive B-cell Crossmatches and Correlation with Graft Outcomes Luminex Defined Donor-specific Antibodies as A Predictor of Graft Outcomes Solid Phase Antibody Analysis: Virtual Crossmatch Prediction of Flow Crossmatch Outcomes by Luminex Assays Improved Transplant Rate by Defining Acceptable Mismatches Cost-Effectiveness of Solid Phase Assays Virtual Crossmatch Possible Weaknesses of Solid Phase Assays Virtual Crossmatch Luminex Antibody Analysis and Luminex Lysate Crossmatch DSA in Transplant Glomerulopathy Significance of Post-transplant Monitoring Summary of Conclusions References Appendix

6 ii. List of Tables Chapter 2 Table 2.1.1: Preparation of PCR Mixture for Sequence Base Typing Table 2.1.2: PCR conditions for Sequence Base Typing Table 2.1.3: Conditions of sequencing amplification Table 2.1.5: Preparation of PCR mixtures for Luminex SSO Table 2.1.6: Condition of PCR amplification for Luminex SSO Table 2.1.7: Hybridisation steps for Luminex SSO Chapter 3 Table 3.4.1: Descriptive statistics of cohorts by the result of B-cell crossmatching Table 3.4.2: Immunosuppresion in T B and T B+ groups Table 3.4.3: Prevalence of cellular, vascular and glomerular rejection in All grafts, T B and T B+ groups Table 3.4.4: Graft rejection at 6 months according to the antibody specificity in T B Table 3.4.5: Risk of cellular rejection according to the antibody specificity in T B Table 3.4.6: Risk of vascular rejection according to the antibody specificity in T B Table 3.4.7: Risk of glomerular rejection according to the antibody specificity in T B Table 3.4.8: egfr at 6, 12 and 36 months by B-cell crossmatch and antibody specificity, Table 3.4.9: Graft survival rates and hazard ratios for graft loss by antibody specificity Table : Prevalence of cellular, vascular and glomerular rejection in T B+ subgroups Table : Odds ratios of graft rejection at 6 months in T B+ subgroups Table : egfr at 6, 12 and 36 months post-transplant in T B+ subgroups Table : Graft survival rates and hazard ratios for graft loss in T B+ subgroups Table : Graft outcomes at 5 years post-transplant Chapter 4 Table 4.4.1: Descriptive statistics of cohort and univariate analyses of risk factor of TG Table 4.4.2: Immunosuppressants at induction and prior to diagnosis of TG Table 4.4.3: Cellular, vascular and glomerular rejection as a risk factor of TG Table 4.4.4: Univariate effects of HLA presensitization, defined by CDC or Luminex techniques, in the development of TG Table 4.4.5: Univariate effects of anti-hla antibodies in the development of TG Table 4.4.6: In multiple regression analysis, DSA was the only independent predictor of TG Table 4.4.7: Graft survival rates in patients with TG, according to antibody specificity Table 4.4.8: Graft survival and risk of graft loss in TG patients, according to antibody specificity, in 20 years of follow up Chapter 5 Table 5.3.1: Patient recruitment Table 5.4.1: Descriptive statistics of study cohort, compared to all other grafts transplanted within the same period in participating centres Table 5.4.2: Immunosuppressants at induction Table 5.4.3: Risk of cellular, vascular and glomerular rejection in highly sensitized patients Table 5.4.4: Graft function, egfr, of highly sensitized patients Table 5.4.5: Total graft survival rates in highly sensitized patients, compared to all grafts Table 5.4.6: Demographics of T B and T B+ groups Table 5.4.7: Immunosuppressants at induction for T B and T B+ groups Table 5.4.8: Correlations of T B+ with cellular, vascular and glomerular rejection Table 5.4.9: Comparison of graft function (MDRD egfr) between T B and T B+ groups Table : Correlation of T B+ and total graft survival Table : Demographics of DSA and Non-DSA groups Table : Immunosuppressants at induction for DSA and Non-DSA groups Table : Correlations of DSA with cellular, vascular and glomerular rejection Table : Comparison of graft function (MDRD egfr) between Non-DSA and DSA groups Table : Correlation of DSA and total graft survival Table : BXM results and specificity of DSA in highly sensitized patients Table : Independent predictors of graft rejection

7 Chapter 6 Table 6.1.1: A summary of studies investigated clinical significance of positive BXM in renal transplantation Table 6.2.1: A review of clinical significance of low titer DSA detected by Luminex or Flow crossmatch in sera with negative CDC / AHG-CDC

8 iii. List of Figures Chapter 1 Figure 1.2.1: Human MHC gene map Figure 1.2.2: Timescale of MHC evolution Figure 1.3.1: The number of antigens and alleles from 1968 to September Figure 1.3.2: HLA gene map illustrates clustering of immune system Figure 1.5.1: Structure of HLA class I molecules Figure 1.5.2: Ribbon structure of class I peptide binding groove Figure 1.5.3: Structure of HLA class II molecule Figure 1.5.4: Peptide in HLA peptide-binding groove Figure 1.5.5: Amino acid pockets in the floor of binding groove determine the binding of peptide Figure 1.5.6: HLA molecule epitopes Figure 1.6.1: Roles of peptide loading complex in MHC class I assembly and peptide loading Figure 1.6.2: Endogenous peptide presentation pathway Figure 1.6.3: Exogenous antigen presentation pathway Figure 1.6.4: Positive and negative selection of T-cell receptor repertoire Figure 1.6.5: Regulation of NK activity by HLA class I molecules Figure 1.7.1: Effector mechanisms of rejection in graft carrying mismatched HLA class I antigen Figure 1.7.2: Effector mechanism of rejection in graft carrying mismatched HLA class II antigens Figure 1.7.3: Effector mechanisms of rejection in graft carrying mismatched HLA class I and class II antigens Chapter 3 Figure 3.4.1: T B+ as a group was not associated with inferior graft survival Figure 3.4.2: Graft survival according to antibody specificity in T B Figure 3.4.3: Significance of DSA class in graft survival Figure 3.4.4: 5 years graft survival in T B+ subgroups stratified by peak/current status Chapter 4 Figure 4.4.1: Cumulative graph for biopsy events in relation to time post-transplantation Figure 4.4.2: Graft survival in the controls and patients Figure 4.4.3: Graft survival before and after diagnosis of TG Figure 4.4.4: Cumulative TG diagnosis events in relation to time post-transplant, according to presence of DSA Figure 4.4.5: Graft survival in patients with TG according to presence of DSA Figure 4.4.6: Graft survival in TG subgroups, compared to the control group Chapter 5 Figure 5.4.1: Graft function of highly sensitized patients, compared to all grafts Figure 5.4.2: Total graft survival of highly sensitized patients, compared to all grafts Figure 5.4.3: Correlation of graft function and BXM Figure 5.4.4: Significance of positive BXM in relation to graft survival Figure 5.4.5: Specificity of HLA class I donor-specific antibodies in highly sensitized patients Figure 5.4.6: Specificity of HLA class II donor-specific antibodies in highly sensitized patients Figure 5.4.7: A trend of poor graft function in DSA group within 1 year post-transplant Figure 5.4.8: Graft survival in the presence of DSA

9 iv. Summary Identification of the complement C4d fragment in peritubular capillaries as a specific marker for antibody mediated rejection in renal transplantation revealed the critical role of antibodies in graft survival. In this thesis, I document the design and findings of studies performed to investigate the clinical impact of anti-hla antibodies present before and/or after transplantation. Over time, the detection techniques for anti-hla antibodies has evolved from the less sensitive complement-dependent lymphocytotoxicity (CDC) crossmatching (XM) to more sensitive solid phase assays such as Luminex. Studies have been conducted to compare the predictive value of different antibody detection techniques. The first result chapter presents antibody specificity in positive CDC B-cell crossmatch (BXM), analysed with highly specific Luminex assays. The study also investigates the predictive value of BXM in the general transplant population. I found that donor-specific anti-hla antibodies (DSA) are only present in one third of positive BXM and are associated with poor outcomes. The novel finding is that >80% of the DSA detected by BXM are complement-fixing IgG 1 and IgG 3 subclasses. Transplant glomerulopathy (TG) is type of chronic renal graft rejection. The pathogenesis of TG is unclear. In the second result chapter, I report risk factors and involvement of anti-hla antibodies in the development of TG. This study shows that glomerular rejection, delayed graft function, HLA presensitization and DSA have a univariate effect on TG development. Multivariate analysis revealed that DSA are an independent predictor of TG, after adjustment for other risk factors. I have further investigated the role of BXM in a unique, well-matched, highly sensitized patient group transplanted under the national renal exchange programme. I compared Luminex antibody analysis with BXM in predicting transplant outcomes. In highly sensitized patients, DSA are found in two thirds of positive BXM. In univariate analyses, BXM is associated with humoral rejection whereas DSA defined by Luminex are associated with total and all rejection types. The major finding is that, by multivariate analysis, DSA defined by Luminex are an independent predictor of total and humoral rejection, but BXM are not. These interesting findings are reported in the third result chapter. 9

10 Studies reported in this thesis define the clinical significance of anti-hla antibodies in renal transplant outcomes. Method comparison studies provide useful information on antibody specificity and their impact on graft survival. Collectively, a better understanding of alloantibodies associated with graft rejection and limitation of antibody detection methods may facilitate donor selection and choice of immunosuppressants, and consequently improve transplant outcomes. 10

11 v. Dedication For my father Kheng Choo, mother Sew Kee, siblings Hooi Ling, Inn Soo, Hooi Ping, Soo Tchuen, Chuen Sin, Sin Zheng and husband Lee Khuan. 11

12 vi. Acknowledgements A thesis, with multiple research studies and high quality collaboration, such as this can only be completed with a brilliant and exceptional supportive team of supervisors, collaborators, and team assistance. I have had the privilege to be supervised by Professor Graeme Russ, Dr. Toby Coates and, and Dr. Peter Bardy, without whom, my study would not be possible. I am so honoured and most grateful to have the opportunity to work in such an amazing team environment. My most heartfelt thanks for all your professional advice, your guidance, and unwavering support. To my wonderful collaborators, of whom I am most appreciative and honoured to have work with. Firstly Dr. Brian Tait (Royal Melbourne Hospital) and his team for their involvement in the National B-cell Cross-match Study, through data and serum collection, and presentation and further learning and networking opportunities within the local and international research and clinical community. Many thanks, it is such an honour to work with such a gracious and encouraging individual. My thanks to all the participating hospitals, and tissue-typing laboratories, who have collected serum and data for contribution to this research. Your professional accurate and timely support has made this research possible. To the ANZDATA team who has worked tirelessly, sometimes with unimaginable deadlines, and overtime hours, for such high quality data analysis. Special thanks to Ms. Hannah Dent, Mr. Brian Livingstone, and Dr. Sean Chang. Thank you to my colleagues at the South Australian Tissue Typing Laboratory (Pirie Street, Adelaide) of National Transplantation Immunology Services, Australian Red Cross Blood Service, and The Queen Elizabeth Hospital s Transplantation Immunology Team. The support, assistance and encouragement, and positive learning and working environments they have provided have been enjoyable and exceptional, and fundamental to my completion of this work. Finally I would like to thank the organisations who have made many educational, financial and facility contributions, to this thesis possible. Thank you especially to the National Transplantation Services of Australian Red Cross Blood Service, the Australian Department of Education, Science and Technology, and University of Adelaide, and Department of Nephrology and Transplantation services Queen Elizabeth Hospital, Woodville, South Australia for their support. 12

13 vii. Declaration This work contains no material which has been accepted for the award of any other degree or diploma in any university or other tertiary institution and, to the best of my knowledge and belief, contains no material previously published or written by another person, except where due reference has been made in the text. I give consent to this copy of my thesis when deposited in the University Library, being made available for loan and photocopying, subject to the provisions of the Copyright Act I also give permission for the digital version of my thesis to be made available on the web, via the University s digital research repository, the Library catalogue, the Australasian Digital Theses Program (ADTP) and also through web search engines, unless permission has been granted by the University to restrict access for a period of time Hooi Sian Eng 13

14 viii. Thesis Structure Chapter 1 is a review of Human Leukocyte Antigen (HLA) system consists of the genetics, structures and functions of HLA molecules. It also describes role of HLA antigen in renal graft rejection, and current tissue typing and antibody detection techniques. Chapter 2 contains the materials and methods used to perform the studies described in Chapter 3 to 5. The steps of tissue typing and Luminex assays and reagent preparation are documented. Statistical tests for patient demographics and transplant outcomes comparisons are recorded in this chapter. This thesis consists of three studies performed on patients with different transplant characteristics. The first study investigated the clinical significance of positive B-cell crossmatch in an Australian renal transplant population using Luminex assays. This study also examined the specificity of antibody causing positive B-cell crossmatch and correlation with renal transplant outcomes. The design and results of the study are described in Chapter 3. The pathogenesis of transplant glomerulopathy remains unknown. The second study described in Chapter 4 reported the involvement of anti-hla antibodies in development of transplant glomerulopathy. This study also indentified the influence of anti-hla antibodies on the progression of transplant glomerulopathy. The Australia National Renal Exchange Programme commenced in October 2004 to improve the HLA antigen matching and transplant rate for highly sensitized patients. Chapter 5 reported the overall performance of patients transplanted under this programme in comparison to general transplant population. Clinical predictive value of B-cell crossmatch and Luminex antibody analysis in highly sensitized patients were also compared. Chapter 6 brings together the findings of the studies reported in Chapter 3 to 5, and examined the uses of BXM and Luminex assays in current tissue typing protocol. The results are compared to previous studies and appropriate comments are included. A list of references and then an appendix are attached after Chapter 6. The appendix consists of the published journal article that has arisen from the findings presented in this thesis. 14

15 ix. Abbreviations AAT ADCC ELISA AHR APC BXM CAN CDC CN/HP CNX CREGs CRT CTLRs ctls DISC DS/SAPE DSA DTH DTT EPO ER ERAAP or ERAP 1 ERp57 FasL Fc-receptors H HLA IFN IL IMGT IR ITIMs KIR LD Li MESF MFI MHC MIC MMF Mya NK amini acid triplets antibody mediated cell cytotoxicity enzyme-linked immunoabsorbent assay acure humoral rejection antigen- presenting cells B-Cell crossmatch chronic allograft nephropathy complement-dependent lymphocytotoxicity current serum negative/ historic serum positive calnexin cross-reactive groups chaperone calreticulin C-type lectin receptors cytotoxic T-lymphocytes death-inducing signal complex dilution Solution/PE Streptavidin donor specific antibodies delayed hypersensitivity reaction dithiothreitol eosinophil persoxide endoplasmic reticulum ER associated aminopeptidase thiol oxidoreductase Fas-ligand fragment, crystallisable histocompatibility human leukocyte antigen interferon interleukin ImMunoGeneTics Immune Response immunonoreceptor tyrosine-based inhibitory motifs killer like receptor linkage disequilibrium Invariant chain Molecules of Equivalent Soluble Fluorochrome Median Fluorescence Intensity Major Histocompatalibity complex MHC class 1 chain-related genes mycophenolate mofetil million years ago natural killer 15

16 NO PBS PBS/citrate PFD RSA TAP TCR TG TH1 TN/BP TNF TPSN TXM UNOS XM nitric oxide phosphate buffered saline trisodium citrate probability of finding a donor recurrent spontaneous abortions transporter associated with antigen processing T-cell receptor transplant glomerulopathy T helper negative T-cells/positive B-cells crossmatches tumour necrosis factor Tapasin T-cell crossmatch United Network for Organ Sharing Registry crossmatching 16

17 x. Publications 1. HS Eng, G Bennett, I Humphreys, E. Tsiopelas, M Lake, SH Chang, P Bardy, PTH Coates and G. Russ. Luminex Investigation: Low titer HLA donor-specific antibodies in positive B-cell crossmatches predict late graft loss. Nephrology 2007:Supp 1;A HS Eng, G Bennett, I Humphreys, E. Tsiopelas, M Lake, SH Chang, P Bardy, PTH Coates and G. Russ. Luminex Investigation of Positive B-cell Crossmatches: low titler HLA donor-specific antibodies associated with inferior outcomes. Human Immunology 2007:68 Supp 1:S HS Eng, G Bennett, I Humphreys, E. Tsiopelas, M Lake, SH Chang, P Bardy, PTH Coates and G. Russ. Anti-HLA donor-specific antibodies detected in positive B-cell crossmatches by Luminex predict late graft loss. American Journal of Transplantation (11): Hooi Sian Eng, Scott Campbell, Graeme Russ, P. Toby H. Coates, Peter Bardy, Greg Bennett, Sean Chang, Brian Tait, On behalf of the Renal Transplant Advisory Committee (RTAC). A retrospective study of B-cell crossmatch with Luminex technology in well-matched highly sensitized patients from the AUSTRALIAN interstate exchange. Nephrology 2008:13;A Hooi Sian Eng, Greg Bennett, Eleni Tsiopelas, Sean Chang, Peter Bardy, Graeme Russ, P. Toby H. Coates. Alloantibody involvement and risk factors in transplant glomerulopathy. Nephrology 2008:13;A Hooi Sian Eng, Scott Campbell, Graeme Russ, P. Toby H. Coates, Peter Bardy, Greg Bennett, Sean Chang, Brian Tait, On behalf of the Renal Transplant Advisory Committee (RTAC). Predictive value of B-cell crossmatch and Luminex antibody analysis in well-matched highly sensitized patients from the Australian National Interstate Exchange (ANIE). Human Immunology 2008:69;S13. 17

18 7. Hooi Sian Eng, Greg Bennett, Eleni Tsiopelas, Sean Chang, Peter Bardy, Graeme Russ, P. Toby H. Coates. Donor specific anti-hla antibodies in transplant glomerulopathy. Human Immunology 2008:69;S Hooi Sian Eng, Greg Bennett, Peter Bardy, Patrick Coghlan, Graeme R. Russ, P. Toby H. Coates. Clinical significance of anti-hla antibodies detected by Luminex : Enhancing the interpretation of CDC-BXM and important post-transplant monitoring tools. Human Immunology 2009 In press. 18

19 xi. Presentations 1. HS Eng, G Bennett, I Humphreys, E. Tsiopelas, M Lake, SH Chang, P Bardy, PTH Coates and G. Russ. Clinical relevance of a positive B-cell crossmatch on renal transplantation: A single centre study. The Queen Elizabeth Hospital Research Day 20 October 2006, Adelaide, Australia. 2. HS Eng, G Bennett, I Humphreys, E. Tsiopelas, M Lake, SH Chang, P Bardy, PTH Coates and G. Russ. Analysis of the specificity and properties of alloantibodies associated with rejection of renal transplant. The 3 rd International Summer School on Immunogenetics November 2006, Bangkok, Thailand. 3. HS Eng, G Bennett, I Humphreys, E. Tsiopelas, M Lake, SH Chang, P Bardy, PTH Coates and G. Russ. B-cell crossmatching and kidney allograft outcomes in cadaveric transplant recipients. The 30 th Annual Scientific Meeting of The Australasia and South East Asia Tissue Typing Association November 2006, Chiang Mai, Thailand. 4. HS Eng, G Bennett, I Humphreys, E. Tsiopelas, M Lake, SH Chang, P Bardy, PTH Coates and G. Russ. Luminex investigation of positive B-cell crossmatches in renal transplantation. The 31 st Annual Scientific Meeting of The Transplantation Society of Australia and New Zealand March 2007, Canberra, Australia. TSANZ Young Investigator Award HS Eng, G Bennett, I Humphreys, E. Tsiopelas, M Lake, SH Chang, P Bardy, PTH Coates and G. Russ. Luminex Investigation: Low Titer HLA Donor-specific Antibodies in Positive B-cell Crossmatches Predict Late Graft Loss. The 43 rd Annual Scientific Meeting of The Australia and New Zealand Society of Nephrology (ANZSN) 8-12 September 2007, Queensland, Australia. Kidney Health Australia (KHA) Award for The Best Clinical Science Presentation 19

20 6. HS Eng, G Bennett, I Humphreys, E. Tsiopelas, M Lake, SH Chang, P Bardy, PTH Coates and G. Russ. Luminex Investigation of Positive B-cell Crossmatches: Low Titer HLA Donor-specific Antibodies Associated with Inferior Outcomes. The 33 rd Annual Meeting of American Society For Histocompatibility and Immunogenetics (ASHI) 8-12 October 2007, Minnesota, USA. 7. HS Eng, G Bennett, I Humphreys, E. Tsiopelas, M Lake, SH Chang, P Bardy, PTH Coates and G. Russ. Low Titer Donor-specific Antibodies in Positive B-cell Crossmatches Predict Late Graft Loss. The 31 st Annual Scientific Meeting of Australasia and South East Asia Tissue Typing Association November 2008, Perth, Australia. 8. Hooi Sian Eng, Scott Campbell, Graeme Russ, P. Toby H. Coates, Peter Bardy, Greg Bennett, Sean Chang, Brian Tait, On behalf of the Renal Transplant Advisory Committee (RTAC). Clinical significance of B-cell positive crossmatches and Luminex defined donor-specific antibodies in well-matched highly sensitized patients from the AUSTRALIAN interstate exchange. The XXII International Congress of The Transplantation Society August 2008, Sydney, Australia. Top 10 Abstracts Young Investigation Award of The XXII International Congress of The Transplantation Society August 2008, Sydney, Australia. Nominee for The President s Award of The Transplantation Society of Australia and New Zealand (TSANZ). TSANZ Young Investigator Award Amgen Young Investigator Award 20

21 9. Hooi Sian Eng, Greg Bennett, Eleni Tsiopelas, Sean Chang, Peter Bardy, Graeme Russ, P. Toby H. Coates. Risk factors and associations with alloantibodies in transplant glomerulopathy. The XXII International Congress of The Transplantation Society August 2008, Sydney, Australia. 10. Hooi Sian Eng, Scott Campbell, Graeme Russ, P. Toby H. Coates, Peter Bardy, Greg Bennett, Sean Chang, Brian Tait, On behalf of the Renal Transplant Advisory Committee (RTAC). A retrospective study of B-cell crossmatch with Luminex technology in well-matched highly sensitized patients from the AUSTRALIAN interstate exchange. The 44 th Annual Scientific Meeting of Australia and New Zealand Society of Nephrology (ANZSN) 6-10 September 2008, Newcastle, Australia. Finalist for The Clinical Science Young Investigator Award 11. Hooi Sian Eng, Greg Bennett, Eleni Tsiopelas, Sean Chang, Peter Bardy, Graeme Russ, P. Toby H. Coates. Alloantibody involvement and risk factors in transplant glomerulopathy. The 44 th Annual Scientific Meeting of Australia and New Zealand Society of Nephrology (ANZSN) 6-10 September 2008, Newcastle, Australia. 12. Hooi Sian Eng, Scott Campbell, Graeme Russ, P. Toby H. Coates, Peter Bardy, Greg Bennett, Sean Chang, Brian Tait, On behalf of the Renal Transplant Advisory Committee (RTAC). Predictive value of B-cell crossmatch and Luminex antibody analysis in well-matched highly sensitized patients from the Australian National Interstate Exchange (ANIE). The 34 th Annual Meeting of American Society For Histocompatibility and Immunogentics (ASHI) October 2008, Toronto, Canada. 13. Hooi Sian Eng, Greg Bennett, Eleni Tsiopelas, Sean Chang, Peter Bardy, Graeme Russ, P. Toby H. Coates. Donor specific anti-hla antibodies in transplant glomerulopathy. The 34 th Annual Meeting of American Society For Histocompatibility and Immunogentics (ASHI) October 2008, Toronto, Canada. 21

22 xii. Awards 1. Young Investigator Award of Amgen-The Transplantation Society of Australia and New Zealand. The 31 st Annual Scientific Meeting, March 2007, Canberra, Australia. 2. Kidney Health Australia For The Best Clinical Science Presentation Award. The 43 rd Annual Scientific Meeting of The Australia and New Zealand Society of Nephrology (ANZSN) 8-12 September 2007, Queensland, Australia. 3. Top 10 Abstracts Young Investigation Award of The XXII International Congress of The Transplantation Society August 2008, Sydney, Australia. 4. Young Investigator Award of Amgen-The Transplantation Society of Australia and New Zealand. The XXII International Congress of The Transplantation Society August 2008, Sydney, Australia. 5. Young Investigator Award of The Transplantation Society of Australia and New Zealand. The XXII International Congress of The Transplantation Society August 2008, Sydney, Australia. 6. Nominee for The President s Award of The Transplantation Society of Australia and New Zealand. The XXII International Congress of The Transplantation Society August 2008, Sydney, Australia. 7. Finalist for The Clinical Science Young Investigator Award The 44 th Annual Scientific Meeting of Australia and New Zealand Society of Nephrology (ANZSN) 6-10 September 2008, Newcastle, Australia. 22

23 Chapter 1 Literature Review 23

24 1.1 Overview of Human Major Histocompatibility Complex (MHC) Gene Discovery Major Histocompatibility Complex (MHC) genes were first described in mice by scientists who studied genetic and antigenic basis of tumour transplantation [1-6]. Gorer et. al.. and Snell et. al.. reported that several antigens were responsible for rejection of tumors as well as normal tissues in mice. The loci coding for these antigens were named as Histocompatibility or H genes. One of the loci (H-2) appears to cause the greatest degree of rejection, and later known as Major Hiscompatibility locus [1-7]. In human, MHC is called Human Leucocyte Antigen (HLA). The first HLA antigen, MAC (later known as HLA-A2), was discovered by Jean Dausset as a result of reacting sera from multiple transfused patients with leucocytes from healthy blood donors [8]. This was followed by discoveries of supertypic antigens 4a4b (Bw4 and Bw6) by Van Rood et. al.. and antileucocytes antibodies in multiparous women by Rose Payne [9, 10]. In 1960s and 1970s, Green, Paul and Benacerraf showed that, in inbred mice, immune response to selected antigens was controlled by genes [11]. These genes were named as Immune response (Ir) genes located within the MHC region [12]. Concordantly, Mitchell and Miller Jacques reported T-cells had regulatory functions over B-cells [13, 14]. MHC molecules were involved in the interaction between T-cells and B- cells for production of antibody [15]. In 1974, Zinkernagel and Doherty discovered a critical role of MHC genes in which antigens primed T-cells only recognised a foreign antigen when it was expressed by target cells sharing the same MHC antigens of the initial antigen presenting cell [16]. This phenomenon is known as MHC restriction. Zinkernagel and Doherty were awarded the Nobel Prize for Physiology and Medicine in 1996 for their findings of the nature of the cellular immune defence. 24

25 1.1.1 Chronology of HLA Discovery 1958 First HLA antigen, MAC (HLA-A2, -A28) 1963 Identify 4a/4b (Bw4/Bw6) 1964 Identify LA1, LA2 and LA3 (HLA-A1, HLA-A2 and HLA-A3) 1964 Acceptance of cytotoxicity over agglutination 1965 Proposal of allelism of HLA antigens 1967 Segregation of Alleles demonstrated in families 1970 Recognition of 2 loci, HLA-A and HLA-B world populations typed by 75 laboratories 1975 Recognition of third locus, HLA-C 1977 Identify HLA-DR 1984 Studies of HLA and Disease associations Studies of gene structure Definition of MB (HLA-DQ) 1987 Introduction of DNA techniques with serological, biochemical and cellular methods Definition of HLA-DP and HLA-DQ 1992 Use of Polymerase Chain Reaction - eg sequence-specific oligonucleotide probes 1996 Molecular definition of HLA class I Studies of HLA-G, -E, -DM, Tap & LMP Source: The official website of The Australasia and South East Asian Tissue Typing Association Genetics of MHC The MHC is a group of linked genes found in various jawed vertebrates, but not in jawless vertebrates or invertebrates [17-19]. In 1999, The MHC Sequencing Consortium published the first complete sequence and gene map of human MHC (Figure 1.2.1) [20]. The entire human MHC is 3.6Mb in length with over 260 genes, locates on Chromosome 6p21.3. Approximately 40% of the expressed genes in this region have immune system function: (i) antigen processing and presentation by the 25

26 class I and class II genes (ii) innate immunity, inflammation and regulation of immunity by class III genes, (iii) intercellular interactions via MHC receptors and ligands. Genetics, population frequencies and functions of MHC genes have been well-characterized. Human MHC is the reference for comparative genomic analysis in various species. 26

27 Figure 1.2.1: Human MHC gene map The MHC has been divided into Class I, III and II subregions from telomere to centromere. The white, grey, striped and black boxes show expressed genes, gene candidates, noncoding genes and pseudogenes. Source: Modified from Shiina T. Tissue Antigens 2004:64; [21]. 27

28 1.2.1 Evolution of MHC Genes Genomic sequences of MHC genes in various species have been characterized (Figure1.2.2) Non-human primates such as chimpanzees (man s closest relative, 5 million years ago, Mya) have an almost identical MHC organization compared with humans [22-24]. The major difference is the location of polymorphisms and presence of a unique HLA-A related gene in Chimpanzees. In pig or swine MHC (SLA complex, 60Mya), class II region is located in the q- arm (7q1.1) whereas class I and III is located on the p-arm (7p1.1) [25-27]. The linkage of MHC class I, II and III is lost in swine. Rat and mouse (100Mya) MHC are very similar to human MHC, consists of three distinct regions, class II, class III and class I from centromere to telomere [28-31]. However, one major difference is that rat and mouse lack MIC-related genes within MHC [28-31]. Chicken B locus (300Mya) represents the minimal essential MHC within nonmammalian vertebrate, which only consists 19 genes [32]. Most of the genes have a counterpart in the human MHC [32]. In avian MHC, the linkage between the class I and class II regions is mostly retained, but the genomic organization of MHC is different to its mammalian counterparts. In avian MHC, human equivalents of the chicken natural killer (NK) receptor-like genes are present in MHC region, but in human, KIR genes are independent of the human MHC and found on different chromosome [18]. Linkage of some MHC class I, II and III genes are found in cartilaginous fishes (500Mya, million years ago), which are the earliest jawed vertebrates to diverge from a common ancestor with human [17]. This suggests that the common ancestor of all jawed vertebrates should have an organised MHC[17]. Comparative genomic analysis of MHC genes in various species reflects the emerging picture of the MHC region arose alongside the emergence of adaptive immune system. From species to species, the number and organization of genes can differ greatly. MHC changes rapidly through gene polymorphism, duplication and gain or loss of some loci. MHC is highly polymorphic, with marked linkage disequilibrium and great gene diversity. 28

29 Figure1.2.2: Timescale of MHC evolution Key events and inter-species differences in MHC genes. Source: Modified from Kelley J et. al.. Immunogenetics 2005:56; [17] MHC Gene Polymorphism MHC genes are a key coordinator of specificity in adaptive and innate immune systems in vertebrates. Distinctive features of MHC, such as high gene density and diversity, allows recognition of a large number of pathogens and also determination of susceptibility and resistance to infectious, autoimmune and other diseases. The nature of MHC genes are (i) highly polymorphic, (ii) large number of alleles, and (iii) greater rate of non-synonymous than synonymous substitutions [33-43]. Mechanism underpins selection of MHC genes remains unknown. However, several factors have been correlated with polymorphism of MHC genes such as parasite and pathogen mediated selection, and reproduction and mating preference. 29

30 According to the pathogen driven selection theory, specific alleles which can provide protection from a pathogen are favoured and will rise in frequency. Doherty and Zinkernagel suggested that heterozygotes have higher fitness than homozygotes because individuals carrying two alleles have greater repertoire of peptides that can be presented to T-cells [44]. This results in heterozygote advantage or overdominant selection. Black et. al. showed that the number of persons with homozygous HLA haplotypes in several groups of South American Indians was 39% less than that expected assuming unmodified equilibrium. In a subpopulation of 122 persons whose parents' HLA constitutions were known, there were 56% fewer homozygous persons than expected [33]. Overdominant selection increases the percentage of heterozygotes and contributes to HLA polymorphism. MHC presents peptides from parasites and pathogens, thus the ability of transformation in order to recognise novel pathogen strain is crucial. Human MHC gene map reveals that the class I and class II regions consist of a high level of pseudogenes [20]. The class I and II regions appear to have duplicated numerous times, generating novel gene family members which have then diverged into new functions. This phenomenon is known as gene conversion, a possible mechanism of generating new alleles [20]. The findings offered an explanation for the large allele number in MHC region. In 1976, Yamazaki et. al. reported the first observation of MHC based negative assortative mating, in which male mice from inbred strains preferred to mate with females with different MHC types [45]. The work has been extended by several groups, where similar findings were reported [46, 47]. Female mice were more attracted to odours of male mice with different MHC haplotype, suggested that olfaction may be a mechanism in MHC-based mating preference [48]. The hypothesis has been tested in human. Wedekind et. al. showed that females prefered odour from male with differ HLA type [49], and in turn, men also prefer odour of females with differ HLA type [50]. In the study of the Anabaptist Hutterite religious community, a close community with a high fertility rate, consistent findings with avoiding spouse who have the same HLA haplotypes were reported [46]. MHC may mediate kin recognition or discrimination to prevent inbreeding [51, 52]. Data shows that individuals choose mates to maximise MHC heterozygosity and to prevent inbreeding. However, due to a lack of repeatability 30

31 and contradictive findings in other studies, the associations of MHC and mate choice in human and ruminants remain inconclusive [53-55]. Maternal-foetal interaction may contribute to HLA polymorphism. Studies showed an increased sharing of HLA alleles in couples that have experienced recurrent spontaneous abortions (RSA) [41, 56]. There was a significant excess of HLA-DR sharing in couples with RSA and a significant excess of HLA-DQ sharing in couples with unexplained infertility who failed treatment by In-vitro Fertilisation [56]. In addition, an excess of HLA heterozygotes over Hardy-Weinberg expectation has been demonstrated in a study of South Amerindian populations [57]. The excess of heterozygotes may increase the overall progeny fitness by maximising immunological capability. 1.3 Human Leukocyte Antigen (HLA) System In 1967, Histocompatibility Workshop named the major system of leukocyte antigens in human as HL-A [58]. The 1998 WHO nomenclature committee officially recognised forty-two HLA genes, and grouped them into three regions, called the class I, III and II regions based on physical location, function and structure, from centromere to telomere [20, 59]. The MHC region contains approximately 180 genes which include a large number of polymorphic multicopy genes arose from repeated gene duplication during evolution [60-62]. Up to September 2008, more than 3300 alleles in HLA regions have been defined, and the sequences have been published on the ImMunoGeneTics (IMGT)/HLA database [Anthony Nolan Website Of these, in class I region: 697 alleles have been found in HLA-A locus,1109 in -B locus and 381 in -C locus; in class II region: 690 alleles have been found in -DRB1 locus, 95 in -DQB1 and 131 in -DPB1 [Anthony Nolan Website HLA is the most polymorphic genetic system thus far, compared to 1496 alleles in 2003, the number of alleles have increased significantly [63]. New alleles are discovered at the rate of 5 to 7 every month (Figure 1.3.1). 31

32 Figure 1.3.1: The number of antigens and alleles from 1968 to September MHC are the most polymorphic genes. New alleles are discovered every month. Source: Modified from IMGT Anthony Nolan website 32

33 1.3.1 HLA Gene Inheritance and Frequency Most of the HLA genes are in linkage disequilibrium and are therefore inherited en bloc [63]. Linkage disequilibrium (LD) refers to the non-random association between alleles at adjacent genetic loci. LD is more prominent among the HLA class II genes; in particular the association of -DQ, -DRB3, 4, 5 with some -DRB1 may approach 100% in some ethnic groups. A haplotype is defined as a combination of lined HLA genes transmitted on a single parental chromosome [7]. Individuals inherit one set of HLA genes from each parent. Recombination between HLA genes is infrequent, 1-3%, compared to the whole genome. Strong association among HLA genes suggested that certain MHC genes on a particular haplotype are tuned to work together. Low recombination rate also suggests that maintenance of allele combinations with favoured immunological function through selection against unfavourable recombinant haplotypes [7, 18, 37, 38, 41, 47]. The distribution and frequency of HLA alleles and specific HLA haplotypes varies among different ethnic groups. The difference can be due to different selection pressures on population evolving in geographically distinct environments. Some HLA alleles occur in all populations, but at different frequencies. There are alleles or haplotypes that only present in a specific population [7, 18, 29, 37, 38] HLA Class I, II and III Regions The HLA class I region is 1.8Mb in length, from MICB to the HLA-F gene at the telomeric end of the HLA region [64]. The HLA class I and class II antigens involved in genetics control of the immune regions, whereas the class III region contains genes for complement components such as C2 and C4 (Figure 1.3.2). Six genes involve in peptide presentation are known as classical genes: HLA-A, -B and -C in the class I region, and HLA-DP, -DR and -DQ genes in the class II region [41]. The class I region is the most divergent, the class II region is intermediate and the class III region is conserved in gene diversity and organization [65]. 33

34 Figure 1.3.2: HLA gene map illustrates clustering of immune system Source: Modified from Tragerne JA. International Journal of Immunogenetics 2008:35; [66]. Class I region consists 18 HLA genes, of which, 6 are coding and 12 are pseudogenes, and 7 MHC class I-chain-related genes (MIC, 2 coding and 5 pseudogenes) [18]. There are also more than 50 non-hla genes found in the Class I region [64]. Some of the non-hla genes are involved cell growth, DNA replication and repair, and regulation of transcription [64, 67]. The HLA class I and MIC genes are organised together as a repeating unit within three duplication blocks, known as the α, β and К blocks [18]. Microsatellites are polymorphic DNA loci repeating units that usually consist of 1-6 base pairs in length. A total of 758 microsatellite repeats have been found in the class I region with the density of one microsatellite every 2.3kb. These microsatellite are well dispersed throughout class I region, and have been used as a genetic marker for mapping of genes associated with diseases such as Behcet s and psoriasis [7, 18, 20, 63]. The HLA class II region is located at the centromer end of the MHC and covers approximately 1Mbp. The precise length differs, depending on the haplotype, the 34

35 main differences being due to the number of DRB genes [68]. It has been divided into four subregions, knows as DP, DM, DQ and DR from centromere to the telomere [68]. There are approximately 19 HLA class II genes: 11 are coding and 8 are pseudogenes [18]. The class II genes, except for DO, are all arranged as related pairs of α and β genes (eg DQA1 next to DQB1, DPA1 next to DPB1) [68]. Almost all class II genes are involved in immune response, except RING3. Gene with related function clustering in class II region may facilitate regulation of expression and function (Figure 1.3.2) [68]. The HLA class III region is located centrally between the class II and I regions. There are genes found in this region, spanning kb of genomic DNA [18, 69]. It is the most gene dense region within the human genome [18]. The class III genes, C4, factor B (BF) and C2, encode subunit protein for the C3 and C5 convertases that are essential for complement activation [69]. Some of the genes in this region encode cytokine (LTB, TNF, LBA), heat shock protein, and putative regulation genes (IFKB, BAT1) involved in inflammation and regulation of immunity [18, 69]. Other genes such as PPT2 and CYP21B encode enzymes related to the metabolism of lipids and steroids [69]. 1.4 Nomenclature Current nomenclature was recommended during the 10 th International Histocompatibility Workshop in 1987, and other rules or modification have been subsequently added when necessary [7]. Each HLA allele name is unique; formed by four, six or eight digit number. Each allele has been given at least a four digit name designates specificity. The six and eight digit names are only assigned when necessary. The length of the allele designation is dependent on the sequence of the allele and that of its nearest relative. A suffix is added at the end of the name to describe allele expression. HLA alleles are named according to the following rules: 1. HLA followed by a hyphen designated the MHC 2. Capital letters designated HLA loci 3. First two digits describe the type, which often corresponds to the serological antigen 35

36 4. Third and fouth digits describe the subtypes; numbers have been given in the order where DNA sequences had been determined. Alleles whose numbers differ in the first four digits must differ in one or more nucleotide substitutions that change the amino acid sequence of the encoded protein. 5. Fifth and sixth digits are given to alleles that differ only by synonymous nucleotide substitutions (also called silent or non-coding substitutions) within the coding sequence. 6. Seventh and eight digits are given to alleles that differ by sequence polymorphisms in the introns or in the 5' or 3' untranslated regions that flank the exons and introns. 7. Suffixes for expression status: N Null alleles, alleles not expressed L - 'Low' cell surface expression when compared to normal levels S - A protein which is expressed as a soluble 'Secreted' molecule but is not present on the cell surface C - An allele product which is present in the 'Cytoplasm' but not on the cell surface A - 'Aberrant' expression where there is some doubt as to whether a protein is expressed Q - The expression of an allele is 'Questionable' given that the mutation seen in the allele has previously been shown to affect normal expression levels 36

37 1.5 Structure of HLA Molecules MHC molecules are glycoprotein, formed by two types of protein chain (heterodimers) (Figure 1.5.1, Figure 1.5.2). HLA class I molecules contain a heavy protein chain (alpha, MW=45,000) that is noncovalently associated with a light nonpolymorphic β-2 microglobulin (β 2 m, MW=12,000) [7]. The α chain is coded by class I genes, the β chain is coded by the β-2 microglobulin gene on chromosome 15. The α chain has five domains, two peptide binding domains (α1 and α2), one immunoglobulinlike domain (α3), the transmembrane region and the cytoplasmic tail [70]. The α1, α2 and α3 domains are the extracellular portion of the heavy chain, each approximately 90 amino acids long and encoded on separate exons [71]. The amino acid differences that account for molecular diversity in the class I region occur mainly in the α1 and α2 domains within any of the seven hypervariable sequences (9-12, 40-45, 62-83, 94-97, , , ) [63]. The α3 domain and β2m are relatively conserved [71]. Eight β-pleated sheets formed by the amino-terminal segment of the α1 and α2 domains form a platform bound by two helices that create the sides of a cleft (Figure 1.5.2). The floor and sides of the cleft interact principally with the peptide, whereas the top of the helices and areas adjacent to the peptide-binding groove interact with the T- cell receptor (TCR). For class I, the sides of the peptide-binding groove restrict the bound peptide at its two ends and thus can only bind peptides of 8 10 residues [72]. Both α3 domain and β 2 m are β-sandwich structures composed of two antiparallel β- pleated sheets, one with four β-strands and one with three β-strands, connected by a disulphide bond [71]. 37

38 Figure 1.5.1: Structure of HLA class I molecules Class I molecules are formed by α chain and β2m. The α1 and α2 domains form the peptide binding groove. The transmembrane region and cytoplasmic tail hold the HLA molecule on cell surface. The α3 domain and β2m composed of two antiparallel β- pleated sheets, one with four β-strands and one with three β-strands, connected by a disulphide bond. Source: Modified from med.unne.edu.ar/catedras/bioquimica/hla.htm. Figure 1.5.2: Ribbon structure of class I peptide binding groove. Eight β-pleated sheets formed by the amino-terminal segment of the α1 and α2 domains form a platform bound by two helices. Source: Modified from depts.washington.edu/rhwlab/resmat/dq/dq.html. 38

39 HLA class II molecules contain 1 alpha (α) and 1 beta (β) polypeptide chains encoded by class II genes (Figure 1.5.3). Each of the α and β chains has four domains: the peptide-binding domain (α1 or β1), the immunoglobulin-like domain (α2 or β2), the transmembrane region and the cytoplasmic tail [70]. The peptide-binding site of class II molecules is much similar to that of class I, where the amino-terminal portions of the α1 and β1 domains fold into β-pleated sheets and the carboxyl terminals form the helices. However, subtle changes in the helical regions produce a binding groove with open ends, which allows peptides to hang out of the groove at both ends and thus accommodate a larger peptide, amino acids, than the class I molecules [72]. Figure 1.5.3: Structure of HLA class II molecule The molecule is formed by 1 α and 1 β polypeptide chains. Source: Modified from med.unne.edu.ar/catedras/bioquimica/hla.htm. 39

40 HLA Peptide-Binding Grooves The HLA peptide-binding groove consists of a floor (β-pleated sheet) and two walls (helixs) (Figure 1.5.4). Amino acid pockets in the floor determine which peptide bind to a HLA molecule. Most of the HLA class I molecules have six pockets (A through F) distributed along the groove (Figure 1.5.5). Only two or three of the pockets are crucial for peptide specificity. These pockets are known as peptide anchors. The residues lining the pockets are encoded in genetic segments defined by the individual alleles of the HLA genes. HLA molecules, products of a specific HLA gene, are able to bind a large number of peptides. These peptides differ in their sequences, but share two or three amino acid residues (a motif) that fit into the anchor pockets. Peptides that bind to different allelic products are distinguished by their motifs [70]. Figure 1.5.4: Peptide in HLA peptide-binding groove The residues lining the pockets determine the specificity of peptides. Source: Modified from Dr. Nattiya Hirankarn s power point HLA Overview (slide 8), The 3 rd International Summer School on Immunogenetics,

41 Figure 1.5.5: Amino acid pockets in the floor of binding groove determine the binding of peptide HLA molecules can bind to a large number of peptides sharing two or three amino acid residues that fit into the anchor pockets. Source: Modified from Klein J et. al. The New Engl J Medicine 2000:343; [70]. 41

42 1.5.2 HLA Epitopes An epitope, also known as antigenic determinant, is the part of a HLA molecule that is recognized by the immune system, specifically by antibodies, B cells, or T cells. A HLA molecule can have more than one epitope [73]. Epitopes come in a huge variety of different shapes with different binding affinity (Figure 1.5.6). Some epitopes are continuous linear amino acids sequence, and others are formed by a few key amino acids arise from different positions in the linear amino acids sequence due to folding of protein [74]. HLA molecule epitopes are defined with monospecific antibodies or monoclonal antibodies. These epitopes are mostly located in and around the exposed peptide groove. HLA epitopes fall into two groups: private and public epitopes. Epitopes that only found on a single type of HLA molecule are referred as private epitopes. Epitopes that found in more than one type of HLA molecules are referred as public epitopes. In serology, antibodies directed to public epitopes have been used to categorise HLA types into major cross-reactive groups (CREGs) [7]. CREG matching has been considered a feasible alternative to HLA matching. However, the significance of CREG matching in kidney transplantation is controversial [75, 76].. Figure 1.5.6: HLA molecule epitopes HLA molecule epitopes (yellow) are a cluster of amino acids that recognised by immune system. Most of them locate around peptide binding groove. Peptide is coloured brown. Source: Rene Duquesnoy s power point file entitled HLAMATCHMAKER Algorithm, slide

43 1.5.3 HLA Expression HLA class I antigens are expressed on most of the nucleated cells. The level of expression is dependent on cell type: high expression in lymphoid and myeloid cells, less on liver, lung and kidney and only sparsely on brain and skeletal muscle [70, 77]. HLA-G is mainly expressed on the surface of the placental extravillous cytotrophoblast. Class II molecules only found in restricted cell types, such as B-cells, dendritic cells, macrophages and thymic epithelium. In the presence of certain cytokines such as interferon-gamma, other cell types may express class II molecules or increase expression of class I molecules [70, 77]. 1.6 Roles of HLA Molecules The major biological functions of HLA molecules are: antigen presentation, T- cell receptor repertoire selection, regulation of NK cell activity and protection of foetal tissue from maternal immune system during pregnancy Antigen Presentation Endogenous Presentation Pathway In endoplasmic reticulum (ER) lumen, the MHC class I heavy chain (α-chain) binds to chaperone, calnexin (CNX) upon synthesis (Figure 1.6.1) [78]. Folding and intrachain disulfide bond formation in α-chain occurs at this stage [70, 78]. Upon dissociation from chaperone, the α-chain binds β 2 m and is incorporated into the peptideloading complex. The other constituents of the complex are two subunits of the transporter associated with antigen processing (TAP1 and TAP2), the transmembrane glycoprotein tapasin (Tpsn), the soluble ER chaperone calreticulin (CRT), and soluble thiol oxidoreductase (ERp57) [78]. Tapasin associated with TAP via its transmembrane domain, and stabilises the TAP1/TAP2 heterodimer [79-82]. CRT is required for stable binding of the MHC α-chain - β 2 m heterodimer [83-85]. ERp57 is a disulfide isomerise 43

44 that cooperates with CRT or CNT in assisting the folding of newly synthesized glycoprotein [86-88]. Endogenously derived protein are excreted into cytosol, and chopped into smaller peptides by proteasomes (Figure 1.6.2) [89]. Proteasomes are assembled from smaller protein subunits, contain several proteolytic sites. Peptides are transported into the ER from the cytosol via TAP, and, if necessary, they are trimmed by an ER associated aminopeptidase (ERAAP or ERAP1) to 8-10 amino acids, the size of peptides that can fit into class I molecules peptide-binding groove [90-92]. There are insufficient data to explain how binding of peptide to the MHC class I heavy chain - β2m heterodimer will dissociate the heterodimer from peptide-loading complex. However, it is believed that the class I molecules dissociate, together with CRT, after binding peptides. The fully assembled class I molecule then leaves the ER and travels via the Golgi apparatus to the plasma membrane, where it is accessible to CD8+ T-cells [78]. Figure 1.6.1: Roles of peptide loading complex in MHC class I assembly and peptide loading. Binding of newly synthesized α-chain to CNX initiates folding and intrachain disulfide bond formation in the α-chain. Upon dissociation from CNX, the α-chain and β 2 m are joined and incorporated into the peptide loading complex. The peptide loading complex is formed by TAP, tapasin, CRT and ERp57. Peptides are transported into the ER via TAP, and then fit into class I molecule peptide binding groove. Source: Modified from Cresswell et. al. Immunological Reviews 2005;207: [78]. 44

45 Figure 1.6.2: Endogenous peptide presentation pathway Peptide is loaded to HLA class I molecule in endoplamic reticulum and travel to plasma membrane via golgi apparatus. Source: Modified from Klein J and Sato A. N England J Med 2000;343: [70] Exogenous Presentation Pathway Newly synthesized class II α and β chains assembled with a protein produced in ER, called invariant chain (Ii). The Ii chain maintains the newly developed class II chains in a partially folded state, and a portion of Ii chain called CLIP act as a stopper for the peptide-binding groove, and to prevent premature loading of peptides [7, 70]. Enclosed in membranous vesicles, Ii directs the molecules through the ER membrane, into the cytosol, and to membrane-bound endosomal compartments, where exogenous peptides will be loaded. Exogenous antigens entered cells via endocytosis, which includes pinocytosis (uptake of fluid) and phagocytosis (uptakes of particle) (Figure 1.6.3). The material taken up is encased within the plasma membrane. Endosomes containing the exogenous material eventually fuse with lysosomes within cytosol and the material is partially degraded. Proteolysis of protein produces peptide in various sizes [7, 70, 93, 94]. The endolysosomes eventually fuse with the MHC class II transporting vesicles to 45

46 form the MHC class II compartment, and leads to degradation of invariant chain by protease. This process leave behind CLIP which remains bound in the peptide-binding groove. Binding of HLA-DM, a non-classical HLA molecule, to class II molecule releases CLIP peptide and allows other peptides to bind to the empty groove. Upon dissociation from DM, fully folded class II peptide complex is incorporated into cell membrane [7, 70, 93, 94]. Protein processing and loading of peptides onto class I molecules are occurring all the time in most cells. Whereas the processing of exogenous proteins and the loading of peptides onto class II molecules are normally restricted to B-cells, macrophages and dendritic cells [70]. Even though most of the class I and class II molecules process and present peptides derived from endogenous and exogenous proteins, respectively, cross-presentation may occur. Dendritic cells are the dominant cell type that perform this function, however, the mechanism of cross-presentation remains unclear [78]. Figure 1.6.3: Exogenous antigen presentation pathway Peptide loading of class II molecules occurs outside ER. Peptide binding groove of immature HLA class II chains bind to Ii. Ii is degraded by protease in lysosomes, and leave behind CLIP which remains bound in peptide-binding groove. Binding of HLA-DM release CLIP and allows other peptides to be loaded into the empty groove. Source: Modified from Klein J and Sato A. N Engl J Med 2000;343: [70]. 46

47 1.6.2 T-cell Receptor Repertoire Selection The T-cell receptor or TCR is a molecule found on the surface of T-cell that is, in general, responsible for recognizing antigens bound to major histocompatibility complex (MHC) molecules [95]. It is a heterodimer consisting of an alpha and beta chain in 95% of T-cells, whereas 5% of T cells have TCRs consisting of gamma and delta chains. Engagement of the TCR with antigen and MHC results in activation of T- cells. The selection of a final TCR repertoire is largely determined by the type of MHC/peptide complex the T-cells exposed to during its maturation in thymus [70]. In general, the thymocytes maturation can be divided into 3 stages based on expression of coreceptor, CD4 and CD8, on cell surface. Stage 1: early double negative CD4- and CD8- cells, 2: a predominant double positive cells CD4+ and CD8+ and 3: mature CD4+ or CD8+ single positive cells [96]. Immature CD4- and CD8- thymocytes upregulate coreceptors upon TCR β locus rearrangement. This is followed by TCR α chain rearrangement and expression of TCR αβ heterodimers on the cell surface at Stage 2. These thymocytes have now become eligible for both positive and negative selection. The T-cells progenitors entering the thymus are programmed to die unless they receive survival signal to prevent apoptosis and to differentiate further. These signals are generated from the interaction between HLA-peptide complexes with both TCR and CD4 and CD8 coreceptors [97, 98]. The TCR engages the peptide and the peptide binding part of the HLA molecules, whereas the coreceptors interact with parts outside the peptide binding groove. Interaction of these T-cells with HLA molecules results in down regulation of one and the up-regulation of the other coreceptor. T-cells express MHC class II restricted receptors are positively selected to the CD4 lineage, while T-cells expressing class I restricted TCRs are usually selected to the CD8 lineage. Consequently, mature T-cells are CD4+CD8- type or CD4-CD8+ type [70, 96]. Thymocytes enter the thymus and proceed toward the center, dividing and differentiating along the way. Each clone of thymocytes produces TCR specific for a different ligand according to its TCR genes. As they enter the thymic cortex, thymocytes are trying to match their newly formed receptors with the abundant HLA-peptide complexes on cortical epithelial cells. Most of the thymocytes do not find ligands that fit the binding sites of their receptors, consequently they do not receive signals to justify 47

48 their survival, and they undergo apotosis [70, 96-98]. A minority of the thymocytes manage to weakly engage their receptors with HLA-peptide complexes. These thymocytes have been positively selected, and received survival signals that block the pathway to apotosis. The binding of the thymocytes and HLA-peptide complexes are weak, therefore they are capable to dissociate from cortical epithelial cell and move deeper into the thymus (Figure 1.6.4) [7, 96]. Thymic medulla, thymocytes that survived in the positive selection in cortical zone encounter macrophages and dendritic cells that displayed HLA-peptide ligands. The thymocytes once again try to match their receptors with the HLA-peptide ligands. This stage is known as negative selection, in which a high affinity match between thymocyte TCR and HLA-peptide ligands generates signals that direct the thymocytes to undergo apoptosis. Negative selection diminishes thymocytes that has high affinity against self-cells, and therefore prevent initiation of an autoimmune response (Figure 1.6.4) [7, 70, 95-98]. Cells that survive both positive and negative selection leave the thymus and enter the periphery as naïve T-cells. Of all the progenitor cells enter thymus and proliferate in it, less than 1 percent mature into T-cells. In periphery, the naïve T-cells may interact in low affinity with HLA molecules and self peptide to support their existence. During an infection, T-cells that bind to complexes of HLA molecules and pathogen peptides with high affinity are stimulated to initiate an immune response [70, 96]. 48

49 Figure 1.6.4: Positive and negative selection of T-cell receptor repertoire In thymic cortex, the thymocytes are positively selected if they bind to ligands on cortical epithelial cells. The selected thymocytes differentiate and move into thymus medulla. At this stage, negative selection induces apoptosis in cells which bind to ligands presented by macrophages and dendritic cells. Source: Modified from Sebzda E et. al. Annual Rev Immunol 1999:17; [96]. 49

50 1.6.3 Regulation of NK Cell Activity Natural killer (NK) cells are a type of lymphocytes that can kill cellular targets including tumors and infected cells without antigen presentation [99, 100]. This means that NK cells have the potential to attack normal self cells if the regulation is not well synchronised. NK cells do not express antigen receptors encoded by genes that active in recombination [101]. Thus, NK cells are considered part of innate immune system. NK cells release small cytoplasmic granules of proteins called perforin and granzyme that cause the target cell to die by apoptosis or necrosis [102]. The effector functions of NK cells are regulated by normal expression of MHC class I molecules on cells, known as missing-self recognition (Figure 1.6.5) [103]. There are two types of NK receptors, the C-type lectin receptors (CTLRs) and the Ig-like receptors. Both receptor types include inhibitory, that recognize MHC class I molecules and activating receptors, that recognize molecules collectively present on all cell surfaces [104]. Those that are inhibitory contain immunoreceptor tyrosine-based inhibitory motifs (ITIMs) within their cytoplasmic tails. The ITIMs inhibitory receptors exert an inhibitory function within the cell by recruiting phosphatases, such as SHP1 and SHP2 that can antagonize signal transduction events that would otherwise lead to release of NK cytotoxic granules or cytokines. Activating receptors are associated with accessory proteins, such as DAP-12 (DNAZ activation protein 12) that contain positively acting ITAMs within their cytoplasmic tails and promote events leading to NK mediated attack. Upon engagement with MHC class I molecules, inhibitory receptors suppress signals that would otherwise lead to NK cell activation. Cells that lack of MHC class I molecules are therefore unable to engage the inhibitory receptors and are likely die. Balance between inhibitory and activating stimuli justify whether NK mediated killing will occur (Figure 1.6.5) [102, 104, 105]. In addition to missing-self regulation, NK cells also use their receptors to recognise pathogen components or MHC class I-like proteins, such as the MHC class I chain-related A chain (MICA), that are normally poorly expressed on normal health cells. MICA and related ligands are often upregulated on transformed or infected cells and this may generate signals to activate NK cells. This phenomenon is known as induced-self recognition (Figure 1.6.5) [102]. 50

51 Figure 1.6.5: Regulation of NK activity by HLA class I molecules The balance of inhibitory and stimulatory signals received by a natural killer cell determines the outcome of interactions with target cells. Normal expression of MHC class I antigens generates inhibitory signals that balanced the stimulatory signals, and protect the target cell from killing by natural killer cells. Stimulatory signals delivered by target cells are unopposed when class I expression is down-regulated (cell transformation or infection), resulting in NK-cell activition and cells lysis. Source: Modified from Raulet D and Vance R. Nat Rev Immunol 2006:6; [102]. 51

52 1.7 Significance of HLA in Renal Transplantation Even though potent immunosuppressants have been widely used in posttransplantation treatment, allograft rejection remains a major cause leading to graft dysfunction or graft loss. Renal allograft rejection may occur at any time after transplantation. They can be categorized into hyperacute (early), acute (short-term) and chronic (long-term) rejections, and this also reflects the different underlying mechanisms. Hyperacute rejection may occur a few minutes to hours following revascularization of the graft. It usually caused by powerful preformed circulating cytotoxic anti-hla antibodies. It also occurs minutes before closure of the incision or immediate after closure of the wound, and this causes the kidney to never function [106, 107]. Acute rejection occurs within a few weeks or months following kidney transplantation. Early detection of renal acute rejection is important because prompt graft rescue treatment may reverse renal damage. Chronic rejection is seen after months or years of good renal function. It leads to a slow progressive loss of graft function and ultimately causes late graft loss [106, 107] Mechanism of Renal Allograft Rejection The primary event that leads to graft rejection is recognition of alloantigens by naive host T-cells. Host T-cells recognise MHC antigen through two distinct, yet not mutually exclusive pathways: direct and indirect recognition [108, 109]. In direct pathway, T-cells recognise intact allo-mhc molecules on the surface of donor antigenpresenting cells (APC), such as dendritic cells. Whereas in indirect pathway, T-cells recognise processed alloantigen presented as peptides by self-apcs [108, 109]. The MHC class I antigens are recognised by the T-cell receptor (TCR) of CD8+ T-cells, whereas the MHC class II antigens are recognised by the TCR of CD4+ T-cells [110]. Interaction between costimulatory molecules present on activated DC (eg B7) with their respective counter-receptors on T-cell membranes generates intracellular signals and induces production of interleukin (IL)-2 by naive CD4+ T-cells. Once activated by both TCR and costimulatory signals, T-cells expansed [111]. CD4+ T-cells 52

53 can differentiate into two different subsets whose functional properties are characterised by the cytokines they secrete. T helper (Th)1 cells produce interferon (IFN)-γ and IL-2, which will result in the activation of CD8+ cytotoxicity, macrophage-dependent delayed-type hypersensitivity, and antibody production by B-cells [110, ]. In addition, Th1 cells express the death inducing molecules: Fas-ligand (FasL) [116, 117]. Th2 cells secrete IL-4, IL-5, IL-9, IL-10 and IL-13, cytokines trigger eosinophil activation [ ]. Th2 cells do not express FasL, thus do not mediate direct cytotoxicity. Cytokines are crucial in Th1 or Th2 expansion during the initial steps of CD4+ T-cell activation. IFN-γ has antiproliferative effect, it downregulates Th2 expansion via interaction with IFN-γ receptors present on Th2 cells. Th1 expansion is not affected by IFN-γ because Th1 cells do not express INF-γ receptors. Activation of CD8+ T cell by donor MHC class I molecules prevents generation of Th2 cells. IFN-γ synthesis and CD8+ T-cell activation skew T-cell differentiation toward Th1. In the absence of IFN-γ and CD8+ activation, alloreactive CD4+ T-cells will give rise to a mixture of Th1 as well as Th2 cells [ ] Isolated MHC Class I Disparity Recognition of intact MHC class I molecules by host CD8+ alloreactive T-cells presented by donor dendritic cells initiates graft rejection (direct recognition pathway) (Figure 1.7.1) [110, ]. Host CD4+ T-cells are activated by recognition of donor MHC class I-derived allopeptides presented in recipient MHC class II molecules (indirect recognition pathway) [110]. Since MHC class II molecules are shared by recipient and donor DC, the donor MHC class I-derived peptides are presented to CD4+ T-cells by DC of both donor and recipient origin. CD8+ cells receive help, via IL-2, from activated alloreactive CD4+ T-cells. In turn, CD8+ T-cells produce IFN-γ that will skew the indirect alloreactive response toward the Th1 type. B-cells bind HLA molecules by its surface immunoglobulins (Ig), and then internalized the HLA-Ig complex. The HLA-Ig complex is cleaved into allopeptides, loaded into B-cell MHC class II molecule groove, and returned to the B-cell surface. B- cells require T-cell help to divide and produce antibodies. Only CD4+ cells belong to the 53

54 subset involved in the indirect recognition of MHC alloantigens are able to recognise the allopeptides presented on B-cells [112, 128, 129]. Once coated by antibodies, graft cells can be killed via activation of complement cascade or by NK cells. Activation of the complement cascade leads to formation of membrane attack-complex that lyses cells [112, 128, 129]. Whereas for NK cells, cross-linking of NK cells Fc receptors to IgG coated cells triggers perforin/granzyme-mediated NK cytotoxicity, a process called antibody-mediated cell cytotoxicity (ADCC) [110]. Activated CD8+ T cells acquire cytotoxic properties in the presence of Th1 cytokines. Cytotoxic CD8+ T-cells kill target cells via perforin/granzyme pathway. During the 48 hours after TCR engagement, activated CD8+ T-cells synthesize perforin and granzymes. When a cytotoxic T-cell recognises the allo-mhc molecule, it forms a tight junction with the allogeneic cell, allowing cytotoxic granules to fuse with the target cell membrane. Perforin molecules insert within the allogeneic cell membrane and form polymers that create channels, through which granzymes A and B penetrate into the cytoplasm. Granzymes in cytoplasm cleave cytoplasic pro-caspases into caspases, whereas granzyme B within in mictocondria release cytochrome C which triggers the caspase system. Caspase activation enhances nuclease activity that finally triggers DNA fragmentation and leads to apoptosis [ ]. CD4+ and CD8+ derived IFN-γ production causes macrophage activation and delayed hypersensitivity reaction (DTH). DTH reactions characterized by tissue swelling, increased vascular permeability and presence of inflammatory infiltrate rich in T-cells, macrophages and neutrophils. Activated Th1 release IFN-γ and tumour necrosis factor, TNF-α, which trigger macrophages to produce toxic molecules such as nitric oxide (NO), oxygen intermediates, and TNF-α. NO is cytotoxic at high concentrations. TNF-α binds to TNF receptors and induce target cell apoptosis or necrosis through caspase cascade. Activated neutrophils release myeloperoxidase, leading to production of toxic metabolites such as oxygen species and H 2 O 2 [130, 131]. 54

55 Figure 1.7.1: Effector mechanisms of rejection in graft carrying mismatched MHC class I antigen. Donor dendritic cells presents intact MHC class I molecules to host CD8+ alloreactive T- cells. Recognition of donor MHC class I-derived allopeptides presented in recipient MHC class II molecules activates CD4+ T-cells. CD8+ T-cells will receive help, basically IL-2, from activated alloreactive CD4+ T-cells. In turn, CD8+T-cells produce IFN-γ that will skew the indirect alloreactive response toward the Th1 type. Alloreactive B cells interact with CD4+ T-cells primed by the indirect pathway to produce alloantibodies. The effector mechanisms of rejection includes: (a) perforin/granzyme-dependent CD8+ T-cell cytotoxicity; (b) CD4+ and CD8+ derived IFN-γ production and delayed-type hypersensitivity, where macrophages release toxic molecules, such as NO, TNF-γ, and oxygen species; (c) antibodies activate complement cascade, resulting cell lysis. Source: Modified from Le Moine A et. al. Transplantation 2002;73: [110]. 55

56 Isolated MHC class II Disparity Graft rejection in allografts bearing mismatched MHC class II antigen is initiated by direct or indirect recognition of MHC-derived peptides by host CD4+ T-cells (Figure 1.7.2) [132, 133]. No CD8+ T-cells activation and IFN-γ production by CD8+ T- cells occurs because the allograft does not display MHC class I alloantigens. Thus, alloreactive CD4+ T-cells will give rise to a mixture of Th1 and Th2 cells [122, 123]. A few hours after T-cell activation, Th1 acquire cytotoxic effect by expressing Fas-ligand (FasL) on cell surfaces. Fas is a type of death receptor in TNF family. Fas is expressed on most cell surface. On the Th1 cell surface, FasL is rapidly cleft by a metalloproteinase and binds Fas on the target cell surface. Interaction of Fas and FasL results in the death-inducing signal complex (DISC) formation and activation of caspase cascade, which will ultimately induce target cell apoptosis similar to that induced by the perforin/granzyme system [116, 117]. Alloreactive Th2 release IL-4, IL-5 and IL-13 which causes eosinophil infiltration within the allograft. IL-4 and IL-13 upregulate the expression of VCSM-1 on endothelial cells to improve adherence of eosinophils that express the counterreceptors, VLA-4, on their membrane. IL-4 and IL-13 also stimulate synthesis of eptaxin, a chemokine that works with IL-5 to activate eosinophils within inflamed tissues. Activated eosinophils release granules that contain harmful enzymes such as eosinophil peroxidise (EPO) that causes tissue destruction [ ]. Alloreactive B-cells interact with CD4+ T-cells primed via indirect pathway to produce alloantibodies. Antibodies react against donor MHC causes antibody-mediated rejection [110, 112, 113, 116, , 123, 125, , 132]. CD4+ T-cells derived IFN-γ production causes macrophage activation and delayed hypersensitivity reaction (DTH) in MHC-CII mismatched graft [110, 130]. 56

57 Figure 1.7.2: Effector mechanism of rejection in graft carrying mismatched HLA class II antigens. CD4+ T-cells recognise allogenic MHC class II molecules via direct or indirect pathways. No CD8+ T-cells activation and IFN-γ production, thus CD4+ T-cells differentiate into Th1 and Th2 cells. The effector mechanisms of rejection include: (a) Th1 type CD4+ T- cell cytotoxicity mediated by Fas/FasL interactions; (b) eosinophil recruitment, activation, and degranulation induced by Th2-derived IL-5 and IL-4 production. Activated eosinophils release granules that contain several harmful enzymes responsible for tissue destruction; (c) alloreactive B cells interact with CD4+ T-cells primed by the indirect pathway to produce alloantibodies. (d) CD4-derived IFN-γ production and delayed-type hypersensitivity, where macrophages release toxic molecules, such as NO and oxygen species. Source: Modified from Le Moine A et. al. Transplantation 2002;73: [110]. 57

58 MHC Class I and II Disparity Graft rejection is initiated by alloantigen recognition by T-cells via direct and indirect pathways (Figure 1.7.3) [110]. CD8+ T-cells that recognise MHC class I molecules receive help, IL-2, from activated CD4+ T-cells. In turn, CD8+ T-cells produce IFN-γ that skew direct and indirect immune response toward the Th-1 type. Activated CD4+ and CD8+ T-cells become cytotoxic via Fas/FasL and perforin/granzyme pathways, respectively. CD4+ and CD8+ T-cells derived IFN-γ production causes macrophage activation and delayed hypersensitivity reaction that induces release of toxin and results in cell death. Alloreactive B-cells interact with CD4+ T-cells primed via indirect pathway to produce alloantibodies. Antibodies react against donor MHC causes antibody-mediated rejection [110, 112, 113, 116, , 123, 125, , 132]. 58

59 Figure 1.7.3: Effector mechanisms of rejection in graft carrying mismatched MHC class I and class II antigens The rejection process is initiated by alloantigen recognition by host alloreactive T-cells. Host CD4+ T-cells recognize intact MHC molecules on donor DC via direct or indirect pathways. CD8+ T-cells that recognize donor MHC class I molecules will receive help, basically IL-2, from activated alloreactive CD4+ T-cells. Activated CD8+ T-cells produce IFN-γ that will skew T-cell differentiation toward Th1 type. Alloreactive B cells interact with CD4+ T-cells primed by the indirect pathway to produce alloantibodies. The effector mechanisms of rejection include: (a) CD4+ and CD8+ T-cell cytotoxicity, (b) CD4+ and CD8+ derived IFN-production and delayed-type hypersensitivity, where macrophages release toxic molecules, such as NO, TNF-γ, and oxygen species; (c) antibodies mediate complement activation resulting cell lysis. Source: Modified from Le Moine A et. al. Transplantation 2002;73: [110]. 59

60 1.8 Pre-transplantation HLA Antibody Detection and Crossmatching Exposure to foreign HLA antigens during pregnancies, blood transfusion or transplantation can induce anti-hla antibody production. In 1969, Terasaki and Patel reported association of hyperacute renal allograft rejection with preformed lymphocytotoxic antibodies [106]. Morris conducted a study to examine presence of anti-hla antibodies after graft failure. Antibodies were found in 11 (38%) of 29 patients who had rejected their grafts [134]. Following this, a substantial number of studies have been conducted to determine the prevalence of antibodies develop before and/or after organ transplantation, and the significance of these antibodies in relation to allograft rejection and graft loss [ ]. Prior to transplantation, patient serum is tested against potential donor lymphocytes to detect preformed cytotoxic antibodies. Pre-transplant crossmatching minimizes risk of hyperacute rejection and early graft loss due to immunological responses. Crossmatches are performed on separated donor T- and B-cells due to differential expression of HLA class I and class II antigens on these cells. Even though a positive crossmatch is considered a contraindication for transplantation, some positive crossmatches are acceptable Negative T-cells / Positive B-cells Crossmatches (TN/BP) Although anti-hla class I antibodies can be detected by both T- and B-cell crossmatches (TXM and BXM), only B-cells can detect anti-hla class II antibodies. However, a positive BXM with a concomitant negative TXM (TN/BP) is not necessarily due to anti-hla class II antibodies. This scenario may happen in the presence of class I antibodies that falls below TXM detectable level. The low titer class I antibodies may be detected by BXM due to a higher expression of HLA class I antigens on B-cells [139]. TP/BN may caused by non-hla antibodies which the respective antigens are expressed by B-cells, or non-specific binding of IgG antibodies to Fc-receptors (Fragment, crystallisable) on B-cells. 60

61 Early studies showed that positive BXM performed with complementdependent lymphocytotoxicity (CDC-BXM) technique correlates with early graft loss [ ]. CDC assay only detects cytotoxic antibodies. Similar findings were reported by other groups using a different crossmatch technique, flow cytometry [143, 144]. Ladza el al. demonstrated a poorer 1-year graft survival among patients transplanted across a strong positive BXM (>50 channel shift). The patients also experienced more rejection episodes (per patient), and they have a greater incidence of irreversible rejection compared to the patients with <50 channel shift [144]. Deleterious effects were not seen in patients that have weak positive BXM. The findings indicate not all antibodies cause rejection. The consequence of transplanted across a positive B-cell crossmatch is dependent on the level of antibody [144]. A few studies found no significant differences in the 1-, 5- and 10-year graft survival rates between the TN/BP and TN/BN groups [145, 146]. In Fergundes s study, only positive BXM caused by IgG antibodies correlated with graft loss. Other irrelevant positive BXM, such as IgM antibodies, do not affect graft outcomes [145]. Inadequate description of antibody specificity in some studies may conceal the significance of positive BXM as a predictor of graft outcome. Clinical significance of a positive B-cell crossmatch with a concomitant negative T cell crossmatch reminds controversial [ , ]. There is little consistency among transplant centres in performing BXM as part of recipient selection protocol. Many transplant centres either do not perform a B-cell crossmatch or consider the results to be irrelevant [7]. A clear definition of antibodies binding to B-cells and analyses of clinical impact of these antibodies according to their specificity and property may resolve the position of BXM in predicting renal transplant outcomes Crossmatch: Current Serum Negative / Historic Serum Positive (CN/HP) Individuals may develop secondary immune responses (also called the memory or anamnestic immune responses) if they have been re-exposed to antigens which they had previously developed antibodies. However, several studies showed that majority of the allograft recipients who lost HLA antibodies and subsequently received 61

62 an organ with the respective antigens did not have graft rejection [ ]. When perform crossmatching, some transplant centres include the historically highest Panel Reactive Antibody level (PRA %) serum (even though the serum was drawn several years ago), in additional to current serum. Whereas other centres only includes one or two sera that are less than 6 months in addition to current serum. Individuals who have been exposed to an antigen for the first time would experience a relatively weak and transient response, which is known as primary immune response. After presentation of antigen to the antigen-specific T- and B-cells, an antibody against the antigen appears in the serum. At this stage, the antibody is predominantly of IgM type, but gradually, IgG-type antibody appears and supercedes the IgM response. After a few more days, the IgG antibody s titer in serum increases logarithmically. The concentration of IgG antibody reaches plateau and then falls to undetectable levels without a further stimulation of the respective antigen. When the antigen is present again, the kinetic pattern of the secondary immune response is markedly different from the primary immune response. Firstly, in the secondary immune response, the lag period before an antibody is detected is shorter. Secondly, IgG-type antibody, not IgM, is predominantly present in the serum in the secondary response. Furthermore, in secondary immune response, IgG antibody more rapidly rises in titer, has a longer plateau and a slower decline and a higher affinity than during the primary immune response [161]. Consequently, in patients who had made antibodies reactive with a given graft, the persistence of memory B-cells may evoke a secondary immune response that results in graft loss, even though the antibodies may not be detectable prior to transplantation. However, studies showed no significant difference in graft survival in recipients transplanted across T-cell crossmatch CN/HP compared to patients whose sera tested negative [ , 162]. These studies did not report specificity (allo- and auto-antibodies), immunologlobulin class (IgG and IgM) and PRA% of the antibodies that were detected in historic sera. In contrast to the above negative reports, a few studies that investigate the specificity and properties of antibodies causing positive reactions in historic sera established the correlations of this crossmatch scenario with transplant outcomes. Taylor et. al. showed that in a small group (n = 9) of recipients who had transplanted across CN/HP with IgG anti-hla antibodies detected, 8 of them lost their grafts within 1 62

63 year post-transplantation [163]. The findings have been confirmed by larger scale studies [140, 164]. Most of the centres avoid transplantation across a CN/HP, especially in high risk patients. Study of CN/HP using newer techniques such as B-cell tetramer that detects HLA-specific memory B-cells may resolves the controversy [165] Crossmatch: DTT-Reducible Positivity Autoantibodies broadly react with the recipient s own cells as well as the cells of most individuals with uncertain specificity. Patients with autoimmune disease, eg systemic lupus erythematosus, Sjogran s Syndrome and those who are having heart or blood pressure medication often produce autoantibodies in their sera [166, 167]. It is important to exclude the presence of autoantibodies by performing a recipient s autocrossmatch (reacting the recipient s serum with his/her own lymphocytes). This prevents false positive interpretation in donor-recipient crossmatches. Most autoantibodies are IgM and their reactivity can be diminished by treating serum with chemicals that break disulfide bonds, such as dithiothreitol (DTT). DTT is capable of splitting the 19S subunits of the IgM into 7S monomers, which causes IgM antibodies fail to activate complement cascade [168]. However, the concentration of DTT used in crossmatching is crucial, where a lower concentration of DTT may not eliminate all IgM reactivity. A higher concentration of DTT might also eliminate part of IgG reactivity [142]. Even though both IgG and IgM antibodies are efficient in activating complement cascade, studies showed that positive crossmatches caused by IgM antibodies usually do not affect graft outcomes [166, 167, ]. However, most of the studies do not distinguish alloreactive and autoreactive IgM antibodies and therefore no separate correlation analysis had been performed. The presence of IgM antibodies is associated with the prominent presence of naïve cytotoxic T-lymphocytes (ctls), which are sensitive to cyclosporine A. In contrast, the presence of IgG antibodies is associated with the primed ctls that are resistant to cyclosporine A [172]. These findings explained the contrasting effects produced by IgG antibodies and IgM antibodies on graft survival, with the latter generally being considered harmless [171]. It is interesting to determine the effect of IgM alloantibodies 63

64 on transplant outcomes, especially among recipients who do not have cyclosporine as part of immunosuppressive treatment. 1.9 Post-transplantation Monitoring Donor-Specific Antibodies The degree of HLA incompatibility between donor and recipient tissues is defined by the number of HLA antigen mismatches. The number of HLA mismatches correlates well with graft outcomes [173, 174]. However analysis of mismatches at different HLA loci suggests that not all HLA loci are equally important. The formation of donor-specific antibodies (DSA) is due to the humoral response of allograft recipient directed against mismatched donor cell antigens. DSA in recipient serum can be detected by reacting the donor T- or B-cells with the recipient serum. Presence of de novo DSA antibodies has been significantly correlated with acute humoral rejection (AHR) [ ]. Supon et. al. reported that, 80% of the patients who developed de novo DSA antibodies experienced either acute or chronic rejection. However, these antibodies do not cause irreversible rejection as the majority of these patients retained functioning grafts 2-7 years later [177]. In Worthington s study, 57% of the 112 patients whose graft failed in his study developed de novo DSA antibodies compared to only 1.6% in the control group with functioning graft [178]. In some graft failures, DSA were only detected after transplant nephretomy. This phenomenon may caused by attachment of DSA onto the graft. DSA are only detected in serum when DSA production is overloaded or after the graft has been removed [107]. The choice of immunosuppressive regimen may affect alloantibodies production. Mycophenolate Mofetil (MMF) and tacrolimus appears to be effective for reducing the incidence of HLA antibodies [179]. The correlation of use of MMF and tacrolimus with development of transplant glomerulopathy has been examined in this thesis, and the results are presented in the Chaper 4. Post-transplantation monitoring for de novo DSA is important because early detection may allow tailored immunosuppressive treatment and prevent graft loss. 64

65 1.10 MHC Class I Chain Related (MIC) Genes: MICA and MICB The role of HLA alloantibodies in solid organ transplantation has been recognized since the 1960s. A number of studies showed that almost all patients who rejected a kidney graft had HLA antibodies [180]. However, rejection may occur in the absence of detectable HLA antibodies. Lee et. al. reported that 20% of 80 patients had graft failure without HLA antibodies detected [181]. In addition, 11% of 64 patients in Worthington s study had graft failure even though they were HLA antibody-negative [182]. These findings suggest that non-hla systems may also contribute to graft rejection. In 1994, two new polymorphic families of MHC class I-related genes were discovered and termed as MHC class I-related chain A (MICA) and MHC class I-related chain B (MICB) genes [183]. MIC genes locate in the MHC class I region of chromosome 6, which is in close proximity to the HLA-B locus [183]. Recently, the HLA nomenclature committee recognized approximately 60 alleles of MICA and 25 alleles of MICB. MICA and MICB genes encode for cell surface glycoproteins, which share limited sequence homologies with HLA class I molecules [Refer Section 1.3.2]. MIC products are expressed on gut epithelial cells, endothelial cells, skin-derived fibroblasts, keratinocytes and monocytes. MIC molecules are absent on the surface of normal CD4+ T-cells, CD8+ T-cells and B-cells, thus the route of exposure to foreign MICA/B antigen remain unclear [184]. Vascular damage is often seen in rejected graft. Endothelial cells in vascularized allografts are the first site that have contact with the recipient s blood stream, and therefore it is likely that this site would be the first to experience the damage by the recipient s immune response [185]. The expression of MIC molecules on the surface of endothelial cells makes these polymorphic molecules possible targets for both humoral and cellular immune responses during graft rejection. MICA-specific antibodies have been found in the eluates from allograft nephrectomy specimen suggests that anti-mica antibodies may cause graft rejection [186]. A better understanding of development, frequency and correlation of MIC antibody with transplant outcomes is useful in making MIC crossmatch a routine test in tissue typing laboratories. 65

66 1.11 HLAMATCHMAKER Amino Acid Triplet Program HLAMatchmaker assesses HLA compatibility at the structural level [73, ]. This computer algorithm performs intralocus and interlocus comparisons of donorrecipient amino acid triplet sequences located in antibody-accessible positions of HLA molecules [73, ]. HLAMatchmaker applies two concepts. Firstly, multiple antigenic epitopes on HLA molecule are defined as a string of amino acid triplets (AAT) that recognized by distinct antibodies. Polymorphic AATs potentially induce humoral rejection if they are located in the alloantibodies accessible position. Secondly, patient cannot produce antibodies against AAT mismatch on foreign HLA molecule if such AAT is present in the same sequence location of the patient s own HLA molecule. HLAMatchmaker identifies HLA antigens that share all their triplets with the patient s own HLA antigens. Therefore, such HLA antigens are considered fully compatible with the patient s HLA at the AAT level [73, ]. HLAMatchmaker is especially useful for identifying compatible donors for highly sensitized patients. Dequesnoy et. al. demonstrated that, in 54 highly sensitized patients whose PRA >85%, the median value of probability of finding a donor (PFD) were increased from 0.009, if only zero antigen mismatches were accepted, to if included 0 AATMM donors; to if included 0 and1 AATMM donors and to if included 0, 1 and 2 AATMM donors. The chance of finding a donor for highly sensitized patients will further increase if antigens mismatched that consistently gave negative reactions in serum screening and their respective AATMMs are accepted for transplant [187]. Through the collaborative serum screening work during the 12 th International Workshop, Dequesnoy et. al. published the estimated relative immunogenicity of AAT based on the ratio of positive and negative correlation frequencies. A positive correlation means a patient produces specific antibodies against a given AATMM. Thus, a high frequency of positive correlation indicates that most patients who exposed to a given AATMM had produced the specific antibodies. Therefore, such AAT is very immunogenic. In turn, a high frequency of negative correlation indicates that most patients did not produce specific antibodies to a given AAT and therefore, such AAT may be less immunogenic. By determining the ratio of positive and negative correlation, 23 AATs were classified as high (ratio >2), 33 AATs as intermediate (ratio <0.3 to <2) 66

67 and 23 AATs as low (ratio <0.3) immunogenic AATs. Acceptance of low immunogenic AATMMs for transplantation provides additional opportunities of identifying donors with acceptable HLA mismatches for highly sensitized patients [188]. The unequal immunogenicity of AATs hypothesis was supported by Lobashevsky et. al., that they reported a significance correlation between the numbers of total or high immunogenic AATMM with flow cytometry crossmatching results [190]. The associations of numbers of AATMM with transplant outcomes and antibodies production were investigated. Dequesnoy et. al. found that HLA-DR matched patients who have zero, one or two mismatches at the triplet level of HLA-A and B loci had comparable graft survival rates with the patients who have zero-hla-a, B antigen mismatches. Poorer graft survival rates were seen in recipients with 3-10 AATMM, and >10 AATMM [189]. However, similar correlation on graft survivals was not found in Laux et. al. s study. Laux et. al. claimed that the significant correlation between the number of AATMM and graft survival rates reported by Dequesnoy et. al. was a defect because the significant correlation was loss after adjusted for the number of antigen mismatched. Thus, the effect described by Duqeusnoy et. al. was indeed mainly caused by the conventional HLA class I antigen mismatched rather than by the AATMM [191]. In general, the more HLA antigen mismatched, the greater the chance a patient produces antibodies. In HLAMatchmaker, it is believed that the higher the number of AAT mismatched, the greater the chance a patient produces antibodies. Dankers et. al. showed that no antibodies were detected in zero AATMM recipients, whereas antibodies were detected in 94% of the 11 or 12 AATMM recipients [192]. HLAMatchmaker increases the chance of finding a well-match donor for highly sensitized patients without extensive serum screening. Therefore it is worth to continue to investigate the reliability and the application of this algorithm, both amino acid triplet and Eplet versions, in local transplant programmes Antibody Detection Techniques Complement-dependent Lymphocytotoxicity Assay Complement-dependent lymphocytotoxicity testing (CDC) was developed in the 1960s. It is widely used to perform HLA typing by serology, including antibody 67

68 screening and panel reactive antibody testing. CDC assay mimics in vivo reaction. In CDC, lymphocytes (T-cells for HLA class I, B-cells for HLA class II) are incubated with serum and then with fresh-frozen rabbit serum as a source of complement. The antigen-antibody reaction that occurs on the lymphocyte membrane leads to the activation of complement, which injures the cell membrane. The injury is visualized microscopically by the uptake of a vital dye, such as acridine orange, by the injured/killed lymphocyte. Positive results are read as a percentage of the injured/killed lymphocytes from all lymphocytes and therefore indicating the presence of anti-hla antibodies. In some laboratories, antihuman globulin (AHG) is added before complement in the CDC test to improve its sensitivity. Compared to other techniques, CDC has advantages for detecting non-hla antibodies specific for some unknown antigens mismatched between the donor-recipient pair. However, CDC is less sensitive compared to newer techniques, such as ELISA, Luminex and flow cytometry in detecting low titer but potentially clinically significant antibodies. CDC only detects antibodies that are cytotoxic or capable to activate complement for mediating cell lysis, such as IgG and IgM [7] Flow Cytometry Flow cytometry was introduced in 1980s. The use of this technique is increasing due to its greater antibody detection sensitivity compared to other detection techniques, such as CDC [ ]. Flow cytometry is a powerful analytical tool because it is capable of simultaneously measuring multiple characteristics or parameters of a cell. The major components of flow cytometry consist of a cell flow chamber, a light beam, a series of mirror and filters and photomultiplier tubes or diode detectors. In brief, when a single cell travels in a fluid stream, it is scanned by a light beam. The light scattered by the cell gives information about the cell s size, granularity and chemical composition. Flow cytometry is used to quantitatively measure the emission of fluorescent light from cells, which indicates the antigen-antibody reactivity occurrence. 68

69 The fluorescence emission is commonly derived from fluorescent dyes that have been conjugated to antibodies that, in turn, have bound to cell membrane molecules. A positive flow cytometry test is defined with channel shifts (the difference between mean or median fluorescence channel of the negative control and the test sample) or relative fluorescence (or fluorescence ratio). If an appropriate secondary fluorescence-labeled antibody is used, flow cytometry can detect non-complement fixing antibodies, such as IgA. However, flow cytometry only detects the binding, but not the cytotoxic action of antibody to its respective antigen. Therefore, in flow cytometry crossmatching, it may detect antibodies that are not pathogenic to donor cells and therefore indirectly reduces the possibility of kidney allocation for a sensitized patient [ ] Solid Phase ELISA In the 1990s, a new HLA antibody detection technique, purified HLA antigenbased enzyme-linked immunoabsorbent assay (HLA-Ag-ELISA) was introduced. In the ELISA technique, test serum is incubated with a panel of HLA antigens which have been pre-coated onto plates. If antigen-antibody reaction occurs, the bound antibodies are detected by peroxidase-conjugated antihuman immunoglobulin antibodies. Absorbance is read using an ELISA reader and the assay results are analyzed by computer [196]. More recently, with the use of single HLA class I and class II antigens, instead of purified antigen mixture, the identification of HLA antibodies has become easier and more specific. PRA% and HLA antibodies detected using ELISA technique often significantly correlate with transplant outcomes [197]. However, the assay is more time consuming and less flexible compared to the Luminex technique that enables different tests to be set-up at once and only needs 30 seconds per sample acquisition Solid Phase Luminex In Luminex, the antigen-antibody reaction takes place on the surface of microsphere sets that have been colour-coded and precoated with antigen. Therefore, each of these microspheres has a distinct colour code and defined antigens precoated 69

70 on its surface. The microspheres are incubated with test serum to allow the binding of antibody, if present, to the precoated antigens. When the microspheres pass through 2 lasers, one of the lasers reports (classification laser) the microsphere s identity according to its colour code, whereas the other laser (reporter laser) reports the occurrence of antigen-antibody reaction [198]. Luminex has provided alternative methods to the CDC assay for HLA antibody detection. The CDC-PRA% was calculated by react the patient serum against a panel of 50 to 100 cells. To define the Luminex PRA%, patient serum is reacts against a set of microbeads coated with different HLA antigens [198]. This new technique has the ability to measure multiple analytes simultaneously in a single-reaction well with high specificity and sensitivity. Luminex can run up to 96 samples in a single run and determine the antibody specificities all on one tray. Therefore it is very time-efficient Method Comparison In Herczyk s study, a total of 66 samples were tested to compare the sensitivity and specificity of the four techniques described above. CDC appeared to have the lowest sensitivity, followed by ELISA and Luminex while flow cytometry having greatest sensitivity [198]. With CDC technique, only 18% of the sera were antibodypositive. The antibody-positive sera percentage increased to 48% with ELISA, to 67% with Luminex and to 73% with flow cytometry. The difference in percentages between Luminex and flow cytometry in detecting positive reaction was only 6%, as compared to the 55% difference between CDC and flow cytometry and 25% difference between ELISA and flow cytometry [198]. Thus, it is be important to conduct antibody screening with several different techniques in high risk group. Due to the lack of highly sensitive and specific techniques, majority of the previous studies do not investigate the specificity and properties of antibodies, such as HLA classes, immunoglobulin types, and non-hla antibodies that cause positive crossmatches and/or correlate with inferior transplant outcomes. Therefore, it is worthwhile to continue to investigate the significance of alloantibodies with new powerful 70

71 tool, such as Luminex in order to identify the pathogenic antibodies that lead to inferior graft outcomes. 71

72 2 Chapter 2 Materials & Methods 72

73 2.1 HLA Antigen Typing Dynabead T-cell Isolation ml of blood was collected into an ACD tube and was spun at 200g for 10 mins. 2. Platelet rich plasma (supernatant) was discarded and reconstituted with chilled phosphate buffered saline (PBS) with 0.6% trisodium citrate (PBS/citrate) drop (50µL) of Dynabeads HLA Class I was added and mixed for 5 mins. 4. Sample was placed on Dynal magnet for 5 mins and supernatant was poured off 5. PBS/citrate was added, placed on the magnet for 2 mins and supernatant was discarded. 6. Wash step. PBS/citrate was added, left for 2 mins and supernatant was discarded. Wash step was repeated 2 times for a total of 3 washes. 7. Reconstituted with 2mL of FCS/RPMI (Foetal Calf Serum and Roswell Park Memorial Institute 1640 Media), incubated for 2 mins, and drained thoroughly. 8. Resuspended in 0.25mL FCS/RPMI. 9. T-cell and dynabead rosettes were checked on haemocytomoter and volume was adjusted accordingly (7-15 rosettes per high field view) Dynabead B-cell Isolation ml of blood was collected into an ACD tube and was spun at 200g for 10 mins. 2. Platelet rich plasma (supernatant) was discarded and reconstituted with chilled PBS/citrate drop (50µL) of Dynabeads HLA Class II was added and mixed for 10 mins. 4. Sample was placed on Dynal magnet for 5 mins and supernatant was poured off 5. PBS/citrate was added, placed on the magnet for 2 mins and supernatant was discarded. 6. Wash step. PBS/citrate was added, left for 2 mins and supernatant was discarded. Wash step was repeated 2 times for a total of 3 washes. 7. Reconstituted with 2mL of FCS/RPMI, incubated for 2 mins, and drained thoroughly. 73

74 8. Resuspended in 0.25mL FCS/RPMI. 9. B-cell and dynabead rosettes were checked on haemocytomoter and volume was adjusted accordingly (7-15 rosettes per high field view) Serology HLA Antigen Typing 1. Wells of 72-well microplate were filled with 5µL paraffin oil and 1µL of antiserum was added to the assigned well. 2. 1µL of T-cells for HLA class I typing or B-cells for HLA class I and II typing (isolated by dynabeads technique) was added and mixed. HLA class I typing was usually performed using a set of 144 (2 trays) or 288 antisera (4 trays). HLA class II typing was usually performed using a set of 72 antisera (1 tray). 3. The tray was incubated at 22ºC, 30 mins for class I typing or 40 mins for class II typing and then 5µL of rabbit serum (source of complement) was added and mixed. 4. Incubated at 22ºC, 35 mins for class I typing or 40 mins for class II typing. 5. 5µL of AO/EB/haemoglobin mixture (acridine orange / ethidium bromide) was added and mixed. 6. The tray was viewed and percentage of cell death in each well was recorded using a 0 to 8 score system DNA Quantitation 1. 2µl of sterile water was added to Nanodrop probe to set blank reference. 2. 2µl of DNA sample was added to Nanodrop probe to measure absorbance. 3. Concentration (A 260 ) and purity (A 260 /A 280 ratio) of DNA sample was recorded. The optimum DNA purity is Probe was cleaned using alcohol after sample measurement Gel Electropheresis g agarose powder and 75mL 0.5X TBE buffer were added and mixed. 2. The mixture was dissolved in a microwave and then was placed in a 56ºC waterbath for 10 mins.. 74

75 3. 3µg of ethidium bromide was added, and the mixture was poured into the gel mould, and then combs were inserted. 4. The gel was left for 40 mins or until it set. 5. The combs were removed, and the gel was placed in electrophoresis tank. 6. The tank was filled with 0.5X TBE buffer Visualization of PCR Product 1. 4µL of gel loading buffer was added to each well of a microtitre tray. 2. 4µL of PCR product or DNA ladder was added and mixed with the loading buffer. 3. 8µL of PCR product and loading buffer mixture were added to the assigned well of gel and left for 40 mins at 200V. 4. The gel was placed on the transilluminator surface and a photo was taken using poraloid camera Sequence Base Typing Prepare PCR Mixture 1. DNA samples were diluted with sterile distilled water to µg/mL. 2. Conexio PCR mastermix, platinum taq polymerase were thawed, vortex for mixing and then quick spun to bring down solution from wall and cap. 3. Reagents (n+1) were added and mixed in a 1.5mL sterile tube as per Table Table 2.1.1: Preparation of PCR Mixture for Sequence Base Typing HLA Genes A B C DRB1 Conexio Mastermix (µl) Taq polymerase (µl) DRB3 /4/5 DQB1 (a, b) DQB1 Exon 3 DPB Aliquot (µl)

76 4. PCR mixture was aliquot to the assigned well. There were 3 DQB1 amplification mastermix: DQBa amplifies DQB1*02, 03 and 04 DQBb amplifies DQB1*05 and 06 DQB Exon 3 resolves DQB1*0301/9 and 0201/ DNA sample (2µL for HLA A, B and C; 3µL for HLA DRB, DQB and DPB) was added to the assigned well and mixed well with the PCR mixture and then placed in thermalcycler. 6. Amplification steps of DNA template was showed in Table Table 2.1.2: PCR conditions for Sequence Base Typing HLA Genes A and C B DRB1 DRB345 Initial Denaturation 96ºC, 6 mins 96ºC, 6 mins 95ºC 15 mins 95ºC 15 mins DQB1 a, b, Exon 3 95ºC 15 mins DPB1 95ºC 15 mins Total Cycles Denaturation 96ºC, 96ºC, 95ºC, 95ºC, 95ºC, 95ºC, 30s 30s 20s 20s 20s 20s Annealing 70ºC, 63ºC, 63ºC, 62ºC, 62ºC, 62ºC, 30s 30s 10s 10s 10s 10s Extension 72ºC, 72ºC, 72ºC, 72ºC, 72ºC, 72ºC, 2 mins 2 mins 90s 90s 90s 90s Repeat Cycles Total Cycles 5 Denaturation 95ºC 20s Annealing 57ºC, 10s Extension 72ºC, 90s Repeat Cycles 4 Final 72ºC, 72ºC, Elongation 10 mins 10 mins HOLD 15ºC 15ºC 15ºC 4ºC 4ºC 4ºC 76

77 PCR Product Visualization 1. A 2% agarose gel was prepared (Refer to Gel Electropheresis) 2. Electopheresis was performed at 200V for 40mins. A DNA ladder 100bp was included for determination of size of PCR product AMPure PCR Purification 1. 10µL of amplified PCR products were transferred to a new plate and 18µL of AMPure (magnetic particles) was added and mixed. 2. The plate was quick spun, incubated at room temperature for 5 mins and then was placed on the SPRIPlate magnet for 5 mins and then supernatant was discarded. 3. Wash step. 200µL of 70% ethanol was added, incubated at room temperature for 30 seconds, and then supernatant was discarded. The wash step was repeated once for a total of 2 washes. 4. The plate was left in vacuum dryer for 4 mins to remove ethanol. 5. Elution step. For Class I, 50µL of sterile distilled water was added to each well. For Class II, 90µL of sterile distilled water was added to each well. 6. The plate was vortex for 30 sec; pulse spun and placed on magnet 77

78 ABI Big Dye Terminator Ready Reaction Kit 1. Conexio sequencing primers were thawed. DRB1 STAMPF (Forward) STAMP R (Reverse) DRB3/4/5, DQB1, DPB1 M13F M13R DQB1 Exon 3 AcF AcR HLA-A Exon 2, Exon 3 5AIN1-46 3AIN3-66 ASEQ 5 ASEQ 3 HLA-B Exon 2, Exon 3 M13F M13R BF1B BIN 2R HLA-C M13F M13R CF1C CIN2R 78

79 2. The Forward and Reverse primer reaction mixes for each locus was prepared (n+1) For each reaction: Big Dye Terminator (dye) 5x dye terminator buffer Sequencing primer Sterile distilled water 0.7µL 3.7µL 2.0µL 11.6µL 3. 18µL of primer reaction mix and 2µL of AMPure cleaned PCR product were added to the assigned well and mixed well. 4. The plate was placed in thermacycler. 5. Sequencing reaction of DNA template was showed in Table HLA Genes Table 2.1.3: Conditions of sequencing amplification A, B, C and DRB1 DRB3/4/5, DQa, b, Exon 3, DPB1 Method Name BDTSEQ DQSEQAmp Initial Denaturation 96ºC, 1 min 96ºC, 1 min Denaturation 96ºC, 10s 96ºC, 10s Annealing 50ºC, 5s 50ºC, 5s Extension 53ºC, 4 mins 60ºC, 4 mins Cycles HOLD 22ºC 22ºC CleanSEQ Dye Terminator Removal 1. 85% ethanol was prepared by adding 23mL of absolute ethanol to 4mL of sterile distilled water (for 100 wells) µL of CleanSEQ and 62µL of 85% ethanol were added and mixed. 3. The plate was placed on magnet for 3 mins and supernatant was discarded. 79

80 4. Wash step. 100µL of 85% ethanol was added, left for 30 seconds and the supernatant was discarded. The wash step was repeated once for a total of 2 washes. 7. The plate was left in vacuum dryer for 4 mins to remove ethanol. 8. Elution step. 40µL of sterile distilled water was added and left in room temperature for 5 mins. 9. The plate was sent into sequencer Luminex Single Stranded Oligonucleotide (SSO) HLA Typing Prepare PCR Mixture 1. DNA samples were diluted with sterile distilled water to µg/mL. 2. LIFECODES PCR mastermix, platinum taq polymerase were thawed, vortex and then quick spun. 3. Reagents (n+1) were added and mixed in a 1.5mL sterile tube as per Table Table 2.1.4: Preparation of PCR mixtures for Luminex SSO Reagents Amount per reaction LIFECODES Mastermix 15µL Taq polymerase (5U/µL) Nuclease free water 0.5µL (2.5U) 32.5µL 4. 50µL of PCR mixture and 2µL of DNA sample were added to the assigned well. 5. Amplification of DNA template was showed in Table

81 Table 2.1.5: Condition of PCR amplification for Luminex SSO PCR DNA Amplification Initial Denaturation All HLA Loci 95ºC, 5 mins Total Cycles 8 Denaturation Annealing Extension 95ºC, 30s 60ºC, 45s 72ºC, 45s Repeat Cycles 7 Total Cycles 32 Denaturation Annealing Extension 95ºC, 30s 63ºC, 45s 72ºC, 45s Repeat Cycles 31 Final Elongation 72ºC, 15 mins HOLD 22ºC Visualization of PCR Product 1. A 2% agarose gel was prepared (Refer to Gel Electropheresis) 2. Electopheresis was performed at 200V for 40mins. A DNA ladder 100bp was included for determination of size of PCR product Hybridisation 1. LIFECODES probe mix was left in a 55ºC - 60ºC heated block for 10 mins and then vortex for 15s. 2. 5µL of PCR product and 15µL of probe mix were added to the assigned well and placed in a thermalcycler. 81

82 3. Hybridisation steps were showed in Table Table 2.1.6: Hybridisation steps for Luminex SSO Steps Initial Denaturation Hybridisation All HLA Loci 95ºC, 5 mins 47ºC, 30 mins 56ºC, 10 mins HOLD 56ºC 4. While the plate was undergoing hybridisation, a 1:200 Dilution Solution/PE- Streptavidin (DS/SAPE) mixture was prepared by adding170µl DS to 0.85µL of SAPE per sample and was kept in dark. 5. After 10 mins at the 56ºC hold, while sitting in thermal cycler, 170µL of DS/SAPE mixture was added to each sample and mixed well. It is critical to add DS/SAPE to all samples within 5 mins. 6. The samples were acquired by Luminex instrument. 2.2 Crossmatch Direct Crossmatch µL of patient s serum was spun at rpm for 10 mins. 2. 5µL of paraffin oil and 1µL of serum were added to wells. 3. A positive control and a negative control were included in each tray. 4. 1µL of Dynabead isolated donor s T-cells were added in HLA class I crossmatch whereas 1µL of Dynabead isolated donor s B-cells were added in HLA class II crossmatch. 5. The tray was incubated in room temperature for 40 mins. 6. 5µL of complement (source: rabbit serum), was added to the well and incubated for 40 mins. 7. 5µL of AO/EB/haemoglobin mixture was added. 82

83 8. Tray was viewed and percentage of cell death in each well was recorded using a 0 to 8 score system. 0 = no cells, 1= no cell death (negative), 2-8 = cell death>10% (positive). 9. An autocrossmatch was performed concurrently. If autocrossmatch is positive, the donor-recipient crossmatch should be repeated with DTT treatment DTT Crossmatch µL of patient s serum was spun at rpm for 10 mins. 2. 5µL of paraffin oil and 1µL of serum were added to wells in duplicate. 3. The same serum was tested: - As neat serum - 1µL of PBS added (diluted serum N/2 in filtered PBS) - 1µL of DTT added (diluted serum N/2 in DTT) 4. A positive IgM control and a positive IgG control were included in each tray. 5. A negative control (AB serum) was placed between each pair of duplicate tests. 6. The plate was incubated at 37ºC for 30 mins and then left in room temperature for 10 mins. 7. 1µL of L-cystine was added to wells contained DTT and PBS. 8. 1µL of Dynabead isolated donor s T-cells were added in HLA class I crossmatch whereas 1µL of Dynabead isolated donor s B-cells were added in HLA class II crossmatch. 9. The tray was incubated in room temperature for 40 mins µL of complement was added and incubated for 40 mins µL of AO/EB/haemoglobin mixture was added. 12. Tray was viewed and percentage of cell death in each well was recorded using a 0 to 8 score system. 0 = no cells, 1= no cell death (negative), 2-8 = cell death>10% (positive). 13. In a valid test, the well containing IgM positive control was positive and well containing IgM positive control and DTT was negative (antibody reactivity removed). Well containing PBS should retain antibody reactivity. If this is not the case, dilution cannot be excluded as a cause of negative results (antibody removal). 83

84 2.3 Anti-HLA Antibodies Screening SeraClean 1. Seraclean beads were vortex thoroughly prior to use. 2. 4uL of Seraclean beads and 20uL of the high background serum were added, mixed well and vortex for 30 seconds. 3. The serum was left on a rotator for 30 mins at room temperature. 4. The serum was spun for 3 mins at 15,000 xg and the treated serum (supernatant) was transferred to clean tube for testing Luminex Screen and PRA 1. The kit was brought to room temperature µL of patient s serum was spun at rpm for 10 mins µL of distilled water was added to wells of a milipore filter plate and left for 3 mins and then was aspirated µL of wash buffer and 12.5µL of serum were added. 5. Luminex beads were vortex for 1 min, 5µL of beads was added, and then incubated at room temperature on rotating platform, 200 rotations per minute, for 30 mins. 6. While the plate was incubating, working solution of PE conjugated anti-hla IgG antibody was prepared by adding 5µL of concentrated conjugate to 45µL of wash buffer for each well µL of wash buffer was added and then aspirated. 8. Wash step. 250µL of wash buffer was added and then aspirated. The wash step was repeated 2X for a total of 3 washes µL of working conjugate was added to the well and the plate was incubated at room temperature on rotating platform, 200 rotations per minute, for 30 mins µL of wash buffer was added to the well and mixed well with the sample. 11. Sample was sent into Luminex machine for acquisition. 84

85 2.3.3 Luminex Single Antigen Bead 1. The kit was brought to room temperature µL of patient s serum was spun at rpm for 10 mins µL of distilled water was added to a well of a milipore filter plate and left for 3 mins and then aspirated. 4. Luminex beads were vortex for 1 min, 40µL of beads and 10µL of serum were added. 5. The plate was incubated at room temperature on rotating platform, 200 rotations per minute, for 30 mins. 6. While the plate was incubating, working solution of PE conjugated anti-hla IgG antibody was prepared by adding 5µL of conjugate to 45µL of wash buffer for each well µL of wash buffer was added and then aspirated. 8. Wash step. 250µL of wash buffer was added and then aspirated. The wash step was repeated 2X for a total of 3 washes µL of working conjugate was added and the plate was incubated at room temperature on rotating platform, 200 rotations per minute, for 30 mins µL of wash buffer was added to the well and mixed well with the sample. 11. Sample was sent into Luminex machine for acquisition Luminex Single Antigen Beads Anti-HLA antibodies IgG subclass 1. The kit was brought to room temperature µL of patient s serum was spun at rpm for 10 mins µL of distilled water was added, left for 3 mins and then aspirated. 4. Luminex beads were vortex for 1 min, 40µL of beads and 10µL of serum were added. 5. The plate was incubated at room temperature on rotating platform, 200 rotations per minute, for 30 mins µL of wash buffer was added and then aspirated. 7. Wash step. 250µL of wash buffer was added and then aspirated. The wash step was repeated 2X for a total of 3 washes. 85

86 8. 50µL of IgG subclass conjugate was added and then incubated at room temperature on rotating platform, 200 rotations per minute, for 30 mins µL of wash buffer was added to the well and mixed well with the sample. 10. Sample was sent into Luminex machine for acquisition. 2.4 Database National Organ Matching System (NOMS) Principle The NOMS system is a windows-based client server computer system. NOMS provides the Tissue Typing laboratories in each state of Australia with facilities to manage all the relevant personal, medical and scientific data for patients awaiting transplant and donors as they become available. Uses Management of patient, donor and physician data Processing of laboratory specimens Production of crossmatch and screening trays for large number of patients Tissue matching using the current algorithms as defined by the Renal Transplant Advisory Committee and state physician committees of patients with organ donors as they become available. Reporting of patient and donor laboratory results and transplant waiting lists to individual patients doctors, dialysis centres, transplant centres and the ANZDATA Transplant Registry. Provision of statistical analyses of transplant waiting lists on request. Efficient replication of data between all organ matching centres in Aus in a timely and inexpensive manner. 86

87 Protocol Find Person 1. Person Search screen was opened by clicking on the icon of face. 2. The patient s centre and transplant organ were selected from the drop down boxes. 3. The name of patient was entered and clicked Retrieve button. (I) Person Tab Description: This tab displays the person (patients, donor and others) information such as name, contact, gender ethnic, ABO blood group, date of birth, living or deceased, date of death, patient hospital etc. (II) Specimen Tab Description: This tab displays a list of specimens received for the person. Each specimen has been allocated with a unique laboratory number. The type of specimen, collection date and amount of specimen (number of backmen and complement tube) were recorded in NOMS for inventory purpose. HLA class I frequency (PRA) was displayed to facilitate the selection of peak serum for crossmatch. The location of specimen was recorded in NOMS. (III) Antibody Tab This tab displayed antibody results for the person. It recorded the specificity of antibodies, frequency (PRA%), HLA class, IgG class, testing method (CDC, ELISA or Luminex ) on selected specimen. A tick in authorised box indicated that a particular antibody specificity has been authorised to be used for exclusion in the matching process. The authorised antibody frequency is also used by the matching algorithm to assign a matching score. (IV) HLA Tab The HLA tab page displayed HLA typing results and the typing method. (V) Score Tab 87

88 For patient, the score tab page shows all the matches the patient has ever had with any donors. For donor, this page shows matching score, rank of the patient on the matching list, number of transplant at the time the organ was matched and state/national allocation algorithm. (VI) Crossmatch tab For patients, this page displayed results of patient autocrossmatch and other crossmatches the patient has ever had with any donors. For donors, this page shows crossmatching outcomes with patients. The details of crossmatch such as tray and well identity, the type of crossmatch (DTT or direct), cell class (T- or B-cells), the source of cell (patient or donor name), the specimen reference, the laboratory performed the crossmatch, peak or current serum and test date. (VII) Transplant tab Official definition of a transplant (as defined by ANZDATA in December 2007) An operation that went so far as the donor came into contact with the recipient. This does not have to include an anastomosis, simply coming out of the esky and touching the patient is enough. For patients, it displayed all the transplants a patient has had. Other transplant information includes: donor name, organ, transplant hospital, date and time of the transplant, transplant outcomes and date and cause of graft failure. 2.5 Statistics Comparisons between two parametric continuous variables were performed using the Student s t-test. When comparing more than two categories, one-way ANOVA test was performed. Fisher s exact test was used in proportion comparisons where the expected numbers were less than 5 in any cell. In cases with expected numbers greater than 5 in all cells, Chi-square test was used for comparison of proportions. Odds ratios (OR) were calculated using standard formulae and tabulated as OR [95% confidence intervals]. A probability (p) value of <0.05 was considered statistical significance. In multivariate analysis, we modelled the probability of rejection using logistic regression 88

89 which produces odds ratios (binary outcome yes/no). Graft loss is a time-to-event outcome, we used a Cox regression model to obtain hazard ratios. 2.6 Reagent Preparation AB Serum 1. Fresh serum. The sample/pack blood was requested from known HLA-AB donors. Frozen stock. AB serum was thawed in a 37ºC waterbath. 2. For heat inactivation, serum was placed in a 56 ºC waterbath for 30 mins. 3. AB serum was spun at 12600g for 10mins at 21ºC and fat deposits were discarded. AOEB/Hb Stain Quenching Agent (stock) gm of bovine haemoglobin (Hb) powder and 1200mL of Dynabead phosphate buffer (with citrate) were added, and then stirred. 2. 8mL of 0.1mol/L sodium azide solution was added and mixed. 3. The mixture was spun at 500g for 45 mins at 10ºC 0.2L AO/EB (stock) 1. 60mg of acridine orange powder and 200mg of ethidium bromide tablets (or equivalent in powder) were added to 4mL of ethanol, and mixed. 2. The dissolved mixture was added to 196mL of phosphate buffered saline. Ethylene Diamine Tetra Acetic Acid (EDTA): 0.27mol/L, 0.5M (stock) g of disodium EDTA powder was added to 1000mL of filtered sterile water. 2. The mixture was left in 37ºC waterbath for approximately 2 hours and stirred until EDTA powder dissolved. 89

90 Working solution (AO/EB Hb stain) 1. 10mL of stock AO/EB solution and 120mL of Hb quenching agent were thawed mL of 0.27mol/L EDTA solution was placed in a 37ºC waterbath until solution heated through. 3. The EDTA solution and 30mL of Dynabead phosphate buffer were filtered. 4. The AO/EB solution, Hb quenching agent, EDTA solution and Dynabead phosphate buffer were mixed well in an amber container. Complement for CDC test 1. Commercial rabbit serum was used as a complement source. Frozen complement stock should not be completely thawed in the waterbath as the heat will deactivate the complement. Thaw until an ice cube remains, approximately 1/5 of the total volume. 2. The complement was spun at g for 10 mins at 4ºC and fat deposits were discarded. Dithiolthreitol (DTT) g (154.3 M.W.) of DTT was added to 10mL of PBS (ph 7.4) and stirred. 2. The stock solution was aliquot into 1mL and stored below 0ºC. 3. For working solution, an 1mL aliquot of DTT stock solution was added to 9mL of PBS (ph7.4) to produce 0.01M DTT solution for use. Dynabead Phosphate Buffer 1. These chemicals were added to 1L of filtered water: NaH 2 PO 4.H 2 O 0.175g Na 2 HPO 4.12H 2 O 1.98g NaCl 8. 10g Na-citrate 6.0g 2. The mixture was placed in a 37ºC waterbath for approximately 2 hours. 3. ph of the reagent was adjusted to 7.4 using an alkaline solution (eg sodium bicarbonate) or an acid (eg HCl) and make up the final volume to 1000mL. 90

91 10mg/mL Ethidium Bromide g of ethidium bromide powder was added to 5mL of sterile distilled water. 2. The mixture was left in dark for 2-3 days with constant agitation until the dye was fully dissolved X dilution to produce working solution. Foetal calf serum (FCS) 1. Frozen stock of commercial FCS was thawed in a 37ºC waterbath. 2. For heat inactivation, FCS was placed in a 56 ºC waterbath for 30 mins. 3. FCS was spun at 12600g for 10mins at 21ºC and fat deposits in the serum were discarded. Gel loading buffer 1. These chemicals were added to 50mL of filtered water: 0.25g bromophenol blue 50ml glycerol 2. The mixture was allowed to mix for 2-3 days with constant agitation until the bromophenol blue is fully dissolved. 3. The solution was filtered through Whatman#1 filters and stored at room temperature. IgM positive control 1. The reagent was tested for reactivity (100% cell death with neat sera) and determined dilution factor of the stock. 2. Stock solution was aliquoted to 1mL vials and stored at 75 ºC. 3. To prepare working solution, 1mL stock vial was diluted with filtered Dynabead phosphate/citrate buffer. 4. Aliquots of working solution were stored at 75 ºC. L-cystine g of L-cystine powder was added to 10mL of filtered water and mixed with a rotator/mixer for >1 hour. 91

92 2. The solution was aliquoted in 100µL. It is important to continually agitate the solution whilst aliquoting to ensure that undissolved L-cystine is distributed evenly. 3. The solution was stored at 75 ºC. Phosphate Buffered Saline (PBS) 1. These chemicals were added to <1L of filtered water NaCl...8g KCl...0.2g Na 2 HPO 4.12H 2 O g KH 2 PO g 2. The mixture was placed in a 37ºC waterbath for approximately 2 hours. 3. ph of the reagent was adjusted to 7.4 using an alkaline solution (eg sodium bicarbonate) or an acid (eg HCl) and make up the final volume to 1L. Physiological Saline 1. 27g of sodium chloride was added to 3L of filtered water and was dissolved by agitation. 2. The solution was stored at room temperature. 4X Physiological Saline g of sodium chloride was added to <500mL of filtered sterilised water. 2. The mixture was placed in a 37ºC waterbath until the salt was fully dissolved. 3. The solution was top up to 500mL with filtered sterilised water. RPMI 1640 Medium 1. RPMI powder was added to <1L of filtered distilled water and dissolved by agitation. 2. 2mL of penicillin/streptomycin was added and mixed well. 4. ph of the reagent was adjusted to 7.4 using an alkaline solution (eg sodium bicarbonate) or an acid (eg HCl) and make up the final volume to 1L. 92

93 Anti-HLA IgG positive control 1. Multispecific sera were pooled, reconstituted with filtered sterile water and mixed well. 2. 1mL of reconstituted sera was diluted 1/10 with 10% FCS/RPMI. 10X TBE Buffer ph These chemicals were added to 800mL of distilled water 107.8g Tris 55g Boric Acid 3.73g EDTA 0.4g NaOH 2. ph of the reagent was adjusted to 8.3 with 10M NaOH and top up to 1L with distilled water. 3. To prepare 0.5X working solution of TBE buffer for electrophesis, 10X TBE buffer stock was diluted 1 in 20 with distilled water. 50mL 10X TBE buffer was added to 950mL of distilled water and mixed well. 93

94 2.7 Equipments 1µL and 5µL syringes Model: Hamilton Co Reno Nevada Microtiter 705 and ºC freezer Model: Sanyo Biomedical Freezer 4ºC refrigerator Model: Thermoline Ltm Pharmaceutical Refrigerator 80ºC freezer Model: Revco Autoclave machine Model: Siltex 250D Automated tray reading system Model: Lambda Scan TM Plus III Bench top centrifuge Bench top centrifuge Model: Sigma Laboratory centrifuge 3-15 Model: Sigma Laboratory centrifuge 4K15 Blood tube magnet DNA Extraction machine Model: Maxwell Promega 16 Electropheresis power pack Model: Pharmacia LKB-GPS 200/400 Electropheresis tank Model: Owl Separation System D3 Gel camera Model: Polaroid MP4 Land Camera Heating block Incubator 37ºC Inverted phase contrast Model: Nikon Eclipse TS1000 microscope Lambda jet Model: Lambda Jet III MT Laminar flow safely cabinet Light microscope Model: Olympus CX41 Luminex 100 XY platform Model: Dynal MP6 Magnetic Particle Concentrator Model: Ratek Instruments Dry Block Heater Model:Contherm Series Five Incubator Model: Gelman Sciences Biohazard Protection CII Model: LifeMatch Fluoroanalyzer, Luminex XMAP Technology Magnetic bed mixer Microwave Multichannels pipette Model: Finnpipette Neubauer haemocytometer Model: Weber England 748 Oiler Model: Greiner Orbital Mixer Model: Ratek Instruments ph meter Rotator Model: Robbin Scientific Sequencer Model: ABI Prism 3130X1 Single channel pipettes (0.2- Model: Gibson 2µL, 2-20µL, µL, µL) Spectrophotometer Model: Nanodrop ND1000 Static discharger Thermal cycler UV platform Model: UVi Tec Vacuum pump Vortex Model: Chiltern MT19 Waterbath Weighing machine Model: Syborn Mode Model: Panasonic Inverter System Inside Model: Shindengen ISFET KS723 Model: Electro Technic Products INC Mode BD 20V Model: Eppendorf Mastercycler Model: Gast 0522-V3-G21DX Model: Ratek Instruments Shaking Incubator Model: Sartoris CP423S 94

95 2.8 Kits AmPure Supplier: Beckman Coulter Agencount Big Dye Terminator Supplier: Applied Biosystem CleanSeq Supplier: Beckman Coulter Agencount DNA purification Supplier: Promega Maxwell 16 Blood DNA Purification kit Dynabead class I Supplier: Invitrogen Dynal Oslo, Norway Dynabead class II Supplier: Invitrogen Dynal Oslo, Norway Luminex antibody detection (screen, ID, SAG) Supplier: Tepnel LifeMatch Luminex SSO kits (HLA-A, -B, -C, -DQB1, -DQA, -DR, -DP) Supplier: Tepnel LifeMatch PCR polymerase Supplier: Invitrogen Sequencing primers (HLA-A, -B, C, -DQB1, -DQA, -DR, -DP) Supplier: Conexio SeraClean Supplier: Tepnel LifeMatch Streptomycin/PE Supplier: Tepnel LifeMatch 2.9 Chemicals 70% ethanol Supplier: Orion Laboratory AB serum (negative Supplier: In house control) Absolute ethanol Supplier: Chem-supply Acridine orange Supplier: Sigma Aldrich Chemical Co Agarose powder Supplier: Omnigel-Low Edwards Unstrument Co Bleach Supplier: Dominant Chemicals Boric acid Supplier: Sigma Aldrich Chemical Co Bovine haemoglobin Supplier: Sigma Aldrich Chemical Co powder Bromophenol blue Supplier: Sigma Aldrich Chemical Co Dithiothreitol (DTT) Supplier: Sigma Aldrich Chemical Co EDTA Supplier: Merck Ethidium bromide Supplier: ICN Chemicals Foetal calf serum Supplier: GIBCO Glycerol Supplier: Univar Hydrochrolic acid (HCl) Supplier: AnalaR IgG anti-hla antibody Supplier: In house positive control IgG subclass Supplier: Tepnel LifeMatch IgM anti-hla antibody Supplier: In house positive control KH 2 PO 4 Supplier: Univar L-cystine Supplier: Sigma Aldrich Chemical Co Light liquid paraffin Supplier: Gold Cross Luminex running buffer Supplier: Tepnel LifeMatch Na 2 HPO 4 12H 2 O Supplier: Sigma Aldrich Chemical Co 95

96 NaH 2 PO 4 H 2 O Supplier: AnalaR Penicillin/streptomycin Supplier: Sigma Aldrich Chemical Co Phenol red Supplier: MP Biomedicals Potassium Chloride (KCl) Supplier: Univar Rabbit serum complement RPMI 1640 medium Supplier: Sigma Aldrich Chemical Co Sodium azide (NaN 3 ) Supplier: Asia Pacific Specialty Chemicals Sodium bicarbonate Supplier: Sigma Aldrich Chemical Co Sodium chloride (NaCl) Supplier: Univar Tris (Hydroxymetyl Supplier: ICN Biomedicals methylamine) Tri-sodium citrate Supplier: Asia Pacific Specialty Chemicals (Na 3 C 6 H 5 O 7 2H 2 O) Supplier: AMSTERDAM or Bill Cannady Serologicals 2.10 Consumables 60 well Terasaki microtiter tray Supplier: Labquip 72 well Terasaki microtiter tray Supplier: Labquip 96 well filter plate Supplier: Millipore 96 well plate (gel loading) Supplier: Cooke Mictrotiter System 96 well plate (SSO) Supplier: Corning Incorporated Costar 96 well sequencing plate Supplier: ABgene 1.5mL centrifuge tube Supplier: Quality Scientific Plastics Beckmen tubes (0.2mL and 0.4mL) Supplier: Quality Scientific Plastics Freezing tube (5mL) Supplier: Greiner Bio-one Photo negative Supplier: Poraloid Pipette tips (0.2-2µL, 2-20µL, µL, µL) Supplier: Pagoda TM Plastic tubes (5mL and 10mL) Supplier: Sarstedt Plastic sealar Supplier: MicroAmp Applied Biosystem 96

97 3 Chapter 3 Clinical Relevance of A positive B-cell Crossmatch on Renal transplantation 97

98 3.1 Introduction Complement-dependent lymphocytotoxicity (CDC) crossmatching (XM) is routinely performed to detect the presence of preformed donor-specific anti-hla antibodies (DSA) in clinical transplantation [106]. Preformed DSA are a significant barrier to renal transplant as they predispose to hyperacute antibody-mediated rejection [106]. Over time, the detection techniques for anti-hla antibodies has evolved from the less sensitive CDC and CDC with anti-human globulin to more sensitive solid phase assays. Solid phase Luminex is a recently developed assay in which purified HLA glycoproteins are bound to a solid matrix (bead) and binding of alloantibody detected using a dual laser flow cytometer. Luminex has greater specificity than CDC in detecting anti-hla antibodies, is not influenced by the presence of non-hla antibodies, autoreactive antibodies or lymphocytotoxic therapeutic antibodies [199, 200]. Generic Luminex assays detect HLA antibodies of all IgG subclasses (IgG 1, IgG 2, IgG 3 and IgG 4 ), whereas CDC detects complement fixing IgG 1 and IgG 3 antibodies only. CDC remains the most frequently employed crossmatch technique in clinical renal transplantation. A positive CDC T-cell crossmatch (TXM) reflects the presence of anti-hla class I antibodies and is a contraindication to transplantation. A positive B-cell crossmatch (BXM) may reflect the presence of anti-hla class II antibodies, or weak anti-hla class I antibodies due to the higher expression of HLA class I antigens on B- cells [139]. Importantly, a negative pre-transplant CDC crossmatch significantly reduces the risk of hyperacute rejection [201]. Nevertheless, early antibody-mediated rejection does occur in the presence of a negative CDC crossmatch indicating that the CDC crossmatch may lack the sensitivity to detect low level anti-hla antibodies [202]. The clinical significance of a positive BXM with a concomitant negative T-cell crossmatch (T B+) is controversial [ , , 203]. Thus investigating antibody specificity in T B+ crossmatches to determine presence of DSA, using more sensitive and specific solid phase techniques may be useful in resolving the place of the BXM in deceased donor kidney transplantation. 98

99 3.2 Objectives This chapter aims to investigate the nature of anti-hla antibodies in positive B- cell crossmatches using solid phase Luminex techniques. Luminex HLA class I and class II PRA and Single Antigen Bead (SAB) tests were performed on sera from patients who had been transplanted across a CDC T B+ crossmatch and transplant outcome studied in relation to presence of DSA. 99

100 3.3 Materials and Methods Study Cohort Four hundred and seventy one (n=471) deceased donor renal allograft recipients with complete crossmatch records and transplanted between 1 st January 1987 and 31 st December 2005 in the single South Australian Renal Transplant Service were studied. Crossmatches were performed on enriched donor T- and B-cells using CDC extended incubation methods by an ASHI accredited protocol [169]. Serum was collected monthly from all patients waiting for kidney transplantation. Donor T-cells and B-cells were isolated by the NIH method until 1995 (n=18, BXM+ group) and subsequently by the immunomagnetic particle technique (Dynabeads, Dynal, Oslo, Norway, n=67, BXM+ group). Crossmatch was performed with a patient s current and peak sera. More than 95% of the patients current sera were collected within 1 month prior to the transplant day. Peak serum was that with the highest class I CDC panel reactive antibodies level (%PRA). Patients were identified as control (T B ) or T B+ based on their CDC crossmatch results HLA Typing All donor-recipient pairs were ABO blood group compatible. Donor and recipient HLA-A, -B and -DR typing was performed by the routine lymphocytotoxicity method as discussed in Chapter 2 Materials and Methods, and by Sequence Based Typing (SBT) when necessary [204] Screening for Autoantibodies Autocrossmatching was performed by the CDC method (patient serum tested against patient own T-cells and B-cells). A positive autocrossmatch was dithiothreitol (DTT) treated to identify IgM autoantibodies. Patient sera with autoantibodies were excluded from the study to avoid misinterpretation of antibody specificity. 100

101 3.3.4 Luminex Recipient sera were assessed for HLA class I and class II allo-antibodies using the Tepnel Lifecodes Luminex PRA assay. Patients with multispecific antibodies had DSA confirmed by SAB. In brief, HLA class I or II antigen coated Luminex beads were incubated with 12.5μL of patient serum in 96-well Millipore multiscreen filter plates, in conjunction with manufacturer s reagents. The mixture was incubated for 30 minutes in the dark, at room temperature, on a rotating platform. The wells were then washed 3 times to remove unbound excess serum using a vacuum system. A phycoerythrin (PE)- conjugated anti-human IgG all (a mixture of IgG 1, IgG 2, IgG 3 and IgG 4 ) was then added to each well and incubated for 30 minutes. After incubation, the plate was ready for data acquisition using Lifematch Fluoroanalyzer. Data acquired by the Luminex software was imported into the Lifematch QuickType TM Analysis software and results were analyzed as per manufacturer recommendations. (PE)-conjugated anti-human IgG subclasses were supplied by from Tepnel Lifecodes. The IgG subclass of anti-hla antibodies was analysed using a modified Luminex assay where (PE)-conjugated antihuman IgG all was replaced by a (PE)-conjugated anti-human IgG 1 and IgG 3 mixture prior to the second incubation Antibody Specificity Luminex data was analyzed by two independent observers. Previous organ transplantation, pregnancy and antibody screening history were considered in defining the specificity of antibodies. Antibody specificity was then compared with donor HLA type to determine if donor-specific antibody was present Clinical Data Clinical data including donor-recipient demographics, number of transplants, HLA class I PRA sensitization level, histocompatibility matching (number of mismatches), rejection episodes, graft loss, post-transplant serum creatinine levels and 101

102 estimated glomerular filtration rates (egfr) were collected from Australia and New Zealand Dialysis and Transplant Registry (ANZDATA Registry) [205]. The ANZDATA Registry has collected details of rejection episodes within 6 months from the date of transplant since 1 April For each episode, the treating physician reported whether the rejection was biopsy proven. For the transplants performed prior to April 1997, the 6 month rejection data was retrieved from biopsy reports and case notes. Rejection episodes were confirmed by biopsy in 94% of cases. The egfr was calculated using the Modification of Diet in Renal Disease (MDRD) equation [206]. Graft loss was defined as loss of graft function (return to dialysis or re-transplant) or death of patient. Patients were followed up until graft loss or 31 December 2006 [205] Statistics Graft survivals were compared using the Kaplan-Meier method. Continuous outcomes were compared using t-tests or analysis of variance (ANOVA) as appropriate. Categorical variables were compared using chi-square or Fisher s exact tests. 102

103 3.4 Results Characteristics of Patients with T B+ crossmatch There were 471 grafts available for analysis, of which 85 (18%) patients were transplanted across T B+ crossmatch on current and/or peak sera. In T B+ patients, 35 (41%) had negative or low T-cells antibodies (PRA <10%), while the remaining 50 (59%) were sensitized (PRA >10%) at the time of transplant. Patients (n=386, 82%) who had negative crossmatches against T- and B-cells (T B ) formed the comparator group. Demographic data of T B and T B+ patients were comparable for most of the parameters (Table 3.4.1). However, the T B+ group had more indigenous Australian (26% vs. 13%, p=0.003) and sensitized (PRA>10%) recipients than the T B group (59% vs. 37%, p=0.001). There were no significant differences in immunosuppression regimen at induction between T B and T B+ patients. Approximately 96% of the patients in both groups received calcineurin inhibitors. More than 50% of them had Mycophenolic acid and prednisolone. IL-2 receptor inhibitors were given to approximately 10% of the patients. No more than 5% of the patients received ATG/OKT3 (Table 3.4.2). 103

104 Table 3.4.1: Descriptive statistics of cohort by the result of B-cell crossmatching T B (n=386) T B+ (n=85) p-value Recipient age (mean [95% CI], years) 46.1 [ ] 44.0 [ ] 0.16 Donor age (mean [95% CI], years) 38.6 [ ] 36.4 [ ] 0.29 Male gender 246 (64%) 50 (59%) 0.40 Indigenous 51 (13%) 22 (26%) Diabetes as cause of end-stage renal disease 39 (10%) 13 (15%) 0.17 >1 Transplant 52 (13%) 11 (13%) 0.90 HLA Mismatch 0-2MM 86 (22%) 0-2MM 10 (12%) MM 300 (78%) 3-6MM 75 (88%) Cold ischemia time >18 hours 134 (35%) 26 (31%) 0.34 Peak PRA >10% (T-cell) 140 (37%) 50 (59%)

105 Table 3.4.2: Immunosuppresion in T B and T B+ groups Immunosuppresion (at induction) T B (n=386) T B+ (n=85) p-value Calcineurin inhibitors CsA 318 (83%) CsA 67 (80%) 0.65 Tacrolimus 49 (13%) Tacrolimus 14 (16%) None 15 (4%) None 3 (4%) Anti-metabolites Azathioprine 131 (34%) Azathioprine 27 (32%) 0.92 Mycophenolic Acid 205 (54%) Mycophenolic Acid 46 (55%) None 46 (12%) None 11 (13%) Steroid Prednisolone 241 (64%) Prednisolone 56 (67%) 0.70 None 136 (36%) None 28 (33%) IL-2 receptor antagonist 41 (11%) 10 (12%) 0.76 None 345 (89%) None 75 (88%) ATG/OKT3) 18 (5%) 2 (2%) (95%) 83 (98%) 105

106 3.4.2 Luminex Specificity of Anti B-cell Antibodies A total of 154 peak and/or current sera from 83 patients were screened for IgG HLA antibodies using Luminex techniques. Sera were not available for 2 patients. Nine (11%) patients had peak serum only, 3 (4%) had current serum only and 71 (84%) had both peak and current sera screened. Twenty seven patients (33%) had identified DSA. Fourteen patients had DSA directed against mismatched donor class I antigen(s) only (DSA-CI group). Six patients had DSA directed against mismatched donor class II antigen(s) only (DSA-CII group). Seven patients had DSA directed against mismatched donor class I antigen(s) and class II antigen(s) (DSA-CI&CII group). Of the 27 patients with DSA, 18/21 (86%) were tested and confirmed as complement fixing IgG1 and/or IgG3 subclass antibodies. Anti-A2 was the most common DSA, reflecting the high frequency of this antigen in the Caucasian population. A further 21 (25%) patients had non-dsa anti-hla and 5 (6%) had weakly positive Luminex reactions with no defined antibody specificity. Five patients were excluded from analysis due to weak borderline screen results. Importantly, 30 (36%) T B+ patients had no IgG HLA antibodies using Luminex CI and CII PRA tests Early Graft Rejection The overall rejection rate in all patients was 37%, with 175 of 471 subjects experienced at least one rejection episode in the first 6 months post-transplant. Biopsyproven episodes of cellular, glomerular and vascular rejection were reported. Collection of humoral rejection data began in 1 st January 2005, and is not included in this study. The overall prevalence of cellular, vascular and glomerular rejection in all grafts was 28% (134/471), 22% (102/471) and 9.8% (46/471), respectively. 106

107 Correlation of T B+ and Early Graft Rejection The rejection rates in T B group was 36% (139/386) and T B+ group was 42% (42/85), (p=0.27). T B+ prior to 1995 (cell isolation by the NIH method, 7/18, 39%) had similar rejection rate with T B+ post 1995 (cell isolation by Dynabead, 29/67, 43%, p=0.74). No significant differences were seen in cellular (28%, 109/386 vs. 29%, 25/85, p=0.83) and glomerular (9%, 34/386 vs. 14%, 12/85, p=0.14) rejection when comparing the T B and T B+ groups. The T B+ group had a higher prevalence of vascular rejection compared to the T B group (32% vs. 19%, Odds ratio [95% C.I.] 1.9 [ ], p=0.01), (Table 3.4.3). 107

108 Table 3.4.3: Prevalence of cellular, vascular and glomerular rejection in All grafts, T B and T B+ groups All grafts T B T B+ Odds Ratio* p-value Rejection All 37% (175/471) 36% (139/386) 42% (42/85) 1.3 [ ] 0.27 Cellular 28% (134/471) 28% (109/386) 29% (25/85) 1.1 [ ] 0.83 Vascular 22% (102/471) 19% (75/386) 32% (27/85) 1.9 [ ] 0.01 Glomerular 9.8% (46/471) 9% (34/386) 14% (12/85) 1.7 [ ] 0.14 *Compared T B+ against T B. 108

109 Correlation of Early Graft Rejection and Antibody Specificity in T B+ The T B+ patients were stratified into 3 subgroups according to the antibody specificity defined by Luminex : Five patients were excluded from analysis due to weak borderline screen results. The 3 subgroups are: (i) Patients without IgG HLA antibodies (LxN, n=30) (ii) Patients with IgG HLA antibodies that were not donor specific (Non-DSA, n=21) (iii) Patients with donor-specific anti-hla antibodies (DSA, n=27) The rejection rates of the LxN group (33%, 10/30) and the Non-DSA group (38%, 8/21) were not significantly different from the T B group (36%, 139/386). Patients with DSA had significantly higher risk of graft rejection compared to the T B group (41% vs. 36%, 2.2 [ ], p=0.047). DSA was a significant predictor for vascular (p=0.01) or glomerular (p<0.001), but not cellular (p=0.88) rejection. Risk of vascular rejection in patients with DSA was 2.9 times as high as in control group. Patients with DSA also 6 times higher risk of glomerular rejection compared to controls. In the absence of DSA, T B+ (LxN and Non- DSA groups) was not predictive for any type of rejections. Patients with DSA were categorised according to HLA classes: Patients with DSA-CI only, DSA-CII only or both DSA-CI and DSA-CII. Presence of DSA-CII were significant associated with vascular and glomerular rejection (p=0.016 and p=0.005). Rejection risks for T B+ subgroups compared to T B group are shown in Table 3.4.4, Table 3.4.5, Table and Table

110 Table 3.4.4: Graft rejection at 6 months according to the antibody specificity in T B+, compared with T B Rates Odds ratio [95% C.I.] p-value T B 36% (139/386) T B+ LxN group (Non IgG HLA antibodies) 36% (10/30) 0.9 [ ] 0.77 Non-DSA 33% (8/21) 1.1 [ ] 0.85 DSA 56% (15/27) 2.2 [ ] DSA-CI 50% (7/14) 1.8 [ ] 0.29 DSA-CII 67% (4/6) 3.6 [ ] 0.15 DSA-CI & CII 57% (4/7) 2.4 [ ]

111 Table 3.4.5: Risk of early cellular rejection according to the antibody specificity in T B+ Rates Odds ratio [95% C.I.] p-value T B 28% (109/386) T B+ LxN group (Non IgG HLA antibodies) 30% (9/30) 1.1 [ ] 0.84 Non-DSA 24% (5/21) 0.8 [ ] 0.66 DSA 30% (8/27) 1.1 [ ] 0.88 DSA-CI 29% (4/14) 1.0 [ ] 0.98 DSA-CII 17% (1/6) 0.5 [ ] 0.54 DSA-CI & CII 43% (3/7) 1.9 [ ]

112 Table 3.4.6: Risk of early vascular rejection according to the antibody specificity in T B+ Rates Odds ratio [95% C.I.] p-value T B 19% (75/386) T B+ LxN group (Non IgG HLA antibodies) 27% (8/30) 1.5 [ ] 0.34 Non-DSA 24% (5/21) 1.3 [ ] 0.62 DSA 41% (11/27) 2.9 [ ] 0.01 DSA-CI 36% (5/14) 2.3 [ ] 0.15 DSA-CII 67% (4/6) 8.3 [ ] DSA-CI & CII 29% (2/7) 1.7 [ ]

113 Table 3.4.7: Risk of glomerular rejection according to the antibody specificity in T B+ Rates Odds ratio [95% C.I.] p-value T B 9% (34/386) T B+ LxN group (Non IgG HLA antibodies) 3% (1/30) 0.4 [ ] 0.32 Non-DSA 5% (1/21) 0.5 [ ] 0.53 DSA 37% (10/27) 6.1 [ ] <0.001 DSA-CI 21% (3/14) 2.8 [ ] 0.12 DSA-CII 50% (3/6) 10.4 [ ] DSA-CI & CII 57% (4/7) 13.8 [ ]

114 3.4.4 Graft Function at 6, 12 and 36 month Post-transplantation The egfr of the T B group was 52.2mL/min/1.73m 2 at 6 months, 50.0mL/min/1.73m 2 at 12 months and 50.6 ml/min/1.73m 2 at 36 months (Table 3.4.8). No significant difference was found comparing the egfr of the T B group against the T B+ group at any time points (p=0.57 at 6 months, p=0.85 at 12 months and p=0.27 at 36 months). The egfr of the LxN group was also not significantly different from the T B group (p=0.11 at 6 months, p=0.15 at 12 months and p=0.55 at 36 months). The egfr of the DSA (p=0.84 at 6 months, p=0.66 at 12 months and p=0.06 at 36 months) and Non-DSA (p=0.42 at 6 months, p=0.77 at 12 months and p=0.19 at 36 months) groups were compared to the T B group with no statistically significant difference identified. 114

115 Table 3.4.8: egfr at 6, 12 and 36 months post-transplant by B-cell crossmatch and antibody specificity, compared with T B egfr (ml/min/1.73m 3, mean [95% C.I.]) p-value 6 Months T B 50.9 [ ] All T B [ ] 0.57 LxN (Non IgG HLA antibodies) 56.6 [ ] 0.11 Non-DSA 51.8 [ ] 0.84 DSA 47.7 [ ] Months T B 50.0 [ ] All T B [ ] 0.85 LxN (Non IgG HLA antibodies) 55.1 [ ] 0.15 Non-DSA 48.1 [ ] 0.66 DSA 48.8 [ ] Months T B 50.6 [ ] All T B [ ] 0.27 LxN (Non IgG HLA antibodies) 53.2 [ ] 0.55 Non-DSA 41.3 [ ] 0.06 DSA 44.3 [ ]

116 3.4.5 Graft Survival The overall graft survival rates of all subjects were 88% at 1 year, 80% at 3 years and 70% at 5 years. There were 132 graft losses in 5 years, of which 58 occurred in the first year and 74 occurred in 1-5 years. The risk of graft loss in various patient groups was assessed at 1 year and 5 year post-transplantation by Kaplan-Meier analysis Significance of T B+ in Graft Survival The graft survival rates of the T B group were 88% and 70% at 1 and 5 years, respectively. T B+ group has comparable graft survival rates of 88% at 1 year and 70% at 5 years. Thus, as a group, a T B+ did not confer a higher risk of graft loss (0-1 year hazard ratio [95% C.I.] was 0.8 [ ], p=0.66 and 1-5 years was 1.3 [ ], p=0.40) (Table 3.4.9, Figure 3.4.1) Correlation of Antibody Specificity in T B+ with Graft Survival Patients with T B+ were subgrouped according antibody specificity. There were 12 graft losses in 27 patients who have DSA (44%). The graft survival rates of the DSA group were 80% at 1 year and 58% at 5 years post-transplant. The hazard ratio [95% C.I] of graft loss in the presence of DSA relative to the T B group was 1.8 [ ] at 5 years, p= The risk of graft loss in patients with DSA was the greatest during the first year, in which the 0-1 year H.R [95% C.I] was 2.3 [ ], (p=0.042) compared to 1.5 [ ], (p=0.42) for 1-5 years (Table 3.4.9, Figure 3.4.2). Interestingly, IgG 1 and IgG 3 DSA were found in 11 of 15 grafts that were functioning 5 years post-transplant. The 1 year and 5 years graft survival rates of the DSA-CI group were 71% and 49%. The risk of graft loss at 5 year when only DSA-CI present was 2.3 [ ], (p=0.036). In this study, 2 of 6 patients in DSA-CII group and 3 of 7 patients in DSA- 116

117 CI&CII group lost the graft during the 5 year follow-up period. Table shows graft survival rates and graft loss hazard ratios according to antibody specificity in T B+ patients compared to T B patients. Figure shows Kaplan-Meier graft survival of DSA according to HLA classes up to 5 years post-transplantation. Graft survival rates in LxN group were 93% at 1 year and 72% at 5 years. This patient group has comparable graft survival with the control group (1 year p=0.36 and 5 years p=0.84). Presence of third party antibodies (Non-DSA) did not conferred a higher risk of graft loss (1 year 0.4 [ ], p=0.33 and 5 years 0.7 [ ], p=0.41) (Table 3.4.9, Figure and Figure 3.4.3). Risk of graft loss and Kaplan-Meier graft survival curves of various T B+ subgroups were showed in Table and Figure 3.4.1, Figure and Figure

118 Table 3.4.9: Graft survival rates (number of graft failures/total grafts) and hazard ratios for graft loss by antibody specificity Rates Hazard Ratio p-value 1 year T B 88% ALL T B+ 88% 0.9 [ ] 0.87 LxN (Non IgG HLA antibodies) 93% 0.5 [ ] 0.36 Non-DSA 95% 0.4 [ ] 0.33 DSA 74% 2.3 [ ] DSA-CI 71% 2.5 [ ] DSA-CII 83% 1.5 [ ] 0.69 DSA-CI & CII 71% 2.5 [ ] years T B 70% ALL T B+ 68% 1.1 [ ] 0.61 LxN (Non IgG HLA antibodies) 72% 0.9 [ ] 0.84 Non-DSA 78% 0.7 [ ] 0.41 DSA 55% 1.4 [ ] DSA-CI 49% 2.3 [ ] DSA-CII 67% 1.1 [ ] 0.85 DSA-CI & CII 57% 1.8 [ ]

119 Graft survival Years post-transplant Control T-B+ Figure 3.4.1: T B+ as a group was not associated with inferior graft survival 119

120 Graft survival Years post-transplant Control Non-DSA LxN DSA Figure 3.4.2: Graft survival according to antibody specificity in T B+ Presence of DSA in T B+ was associated with 5 years graft loss (1.8 [ ], p=0.045). Non-DSA and LxN groups (no HLA antibodies) did not correlate with graft loss. 120

121 Graft survival Years post-transplant Control DSA CII only DSA CI only DSA CI & CII Figure 3.4.3: Significance of DSA class in graft survival The risk of graft loss at 5 years when only DSA-CI present was 2.3 [ ], p= Presence of DSA-CII (p=0.85) or DSA-CI&CII (p=0.32) in T B+ were not statistically significantly correlated with graft loss. However, the clinical significance of DSA-CII and DSA-CI&CII remain inconclusive due to the small number of patients in this study Historic Positive B-cell Crossmatch Of 85 BXM positive patients, 23 (27%) recipients had a positive BXM with peak serum only, 20 (24%) with current serum only and 42 (49%) with both peak and current sera. Vascular rejection rate was significantly higher in patients with positive BXM in peak serum only. However, this issue should be re-assessed by studies with larger patient numbers. Total, cellular and glomerular rejection rates of various BXM groups were not significantly different from T B (Table and Table ). A historically positive BXM was not associated with poorer graft function (Table ) or graft loss (Table and Figure 3.4.4). 121

122 Table : Prevalence of cellular, vascular and glomerular rejection in T B+ subgroups stratified by peak/current status T B T B+ T B+ T B+ Current serum Peak serum Current and peak sera Rejection All 36% (139/386) 40% (8/20) 52% (12/23) 38% (16/42) Cellular 28% (109/386) 30% (6/20) 30% (7/23) 29% (12/42) Vascular 19% (75/386) 26% (5/20) 39% (9/23) 31% (13/42) Glomerular 9% (34/386) 10% (2/20) 17% (4/23) 14% (6/42) 122

123 Table : Odds ratios of graft rejection at 6 months in T B+ subgroups stratified by peak/current status, compared with T B Odds ratio [95% C.I.] p-value All rejection T B+ Current serum 1.3 [ ] 0.59 Peak serum 1.9 [ ] 0.12 Current and peak sera 1.1 [ ] 0.79 Cellular rejection T B+ Current serum 1.1 [ ] 0.75 Peak serum 1.1 [ ] 0.82 Current and peak sera 1.0 [ ] 0.96 Vascular rejection T B+ Current serum 1.5 [ ] 0.46 Peak serum 2.7 [ ] 0.03 Current and peak sera 1.8 [ ] 0.08 Glomerular rejection T B+ Current serum 1.2 [ ] 0.80 Peak serum 2.2 [ ] 0.18 Current and peak sera 1.7 [ ]

124 Table : egfr at 6, 12 and 36 months post-transplant in T B+ subgroups stratified by peak/current status, compared with T B egfr (ml/min/1.73m 3, mean [95% C.I.]) Analysis of Variance Between groups p-value 6 Months T B 50.9 [ ] 0.82 T B+ Current serum 49.7 [ ] Peak serum 54.3 [ ] Current and peak sera 52.3 [ ] 12 Months T B 50.0 [ ] 0.62 T B+ Current serum 49.2 [ ] Peak serum 55.0 [ ] Current and peak sera 48.5 [ ] 36 Months T B 50.6 [ ] 0.54 T B+ Current serum 49.1 [ ] Peak serum 51.3 [ ] Current and peak sera 45.3 [ ] 124

125 Table : Graft survival rates (number of graft failures/total grafts) and hazard ratios for graft loss in T B+ subgroups stratified by peak/current status, compared with T B Rates Hazard Ratio p-value 1 year T B 88% T B+ Current serum 84% 1.3 [ ] 0.64 Peak serum 87% 1.1 [ ] 0.91 Current and peak sera 90% 0.7 [ ] years T B 70% T B+ Current serum 73% 0.7 [ ] 0.59 Peak serum 69% 1.2 [ ] 0.71 Current and peak sera 64% 1.6 [ ]

126 1.00 p 0.90 Graft survival Years post-transplant Control Peak BXM pos only Curr BXM pos only Both BXM pos Figure 3.4.4: 5 years graft survival in T B+ subgroups stratified by peak/current status T B+ in peak serum, current serum or both were not associated with inferior graft survival. 126

127 3.4.7 BXM as A Predictor of Graft Failure in 5 years Post-transplantation A total of 343 grafts were followed up for at least 5 years (Table ). There were 24 graft losses among the 70 BXM positive patients. Of 273 patients transplanted with a negative BXM, 188 had graft functioning for at least 5 years. The positive predictive value of a BXM for graft failure was 34% (24/70). The negative predictive value of BXM was 69% (188/273). The sensitivity and specificity of BXM were 22% (24/109) and 80% (188/234). Table : Graft outcomes at 5 years post-transplant Graft outcomes (at 5 years) T B T B+ Total Functioning Failure Total

128 3.5 Discussion In this study, the HLA class I and class II antibody profile and clinical outcomes in a large group of patients with a positive B lymphocyte crossmatch (BXM) from a single transplant center were examined. This is the first study to dissect the significance of BXM using Luminex technology. The major finding of this study was that only 33% of the positive BXM were associated with the presence of donor-specific antibodies. BXM was originally introduced as a tissue typing technique to detect the presence of either low level anti-hla class I antibodies or class II antibodies [139, 201]. The BXM is a widely employed tissue typing technique but it lacks specificity [199]. Since 1987, several studies have shown that a positive BXM is associated with acute rejection and graft loss and thus it is a significant risk factor for poor graft outcome [ , 149, 151]. Recent Collaborative Transplant Study (CTS) report showed that positive CDC B-cell crossmatches in the absence of T-cell reactivity were associated with a significant reduction of graft survival in both first and regraft populations [207]. However, none of these studies addressed the issue of antibody specificity in positive B- cell crossmatches [ , 149, 151, 207]. This analysis yielded the finding that positive B-cell crossmatch as a group was not associated with inferior graft survival, and graft survival was not different in patients with and without BXM positivity on current serum (Refer Section 3.4.6). Cell isolation by the NIH method may produce additional positive BXM, not due to HLA antibodies, but did not cause differences in rejection rate. Differences in immunosuppression did not account for these graft survival differences (Table 3.4.2). Nevertheless, the prevalence of vascular rejection was significantly higher in T B+ compared to T B group (Table and Table 3.4.4). When BXM positive patients were examined by Luminex, DSA were found in the sera of only one third of patients which was associated with two fold higher risk for rejection and graft loss (Table and Table 3.4.9). Patients without anti-hla antibodies or those with non-dsa antibodies did not have significantly worse long term graft outcomes suggesting that these patients should not be denied transplantation on the basis of BXM alone. These findings confirmed earlier studies that DSA was a significant risk factor for graft failure [176, 178, 208]. 128

129 More than half of the DSA in T B+ were directed against class I antigens, the presence of which caused a statistically significantly higher rate of graft loss (Table 3.4.9). These findings are consistent with the results reported by Karuppan et. al. [168], where graft survival in patients with DSA-CI only was significantly poorer than patients with class II anti-hla DR antibodies detected in B-cell crossmatches (1 year graft survival of 50% and 75%, respectively, p<0.02) [168]. However, the clinical significance of DSA-CII and DSA-CI&CII remain inconclusive due to the small number of patients in this study. Analysis of larger cohorts would clarify this question. Le Bas-Bernardet et. al.. found DSA in 23% of positive BXM, where T B+ patients with DSA had lower early graft survival and a higher incidence of vascular rejection. Graft survival was comparable in T B+ patients with autoantibodies, T B+ patients without DSA and BXM patients [148]. Positive B-cell crossmatches fall into different groups according to the properties and specificity of the antibodies, with subsequent different consequences on kidney allograft outcome. Autoantibodies may be responsible for some T B+, but it is well documented that their presence does not cause poorer outcomes [167, , 209]. In this study, patients with autoantibodies (positive autocrossmatches) were excluded, thus all positive BXM were potentially only due to alloantibody reactivity. This study showed that one third of the positive BXM sera were negative for IgG HLA antibodies in Luminex CI and CII PRA tests. BXM positivity in these patients could be false positive or due to IgM HLA antibodies, non HLA antibodies, cross-reactive antibodies or low frequency HLA antibodies that were not represented on Luminex beads. One potential weakness of the present study is that this Luminex PRA test does not detect DQα and DP specificities which have recently been suggested to have clinical significance [210]. Studies that examine donor-reactive, IgM HLA alloantibodies are rare because of the low prevalence of these antibodies. Most of studies refute the clinical significance of IgM HLA antibodies [ , 209]. The study defined indigenous background as a risk factor for positive BXM. Indigenous patients wait longer on dialysis, are more sensitized at the time of transplantation and the number of HLA mismatches are greater when allocated deceased donor kidneys [211]. Indigenous Australians are more susceptible to infection and have a higher risk of rejection post-transplantation [211]. 129

130 Several studies have been undertaken to detect non-hla antibodies, such as anti-endothelial and MIC-A/-B antibodies that may be potentially associated with graft rejection. Studies showed that anti-mica antibodies are found in up to 20% of kidney transplant recipient [ ]. BXM does not detect anti-mica/b antibodies which are expressed on endothelial cells and not lymphocytes. Non-HLA antibodies in positive B- cell crossmatches are not investigated in this study. Collectively and importantly, however, this study showed that a positive BXM, in the absence of IgG HLA antibodies, has no detrimental influence on graft survival (Table 3.4.9, Figure 3.4.2). If BXM in this group is due to non-hla antibodies, they did not confer poorer graft survival. It is interesting that despite the presence of pre-existing DSA antibodies of complement fixing IgG 1 and IgG 3 subclass, some grafts failed and others survived for more than 5 years. Antibody-mediated rejection is a pathologic process influenced by several factors: a) the level of DSA in circulation, b) the capability of graft tissue to repair the damage caused by DSA and c) the efficacy of immunosuppressive and anticoagulation therapy in preventing antibody production and coagulopathy. When the damage caused by antibody is reversible by tissue repair and therapy, the graft may continue functioning. In turn, if the damage caused by antibody is irreversible by tissue repair and therapy, the graft may fail [196]. Lazda et. al.. reported that a strongly positive flow cytometry BXM (>50 channel shift) was significantly associated with poorer graft survival, but a positive flow cytometry BXM with channel shift was not [144]. Thus, defining the level of DSA predictive for graft rejection in a reproducible assay such as Luminex may be useful for identifying high risk patients pre-transplant and for monitoring renal transplant patients post-transplant. Such studies should be performed to determine a threshold for significant antibody level. Another possible explanation for graft survival in the presence of DSA is graft accommodation, a condition in which an organ transplant functions normally despite the presence of anti-donor antibodies [215]. The mechanism of accommodation is unclear. Studies suggested that initial exposure to low levels of DSA may initiate a series of protective changes such as down-regulation of antigen on graft epithelial cells or upregulation of anti-apoptotic and anti-inflammatory genes (eg. Bcl-xL, Bcl-2) which enables the graft to resist acute pathological effects of DSA and complement fixation [ ]. 130

131 This study has shown that antibodies responsible for positive B-cell crossmatches are heterogeneous. Only one third of positive B-cell crossmatches was caused by DSA and was associated with poorer transplant outcomes. Thus, the use of BXM to preclude kidney transplantation may potentially disadvantage >60% of patients in whom BXM is not indicative of the presence of DSA. Investigation of antibody specificity in positive B-cell crossmatches using solid phase techniques such as Luminex enhances the interpretation of BXM. In this study, the correlation of strength of BXM and transplant outcomes was not included in the analysis. Luminex antibody testing has not been performed on the large control group (n=386) due to cost limitation. However, the HLA antibody specificity in patients transplanted with CDC T B has been characterized by Luminex in a smaller control group (n=25) in the third study presents in the Chapter

132 4 Chapter 4 Anti-HLA Antibodies and Risk Factors Associated with Transplant Glomerulopathy 132

133 4.1 Introduction Transplant glomerulopathy (TG) is a pathological condition associated with chronic kidney graft dysfunction and poor graft survival [219, 220]. TG is usually diagnosed by biopsies that show glomerular basement membrane duplication, and often accompanied by multilayering of peritubular capillary basement membrane and increased mesangial matrix [221, 222]. Other features include endothelial cell thickening and widening of subendothelial space due to accumulation of electron lucent materials [222]. The prevalence of TG is estimated to range from 3% to 15% in general renal transplant population. The Banff 97 working group defined stages (1-3) of TG according to the extent of double contours, created by mesangial interposition, in the most severely affected glomerulus. Cg0 means no TG, double contours present in less than 10% of peripheral capillary loops in the most severely affected glomerulus. Stage Cg1, 2 and 3 indicate the percentage of double contours increased to <25%, 26-50% and >50%, respectively. TG was first described 4 decades ago, but the pathogenesis remained unclear [223]. Recently, studies showed that presence of anti-hla is a significant risk factor of TG development [224, 225]. In 582 patients who had surveillance biopsies done within 1 year, anti-hla antibodies were found in 76% of the patients who developed TG. The risk of TG was higher in patients with anti-hla class II antibodies (Hazard ratio [95% C.I.] = 4.7 [ ] [224]. The risk of TG was further increased in patients with DSA (9.8 [ ]. Presence of anti-hla class I antibodies alone was not a risk factor of TG [224]. There are significantly more sensitized patients develop TG suggest that antibodies may play a role in TG development. However, in some studies, complement activation residues C4d were detected in minority of TG biopsies only [221, ]. This suggests that other factors may contribute to development of TG. There is a lack of clear definition of TG in the literature. TG is thought to be a renal lesion caused by chronic humoral rejection. The roles and specificity of HLA antibodies associated with development of TG are unclear. Characterization of HLA antibodies by newer techniques such as Luminex is useful to identify patients who at high risk of TG. 133

134 4.2 Objectives In this retrospective study, we investigated: I) Prevalence of TG in a large group of patients followed up in a single South Australian transplant centre II) Risk factors associated with TG development such as anti-hla antibodies III) HLA antibody specificity as a prognostic factor of TG IV) Relevance of development of anti-hla antibodies post-transplant 4.3 Materials and Methods Study Cohorts A total of 1444 renal transplantations were performed in the South Australian transplant centre between January 1990 and December Biopsy reports of the living and deceased donor transplants performed within this period were reviewed by two independent observers. Sixty-one cases of transplant glomerulopathy (patient group) were identified by biopsy reports stated (1) transplant glomerulopathy; (2) glomerular basement membrane reduplication / double contour and (3) mesangial thickening / expansion. Eighty-seven patients transplanted within the same period, who had been biopsy-proven without transplant glomerulopathy formed the control group of this study. All subjects had a negative cytotoxic T-cell crossmatch. B-cell crossmatch results were available at transplantation. Subjects were followed up until graft loss or December

135 4.3.2 HLA Typing All donor-recipient pairs were ABO blood group compatible. Donor and recipient HLA-A, -B and -DR typing were performed by the routine lymphocytotoxicity method, and by Sequence Based Typing (SBT) when necessary [204] Detection of Anti-HLA Antibodies Subject pre-transplant serum was assessed for HLA class I and class II alloantibodies using a Tepnel Lifecodes Luminex Screen assay. Subjects who were screened positive were further investigated to determine the antibody specificity using a Tepnel Lifecodes Luminex PRA assay. Sera with multispecific antibodies were tested by a Tepnel Lifecodes Luminex Single Antigen Bead assay. In brief, HLA class I or II antigen coated Luminex beads were incubated with 12.5μL of patient serum in 96-well Millipore multiscreen filter plates, in conjunction with manufacturer s reagents. The mixture was incubated for 30 minutes in dark, at room temperature, on a rotating platform. The wells were then washed 3 times to remove unbound excess serum using a vacuum system. After incubation, the plate was ready for data acquisition using Lifematch Fluoroanalyzer. Data acquired by the Luminex software was imported into the Lifematch QuickTypeTM Analysis software and the results were analyzed as per manufacturer recommendations. Post-biopsy sera which had been collected peri-biopsy from subjects who did not have DSA pre-transplant were tested by the Tepnel Lifecodes Luminex assays to determine possible DSA development post-transplantation Clinical Data ANZDATA Analysis Clinical data including donor-recipient demographics, number of transplants, rejection episodes and graft loss from 1990 to 2007 were collected from Australia and 135

136 New Zealand Dialysis and Transplant Registry (ANZDATA Registry, [205]. ANZDATA collects a wide range of statistics which relate to the outcomes of treatment of those with end stage renal failure from all renal units in Australia and New Zealand. The ANZDATA Registry has collected details of rejection episodes within 6 months from the day of transplant since 1 April For each episode, the treating physician reported whether the rejection was biopsy proven. For the transplants performed prior to April 1997, the 6 month rejection data was retrieved from biopsy reports and case notes. Rejection episodes were confirmed by biopsy in 95% of cases. Only graft rejections that occurred before diagnosis of TG were included in the analysis. Graft loss was defined as loss of graft function (return to dialysis or re-transplant) or death of patient Statistics Graft survivals were compared using the Kaplan-Meier method. Continuous outcomes were compared using t-tests or analysis of variance (ANOVA) as appropriate. Categorical variables were compared using chi-square or Fisher s exact tests. 136

137 4.4 Results Univariate Risk Factors Demographic Comparison Of 1444 patients, 61 biopsy-proven cases of TG were identified. The estimated prevalence of TG in our renal transplant population was approximately 4% of all grafts transplanted during this period. Univariate analyses of various risk factors were performed by comparing TG patients with the control group (Table 4.4.1). Only delayed graft function were significantly associated with development of TG (38% vs. 21%, 2.3 [ ], p=0.042). Other demographic factors not significantly associated with TG were patient age (30-49 years p=0.17, >50 years p=0.08) gender (p=0.58) and ethnicity (p=0.67), donor age (25-39 years p=0.56, years p=0.26, >50 years p=0.95) and gender (p=0.80). The type of transplantation (deceased or living) (p=0.28), regraft (p=0.16) and total ischaemia time (p=0.25) are not a risk factor of TG. There was a trend of increased risk of TG in patients with more than 1 HLA-A, -B and -DR antigen mismatched. The risk of TG in HLA mismatched grafts were at least 7 times as high as in HLA matched grafts. The analysis of association between HLA antigen mismatches and TG development was restricted by small sample size in this study, the findings should be confirmed by larger scale studies. 137

138 Table 4.4.1: Descriptive statistics of cohort and univariate analyses of risk factor of TG Recipient TG Patients Controls Odds Ratios [95% C.I.] p-value (n=61) (n=87) Age yrs 15 (25%) 12 (14%) yrs 31 (50%) 46 (53%) 0.5 [ ] 0.17 >50 yrs 15 (25%) 29 (33%) 0.4 [ ] 0.08 Gender (Males) 38 (62%) 58 (67%) 0.8 [ ] 0.58 Ethnicity (Indigenous) 9 (15%) 15 (17%) 0.8 [ ] 0.67 Donor Age <25 yrs 12 (20%) 21 (24%) yrs 16 (26%) 21 (24%) 1.3 [ ] yrs 18 (30%) 18 (21%) 1.8 [ ] 0.26 >50 yrs 15 (25%) 27 (31%) 1.0 [ ] 0.95 Gender (Males) 38 (62%) 56 (64%) 0.9 [ ]

139 Continue Table TG Patients Controls Odds Ratios [95% C.I.] p-value (n=61) (n=87) Graft Total HLA Mismatch 0 1 (2%) 10 (11%) (33%) 25 (29%) 8.0 [ ] (43%) 36 (41%) 7.2 [ ] (23%) 16 (18%) 8.8 [ ] Transplant type (Deceased) 52 (85%) 68 (78%) 1.6 [ ] 0.28 >1 Transplant 13 (21%) 11 (12%) 1.9 [ ] 0.16 Cold ischemia time >18 hrs 31 (51%) 33 (38%) 1.5 [ ] 0.25 Delayed graft function 18 (38%) 16 (21%) 2.3 [ ]

140 Immunosuppression Induction immunosuppressant was calcineurin inhibitors in 98% (84/87) of the controls and 93% (57/61) of the patients (p=0.47). In the control group, 85% (73/87) had cyclosporine and 13% (11/87) had tacrolimus. In the patient group, 49/61 (80%) had cyclosporine and 8/61 (13%) had tacrolimus (Table 4.4.2). More control received mycophenolic acid than azathiopine (46% vs. 37%). In the patient group, the use of azathiopine was more frequent than mycophenolic acid (51% vs. 34%, p=0.23). However, the type of anti-metabolites used at induction did not correlate with risk of TG development (p=0.23). Approximately 50% of the controls and patients received prednisolone (p=0.81) and less than 15% of both groups had monoclonal antibody therapy as part of induction treatment (p=0.85). Using the data sequentially collected by ANZDATA on these patients, the issue of change of immunosuppressants as a risk factor for TG was examined (Table 4.4.2). We hypothesised that TG might be precipated by withdrawal of either steroid or calcineurin inhibitors. Therefore, the uses of calcineurin inhibitors and steroid, prior to diagnosis of TG, were compared between the two groups. Time of diagnosis was unavailable for 5 TG patients. These 5 patients were excluded from all analysis involved time of diagnosis. There were 91% of the controls and 84% of the patients receiving cyclosporine or tacrolimus (p=0.76). Prednisolone was not withdrawn in 62% of the controls and 59% of the patients (p=75). No significant association was identified between use of immunosuppressants and TG. 140

141 Table 4.4.2: Immunosuppressants at induction and during diagnosis of TG, correlation with TG development Patients (n=61) Controls (n=87) p-value At induction Calcineurin inhibitors Cyclosprine 49 (80%) 73 (84%) 0.47 Tacrolimus 8 (13%) 11 (13%) None 4 (7%) 3 (3%) Anti-metabolites Azathioprine 31 (51%) 32 (37%) 0.23 Mycophenolic Acid 21 (34%) 40 (46%) None 9 (15%) 15 (17%) Steroid Prednisolone 31 (51%) 46 (53%) 0.81 None 30 (49%) 41 (47%) IL-2 receptor antagonist 4 (7%) 7 (8%) 0.75 CD3 3 (5%) 3 (3%) 0.67 Prior to diagnosis of TG Calcineurin inhibitors Cyclosprine 37 (66%) 61 (72%) 0.76 Tacrolimus 11 (20%) 16 (19%) None 8 (14%) 8 (9%) Anti-metabolites Azathioprine 22 (39%) 28 (33%) 0.38 Mycophenolic Acid 22 (39%) 44 (51%) None 12 (21%) 14 (16%) Steroid Prednisolone 33 (59%) 53 (62%) 0.75 None 23 (41%) 33 (38%) 141

142 4.4.2 Early Graft Rejection We next addressed the question early graft rejection as a risk factor for TG and types of rejection correlated with risk of TG. In the first 6 months, 31/87 controls and 24/61 patients experienced at least one rejection (36% vs. 39%, p=0.65). Cellular rejection were reported in 24 controls and 18 patients (28% vs. 30%, p=0.80). Vascular rejection rates were comparable between the control and the patient groups (23% vs. 18%, p=0.47). Glomerular rejection rate was found significantly higher in the patients than the controls. In the patient group, 12/61 developed glomerular rejection compared to 6/87 controls (p=0.025). The patients who experienced glomerular rejection within 6 months post-transplant were at a 3 times higher risk of TG development compared to those who never have early glomerular rejection (3.3 [ ]). Univariate correlations of various type of rejection and TG development were showed in Table

143 Table 4.4.3: Cellular, vascular and glomerular rejection as a risk factor of TG Patients (n=61) Controls (n=87) Odds Ratios [95% C.I.) p-value Any 24 (39%) 31 (36%) 1.2 [ ] 0.65 Cellular 18 (30%) 24 (28%) 1.1 [ ] 0.80 Vascular 11 (18%) 20 (23%) 0.7 [ ] 0.47 Glomerular 12 (20%) 6 (7%) 3.3 [ ]

144 4.4.3 Significance of Anti-HLA Antibodies in Transplant Glomerulopathy CDC Class I PRA defined Presensitization There was 31% of the patients have CDC class I PRA>10% compared to 24% in the controls (19/61 vs. 21/87, p=0.35) prior to transplantation. In addition, 54% of the patients had peak PRA>10% compared to 44% in the controls (33/61 vs. 38/87, p=0.21). Univariate analyses showed CDC defined historic or current presensitization against HLA antigen did not correlate with TG development (Table 4.4.4) Luminex defined Presensitization Pre-transplant sera from 73/87 controls and 57/61 patients were available for testing (84% vs. 93%). Fourteen controls and 4 patients did not have serum for testing (16% vs. 7%). More than 80% of the sera were collected within 3 months from the transplant day. There were 19/73 controls and 34/57 patients screened positive by the Luminex antibody screen assay (33% vs. 47%). The patients with TG had more preformed antibodies than the control group. Anti-HLA class I antibodies were found in 14/73 controls and 31/57 patients (19% vs. 54%). Anti-HLA class II antibodies were found in 18/73 controls and 18/57 patients (25% vs. 32%). In univariate analyses, only HLA CI presensitization was associated with TG development (5.0 [ ], p<0.001). HLA CII presensitization was not a risk factor of TG (p=0.38) (Table 4.4.4). 144

145 Table 4.4.4: Univariate effects of HLA presensitization, defined by CDC or Luminex techniques, in the development of TG Patients (n=61) Controls (n=87) Odds Ratios [95% C.I.] p-value CDC Class I PRA>10% (T-cells) n=61 n=87 At transplant 19 (31%) 21 (24%) 1.0 [ ] 0.35 Peak (historic) 33 (54%) 38 (44%) 1.0 [ ] 0.21 Luminex Multiantigen Screen n=57 n=73 Class I 31 (54%) 14(19%) 5.0 [ ] <0.001 Class II 18 (32%) 18 (25%) 1.4 [ ]

146 Pre-transplant DSA as A Risk Factor of TG DSA were detected in 10/57 patients and 5/73 controls in pre-transplant serum by the Luminex multiantigen assay (18% vs. 7%, p=0.058). In the patient group, 4 had DSA-CI, 4 had DSA-CII and 2 had DSA-CI&CII. In the control group, 3 had DSA- CI and 2 had DSA CII. Presence of DSA pre-transplant was not significantly correlated with development of TG. Small sample size could be the cause of statistical insignificance (Table 4.4.5) De novo DSA Predict TG Post-biopsy sera from 35 controls and 36 patients who did not have DSA prior to transplantation were tested. DSA were detected in 8/35 controls and 21/36 patients (23% vs. 58%). The prevalence of de novo DSA was significantly higher in the patient group (58% vs. 23%, p=0.002). The presence of de novo DSA confers a risk of 4.7 [ ] of TG development. DSA-CI were found in 19/36 patients and 4/35 controls (53% vs. 11%). DSA- CII were found in 17/36 patients and 7/35 controls (47% vs. 20%). Both DSA-CI and DSA-CII were associated with TG development (CI: 8.7 [ ], p<0.001 and CII: 3.6 [ ], p=0.015) (Table 4.4.5). 146

147 Table 4.4.5: Univariate effects of anti-hla antibodies in the development of TG Patients (n=57) Controls (n=73) Odds Ratios [95% C.I.] p-value Pre-transplant Luminex defined DSA Any 10 (18%) 5 (7%) 2.9 [ ] Class I 6 (11%) 3 (4%) 2.7 [ ] 0.18 Class II 6 (11%) 2 (3%) 4.2 [ ] 0.14 Post-transplant Luminex defined De novo DSA n=35 n=36 Any 21 (58%) 8 (23%) 4.7 [ ] Class I 19 (53%) 4 (11%) 8.7 [ ] Class II 17 (47%) 7 (20%) 3.6 [ ]

148 4.4.4 Multivariate Analysis In univariate analysis, only early glomerular rejection, delayed graft function, Luminex defined HLA CI presensitization and DSA were significantly correlated with TG development only. In multivariate analysis, we combined subjects who had DSA pretransplant with those developed DSA post-transplant to increase the sample size. Presence of DSA was independently associated with TG, 3.8 [ ], p= After adjustment for DSA, HLA CI presensitization (p=0.079), delayed graft function (0.95) and early glomerular rejection (p=0.33) were not associated with TG development (Table 4.4.6). 148

149 Table 4.4.6: In multiple regression analysis, presence of DSA was the only independent predictor of TG after adjustment for the other 3 univariate factors. Factors Odds Ratios [95% C.I] p-values HLA CI presensitization 3.1 [ ] 0.08 Delayed graft function 1.0 [ ] 0.95 DSA (pre- or post-transplant) 3.8 [ ] 0.03 Glomerular rejection within 6 months 2.2 [ ]

150 4.4.5 DSA as A Predictor of TG Time to Biopsy In this study, approximately 50% of the controls and the patients were biopsied within 5 years post-transplant. There was no significant difference in biopsy timing between the two groups (medium time to biopsy, 4.3 years in the control group and 4.0 years in the patient group, p=0.84) (Figure 4.4.1). Figure 4.4.1: Cumulative graph for biopsy events in relation to time post-transplantation Biopsy timing (time to biopsy) were similar for the patient and the control groups 150

151 Association of TG and Graft Survival The total graft survival rates of the control group were 97% at 1 year, 81% at 5 years and 63% at 10 years. In comparison, the patient group had poorer long term graft survival with 98% at 1 year, 66% at 5 years and 33% at 10 years (Table 4.4.7, Figure 4.4.2a). The death-censored graft survival rates of the control and the patient groups were 99% vs. 98% at 1 year, 88% vs. 69% at 5 years and 73% vs. 40% at 10 years (Table 4.4.7, Figure 4.4.2b). The risk of graft loss was significantly higher in the patients with TG (total graft loss = 1.9 [ ], p=0.003 and death-censored graft loss = 2.6 [ ], p<0.001). Median graft survival in the controls and the patients were 11.6 years and 7.2 years, respectively. Similar findings in death-censored graft survival analysis was reported, median death-censored graft survival was 17.3 years in the controls compared to 7.8 years in the patients (Table 28, Figure 4.4.2). Graft survival rates between the groups were comparable before diagnosis of TG, but were significantly different after diagnosis of TG (Figure 4.4.3a). Median of graft survival upon diagnosis of TG were 8.9 years in the control group compared to 3.2 years in the patient group (p=0.002) (Figure 4.4.3b). Thus, the presence of TG was associated with accelerated graft loss. 151

152 Table 4.4.7: Graft survival rates in patients with TG, according to antibody specificity Years post-transplant Total Graft Survival Controls (n=87) 97% 92% 81% 63% ALL Patients (n=61) 98% 90% 66% 33% No DSA pre- and post-transplant (n=14) 100% 93% 93% 51% DSA pre-transplant (n=10) 100% 89% 44% 11% DSA post-transplant (n=21) 95% 95% 52% 29% Death-censored Graft Survival Controls 99% 95% 88% 73% ALL Patients 98% 93% 67% 40% No DSA pre- and post-transplant (n=14) 100% 93% 93% 51% DSA pre-transplant (n=10) 100% 89% 44% 22% DSA post-transplant (n=21) 95% 95% 52% 29% 152

153 Table 4.4.8: Graft survival and risk of graft loss in TG patients, according to antibody specificity, in 20 years of follow up Median Graft Survival Graft Loss Events Hazard Ratios p-value (years) [95% C.I] Total Graft Survival Controls (n=87) ALL Patients (n=61) [ ] No DSA pre- and post-transplant (n=14) [ ] 0.97 DSA pre-transplant (n=10) [ ] <0.001 DSA post-transplant (n=21) [ ] Death-censored Graft Survival Controls ALL Patients [ ] <0.001 No DSA pre- and post-transplant (n=14) [ ] 0.18 DSA pre-transplant (n=10) [ ] <0.001 DSA post-transplant (n=21) [ ] <

154 (a) Total graft survival (b) Death-censored graft survival Figure 4.4.2: Graft survival in the controls and patients. TG patients have significantly poorer graft survival within 20 years of follow up. 154

155 (a) Graft survival of patients prior to diagnosis of TG compared to controls, time from transplant to biopsy. (b) Graft survival after diagnosis of TG Figure 4.4.3: Graft survival before and after diagnosis of TG Graft survival of patients before diagnosis of TG was compatible to the controls (p=0.24). Patients have significantly poorer graft survival upon diagnosis of TG, suggested that TG was associated with accelerated graft loss. 155

156 Pre-transplant DSA Associated with Early TG Development The impact of DSA present pre-transplant and DSA developed posttransplant on TG progress was next examined. In this study, diagnosis time was defined as the duration from transplant to diagnosis of TG. A short diagnosis time is reflecting that patients developed TG soon after transplantation. Median of diagnosis time for the patients who had DSA pre-transplant (pre-tx DSA group) was 1.8 years, which was significantly shorter than the patients who developed DSA post-transplant (post-tx DSA group) or who did not have DSA (No DSA group) (5.6 years, 4.3 years, p=0.049). Diagnosis time was similar between the post-tx DSA and the No DSA groups (p=0.89) (Figure 4.4.4). Figure 4.4.4: Cumulative TG diagnosis events in relation to time post-transplant, according to presence of DSA Patients with preformed DSA were diagnosed significantly earlier than those developed DSA post-transplant or no DSA. 156

157 Significance of DSA in TG Progression Even though the pre-tx DSA group developed TG sooner, the total and deathcensored graft survival was not significantly different from those developed DSA posttransplant (p=0.22 and p=0.63, Table 4.4.7, Table and Figure 4.4.5a, b). Median of both total and death-censored graft survival was 4.4 years in the pre-tx DSA group compared to 5.2 years in the post-tx DSA group (p=0.22). Median of both total and death-censored graft survival for the No DSA group was 15 years. These patients have significantly better graft survival than the pre- and the post-tx DSA groups (total p=0.006, p=0.039; death-censored p=0.006, p=0.039). In the TG subgroups, total graft survival rate was equal to death-censored graft survival rates because almost all of the graft losses were due to graft dysfunction. (a) Total graft survival 157

158 (b) Death-censored graft survival Figure 4.4.5: Graft survival in patients with TG according to presence of DSA Patients without DSA have significantly higher graft survival rate than those with DSA pre and/or post-transplantation. 158

159 4.4.6 DSA are Predictive for Graft Loss in TG In pairwise analysis of 20 years graft survival, TG patients with DSA, regardless of the DSA presentation time, had significantly poorer graft survival compared to the control group (pre-tx DSA <0.001; post-tx DSA 0.002). In contrast, the patients without DSA have comparable graft survival with the control group (p=0.97) (Table 4.4.8, Figure 4.4.6a, b). (a) Total graft survival 159

160 (b) Death-censored graft survival Figure 4.4.6: Graft survival in TG subgroups, compared to the control group Patients without DSA have similar graft survival with control group that have been biopsy-proven without TG. The patients with DSA pre and/or post-transplantation have significantly poorer graft survival than the control group. 160

161 4.5 Discussion In this study, we investigated risk factors and roles of anti-hla antibodies in a large group of patients diagnosed with TG from a single transplant centre. The major finding of this study was that when a sensitive test such as Luminex was used, DSA were detected in more than 50% of the TG cases and were a significant predictor of graft loss in the patients with TG. Graft survival of the TG patients without DSA was comparable with the controls who did not have TG (Figure 4.4.6, Table 4.4.8). Transplant glomerulopathy is a unique pathologic phenotype of late kidney deterioration [223]. TG was originally classified as a variant of chronic allograft nephropathy (CAN), but was recently separated from CAN due to its unknown pathogenesis [229, 230]. Several studies showed that anti-hla antibodies were detected in a substantial percentage of patients (50 70%) who developed TG [221, 225, 228]. Previous studies also identified as association of early glomerular rejection (glomerulitis) with TG [228]. In the present study, glomerular rejection rate was significantly higher in TG patients compared to the control group (20% vs. 7%) and the local renal transplant population (9%) (Table 4.4.3) [231]. However, after adjustment for DSA, glomerular rejection was not independently associated with TG development (Table 4.4.6). Messias et. al. showed that a significantly greater percentage of patients in glomerulitis group experienced delayed graft function compared to the nonglomerulitis group [232]. Hence, we hypothesized that the univariate effect of delayed graft function observed in this study was likely due to its association with glomerular rejection. Immunosuppressant regimen was not significantly different between the patient and the control groups, both at induction and prior to diagnosis of TG (Table 4.4.2). The use of azathiopine was higher than mycophenoric acid in the patient group at induction, but the difference did not achieved statistical significance. Previous studies showed that MMF therapy down-regulates antibody synthesis, and has been used as part of rescue therapy for graft suffering antibody-mediated injury [233, 234]. Other studies reported the use of MMF as prevention of sensitization in transplantation, and removal of MMF may cause late antibody-mediate rejection [235, 236]. Eventhough, in 161

162 this study, we found a significant correlation of antibody production post-transplant and TG development, we did not observed protective effect of MMF on TG development. CDC class I current and historic peak PRA were correlated with TG, but anti- HLA class I antibodies detected by Luminex was a univariate risk factor for TG (Table 4.4.4). CDC techniques detect IgG and IgM antibodies whereas Luminex detects IgG antibodies only. Studies showed that IgM antibodies were not deleterious to grafts [ , 209]. This study showed that presence of DSA confered an increase risk of TG (Table 4.4.8) [224]. However, due to the small sample size, pre-transplant DSA did not reach statistical significance as a risk factor of TG in univariate analysis. Post-transplant DSA caused a 4 times higher risk of TG, both DSA-CI and DSA-CII were associated with TG. HLA class I and class II antigens are expressed in glomerular and peritubular capillaries, thus presence of DSA HLA CI and/or CII may trigger development of TG [237]. Gloor et al and Issa et al reported that HLA CII antibodies confered a higher risk for TG development than CI antibodies [224, 226]. In our study, the risk of TG development was 8.7[ ] in the patients with DSA-CI and was 3.6[ ] in the patients who have DSA-CII. This discrepancy may be due to difference in crossmatch techniques. We studied patients transplanted in the last 2 decades, the T-cell crossmatches were performed by NIH methods and then by conventional CDC method with extended incubation from mid 1995 until now. In comparison, Gloor et al and Issa et al used AHG enhanced CDC T-cell crossmatch which is more sensitive than conventional CDC technique in detecting CI antibodies. Furthermore, B-cell crossmatch which detects CII antibodies was performed prospectively in our centre. Thus, the patients were less likely to be transplanted across a strong positive BXM. TG often associated with poor graft survival (Table 4.4.7). The 5 years graft survival rate of the TG group was only half of the rate of the control group. In this study, 10 years death-censored graft survival of the TG group was 40%, which was consistent with previous reports [238, 239]. Median of graft survival after diagnosis of TG was 3.2 year, similar to that reported by Issa et al [226]. DSA that existed prior to transplant were only detected by Luminex, thus they were relatively low titer antibodies. Some studies showed that low level DSA may be associated with graft accommodation, but Reed et. al. reported that low level DSA may cause endothelial cell proliferation that leads to chronic graft rejection [217, 240, 162

163 241]. In this study, more than half of the group with DSA pre-transplant developed TG in 2 years and lost graft within 5 years. In comparison, of the 14 TG patients who never developed DSA, only 1 patient lost graft within 5 years of follow-up. Our findings suggested that pre-transplant low level DSA may be a significant cause of TG, therefore defining acceptable level of DSA is an important future direction for research. Binding of anti-hla antibodies to HLA molecules may cause endothelial cell injury via complement cascade and/or may induce endothelial cell proliferation and survival [241]. Previous studies showed that binding of antibodies to class I molecules on the surface of endothelial cells results in tyrosine phosphorylation of various intracellular proteins [242]. The two major consequences of class I-mediated phosphorylation are 1) cell proliferation via up-regulation of fibroblast growth factor receptors (FGFR) on surface of endothelial cells [243], and 2) cell survival stimulation by increased endothelial cells expression of anti-apoptotic proteins Bcl-2 and Bcl-xL via P13K/Akt pathway [217, 244]. These observations raise an important question, what are the factors that determine the outcomes of class I-mediated phosphorylation. In vitro studies by Reed s group showed that stimulation of cell proliferation was observed at concentrations of anti-mhc antibodies ranging from µg/ml with maximal cell proliferation at concentrations of 10µg/ml [243]. On the other hand, treatment of endothelial cells with anti-hla class I antibodies for 24 hours induced a prominent increase in Bcl-2 and Bcl-xL protein levels, with maximum increase in protein expression when low concentration antibodies were used (0.01-1µg/mL) [217, 244]. These findings suggested that concentration of anti-hla antibodies is one of the factors contribute to the outcomes of class I-mediated phosphorylation. We hypothesized that development of TG is a consequence of a continuing change in anti-hla antibodies concentration and isotypes which trigger different cell response at different antibodies concentration and isotypes. In the early stage of anti- HLA antibody production, it is present at low concentration and mostly IgM type (Refer to Chapter 1, section 1.8.2), which results in an increased expression of anti-apoptotic proteins that prevent endothelial cell death as observed in early stage of TG [222, 244]. A slight increase in antibody concentration may activate cells evidenced by increased metabolism and protein synthesis, with abundant mitochondria, ribosomes and golgi apparatus in endothelial cells, associated with changes like vacuolation, serration of the subendothelial interface and synthesis of new basement membranes layers as observed 163

164 in sequential untrastructural studies [222, 243]. In latter stage of humoral immune response, the concentration of antibody is high with mostly IgG type. At this stage, the antibodies may activate complement cascade to cause cell injury and apoptosis, resulting graft loss [217, 245]. This correlates the high level of DSA (MFI>10000) detected in post-biopsy sera in most of our TG cases. Even though antibody-mediated complement activation is likely the etiology of TG, the marker of complement activation C4d is not been detected in many TG cases [217, 222, 224, 226, 228, 245]. Mesangial matrix expansion is the most common feature in TG. Mesangial matrix accumulation can be caused by impairment of matrix degradation pathway and/or increase matrix production [246, 247]. Megsin is a serine protease inhibitor which predominantly present in mesangial cell. Binding of Megsin to plasmin impairs the plasminogen acitivator/plasmin cascade results in impairment of matrix degradation [248, 249]. Thus, overexpression of Megsin gene may increase mesangial matrix by lowering matrix degradation [249]. Upregulation of Megsin has been reported in kidney diseases associated with mesangial cell proliferation and/or mesangial matrix expansion such as diabetic nephropathy and IgA nephropathy [248]. However, association of Megsin expression and TG has not been explored. This would be a future direction for research. One of the major findings of the study is that anti-hla antibodies detected by Luminex are strongly and independently associated with TG. Preformed low level DSA detected by Luminex, may be the cause of early TG development. The findings support the role of low titer DSA detected by solid phase assays in chronic graft rejection. Preformed and de novo DSA are significant predictors of graft loss in TG patients. Post-transplantation monitoring of DSA by Luminex may be more costeffective than biopsy because no hospital stay is required. Invasive biopsy protocol may cause graft loss in rare occasion, thus Luminex testing is a safer procedure. Even though biopsy is a standard protocol for diagnosis of TG, we suggest yearly posttransplant DSA monitoring to identify patients who at a high risk of TG and consequently graft lost. 164

165 Chapter 5 A Retrospective Study of B-cell Crossmatch with Luminex Technology in Well-Matched Highly Sensitized Patients from The Australian National Renal Exchange Programme 5 (ANREP) 165

166 5.1 Introduction Exposure to foreign HLA during pregnancies, blood transfusion or previous organ transplantation may induce production of anti-hla antibodies [8, 9, 250, 251]. Transplantation of patients with a high level of anti-hla antibodies is a major challenge in renal transplant programmes. Highly sensitized patients are more likely to have a positive crossmatch than non-sensitized patients, resulting a longer wait time to find a compatible donor [252]. Highly sensitized patients have longer waiting time on dialysis [253]. Dialysis is associated with death risk, especially after 12 months (hazard ratio 1.33 in Australia) compared to transplantation [254, 255]. In recent years, the chance of transplanting highly sensitized patients has been improved by (I) organ sharing programmes [256], (II) definition of acceptable HLA mismatches by various antibody detection assays and HLAMATCHMAKER program [ ] and (IIII) removal of anti-hla antibodies by desensitization [260, 261]. In Australia, the National Renal Interstate Exchange Programme (ANREP) aims to increase the chance of finding negative crossmatch, HLA antigen well-matched donors for highly sensitized patients [256, 262]. The programme commenced in October Under this programme, grafts are allocated interstate to provide wellmatched kidneys to highly sensitized patients. There are 6 matching levels under this programme, with different number of HLA-A, -B and -DR mismatches and levels of sensitization determined by class I CDC panel reactive antibodies level (PRA%). A substantial percentage of the patients received a graft under this programme have peak CDC class I PRA>50% and a maximum of 2 HLA-A, -B and -DR mismatches. Approximately 15-20% of the deceased donor grafts were allocated to highly sensitized patients through this programme [256, 262]. Highly sensitization patients are at a higher risk of rejection [263, 264]. Therefore, a matching protocol with a greater predictive value may improve graft outcomes. B-cell crossmatch (BXM) detects pre-existing class I and II donor-specific anti-hla antibodies (DSA) [231]. BXM is more sensitive than T-cell crossmatch (TXM) in detecting class I antibodies [139, 231]. However, BXM is less favoured than TXM for graft allocation because the clinical significance of BXM is controversial [140, 141,

167 147, 154, 168, 203, 207, 265]. In Australia, graft allocation is based on TXM, BXM is usually performed retrospectively. CDC-BXM is generally considered as less sensitive and less specific compared to solid phase Luminex assays [198, 266]. Luminex is a multiplexed data acquisition and analysis platform for flow cytometric analysis of microsphere-based assays, introduced a decade ago [267, 268] (Refer to Chapter 1, Section 1.11). Many laboratories rely on Luminex to define antibody specificity in highly sensitized patients. Luminex allows detection of low level antibodies that are not detectable by CDC-BXM. However, clinical relevance of the low level DSA remains controversial. Reed et al showed that low level DSA may induce endothelial cell and smooth muscle cell proliferation resulting chronic graft rejection [ ]. Furthermore, low level DSA are associated with transplant glomerulopathy which results in graft loss (Refer Chapter 4) [221, 224]. In contrast, other groups reported the involvement of low level DSA in initiating graft accommodation that promotes graft survival [217, 244]. Luminex detects both complement and non-complement fixing IgG antibodies, whereas CDC-BXM only detects cytotoxic antibodies. Thus, over reliance on Luminex analysis may lead to assignment of irrelevant antibodies as unacceptable and consequently reduces chance of transplantation in highly sensitized patients. Significance of BXM and Luminex defined DSA in renal transplantation is controversial. Further studies may help to resolve the place of BXM and Luminex defined DSA in transplantation. 5.2 Objectives This chapter studies the outcomes of highly sensitized patients transplanted under the Australian National Renal Exchange Programme. The first part of this chapter discusses overall transplant outcomes of highly sensitized patients with epidermiological analysis from ANZDATA, compared to other grafts transplanted within the same period in participating hospitals. The second part of this chapter evaluates clinical impacts of HLA antibodies detected by B-cell crossmatch or Luminex assays: 167

168 I) Predictive value of CDC-BXM for graft rejection, function and survival II) Predictive value of Luminex antibody analysis for graft rejection, function and survival III) Identification of independent predictor of rejection in multivariate analysis 5.3 Materials and Methods Study Cohort We recruited patients transplanted under the top three levels of the program. Level 1 grafts have zero antigen mismatched, patients with PRA>50%. Level 2 grafts have 1 antigen mismatched, patients with PRA>80%. Level 3 grafts have 2 antigen mismatched, patients with PRA>80%. Eighty-two patients were recruited from 28 transplant centres in Australia as listed in Table Table 5.3.1: Patient recruitment (a) Recipient State Number of Patients Western Australia 7 New South Wales 42 Queensland 5 Victoria 22 South Australia 6 Total

169 (b) List of Hospitals and Patient numbers NSW Children's Hospital 1 Westmead John Hunter Hospital 4 Liverpool Hospital 2 Prince of Wales Hospital 6 Royal North Shore Hospital 4 Royal Prince Alfred Hospital 5 St George Hospital 1 St Vincents Hospital 2 The Canberra Hospital 3 Westmead Hospital 14 Total 42 QLD Cairns Base Hospital QLD 1 John Flynn Hospital 1 Princess Alexander Hospital 2 Royal Brisbane Hospital 1 Total 5 SA Alice Springs Hospital 1 Flinders Medical Centre Queen Elizabeth Hospital Royal Adelaide Hospital Total 6 Vic Alfred Hospital 2 Austin and Repatriation Medical Centre 1 Geelong Hospital 2 Monash Mc Adult Unit 5 Royal Children's Hospital 1 Royal Hobart Hospital 1 Royal Melbourne Hospital St Vincent's Hospital 1 Total WA Royal Perth Hospital 5 Sir Charles Gairdner Hospital 2 Total Crossmatch Crossmatches were performed on enriched donor T- and B-cells using CDC extended incubation methods in donor centre. Serum was collected monthly from all patients waiting for kidney transplantation. Donor T-cells and B-cells were isolated by the immunomagnetic particle technique (Dynabeads, Dynal, Oslo, Norway). Crossmatch was performed with a patient s current and peak sera by scientist on duty. Current serum was collected within 1 month before transplant. Peak serum was that with the highest class I CDC %PRA. TXM was performed prospectively, BXM was usually performed retrospectively. 169

170 5.3.3 HLA Typing All donor-recipient pairs were ABO blood group compatible. Donor and recipient HLA-A, -B and -DR typing were performed by the routine lymphocytotoxicity method, and by Luminex Sequence Specific Oligonucleotides, Sequence Specific Primers or Sequence Based Typing (SBT) when necessary by designated tissue typing laboratory. Donor and recipients were typed for HLA-C, -DQA, -DQB1, -DPB1 and - DRB3/4/5 when the respective antibodies were detected Luminex Antibody Analysis Recipient sera were assessed for HLA class I and class II allo-antibodies using the Tepnel Lifecodes Luminex PRA assay (Refer to Chapter 2 Materials and Methods) in The South Australia Tissue Typing Laboratory. The current serum in crossmatch was tested for DSA. The peak serum in crossmatch was tested if DSA were not detected in the current serum. Patients with multispecific antibodies had DSA confirmed by SAB. The cut off value for assigning positive reaction was MFI 3000 for Single Antigen Bead (SAB) assay. This value may change dependents on a patient s antibody profile. In brief, HLA class I or II antigen coated Luminex beads were incubated with 12.5μL of patient serum in 96-well Millipore multiscreen filter plates, in conjunction with the manufacturer s reagents. The mixture was incubated for 30 minutes in dark, at room temperature, on a rotating platform. The wells were then washed 3 times to remove unbound excess serum using a vacuum system. After incubation, the plate was ready for data acquisition using Lifematch Fluoroanalyzer. Data acquired by the Luminex software was imported into the Lifematch QuickTypeTM Analysis software and the results were analyzed as per the manufacturer recommendations. 170

171 5.3.5 Clinical Data Patients were followed up until 30 June Clinical data including donorrecipient demographics, number of transplants, HLA class I PRA levels, and estimated glomerular filtration rates (egfr) were collected from Australia and New Zealand Dialysis and Transplant Registry (ANZDATA Registry) [205]. Rejection episodes, graft survival and serum creatinine levels up to 31 December 2006 were retrieved from ANZDATA. For each rejection episode, the treating physician reported the type of rejection and whether the rejection was biopsy proven. Rejection episodes, graft survival and serum creatinine levels from 1 Jan to 30 June 2007 were collected from transplant hospitals. For each humoral rejection episode, C4d staining was reported. The egfr was calculated using the Modification of Diet in Renal Disease (MDRD) equation [206]. Graft loss was defined as loss of graft function (return to dialysis or retransplant) or death of patient Statistics Graft survivals were compared using the Kaplan-Meier method. Continuous outcomes were compared using t-tests or analysis of variance (ANOVA) as appropriate. Categorical variables were compared using chi-square or Fisher s exact tests. 5.4 Results Characteristics of Study Cohort Eighty-two patients transplanted between October 2004 and 31 December 2006 were recruited for this study. Patient characteristics are shown in Table The patients were compared with all deceased donor transplants performed within the same period in participating hospitals (comparator, also known as All Grafts group in 171

172 this study, n=762). Patients in our study were younger (p=0.019), more females (0.004) and majority of them were re-grafted (<0.001). The comparator group consists of more male recipients (64%) and majority of them had one renal transplant only (89%). The patients were transplanted under the interstate kidney exchange program, thus the grafts have longer cold ischaemia time than the control group (<0.001). All patients were highly sensitized (peak CDC PRA >50%) and received wellmatched HLA type as per the exchange criteria. In the comparator group, only 33% of the grafts were 0-2 HLA mismatched, 67% have >2 mismatches (p<0.001). Half of the patient group had CDC class I PRA >50% at transplant compared to only 2% in the comparator group (p<0.001) (Table 5.4.1). The immunosuppression regimens at induction were different between the patient and the comparator groups. Greater than 80% of the patients had tacrolimus compared to 52% in the comparator group (p<0.001). The use of antibody is higher in the patient group than comparator group (p=0.005) (Table 5.4.2). 172

173 Table 5.4.1: Descriptive statistics of study cohort, compared to all other grafts transplanted within the same period in participating centres Study Cohort (NREP) All Grafts p-value (n=82) (n=762) Age (years ± S.E) 43.8 ± ± Gender (female) 43 (52%) 271 (36%) Ethnicity (indigenous) 5 (6%) 51 (7%) 1.00 Donor Age (years ± S.E) 44.3 ± ± Number of transplants <0.001 First 29 (35%) 676 (89%) Repeat 53 (65%) 86 (11%) Cold ischemia time (hours) n=82 n=748 <0.001 <12 22 (27%) 289 (39%) (47%) 353 (47%) >18 21 (26%) 106 (14%) 173

174 Continue Table Study Cohort All Grafts p-value (n=82) (n=762) CDC class I PRA (At transplant) < % 19 (23%) 698 (92%) 11-50% 22 (27%) 50 (6%) >50% 41 (50%) 14 (2%) CDC class I PRA (Historic peak) < % 0 (0%) 516 (68%) 11-50% 0 (0%) 158 (21%) >50% 82 (100%) 21 (26%) Total HLA Mismatch < (10%) 32 (4%) (90%) 223 (29%) (0%) 275 (36%) (0%) 232 (31%) 174

175 Table 5.4.2: Immunosuppressants at induction Study Cohort (n=82) All grafts (n=762) p-value Calcineurin inhibitors <0.001 Cyclosprine 11 (13%) 325 (43%) Tacrolimus 67 (82%) 393 (52%) None 4 (5%) 38 (5%) Anti-metabolites 1.00 Azathioprine 0 (0%) 8 (1%) Mycophenolic Acid 79 (96%) 721 (95%) None 3 (4%) 33 (4%) Prednisolone 80 (98%) 740 (97%) 1.00 IL-2 receptor antagonist 49 (60%) 466 (61%) 0.81 Antibody 10 (12%) 32 (4%)

176 5.4.2 Graft Outcomes of Highly Sensitized Patients Graft Rejection Of the 82 patients, 27 experienced at least one rejection (33%). The rejection rate was significantly higher in the patient group, compared to the comparator group (22%, p=0.02) (Table 5.4.3). When comparing the two groups according to rejection types, the highly sensitized patients had more vascular (<0.001), glomerular (0.001) and humoral (<0.001) rejection than other patients. Collection of humoral rejection data by ANZDATA commenced in Data was incomplete for 96 patients in the comparator group (all grafts) and 10 patients in the study cohort who were transplanted before These patients were excluded in humoral rejection analysis. Humoral rejection was reported in 14/72 highly sensitized patients, and 21/666 patients in comparator group (19% vs. 3%). Ten of the 14 humoral rejection episodes had positive C4d staining. The risk of humoral rejection was 7 times higher in the highly sensitized patients (7.4 [ ], p<0.001). The risk of vascular or glomerular rejection in highly sensitized patients was 3 times as high as the comparator group. The prevalence of vascular and glomerular rejection in the patient group was >10%, while in the comparator group was <10%. Cellular rejection rates were comparable between the patient and the comparator groups (p=0.56). 176

177 Table 5.4.3: Risk of cellular, vascular and glomerular rejection in highly sensitized patients All grafts Study Cohort Odds Ratios [95% C.I.) p-value (n=762) (n=82) Any 165 (22%) 27 (33%) 1.8 [ ] 0.02 Cellular 129 (17%) 16 (20%) 1.2 [ ] 0.56 Vascular 52 (7%) 15 (18%) 3.1 [ ] <0.001 Glomerular 35 (5%) 11 (13%) 3.2 [ ] Humoral* 21/ 666 (3%) 14/72 (19%) 7.4 [ ] <0.001 *Collection of humoral rejection data commenced in Data was incomplete for 96 patients in the comparator group (all grafts) and 10 patients in the study cohort who were transplanted before These patients were excluded in humoral rejection analysis. 177

178 Graft Function Graft function of the patient group was not significantly different from the comparator group within 1 year post transplant. Mean egfr of the patient group were 51 ml/min/1,73m 3 at 3 and 6 months, and 48mL/min/1,73m 3 at 1 year. In comparison, mean egfr of the comparator group were 51, 52 and 53mL/min/1.73m 3 at 3, 6 and 12 months (p=0.86, 0.69 and 0.06) (Figure 5.4.1, Table 5.4.4). Mean egfr Figure 5.4.1: Graft function of highly sensitized patients, compared to other grafts Table 5.4.4: Graft function, egfr, of highly sensitized patients All grafts (n=762) Study Cohort (n=82) p-value 3 months 51.3 ± ± months 52.2 ± ± year 53.9 ± ±

179 Graft Survival Up to 30 July 2007, there were 9 graft losses in the patients group. Five grafts lost within 6 months, and the remaining 4 graft lost in 7-24 months. Only 3 grafts were lost due to rejection. Other graft losses were due to patient death (2), chronic allograft nephropathy (1), BK virus nephropathy (1), vascular occlusion (1) and other cause (1). The total graft survival rates of the patient group were 94% at 6 months, 90% at 1 year and 84% at 2 years and were found comparable with the comparator group (p=0.48) (Figure 5.4.2, Table 5.4.5) Kaplan-Meier survival estimates 0.90 Graft survival Years post-transplant Number at risk GP SC General population Study cohort Figure 5.4.2: Total graft survival of highly sensitized patients, compared to all grafts 179

180 Table 5.4.5: Total graft survival rates in highly sensitized patients, compared to all grafts Years post-transplant Overall Survival Rates All grafts (n=762) 94% 91% 90% Study cohort (n=82) 94% 90% 84% Number of Failures / Number of Grafts All grafts 51/594 14/440 6/102 Study cohort 5/62 2/47 2/11 Hazard ratio [95% C. I.] 1.3 [ ] p-value

181 5.4.3 Predictive Value of B-cell Crossmatch in Highly Sensitized Patients Prevalence of Positive B-cell Crossmatch Complete BXM data were available for 55 of 82 patients (67%). Twenty-five (45%) patients had negative BXM tested with current and peak sera. Thirty patients (55%) had positive BXM in current and/or peak sera. Of the 30 patients, 10 had positive BXM in peak serum only, 8 in current serum only and the remaining 12 in both current and peak sera Characteristics of Patients with Positive B-cell Crossmatch Patients were grouped according to their BXM results: BXM+ and BXM groups. Demographics of BXM+ group were compared to the BXM group. The groups were not significantly different in age (p=0.17), gender (p=0.46), donor age (p=0.46), number of transplant (p=1.00), cold ischeamic time (p=0.25), CDC class I PRA at transplant (p=0.31) and number of HLA antigen mismatched (p=0.44) (Table 5.4.6). There have more female patients in BXM+ group, but the difference was not statistically significant (57% vs. 36%, p=0.055). No indigenous patient in BXM+ group. Induction immunosuppression regimen was similar between the BXM+ and BXM groups. In both groups, >80% of the patients were treated with tacrolimus (p=1.00), >90% were treated with prednisolone (p=0.02), >50% received IL-2 receptor antagonist (p=0.43) and 10-20% received antibody (p=0.72) as part of Immunosuppression (Table 5.4.7) 181

182 Table 5.4.6: Demographics of T B and T B+ groups T B T B+ p-value (n=25) (n=30) Age (years ± S.E) 45.8 ± ± Gender (female) 7 (36%) 17 (57%) Ethnicity (indigenous) 4 (16%) 0 (0%) Donor Age (years ± S.E) 47.6 ± ± Number of transplants 1.00 First 8 (32%) 10 (33%) Repeat 17 (68%) 20 (67%) Cold ischemia time (hours) 0.25 <12 4 (16%) 11 (37%) (48%) 11 (37%) >18 9 (36%) 8 (26%) 182

183 Continue Table T B (n=25) T B+ (n=30) p-value CDC class I PRA (At transplant) % 7 (28%) 6 (20%) 11-50% 10 (40%) 8 (27%) >50% 8 (32%) 16 (53%) Total HLA Mismatch (16%) 2 (7%) 1 7 (28%) 5 (17%) 2 14 (56%) 23 (76%) 183

184 Table 5.4.7: Immunosuppressants at induction T B (n=25) T B+ (n=30) p-value Calcineurin inhibitors 1.00 Cyclosprine 2 (8%) 3 (10%) Tacrolimus 22 (88%) 26 (87%) None 1 (4%) 1 (3%) Anti-metabolites 0.46 Azathioprine 0 (0%) 0 (0%) Mycophenolic Acid 24 (96%) 30 (100%) None 1 (4%) 0 (0%) Prednisolone 23 (92%) 30 (100%) 0.20 IL-2 receptor antagonist 13 (52%) 19 (63%) 0.43 Antibody 5 (20%) 4 (13%)

185 Correlation of B-cell Crossmatch and Graft Rejection Rejection was reported in 14/30 patients transplanted across BXM+ compared to 7/25 patients with BXM (47% vs. 28%, p=0.16) (Table 5.4.8). BXM+ was significantly correlated with humoral rejection (p=0.032). Nine patients experienced at least one humoral rejection in the BXM+ group compared to only 2 in the BXM group (35% vs. 8%). The risk of humoral rejection of BXM+ was 6 times as high as BXM (6.1 [ ]). There was a trend that vascular and glomerular rejections were more common in the BXM+ group, but the differences were not statistically significant. Vascular rejection rate was 27% in the BXM+ group compared to 8% in the BXM group (p=0.09). Glomerular rejection rate of the BXM+ group was 23% compared to 8% in the BXM group (p=0.16). Cellular rejection rates were comparable between the BXM+ and the BXM groups (24% vs. 30%, p=0.62). 185

186 Table 5.4.8: Correlations of T B+ with cellular, vascular and glomerular rejection T B (n=25) T B+ (n=30) Odds Ratios [95% C.I.) p-value Total 7 (28%) 14 (47%) 2.3 [ ] 0.16 Cellular 6 (24%) 9 (30%) 1.4 [ ] 0.62 Vascular 2 (8%) 8 (27%) 4.2 [ ] 0.09 Glomerular 2 (8%) 7 (23%) 3.5 [ ] 0.16 Humoral* 2 (8%) 9 (35%) 6.1 [ ] *4 T B+ patients transplanted before 2005 were excluded in humoral rejection analysis due to incomplete data collection. 186

187 Graft function B-cell Crossmatch and Graft Function BXM was not associated with impaired graft function. Compared to the BXM group, mean egfr was higher in the BXM+ patients at 3 and 6 months (53.6 vs ml/min/1.73m 3 ), but was similar at 1 year (47.9 vs. 49mL/min/1,73m 3 ) (Figure 5.4.3, Table 5.4.9). Mean egfr Figure 5.4.3: Correlation of graft function and BXM Table 5.4.9: Comparison of graft function (MDRD egfr) between T B and T B+ groups, mean ± S. E ml/min/1.73m 3. T B (n=25) T B+ (n=30) p-value 3 months ± ± months 47.6 ± ± year 49 ± ±

188 Correlation of B-cell Crossmatch and Graft Survival The graft survival rates of the BXM+ group at 6 month, 1 and 2 years were 93%, 90% and 90% compared 84%, 80% and 80% in the BXM group. No significant difference was found when comparing graft survival rates between the two groups at 6 months, 1 and 2 years (p=0.45) (Figure 5.4.4, Table ). Up to June 2007, 4 graft losses were reported in the BXM+ group compared to 5 in the BXM group. In the BXM+ group, only 1 graft lost due to rejection. The other 2 graft lost due to patient death and the remaining graft lost caused by CAN. In the BXM group, rejection causes 2 graft losses. The remaining graft losses were due to vascular occlusion (1), BK nephropathy (1), and other cause (1). Figure 5.4.4: Significance of positive BXM in relation to graft survival No significant difference was found when comparing graft survival rates between the two groups at 6 months, 1 and 2 years (p=0.45) 188

189 Table : Correlation of T B+ and total graft survival Years post-transplant Overall Survival Rates T B (n=25) 84% 80% 80% T B+ (n=30) 93% 90% 90% Number of Failures / Number of Grafts T B (n=25) 4/22 1/16 0/6 T B+ (n=30) 2/29 1/23 0/13 Hazard ratio [95% C. I.] 0.60 [ ] p-value

190 5.4.4 Predictive Value of Luminex defined Donor-specific Antibodies Presence of Donor-specific Antibodies Antibody specificity was defined in 71/82 patients. Sera were not available for testing in 11 patients. DSA were detected in 37 patients prior to transplantation (DSA group). Seventeen patients had DSA-CI only, 10 had DSA-CII only and 10 had both DSA-CI&CII. No DSA were detected in 34 patients (Non-DSA group). Univariate analyses were performed to compare the demographics between patients with and without DSA (Table ). The only factor associated with presence of DSA was the number of HLA antigen mismatched (p=0.026). There was a greater percentage of patients with >2 mismatches in DSA group compared to Non-DSA group (81% vs. 50%). Other factors not significantly associated with DSA were patient age (p=0.28), gender (p=0.64), ethnicity (p=0.67), donor age (p=0.25), number of transplant (p=0.80), cold ischaemia time (p=0.60), and CDC class I PRA at transplant (p=0.58). No significant differences were found when comparing immunosuppression regimens between the two groups (Table ). 190

191 Table : Demographics of DSA and Non-DSA groups Non-DSA DSA p-value (n=34) (n=37) Age (years ± S.E) 42 ± ± Gender (female) 16 (47%) 20 (54%) 0.64 Ethnicity (indigenous) 3 (9%) 2 (5%) 0.67 Donor Age (years ± S.E) 42 ± ± Number of transplants 0.80 First 12 (35%) 11 (30%) Repeat 22 (65%) 26 (70%) Cold ischemia time (hours) 0.60 <12 12 (35%) 10 (27%) (44%) 16 (43%) >18 7 (21%) 11 (30%) 191

192 Continue Table Non-DSA (n=34) DSA (n=37) p-value CDC class I PRA (At transplant) % 9 (27%) 8 (22%) 11-50% 10 (29%) 8 (22%) >50% 15 (44%) 21 (56%) Total HLA-A, -B and -DR Mismatch (18%) 2 (5%) 1 11 (32%) 5 (14%) 2 17 (50%) 30 (81%) 192

193 Table : Immunosuppressants at induction Non-DSA (n=34) DSA (n=37) p-value Calcineurin inhibitors 0.47 Cyclosprine 4 (12%) 2 (5%) Tacrolimus 28 (82%) 34 (92%) None 2 (6%) 1 (3%) Anti-metabolites 0.48 Azathioprine 0 (0%) 0 (0%) Mycophenolic Acid 33 (97%) 37 (100%) None 1 (3%) 0 (0%) Prednisolone 32 (94%) 37 (100%) 0.23 IL-2 receptor antagonist 22 (65%) 22 (59%) 0.81 Antibody 5 (15%) 4 (11%)

194 Specificity of DSA defined by Luminex Most of the DSA in the highly sensitized patients were directed against common antigens such as B7, B8 and B35. For C locus, DSA Cw5 was the most frequent specificity (Figure 5.4.5). For class II, most of the DSA were against DR or DQB1 antigens (Figure 5.4.6). Rare DSA such as DQA and DP were detected in several patients. Specificity of DSA in the highly sensitized patients: DSA Class I Specificity DSA Class II Specificity only only 1 A3 1 DP A3 2 DQ1 3 B18, Cw4 3 DQ3, DRw52, DRw53 4 B35 4 DQ8 5 B35 5 DQA B57 6 DR14 7 B62 7 DR3 8 B7 8 DR3, DRw52 9 B7 9 DR7 10 B7 10 DRw53 11 B8 12 B8 13 Bw6, B45 14 Cw5 15 Cw5 16 Cw5 17 A34, CwI DSA Class I & II Specificity 1 B35, DR15, DP4 2 B62, DR1 3 Cw3, DQ3 4 A2, DR11, DR17, DRw52/ B35, DR13 6 B39, DQ2 7 B55, DR9 8 B62, DQ6 9 B7, DR7 10 Cw12, DR4, DQ3 194

195 Figure 5.4.5: Specificity of HLA class I donor-specific antibodies in highly sensitized patients Figure 5.4.6: Specificity of HLA class II donor-specific antibodies in highly sensitized patients 195