Beta Cell Replacement/Regeneration

Transcription

1 Beta Cell Replacement/Regeneration David M. Harlan, M.D. Captain, U.S. Public Health Service Diabetes Branch/NIDDK July 19, 29 I. Lessons from previous efforts II. Rationale underlying our combined approach III. Current trial plans A. GAD (Diamyd) immunization B. Sitagliptin + lansozaprole C. End points: primary and secondary

2 ACKNOWLEDGEMENTS NIDDK/NIH Bethesda Shira Perl Mahfuzul Khan Klaus Pechhold John Tisdale Kristina I. Rother Kristin Tarbell Eric Liu Lisa Spain Janet Lee Pamela Brooks Terri Wakefield Susanne Pechhold Boaz Hirshberg Benigno J. Digon Radiology Dept, CC/NIH Richard Chang Brad Wood Ziv Nieman Diamyd Medical Anders Essen-Moller Elizabeth Lindner Peter Zerhouni Children s Hospital of Pennsylvania Jake Kushner Geneva M. Stein Lawrence Livermore Lab Bruce Buchholtz University of Washington Jerry Palmer Barbara Brooks-Worrell University of Alberta Wilma Suarez-Pinson Alex Rabinovitch Hospital for Sick Children H. Michael Dosch Amylin Pharmaceuticals Alain Baron Patrick Nelson Kim Chen

3 Experimental autoimmune diabetes (EAD) model RIP B7-1 Pancreas ideal to study the islet victims during autoimmune process Model advantages: 1. Anti-β cell specific autoimmunity inducible so its timing can be controlled 2. Autoimmunity aggressiveness controllable so the β-cell killing pace can be regulated Overlapping Techniques : 1. Confocal immunohistology: gold standard but laborious 2. Flow cytometry based insulin/glucagon/somatostatin/nuclei Glucagon % 75.5% Somatostatin 4.9% Hormone secreting islet cell by type (%) Ins Gcg Sst FCM IHC Ppy

4 Islet cell phenotype changes as diabetes develops Pechhold K, et al. Diabetes 58: , 29 Granularity (SSC) Naïve Diabetic gated on ins + cells diabetic diabetic naïve naïve Conclusions: In diabetic rodents, residual β cells flow cytometric optical Cell size properties (FSC) change: Many β cells increase in size (hypertrophy) Most β cells degranulate (to a finite degree) Many β cells contain less insulin ß cell insulin content 7 gated on ins+ cells Granularity (SSC) diabetic naïve insulin

5 Islet cell subset replication as assessed by BrdU immunofluorescence Almost all BrdU + cells fail to co-stain for insulin or glucagon Virtually no glucagon + cells co-stain for BrdU Most insulin + cells also remain BrdUnegative Conclusions: BrdU incorportation can be detected by fluorescence microscopy Quantification is laborious and complicated by the variable β cell insulin expression level observed in diabetes Few insulin + cells co-stain for BrdU insulin / glucagon / BrdU

6 ß cell proliferation in autoimmune diabetes driven by hyperglycemia treatment after diabetes onset none mice (%) with hyperglycemia Pechhold K, et al. PLoS One. 4: e4827, 29 1 BrdU (i.p.) % R1 CD R R % R7 CD8+ T cell depletion mice (%) with hyperglycemia 1 BrdU (i.p.) R1 R % exog. insulin (pellets) 1 mice (%) with hyperglycemia BrdU (i.p.) insulin / glucagon / somatostatin insulin R BrdU R7

7 Islet Transplantation Procedure Illustration by Giovanni Maki PLoS Medicine, Vol 1, pages , 24

8 Islet transplantation can reliably restore insulin independence to individuals with long standing & brittle T1DM Shapiro et al, NEJM, 2 What did Edmonton do that was different? 1. Improved islet isolation techniques, avoid xenoproteins 2. Infused larger islet dose two donors typically required (>2 in 3%) 3. Immediately infusing the islets post isolation (now thought not to be important) 4. Steroid sparing immunosuppression- FK56, rapamycin, CD25

9 NIH Islet Protocol: Features/Results All pancreata deemed unsuitable for pancreas transplant No more than 2 islet infusions/recipient Measurable islet function in all 6 patients at one year Patient IEQs Infused #1 #2 Total 273K 319K 592K 271K 271K 245K 313K 298K 428K 315K 52K 355K 56K 815K 653K 428K Insulin Independent at 1 year? IEQ/kg 1,85 yes 3,712 never 1,268 x 6 weeks 13,564 yes 11,558 never 6,926 yes Commensurate with the Immune Tolerance Network 1-year follow-up results: 44% insulin independence, 72% with measurable C-peptide (NEJM, 9/28/26) Reference: Solitary islet transplantation for type 1 diabetes mellitus using steroid sparing immunosuppression: The NIH experience. Diabetes Care (December 23).

10 Edmonton Protocol Follow-up Diabetes 25; 54(7): Insulin Independence 1 (percent) Creatinine Clearance 8 (ml/min/1.73 m 2 ) 64% at 1 year Albumin Excretion Rate 6 ( g/min) Serum Triglycerides (mmol/liter) 4 But with immunosuppressive agent toxicity: Pre-Transplant 18 (84 12) 11.87±.7 One Year Post- Transplant 96 (78 18) ±.11 Most Recent 84 (66 12) ±.1 7.5% at 5 years 2 Patient # => Time (years)

11 Solitary Pancreas Transplantation and Patient Survival Journal of the American Medical Association (December 3, 23). Pancreas-Transplant-Alone (n=672) Days Overall (- 146) Days Overall (- 146) Favors Transplantation Favors Conventional Therapy Simultaneous Pancreas-Kidney (n=952) Favors Transplantation Favors Conventional Therapy Relative Risk of Death (± 95% CI) 6.13

12 United Network for Organ Sharing (UNOS) Yearly Pancreas Transplant Frequency Khan M & Harlan DM. Diabetes Care, 29 (in press) Number of Transplants SPK Solitary Pancreas Transplant Year

13 Potential Islet or Islet-like Sources

14 Beta Cell Regeneration Beta Cell Mass Beta Cell Destruction -Autoimmunity -Metabolic abnormalities

15 Beta Cell Regeneration Beta Cell Mass Beta Cell Destruction Conflicting data: Does hyperglycemia promote beta cell growth or impair it? -Autoimmunity -Metabolic abnormalities

16 Does native pancreatic β-cell function recover in transplant recipients? Liu EH, Digon BJ, Hirshberg B, et al. Diabetologia 52: , 29

17 Arginine stimulated C-peptide: Insulin independent patient 26 months post islet transplant Liu EH, Digon BJ, Hirshberg B, et al. Diabetologia 52: , Portal Peripheral 2.5 C-peptide (ng/ml) Time (min) relative to arginine

18 Native pancreas in chronic T1D continues to secrete insulin Lessons/thoughts: Transplant Age Dx Tx Ins BMI recipient Age Age Indep Data suggests that each subject s native pancreas continues to Pancreas 4 Liu EH, Digon BJ, Hirshberg B, et al. Diabetologia 52: , 29 Is [C-peptide] downstream of the native pancreas greater than simultaneously sampled central vein? Islet Yes 2.1 N N A Y Islet Yes 23.7 Y N r 2. Despite up to 1 years immunosuppression, Ythe Y native Y Y g Islet pancreatic 3 57 insulin production No 2.9quite Y small. N i Y Y N Y 3. Pancreas Potential 1 51caveats: Yes 32.8 n Y N Y Y Y Y A. Disease duration? i Pancreas Yes 22.9 N n Y Y Y Y B. Hyperinsulinism? e Pancreas C. Calcineurin phosphatase-based 1 Yes 28.3 immunosuppression? Y N Y Y Y Y 5 of 13 P=.157 N Seconds post infusion make some insulin despite disease duration up to 5 years D. Subject 58 age? 15 1 Yes 28.4 N Y E. Does islet regeneration require a push? Y Y Y N Y N Y N Y 29 of 38 P =.59 Y Y Y N N N N

19 Will exenatide plus immunomodulation (daclizumab) promote β-cell function in chronic T1D Rother KI et al, Submitted Run-In Period (4 months)- intensive insulin therapy Randomization 2 Patients enrolled 16 Patients randomized 4 Discontinued study due to fear of hypoglycemia (n=1),psychosocial issues (n=3) Study Period A (6 months) Daclizumab Study Period B (6 months) No study drug (n=4) Exenatide (n=4) Exenatide (n=4) No study drug (n=4) Exenatide + Daclizumab (n=4) Daclizumab (n=4) (n=2) Exenatide + Daclizumab(n=2) 2 DC d study due to concerns about immunosuppression before 1st dose (n=1), after 2 nd dose (n=1)

20 C-peptide results of patients ON versus OFF Exenatide (irrespective of daclizumab) Rother KI et al, Submitted Lessons from efforts to promote β-cell function in individuals with longstanding T1D: Pancreatic insulin production persists in most Basal individuals with even longstanding T1D..8 Stimulated C-peptide 2. Pancreatic insulin production responds to physiological stimuli, i.e. [ng/ml] A. Exogenous.6 insulin that restores euglycemia decreases pancreatic insulin secretion and B. Mixed meal.4 or arginine stimuli increase pancreatic insulin release BUT 3. We were unable to further promote meaningful pancreatic insulin.2 production New question: Does Screening the adult Run-In human On pancreas Exendin have Off-Exendin the capacity to make new β-cells? T cell assays consistent with ongoing anti-β autoimmunity despite long disease duration, but treatment did not influence the assay results

21 Pine tree β-cell turnover assessed by measuring DNA 14 C content 14 C t 1/2 =573 years! Bomb curve Used to date human neurons to display limited turnover (Spalding KL et al, Cell 25) Later used to date human teeth (Spalding KL et al. Nature 25) Nuclear tests Equilibrium with oceans

22 Human pancreatic islets: quite heterogeneous mini-organs 1% Islet endocrine cell population (%) 6% 2% β-cell range among donors: 28-75% β cells (54%) α cells (35%) δ cells (11%) Also- vascular endothelial cells, dendritic cells, etc Brissova M et al. J Histochem Cytochem 53: , 25

23 Donor β-cell DNA 14 C content 14C 8 Donor => Birth 6 year Hgb A1c # % % β-cell DNA date 195- ' '9 4 Subject # '57 or Non-β-cell DNA b. 1926, date '97 head trauma, 1995-'97 2 HbA1c 5.7% Work by: S. Perl, J. Tisdale, B. Buchholtz, M. Kirby #2 β-cells age approximately 25 years younger than the donor Non β-cell turnover more rapid (newer cells) Beta cell fraction dates to Year

24 Approach 2: Thymidine analog labeling in humans >5 patients received thymidine analogs (BrdU & IdU) in clinical trials Labeling agent (1mg/m 2 ) Radiosensitizing agent, (1mg/m 2 ) Eighteen had autopsies at the NIH CRC. Ten had blocks of the pancreas Nine received IdU One received BrdU

25 Thymidine analog labeling in humans to measure pancreatic beta cell turnover Sex Primary diagnosis Age T2D BW M M glioblastoma multiforme mesothelioma IdU to death 1.5 years 1.8 years Insulin+ cells counted IdU+ & Insulin+ Islets counted These M human osteosarcoma data have been 18 corroborated 53 2 days in non-human 129primate (1%) studies. M astrocytoma days (2%) Lesson: F soft As tissue opposed sarcoma to mice, 31 mature 57 adult 55 days human and non-human primate 4β-cells have M a very mesothelioma limited capacity 45 for cell 74division 71 days and regeneration 31 Hypotheses/new M glioblastoma multiforme directions 55for yes study: 68 8 days 3 1. MEfforts pancreatic designed cancerto promote 7 yes endogenous 83 1 days pancreatic β-cell recovery will 5more likely succeed with younger subjects F esophageal adeno-ca days For individuals with long-standing T1D, exogenous-source β-cell replacement M esophageal adeno-ca years 52 strategies (e.g. xenogeneic, or developed from stem cells) will be required IdU+ cells

26 All cause T1DM mortality improving dramatically Pittsburgh Epidemiology of Diabetes Complications Study Pambianco, G. et al. Diabetes 55: , 26. Cumulative Survival 1% 3.5% 75-8 a mature immune response is characterized 8% 22% by redundancy, raising concern that selective blockade of one potential 5-59 pathway to disease may simply prompt others to take its place. The induction of tolerance seems to be the most promising way forward: we are more likely to win this particular war by gaining the insight 6% needed to negotiate with the immune system than by seeking to bomb it into submission , N = 235, years Edwin A.M. Gale, , M.D. N = 153, years 4% N. Engl. J. Med., , Vol 346: N = , 172, years , N = 167, years , N = 179, years 2% Diabetes Duration (years)

27 Does GAD immunization preserve β-cell function in subjects with recent onset T1D? Ludvigsson J et al. NEJM 359: , 28 Inclusion criteria: 1. Clinically diagnosed T1D within 18 months 2. Age 1-18 years 3. Anti-GAD autoantibody positive 4. C-peptide.1 nmol/l Protocol design: 1. Intensive insulin management: Glycated hemoglobin target 6.5% 2. Placebo controlled, double-blind, randomized trial. 3. GAD-alum (or placebo) 2 µg- administered 1 day and 1 month post randomization 7 randomized 35 GAD-alum 35 placebo Intention to treat analysis 35 analyzed 34 analyzed* * 1 withdrawn due to mononucleosis with icterus Pre-specified study endpoints: Primary: Change in fasting C-peptide 15 months post randomization Secondary: Change in fasting C-peptide relative to baseline at other time points Hgb A1c, glycemia, and insulin requirements

28 Does GAD immunization preserve β-cell function in subjects with recent onset T1D? Ludvigsson J et al. NEJM 359: , 28 Characteristic GAD-alum N= 35 Placebo N= 34 Age (years) Time since diagnosis (months) BMI (kg/m 2 ) Gender (M : F) HLA risk High Moderate Low Fasting C-peptide (nmol/l) Glycated hemoglobin (%) Insulin requirement (U/kg/day) Median anti-gad titer (U/ml) 13.8 ± ± ± : ± ± 1.3%.66 ± ± ± ± : ± ± 1.%.66 ± Baseline characteristics statistically indistinguishable

29 Does GAD immunization preserve β-cell function in subjects with recent onset T1D? Ludvigsson J et al. NEJM 359: , 28 Fasting C- peptide Δ (nmol/l) All protocol participants GAD-alum (N = 35) Placebo (N = 34) P= Months post randomization Patients treated within 6 months of diagnosis P=.1 (N = 11) (N = 14) P= GAD Conclusions: GAD Rx did not achieve primary outcome GAD treatment did preserve more function at 3 mo Subgroup treated within 6 mo intriguing

30 Does GAD immunization preserve β-cell function in subjects with recent onset T1D? Characteristic Age (years) Time since diagnosis (months) BMI (kg/m 2 ) Gender (M : F) HLA risk High Moderate Low Fasting C-peptide (nmol/l) Glycated hemoglobin (%) Insulin requirement (U/kg/day) Median anti-gad titer (U/ml) Ludvigsson J et al. NEJM 359: , 28 GAD-alum N= ± ± ± : ± ± 1.3%.66 ±.3 61 Baseline Placebo N= ± ± ± : ± ± 1.%.66 ± At month 3 GAD-alum Placebo N= 35 N= 34 Lessons/thoughts: 1. IF GAD immunization is effective, need to initiate therapy early after disease onset. 2. Any perceived effect wanes. Booster dose at 3 months? 3. Immunotherapy may not be sufficient?.13 ± ± 1.8%.96 ±.39.9 ± ± 1.2%.96 ±.25 Conclusions part 2: GAD had no effect on insulin requirements or glycated hemoglobin at 15 or 3 mi

31 Combination therapy with a dipeptidyl peptidase-4 inhibitor and a proton pump inhibitor restores normoglycemia in NOD diabetic mice Suarez-Pinzon WL, Cembrowski GS, Rabinovitch A. Diabetologia 29 (in press). 6-8 wk female NOD When non-fasting BG > 1 mmol/l Randomize to 12 weeks treatment with: 1. Vehicle for DPP-41+ PBS 2. DPP-4i (1 mg/kg) by gavage each day 3. PPI 3 mg/kg sq BID 4. Both 2 & 3 Blood Glucose (mmol/l) Vehicles DPP-4i PPI DPP-4i DPP-4I + PPI Restored Euglycemia ( mmol/l) /7 3/8 3/9 6/8 /15 4/14 4/14 17/21 p<.1 Combination Rx also: significantly increased plasma [C-peptide] and pancreatic insulin content

32 Will combining regenerative stimuli and selective immunomodulation preserve β-cell function in new onset T1D Inclusion criteria: 1. Clinically diagnosed with T1D within 4 months 2. Age 16-3 years 3. Anti-GAD autoantibody positive 4. Randon C-peptide.2 nmol/l 5. BMI between 19 and 28 kg/m randomized 41 active Rx 41 placebo Protocol design: 1. Glycated hemoglobin target 6.5% via insulin & web-based BG management 2. Placebo controlled, double-blind, randomized trial. 3. GAD-alum (or placebo) 2 µg- administered at time, & 4 and 12 weeks later 4. Sitagliptin (or placebo): 1 mg/day for adults (pediatric age adjusted) 5. Lansozaprole (or placebo): 3 mg BID (pediatric age adjusted) Endpoints: Primary: β-cell function: calculated by subtracting the mixed meal stimulated C- peptide area under the curve (AUC) at baseline from the 12 month value Secondary: Linkage between HbA1c and total daily insulin requirement, and between HbA1c and mean stimulated C-peptide AUC Immune phenotyping (autoantibody titers, T cell assays)

33 Conclusions Minimal islet function is preserved in most patients with T1D, but no therapy yet shown to promote meaningful islet functional recovery While hyperglycemia promotes mouse beta cells to divide, our data suggests that pancreatic beta cells in the adult human and non-human primate rarely if ever divide. We do NOT know if that suggests they cannot divide (for example, in insulin resistant states or during pregnancy)