In Vivo Generation of Oligoclonal Colitic CD4 T-Cell Lines Expressing a Distinct T-Cell Receptor V

Transcription

1 GASTROENTEROLOGY 2005;128: In Vivo Generation of Oligoclonal Colitic CD4 T-Cell Lines Expressing a Distinct T-Cell Receptor V ANA C. ABADÍA MOLINA,* ATSUSHI MIZOGUCHI, WILLIAM A. FAUBION,* YPE P. DE JONG, SVEND T. RIETDIJK,* MARTINA COMISKEY,* KAREEM CLARKE,* ATUL K. BHAN, and COX TERHORST* *Division of Immunology, Beth Israel Deaconess Medical Center, Boston; and Department of Pathology, Massachusetts General Hospital, Boston, Massachusetts Background & Aims: Transplantation of wild-type (H-2k) bone marrow into tg 26 mice (BM tg 26) induces colitis, characterized by T-helper cell type 1 activation in the lamina propria. Here we determined whether pathogenic T-cell clones could be derived by serial adoptive transfers into healthy tg 26 recipients, starting with the population of CD4 cells in the mesenteric lymph nodes of BM tg 26 mice. Methods: CD4 cells purified from the mesenteric lymph nodes of colitic BM tg 26 mice were adoptively transferred into a second group of healthy tg 26 recipients. Mesenteric lymph node CD4 cells from the second group of mice were then used for consecutive transfers. Lamina propria CD4 cells isolated from each mouse with colitis were analyzed for their cytokine profile and for their T-cell receptor V repertoire. Results: CD4 T cells maintained a dominant T-helper 1 phenotype after multiple transfers (<8) into recipient tg 26 mice. A single T-cell receptor V was enriched (as much as 90%) in 8 CD4 T-cell lines: V 8S3, V 8S1/2, V 10S1, or V 14. Sequence analyses of the T-cell receptor V s showed clonality or the presence of a very restricted number of clones within each line. Adoptive transfers of the oligoclonal lines into either C3H Rag / or severe combined immunodeficiency disease mice (H-2k) also induced colitis, whereas transfers into BALB/c Rag / or severe combined immunodeficiency disease mice (H-2d) did not. Conclusions: Colitis-inducing CD4 T-helper 1 cell clones can be obtained by enrichment through sequential adoptive transfers of CD4 cells from mesenteric lymph nodes. Distinct dominant T-cell receptor V s in each cell line responded to antigens presented by class II major histocompatibility complex. Ulcerative colitis and Crohn s disease, collectively referred to as inflammatory bowel diseases (IBD), are chronic spontaneously relapsing disorders that seem to be immune mediated but also contain genetic and environmental components. Rodent models of chronic intestinal inflammation provide useful tools for dissecting the pathways involved in the pathogenesis of IBD. Targeted disruption or transgenic expression of different genes, manipulation of selected lymphocyte subsets, and spontaneous gene mutations all show that immune-regulatory dysfunction leads to chronic intestinal and systemic inflammation Although both adaptive and innate immune responses are involved in experimental models of IBD, the notion that the adaptive immune system plays a dominant role in the pathogenesis of colitis has been developed. This notion is based on findings that mice that lack an adaptive immune system (eg, severe combined immunodeficiency disease [SCID] and Rag / ) are perfectly capable of coping with intestinal bacteria through innate immune defenses. 14 Colitis develops in most experimental models when effector T helper 1 (Th1) cells are not kept in check by regulatory CD4 CD25 cells (Treg cells). 1,15 Although it is generally accepted that colitis-inducing CD4 T cells recognize bacterial and/or self-antigens that are induced by the presence of bacteria, antigenspecific pathogenic T-cell clones have not been isolated. This is because in vitro activation and cloning of CD4 mesenteric lymph nodes (MLNs) and lamina propria lymphocytes (LPLs), which can transfer disease into a recipient mouse, has proven difficult. We hypothesized that colitis-inducing Th1 clones might be derived in vivo, because repeated exposure to bacteria-derived antigens during consecutive adoptive transfers (ATs) would select a limited number of pathogenic clones in the MLNs. The BM tg 26 murine colitis model is especially suitable to address this concept because of the lack Abbreviations used in this paper: APC, antigen-presenting cell; AT, adoptive transfer; BM, bone marrow; DAI, Disease Activity Index; FITC, fluorescein isothiocyanate; IFN, interferon; IL, interleukin; LPL, lamina propria lymphocyte; MHC, major histocompatibility complex; MLN, mesenteric lymph node; PE, phycoerythrin; TCR, T-cell receptor; Th, T helper; Treg, T regulatory cell; WT, wild type by the American Gastroenterological Association /05/$30.00 doi: /j.gastro

2 May 2005 CLONING OF COLITIS INDUCING CD4 CELLS 1269 of Treg cells in this mouse. 16,17 We have previously shown that transplantation of bone marrow (BM) derived from wild-type (WT) mice into adult tg 26 mice (BM tg 26) results in the development of a highly activated T-cell population in peripheral lymphoid tissues and the colonic mucosa. 12 These Th1 cells induce wasting and colitis in the recipients 4 5 weeks after transplantation. Intestinal inflammation is limited to the colon and is characterized by crypt hyperplasia, extensive mononuclear cell infiltration, and crypt abscesses. Furthermore, CD4 cells from the MLN of BM tg 26 mice with colitis were able to transfer disease to a healthy tg 26 mouse. 12,16 It has also been shown that naive and memory CD4 T cells expand when introduced in lymphopenic animals (ie, into tg 26, SCID, or Rag / mice). 16,18 Thus, tg 26 mice are useful recipients for experiments designed to test the hypothesis that serial ATs will lead to a dramatic enrichment of dominant pathogenic CD4 T cells. In this study we examined whether sequential ATs of a pathogenic MLN CD4 cell population into healthy tg 26 mice would lead to expansion of 1 or more disease-inducing clones. We report that CD4 cells derived from the MLNs of BM tg 26 mice with colitis cause disease in recipient mice with great efficiency. Disease-inducing cell lines can be derived from either splenic or MLN CD4 cells from BM tg 26 mice with colitis as long as subsequent sequential ATs use MLN CD4 cells. The derived cell lines maintain their Th1 phenotype after multiple transfers into lymphopenic tg 26 mice with expansion of 1 dominant disease-inducing clone, as judged by the T-cell receptor (TCR) V. The oligoclonal cell lines adoptively transfer disease into C3H Rag / or C3H SCID mice, but not when transferred into BALB/c Rag / or BALB/c SCID. Materials and Methods Mice tg 26 mice 19 were maintained at the Beth Israel Deaconess Medical Center Research Animal Facility. CBA, BALB/c SCID, and C3H/HeSn SCID mice were purchased from Jackson Laboratories (Bar Harbor, ME). C3H/ HeN Rag2 / and BALB/c Rag2 / were purchased from Taconic Farms (Germantown, NY). The mice were kept under specific pathogen free conditions. Recipient mice were 8 to 12 weeks old. SCID and Rag2 / mice were placed into cages previously used by tg 26 mice for at least 1 week to ensure similarities between recipients in luminal microflora. Cell and Bone Marrow Purification and Cell Transfers tg 26 mice were irradiated with 4.5 gray, and T cell depleted CBA BM cells were injected into the tail vein, as described previously. 12 BM tg 26 mice were killed when signs of diarrhea, hunching, and wasting disease became apparent. MLN and spleen cells were obtained as described previously. 20 Then, CD4 cells were purified by using negative selection columns (R&D Systems, Minneapolis, MN). CD4 cells (0.5 to ) were injected into the tail vein of 450 rad irradiated tg 26 mice. In subsequent transfers, to MLN T cells were injected into healthy tg 26 or C3H Rag2 / or SCID, BALB/c Rag2 / or SCID recipients. Disease Activity Index and Histology Mice were monitored and scored for colitis as described previously. 20,21 Briefly, mice with colitis were killed and scored for Disease Activity Index (DAI). Proximal, median, and distal colon samples were collected to determine the histology score. Histology grades were assigned in a blinded fashion by 1 pathologist (A.K.B.). Cytokine and Flow Cytometric Analysis LPLs 21 ( ) were placed in 96-well plates and stimulated with plate-bound anti-cd3 antibody (10 g/ml; 145.2C11 from PharMingen, San Diego, CA) or antigenpresenting cells (APCs) pulsed with cecal antigen from donor mice 13 for 36 hours. Supernatants were collected, and interleukin (IL)-4, IL-10, and interferon (IFN)- cytokines were detected by ELISA (OptEIA) by following the manufacturer s instructions (BD PharMingen). The limit of detection of the cytokine assay was 125 pg/ml for IL-10 and IFN- and was pg/ml for IL-4. LPLs ( ) were used for analysis of the TCR V family repertoire by flow cytometry in a FACScan (BD Bioscience, San Diego, CA). Cells were doublestained with phycoerythrin (PE) or fluorescein isothiocyanate (FITC)-conjugated anti-cd4 (clone RM4-5) and the following conjugated antibodies: PE V 2 (B20.6), PE V 3 (KJ25), PE V 4 (KT4), PE V 5 (MR9-4), PE V 6 (RR4-7), PE V 7 (TR310), PE and FITC V 8.1/2 (MR5-2), PE V 8.3 (B21.14), PE and FITC V 9 (MR10-2), PE V 10 (B21.5), FITC V 11 (RR3-15), FITC V 12 (MR11-1), PE V 13 (ÌÑ 12.3), FITC V 14 (14.2), FITC V 17 (KJ23), PE CD45Rb (16A), PE CD62L (MEL-14), and PE CD69 (H1.2F3). PE and FITC isotype-matched controls were used as the 0 values. All antibodies were purchased from BD PharMingen. Complementary DNA Synthesis, Polymerase Chain Reaction Amplification, and DNA Sequence Analysis For DNA sequence analysis, LPLs ( ) were frozen in TRIZOL reagent (GIBCO BRL, Grand Island, NY) and stored at 80 C until total RNA extraction was performed by following the manufacturer s instructions. Complementary DNA

3 1270 ABADÍA MOLINA ET AL GASTROENTEROLOGY Vol. 128, No. 5 synthesis used the SuperScript first-strand amplification system (GIBCO BRL) with the following V gene-specific sense oligonucleotide primers: V 8, 5=-GAGGCTGCAGTCACCCAAAGC-3=; V 8S1, 5=-CATTACTCATATAGTCGCTGAC-3=; V 8S2, 5=- TACCCCCTCTCAGACATCAGTGTACT-3=; V 8S3, 5=-TGCT- GGCAACCTTCAAATAGGAGATGTC-3=; and V 10, 5=- GAATTCATGGGCTGTAGGCTCCTAC-3=. A specific C primer was used as antisense: 5=-CTTGGGTGGAGTCACATT- TCTC-3=. Amplified products were cloned with the TOPO TA cloning kit (Invitrogen, Carlsbad, CA). Positive colonies were randomly selected, and cultures were grown in Luria-Bertani medium containing 50 mg/ml ampicillin. Plasmid DNA was isolated from bacterial cultures by using a QIAprep spin miniprep kit (Qiagen, Valencia, CA). Sequencing reactions were performed by cycle sequencing with M13 forward and reverse primers in a Perkin Elmer (Wellesley, MA) automatic DNA sequencer. Statistical Analyses Parametric data are shown as mean SD. Nonparametric data (ie, DAI and histological data) were analyzed with the Mann Whitney test. For statistical purposes, the median and 25th and 75th percentile values were determined. P values.05 were considered significant. Statistical analyses were performed with Prism 3 software (GraphPad, San Diego, CA). Results Serial Adoptive Transfers of CD4 T Cells Isolated From the Mesenteric Lymph Nodes of Colitic Mice Into Healthy Immunodeficient Mice To show that CD4 T cells isolated from the MLNs of mice with enterocolitis induce disease upon serial ATs into healthy tg 26 recipients, the experimental design outlined in Figure 1A was used. For this purpose, mice that developed chronic colitis, as judged by DAI scores and histology (Figure 1B and F), were used 4 weeks after transfer of WT CBA (H-2k) BM into 3tg 26 (H-2k) mice (BM tg 26). CD4 T cells ( ) were purified from the MLNs of these 3 colitic tg 26 mice by using negative selection columns. Because BM tg 26 mice do not develop Tregs, 17 removal of CD4 25 cells was not required. In AT 1, purified MLN CD4 cells were transferred into 1 healthy tg 26 recipient (no. 198). When this tg 26 mouse developed colitis, its MLN CD4 cells were used for a second AT into 2 healthy tg 26 recipient mice (no. 189 and 190), and this process was repeated in AT 3 (no and 2216) and AT 4 (no and 2221). All tg 26 recipients in this series (n 7) developed disease within 2 to 4 weeks (Figure 1C, D, and F), with a median DAI of 6.2 (25th percentile, 5.5; 75th percentile, 7) and a median histology score of 4 (25th percentile, 3.5; 75th percentile, 6.2). When CD4 cells isolated from the lamina propria of each mouse ( LPLs per mouse; n 7) were analyzed for surface expression of TCR V families, an enrichment of V 8S1/2 CD4 T cells was observed. After 4 transfers (AT 4; tg 26), approximately 88% of the CD4 T cells expressed V 8S1/2, whereas the remaining CD4 cells expressed low levels of the other 14 V s (Figure 1E). A similar outcome was detected in line 2 of the first experiment (60% of CD4 cells were V 8S1/2). We concluded from this pilot experiment that serial ATs into healthy tg 26 mice induced colitis; this suggested that oligoclonal cell lines (lines 1 and 2) had been established. Next, cells of lines 1 and 2 were tested for their ability to cause colitis in a different strain of major histocompatibility complex (MHC)-matched mice. During AT 1 and AT 4, the remaining MLN CD4 cells were adoptively transferred into C3H/HeN Rag / recipients ( cells per mouse) with the same class II MHC (H-2k). MHC class II mismatched BALB/c Rag / recipient mice (H-2d) were used as controls. We found that colitis developed in C3H/HeN Rag / recipients either after AT 1 (Figure 1A and C) or after AT 4 (Figure 1D) but that no disease was detected in the allogeneic BALB/c Rag / (n 3) recipients (Figure 1D). The data are indicative of a MHC class II restriction of the CD4 cell lines. To determine whether serial ATs caused deviation of the Th1 phenotype of the BM tg 26 CD4 cells, production of IL-4, IL-10, and IFN- by lamina propria CD4 cells was measured. 13,20 LPL cells from colitic BM tg 26 mice in vitro stimulated either with anti- CD3 or with APCs pulsed with cecal antigens secreted high levels of IFN- and undetectable levels of IL-4 and IL-10 (Figure 1G). CD4 cells isolated from the MLNs of BM tg 26 and AT 4 tg 26 also secreted Th1 cytokines (Figure 2A). Thus, the dominant Th1 cytokine phenotype of MLN and LPL CD4 T cells was maintained after 4 sequential ATs. The same LPL and MLN CD4 cells had a memory/effector phenotype, expressing a high level of CD69 and low levels of CD62L and CD45RB 19 (Figure 2B). Collectively, our findings indicate that sequential ATs of MLN CD4 cells from colitic tg 26 mice into healthy tg 26 recipients lead to enrichment of colitis-inducing CD4 cells (lines 1 and 2). Thus, the oligoclonal cell line maintains a Th1-like response and a memory/effector phenotype and causes disease in a class II MHC restricted fashion, as shown in the Rag / model of colitis. 18 Furthermore, the pilot experiment showed that Treg cells were absent in the adoptively transferred cell lines.

4 May 2005 CLONING OF COLITIS INDUCING CD4 CELLS 1271 Figure 1. Sequential adoptive transfers (ATs) of colitis-inducing CD4 T cells isolated from mesenteric lymph nodes (MLN) of AT experiment A (lines 1 and 2). (A) Schematic outline of the results of sequential ATs into tg 26, C3H/HeN Rag /, or BALB/c Rag / mice. After colitis developed in BM tg 26 mice, CD4 cells purified from the MLNs were transferred into healthy tg 26 mice (AT 1). Transfer of MLN CD4 cells into tg 26 mice was repeated in AT 2, AT 3, and AT 4. Parallel transfers of MLN CD4 cells into C3H/HeN Rag / or BALB/c Rag / mice were also performed in AT 1 and AT 4. (B) Disease activity index (DAI) and histology scores of the 3 BM tg 26 mice with colitis. (C) DAI and histology scores of tg 26, C3H/HeN Rag /, and BALB/c Rag / mice after AT 1 (mouse AT 1 no. 198, 2253, and 2256). (D) DAI and histology scores of tg 26, C3H/HeN Rag /, and BALB/c Rag / mice with colitis after AT 4. ATs of the CD4 cells from mice at AT 3 in lines 1 and 2 induced colitis in tg 26 and C3H/HeN Rag / mice, in contrast to BALB/c Rag / mice, as in AT 1. Statistical significance was calculated by comparing the DAI or histology scores of BALB/c Rag / mice with those of tg 26 mice (DAI, P.008; histology, P.008) or those of C3H Rag / mice (DAI, P.05; histology score, P.05) from AT experiment A. (E) Flow cytometric analyses of line 1 LPL cells showing an enrichment of V 8S1/2 CD4 cells after AT 4. Lamina propria CD4 T cells were analyzed for expression of 15 V family members by flow cytometry, as described in Materials and Methods. Shown here is the enrichment of the major V 8S1/2 population from MLN tg 26 mice of cell line 1. Cells were gated for lymphocyte side scatter vs CD4 expression. (F) Representative histology of inflamed colon tissue. Histology of colon tissue of a healthy WT (CBA C57BL/6) mouse, of a diseased BM tg 26 mouse (DAI score, 6), a tg 26 mouse (AT 1 no. 198; DAI score, 5), and a tg 26 mouse (AT 4 no. 2219; DAI score, 3.5) is shown (original magnification, 10 ). (G) Cytokine production by lamina propria CD4 cells of experiment A. IL-4, IL-10, and IFN- cytokines were measured by enzyme-linked immunosorbent assay in the supernatant of colon LPL from BM tg 26 (n 3) mice with plate-bound anti-cd3 (145.2C11) and APCs pulsed with cecal antigens for 36 hours. Ag, antigen.

5 1272 ABADÍA MOLINA ET AL GASTROENTEROLOGY Vol. 128, No. 5 Figure 2. Cytokine production of MLNs and memory/effector phenotype of MLN and LPL cells from experiment A. (A) Cytokine production by MLN cells. IL-4, IL-10, and IFN- cytokines were measured by enzyme-linked immunosorbent assay in the supernatant of MLNs from BM tg 26 (n 3) and AT 4 (n 2) cultured with plate bound anti-cd3, and APCs were pulsed with cecal antigen for 36 hours. (B) Flow cytometric analyses of MLNs and LPLs showing a memory/effector phenotype of CD4 cells. Flow cytometry diagrams were generated with anti-cd4 and anti-v 8S1/2, -CD62-L, -CD45RB, and -CD69 of colonic lamina propria and MLN CD4 cells isolated from colitic tg 26 mice, as described in Materials and Methods. Representative data are shown of 1 (AT 4 no. 2219; line 1) of 30 mice analyzed in 2 sequential adoptive transfer experiments. The percentage of CD69 ranged from 43% to 90% in LPLs and from 12% to 60% in MLNs. The percentage of CD62L ranged from 1% to 11% in LPLs and from 1% to 16% in MLNs. The percentage of CD45RB hi ranged from 0% to 2% in LPLs and from 0% to 2.5% in MLNs. Ag, antigen. Clonal Expansion in 5 CD4 Cell Lines Obtained by In Vivo Sequential Adoptive Transfers To show that the principle established in the pilot experiment could be extended to other CD4 cell lines, a large serial AT experiment was set up that started with either spleen or MLN CD4 cells isolated from BM tg 26 mice. MLN and splenic CD4 T cells from 15 colitic BM tg 26 mice (DAI: median, 6.1; 25th percentile, 5.7; 75th percentile, 6.7; histology score: median, 5.3; 25th percentile, 2; 75th percentile, 6) were purified and used for the first step (AT 1) of serial AT experiments (Figure 3). In the next stages, AT 2 to AT 7, only MLN CD4 T cells from colitic mice were used for ATs, transferring at cells per recipient. Lamina propria CD4 T cells from each mouse, ranging from 0.1 to LPLs per mouse (n 26), were used for analysis. The genesis of the 5 in vivo derived cell lines (lines 3 to 7) is depicted in Figure 3. The degree of enterocolitis was high in all recipient mice of each cell line, as assessed by the DAI and histology after each AT (DAI: median, 6.2; 25th percentile, 5.2; 75th percentile, 7; histology score: median, 5; 25th percentile, 3.3; 75th percentile, 6; from AT 2 to AT 5; n 18). The outcomes of these ATs clearly showed that colitis-inducing cells are present in both spleen and MLNs of the BM tg 26 mice. Thus, the pathogenic CD4 cells migrate throughout the body and are found in relatively large numbers in lymphoid organs other than those at the site of inflammation. More importantly, because the spleen yielded on average CD4 T cells per mouse (n 26), a total of 10 lines were generated, starting with the spleen-derived CD4 cells. Enrichment of 1 TCR V was studied in greater detail in 5 lines, as indicated in Figure 3. Flow cytometric analyses of lamina propria CD4 cells with a panel of 15 anti-v monoclonal antibodies showed that in each cell line, 1 predominant cell line expressing 1 V developed. V 8S1/2-expressing CD4 T cells were dominant in lines 1, 2, and 3 (Figure 1E and Table 1). In lines 4, 5, and 6, CD4 cells bearing V 8S3 were predominant (Table 1). Because TCR V 10 was

6 May 2005 CLONING OF COLITIS INDUCING CD4 CELLS 1273 Figure 3. Schematic outline of sequential adoptive transfers of colitis inducing CD4 T-cell lines. Upon transfer of WT bone marrow into 15 tg 26 recipients, colitis developed after 4 weeks; MLN CD4 cells were isolated and transferred sequentially as outlined (line 3). CD4 cells from the spleens of the same 15 colitic animals were used for lines 4 to 7. Sequential transfers of CD4 cells from MLNs were performed into individual tg 26 mice as indicated (numbers). CD4 MLN cells were also transferred into tg 26, C3H/HeN SCID, or BALB/c SCID mice in cell line 3 and cell line 4 or into C3H/HeN Rag / and BALB/c Rag / mice in cell line 5. Disease was scored for each mouse as indicated in Materials and Methods. BIDMC, animal research facility, Beth Israel Deaconess Medical center; J, Jackson Laboratories; T, Taconic Laboratories. expressed on the surface of the colitis-inducing CD4 T-cell line (line 7) and because a V 14-expressing oligoclonal cell line (line 8) was found in an independent AT experiment (data not shown), we conclude that a limited diversity of TCR V families is used by the major pathogenic CD4 cells in the tg 26 model. Although the 2 major CD4 cells in line 3 after AT 2 expressed both V 8S1/2 and V 8S3, the V 8S1/2 cell line expanded between AT 2 and AT 5 (83%), and V 8S3-expressing cells represented a small minority (4% at AT 5; Table 1). To refine these analyses, V 8S1-, V 8S2-, and V 8S3-specific polymerase chain reactions were conducted with complementary DNAs obtained from purified colonic CD4 cells from line 3. As judged by CDR3 nucleotide sequence analyses of the polymerase chain reaction products, expansion of 2 V 8S2 clones (V 8S1 gave no product) identified by their diversity region of TCR flanked N-nucleotides and J sequences GQGKD/J 1S2 and GRTGG/J 2S5, seemed to have taken place (Table 2). Remarkably, clone GQGKD was not found in the AT 5 no. 29 and AT 4 (no. 86) recipients, in which the GRTGG/J 2S5 clone was dominant (Table 2). In mouse AT 5 no. 29, 75% of the 2 remaining V 8S2 sequences were GRTGG/J 2S5. Nu- Table 1. Flow Cytometric Analyses of the Main V Families Enriched on the Surface of the Lamina Propria CD4 Cells Isolated From 1 Sequential Adoptive Transfer Experiment, Lines 3 to 7 AT 2 AT 3 AT 4 AT 5 AT 6 AT 7 Experiment B Lines a Mouse b BMT LPL SD #62 #63 #67 #86 #5 #6 #5 #2 #90 #94 #25 #20 #12 #29 #30 #32 #73 #62 F c V 8S1/ ND ND d /5 V 8S /5 V ND /5 BMT, bone marrow transplant; ND, not determined. a Lines generated in this experiment are indicated by numbers. b LPL cells from the mice indicated, numbers match the transfer experiment outlined in Figure 3. c Frequency of a particular enriched V in this experiment. d Enriched V in each line is bold.

7 1274 ABADÍA MOLINA ET AL GASTROENTEROLOGY Vol. 128, No. 5 Table 2. CDR3 Sequence Analysis of TCR V 8S1 and V 8S3 in LPL Cells From the Last 3 Connected Mice in Line 3 Line 3 BM AT 1 #21 AT 2 #62 AT 3 #86 AT 4 #90 AT 5 #29 Mouse AT 3 #86 Mouse AT 4 #90 Mouse AT 5 #29 V NDN J F V NDN J F V NDN J F GQGKD 1S2 13/16 8S2 GRTGG b 2S5 3/6 8S2 GRTGG 2S5 3/16 8S2 GRTGG 2S5 6/8 (8.5%) a PTGG c 2S5 1/6 (50%) (83%) GGQN 2S5 1/6 DSGV 2S1 1/6 DSGV 2S1 2/8 KQA 2S2 5/8 KQA 2S2 9/13 KQA 2S2 22/24 DAAGA 2S3 1/8 DAAGA 2S3 2/13 DAAGA 2S3 1/24 8S3 DAED 2S1 1/8 8S3 8S3 (19%) (17%) EWGR 1/13 (4%) EQGV 1/13 EDSY 2S1 1/8 EDSY 2S1 1/24 NDN, hypervariable diversity region of TCR flanked N-nucleotides region; F, frequency. a Percentage of CD4 LPLs bearing the V indicated. b Predominant CDR3 protein sequence is in bold. c Shared sequence motifs are underlined. cleotide sequence analysis of the CDR3 region showed that KQA/J 2S2 remained the major CD4 clone (22/ 24) of the V 8S3 cells, and other sequences showed the pattern described for V 8S2 in this line. These findings support the notion that oligoclonal cell lines had been selected by the serial AT procedures. To examine whether a similar selection pattern would occur if the initial CD4 cell population in AT 1 was derived from the spleen, cells of lines 4 and 5 were used for analysis. Lines 4 and 5 were related because half of the MLN CD4 cells from AT 2 no. 63 were used to initiate line 4 (AT 3 no. 5) and the other half for line 5 (AT 3 no. 6; Figure 3 and Table 1). Two major CD4 cell populations were isolated from the lamina propria of AT 2 no. 63 expressing either V 8S1/2 or V 8S3. In line 4, the V 8S3-expressing CD4 cells were enriched from 35% of the cell population (AT 3 no. 5) to more than 78% after AT 6 in 3 tg 26 recipients: no. 68 (80%), no. 70 (78%), and no. 73 (90%; Table 1 and data not shown). Only 3% to 9% of the cells after AT 6 expressed V 8S1/2. The dominant CD4 cells in line 5 also expressed V 8S3 AT 6 no. 62 (72%), whereas V 8S1/2-bearing CD4 cells represented the minority (Table 1). Two other lines of cells derived from the splenic CD4 T cells were enriched for either V 8S3 CD4 cells (line 6) or V 10S1 cells (line 7; Table 1). As judged by DNA sequencing, lines 6 and 7 comprised only 1 clone (Table 3). The fact that the major dominant CD4 T cells from these cell lines frequently carry V 8S1/2 or V 8S3 is consistent with the observation that these TCR V families are prominent in the CD4 cell population in the lamina propria of the healthy WT donor mice (average V 8S1/2 CD4 LPLs, 17.1% 4.7%; V 8S3 CD4 LPLs, 11.6% 2.2%; n 5). These same V families are also prominent in the CD4 LPLs of colitic BM tg 26 mice (average V 8S1/2, 21% 17%; V 8S3, 18% 17%; n 12). Taken together, these findings indicate that a dramatic enrichment of a dominant CD4 cell clone can be achieved by a stochastic process that operates within a limited range of TCR V s. Table 3. CDR3 Sequence Analysis of TCR V 8S3 and V 10S1 in LPL Cells From Lines 6 and 7 Line 6 Line 7 BM AT 2 #63 AT3#5 AT 4 #20 BM AT 2 #67 AT3#2 AT 4 #12 Mouse AT 4 #23 Mouse AT 4 #12 V NDN J F V NDN J F 8S3 DAEN b 2S5 15/17 10S1 RKQGDA 1S2 17/17 (79%) a DR 2S5 2/17 (87%) NDN, hypervariable diversity region of TCR flanked N-nucleotides region; F, frequency. a Percentage of CD4 LPLs bearing the V indicated. b Predominant CDR3 protein sequence is in bold.

8 May 2005 CLONING OF COLITIS INDUCING CD4 CELLS 1275 Figure 4. DAI and histology scores of tg 26, C3H/HeN Rag /,or SCID and BALB/c Rag / or SCID mice in lines 3, 4, and 5. CD4 MLN cells from mice at AT 5 in lines 3 and 4 and at AT 6 and AT 7 in line 5 were transferred into tg 26 (n 5 and n 3, respectively) and C3H/HeN Rag / (n 3) or C3H/HeN SCID (n 3) as outlined in Figure 3. C3H/HeN Rag / or SCID mice develop colitis, as do their tg 26 counterparts. *P.05, calculated comparing DAI or histology scores of BALB/c Rag / or BALB/c SCID mice with those of tg 26, C3H Rag /,orc3h SCID mice. Median and independent values for DAI and histology scores are shown. To determine the effect of colitis-inducing clones on the induction of disease, a third experiment was attempted. V 8S1/2 CD4 cells were isolated from the MLNs of 3 BM tg 26 mice by flow cytometry: these comprised 18.4% of all CD4 MLN cells. The V 8S1/2 CD4 cell fraction ( ) and the V 8S1/2-negative CD4 cell fraction ( ) were each injected into 1 healthy tg 26 mouse. Both mice developed colitis, and MLN cells were sequentially transferred 2 more times (AT 2 and AT 3). In the mouse injected with positively selected cells, 99.5% of CD4 cells were V 8S1.2. CD4 LPLs of the mouse injected with the V 8S1/2- negative CD4 cell fraction expressed predominantly V 8S3 at AT 3 (68.5%). Thus, neither depletion nor enrichment of the major V families in the MLN of colitic mice affected the outcome of the disease. Adoptive Transfers of Pathogenic CD4 Cells From tg 26 Mice Induce Colitis When Transferred Into C3H/HeN Rag / and C3H/HeSn Severe Combined Immunodeficiency Disease Mice Some of the MLN CD4 cells isolated from mouse AT 5 no. 29 of cell line 3 were divided in 3 and were transferred into C3H/HeSn SCID or BALB/c SCID mice (AT 6; see outline in Figure 3). Similar ATs were made with cells from line 4, AT 6. All C3H/HeN SCID mice developed colitis (n 3) at the same time as their tg 26 counterparts (n 5), but with a lesser disease severity (Figure 4). In contrast, no disease was found in the class II MHC mismatched BALB/c SCID (H-2d) recipients (n 4). MLN cells of line 5 were transferred into C3H/HeN Rag / and BALB/c Rag / mice in AT 6 and AT 7 (Figure 3). Once again, disease developed only in the C3H/HeN Rag / mice (n 3), but no colitis was observed in BALB/c Rag / recipients (n 3; Figure 4). Cells maintained their Th1 phenotype upon AT into C3H/HeN Rag / or C3H/HeN SCID recipients (Figure 1G, Table 4, and data not shown). The C3H Rag /, C3H SCID, BALB/c Rag /, and BALB/c SCID mice were purchased from 2 different companies (see Materials and Methods). Because the microflora of all these mice was reportedly based on the Schaedler flora and because the mice were kept under specific pathogen free conditions, the bacterial flora in the colon of these mice might be similar. Nevertheless, Rag / and SCID mice were placed in cages previously used by tg 26 mice and were kept there for at least 1 week upon arrival. Because these experiments show that BM tg 26- derived MLN cells transferred into either C3H/HeN Rag / (H-2k) or C3H/HeSn SCID (H-2k) mice caused disease, we conclude that intestinal flora based peptide antigens were shared between C3H/HeN Rag / and C3H/HeSn SCID recipients even though the mice were from different sources. Adoptive transfer of cells into C3H/HeSn SCID or C3H/HeN Rag / (H-2k) mice invariably ended the propagation of the cell line, even when the cells were adoptively transferred in an early phase of the enrichment procedure (AT 1; Figure 1). Because of the notion that colitis-inducing cells respond to a multitude of bacterial antigens, the lack of proliferation of BM tg 26-derived MLN cells in the BALB/c Rag / and BALB/c SCID recipients provides strong evidence that a successful AT is dependent on the presence of the correct MHC class II antigen. Table 4. Detection of IFN-, IL-4, and IL-10 in the Supernatant of LPL Lymphocytes From tg 26 and C3H/HeN Mice With Colitis After Incubation With Plate-Bound Anti-CD3 for 36 Hours Line 5 IL-4 IL-10 IFN- n C3H Rag LPL Tg 26 LPL NOTE. Representative data of 1 out of 3 lines analyzed (lines 3, 4, and 5) in sequential adoptive transfer experiment B.

9 1276 ABADÍA MOLINA ET AL GASTROENTEROLOGY Vol. 128, No. 5 Discussion Here we report a novel method for cloning CD4 cells, which induce colitis, by using serial ATs. CD4 cells purified from the MLNs of BM tg 26 mice with colitis adoptively transfer disease to healthy recipient tg 26 mice with a high degree of efficiency. Seven cell lines were derived by serial ATs of MLN CD4 cells, and each cell line contained 1 distinctly dominant CD4 cell, as judged by TCR V analyses. The cell lines maintained their Th1-like phenotype and induced colitis only when adoptively transferred into class II MHC matched immunodeficient recipients. The findings are indicative of an MHC class II restriction of CD4 /APC interactions during the induction of colitis and support the hypothesis that colitis-inducing clones can be generated by serial ATs of MLN CD4 T cells into healthy tg 26 mice. These results with tg 26 mice contrast with those obtained when CD4 cells were isolated from the MLNs of Rag / mice, in which the naive CD4 cells, ie, CD45RB hi cells, had been transferred to initiate colitis. 10,11 All mice in which CD4 MLN cells from the CD45RB hi recipients had been adoptively transferred developed colitis (AT 1), but BALB/c Rag / or C57BL/6 Rag / AT 2 recipients did not (data not shown). This result supports the observations of Matsuda et al. 18 Although pathogenic Th1-like cells develop in both models, 11,12,20 the presence of B cells and the absence of natural killer cells 19,22 may explain why serial ATs into tg 26 mice cause an enrichment of CD4 TV cell lines. Both B cells and natural killer cells are known immunomodulators A protective role of mature B cells in the IBD-like TCR- / mouse model has also been reported. 25 The finding that in the BM tg 26 mouse B cells do not seem to control development of the pathogenic T-cell clones might be due to a unique role of B cells in the Th2-mediated colitis observed in the TCR- / mouse. It is conceivable that B cells play a different role depending on Th1/Th2 responses. Indeed, an impaired Th1 T-cell response to Salmonella enterica infections was found in B cell deficient mice. 26 A negative regulatory effect on the long-term survival of Th1 clones in Rag / mice by natural killer cells is plausible on the basis of the finding that in the CD45RB hi /Rag / model, an accelerated development of colitis occurs in natural killer cell depleted mice. 27 Taken together, our data suggest that effector/memory cells from MLNs in tg 26 mice with colitis are ready to go into the inflammation site upon transfer and initiate the pathogenic cascade; therefore, colitis develops easily and early. Indeed, a large proportion of MLN CD4 cells from BM tg 26 13,15,20 and MLN tg 26 mice have an activated/memory phenotype. Furthermore, these results provide evidence for the hypothesis that repeated stimulation and cell division, as occur in chronic colitic disorders, contribute to higher numbers of colitis-inducing T-cell clones. In this regard, specific clonal expansions have been observed in LPLs and intraepithelial lymphocytes of the entire bowel of healthy individuals and of IBD patients, as well as in IBD-like rodent models. 18,35 Our data suggest that common antigens drive T-cell stimulation of the isolated clones. Results from germfree rodents in the tg 26 model, in which intestinal inflammation is absent, indicate that bacteria are indispensable contributors to the pathogenesis of chronic immune-mediated intestinal inflammation. 17,36 In humans, IBD inflammation takes place in the highest bacterial concentration, and antibiotics or probiotics seem to be beneficial in the treatment and prevention of chronic IBD. 37 Despite these observations suggesting a central role of normal luminal bacteria in the regulation of intestinal inflammation and mucosal homeostasis, the mechanisms by which bacteria influence the activation of T cells responsible for inducing and perpetuating chronic colitis remain unclear. Several possible mechanisms include (1) adjuvant-like stimulation of the APC/T-cell interaction leading to a Th1-dominated response to an independent luminal or self-antigen, (2) cross-reactive immune response to luminal bacterial and host (self-) antigens (molecular mimicry), and (3) induction of an antigen-specific T-cell response to luminal bacteria. In summary, disease-inducing cell lines can be derived by using either splenic or MLN CD4 cells from BM tg 26 mice with colitis. Here we show that serial ATs of MLNs enrich pathogenic CD4 Th1-like cells that respond to the MHC class II peptide complex. The gut and gut-associated lymphoid tissue are continuously exposed to many antigenic substances that can contribute to the generation of clonal expansions. Therefore, a study of the differences between germ-free and normal conditions in the intestinal TCR repertoire, as well as the variety of antigen challenges, such as food and viral and bacterial microorganisms, might assist in deciphering the antigen-driven TCR V clones involved in the pathogenesis of chronic inflammation. References 1. Read S, Powrie F. CD4( ) regulatory T cells. Curr Opin Immunol 2001;13: Sundberg JP, Elson CO, Bedigian H, Birkenmeier EH. Spontaneous, heritable colitis in a new substrain of C3H/HeJ mice. Gastroenterology 1994;107: Kosiewicz MM, Nast CC, Krishnan A, Rivera-Nieves J, Moskaluk CA, Matsumoto S, et al. Th1-type responses mediate spontane-

10 May 2005 CLONING OF COLITIS INDUCING CD4 CELLS 1277 ous ileitis in a novel murine model of Crohn s disease. J Clin Invest 2001;107: Sadlack B, Merz H, Schorle H, Schimpl A, Feller AC, Horak I. Ulcerative colitis-like disease in mice with a disrupted interleukin-2 gene. Cell 1993;75: Kuhn R, Lohler J, Rennick D, Rajewsky K, Muller W. Interleukin- 10-deficient mice develop chronic enterocolitis. Cell 1993;75: Leach MW, Bean AG, Mauze S, Coffman RL, Powrie F. Inflammatory bowel disease in C.B-17 scid mice reconstituted with the CD45RBhigh subset of CD4 T cells. Am J Pathol 1996;148: Shull MM, Ormsby I, Kier AB, Pawlowski S, Diebold RJ, Yin M, et al. Targeted disruption of the mouse transforming growth factorbeta 1 gene results in multifocal inflammatory disease. Nature 1992;359: Mombaerts P, Mizoguchi E, Grusby MJ, Glimcher LH, Bhan AK, Tonegawa S. Spontaneous development of inflammatory bowel disease in T cell receptor mutant mice. Cell 1993;75: Mizoguchi A, Mizoguchi E, Bhan AK. The critical role of interleukin 4 but not interferon gamma in the pathogenesis of colitis in T-cell receptor alpha mutant mice. Gastroenterology 1999;116: Powrie F, Leach MW, Mauze S, Caddle LB, Coffman RL. Phenotypically distinct subsets of CD4 T cells induce or protect from chronic intestinal inflammation in C. B-17 scid mice. Int Immunol 1993;5: Morrissey PJ, Charrier K, Braddy S, Liggitt D, Watson JD. CD4 T cells that express high levels of CD45RB induce wasting disease when transferred into congenic severe combined immunodeficient mice. Disease development is prevented by cotransfer of purified CD4 T cells. J Exp Med 1993;178: Hollander GA, Simpson SJ, Mizoguchi E, Nichogiannopoulou A, She J, Gutierrez-Ramos JC, et al. Severe colitis in mice with aberrant thymic selection. Immunity 1995;3: Elson CO, Cong Y, Sundberg J. The C3H/HeJBir mouse model: a high susceptibility phenotype for colitis. Int Rev Immunol 2000; 19: de Jong Y, Terhorst C, Weaver CT, Read S, Elson E. Disease induction and prevention in experimental models of inflamatory bowel disease. In: Mucosal immunology. 1994: Read S, Malmstrom V, Powrie F. Cytotoxic T lymphocyte-associated antigen 4 plays an essential role in the function of CD25( )CD4( ) regulatory cells that control intestinal inflammation. J Exp Med 2000;192: Veltkamp C, Tonkonogy SL, De Jong YP, Albright C, Grenther WB, Balish E, et al. Continuous stimulation by normal luminal bacteria is essential for the development and perpetuation of colitis in Tg(epsilon26) mice. Gastroenterology 2001;120: Faubion WA, De Jong YP, Molina AA, Ji H, Clarke K, Wang B, Mizoguchi E, Simpson SJ, Bhan AK, Terhorst C. Colitis is associated with thymic destruction attenuating CD4 25 regulatory T cells in the periphery. Gastroenterology 2004;126: Matsuda JL, Gapin L, Sydora BC, Byrne F, Binder S, Kronenberg M, et al. Systemic activation and antigen-driven oligoclonal expansion of T cells in a mouse model of colitis. J Immunol 2000; 164: Wang B, Biron C, She J, Higgins K, Sunshine MJ, Lacy E, et al. A block in both early T lymphocyte and natural killer cell development in transgenic mice with high-copy numbers of the human CD3E gene. Proc Natl Acad Sci U S A 1994;91: Simpson SJ, Hollander GA, Mizoguchi E, Allen D, Bhan AK, Wang B, et al. Expression of pro-inflammatory cytokines by TCR alpha beta and TCR gamma delta T cells in an experimental model of colitis. Eur J Immunol 1997;27: De Jong YP, Comiskey M, Kalled SL, Mizoguchi E, Flavell RA, Bhan AK, et al. Chronic murine colitis is dependent on the CD154/CD40 pathway and can be attenuated by anti-cd154 administration. Gastroenterology 2000;119: Wang B, Simpson SJ, Hollander GA, Terhorst C. Development and function of T lymphocytes and natural killer cells after bone marrow transplantation of severely immunodeficient mice. Immunol Rev 1997;157: Lanzavecchia A. Receptor-mediated antigen uptake and its effect on antigen presentation to class II-restricted T lymphocytes. Annu Rev Immunol 1990;8: Seaman WE. Natural killer cells and natural killer T cells. Arthritis Rheum 2000;43: Mizoguchi A, Mizoguchi E, Takedatsu H, Blumberg RS, Bhan AK. Chronic intestinal inflammatory condition generates IL-10 producing regulatory B cell subset characterized by CD1a upregulation. Immunity 2002;16: Ugrinovic S, Menager N, Goh N, Mastroeni P. Characterization and development of T-cell immune responses in B-cell-deficient (Igh-6( / )) mice with Salmonella enterica serovar Typhimurium infection. Infect Immun 2003;71: Fort MM, Leach MW, Rennick DM. A role for NK cells as regulators of CD4 T cells in a transfer model of colitis. J Immunol 1998;161: Balk SP, Ebert EC, Blumenthal RL, McDermott FV, Wucherpfennig KW, Landau SB, Blumberg RS. Oligoclonal expansion and CD1 recognition by human intestinal intraepithelial lymphocytes. Science 1991;253: Van Kerckhove C, Russell GJ, Deusch K, Reich K, Bhan AK, DerSimonian H, Brenner MB. Oligoclonality of human intestinal intraepithelial T cells. J Exp Med 1992;175: Blumberg RS, Yockey CE, Gross GG, Ebert EC, Balk SP. Human intestinal intraepithelial lymphocytes are derived from a limited number of T cell clones that utilize multiple V beta T cell receptor genes. J Immunol 1993;150: Gross GG, Schwartz VL, Stevens C, Ebert EC, Blumberg RS, Balk SP. Distribution of dominant T cell receptor beta chains in human intestinal mucosa. J Exp Med 1994;180: Bennet JD, Brown WR, Kotzin BL. Regional variation in the lamina propria T cell receptor V beta repertoire in normal human colon. Clin Immunol 1999;90: Saubermann LJ, Probert CS, Christ AD, Chott A, Turner JR, Stevens AC, et al. Evidence of T cell receptor beta-chain patterns in inflammatory and noninflammatory bowel disease states. Am J Physiol 1999;276(3 Pt 1):G613 G Probert CS, Chott A, Saubermann LJ, Stevens AC, Balk SP, Blumberg RS. Prevalence of an ulcerative colitis-associated CD8 T cell receptor beta-chain CDR3-region motif and its association with disease activity. J Clin Immunol 2001;21: Mizoguchi A, Mizoguchi E, Saubermann LJ, Higaki K, Blumberg RS, Bhan AK. Limited CD4 T-cell diversity associated with colitis in T-cell receptor alpha mutant mice requires a T helper 2 environment. Gastroenterology 2000;119: Sartor RB. The influence of normal microbial flora on the development of chronic mucosal inflammation. Res Immunol 1997; 148: Schultz M, Sartor RB. Probiotics and inflammatory bowel diseases. Am J Gastroenterol 2000;95:S19 S21. Received February 10, Accepted January 12, Address requests for reprints to: Cox Terhorst, MD, Harvard Institute of Medicine, HIM816, 4 Blackfan Circle, Boston, Massachusetts acbadia@ugr.es; fax: (617) Supported by grants DK52510 (to C.T.) and DK43351 (to C.T. and to A.K.B.) from the National Institutes of Health, by the Crohn s and Colitis Foundation of America (to Y.P.d.J.), and by the Spanish Ministry of Education and Culture (to A.C.A.-M.). The authors thank Dr Massimo Morra for critically reviewing the manuscript and Aimée Julien for technical advice and assistance with genotyping mice.