Alloimmune response results in expansion of autoreactive donor CD4⁺ T cells in transplants that can mediate chronic GVHD

Abstract

Chronic graft versus host disease (cGVHD) is considered an autoimmune-like disease mediated by donor CD4⁺ T cells, but the origin of the autoreactive T cells is still controversial. Here, we report that transplantation of DBA/2 donor spleen cells into thymectomized MHC-matched allogeneic BALB/c recipients induced autoimmune-like cGVHD, although not in control syngeneic DBA/2 recipients. The donor-type CD4⁺ T cells from the former but not the latter recipients induced autoimmune-like manifestations in secondary allogeneic BALB/c as well as syngeneic DBA/2 recipients. Transfer of donor-type CD4⁺ T cells from secondary DBA/2 recipients with disease into syngeneic donor-type or allogeneic host-type tertiary recipients propagated autoimmune-like manifestations in both. Furthermore, TCR spectratyping revealed that the clonal expansion of the autoreactive CD4⁺ T cells in cGVHD recipients was initiated by alloimmune response. Finally, hybridoma CD4⁺ T clones derived from DBA/2 recipients with disease proliferated similarly to stimulation by syngeneic donor-type or allogeneic host-type DCs. These results demonstrate that the autoimmune-like manifestations in cGVHD can be mediated by a population of donor CD4⁺ T cells in transplants that simultaneously recognize antigens presented by both donor and host APCs.

Introduction

Chronic graft-versus-host disease (cGVHD) is a multi-system chronic alloimmune and autoimmune disorder that occurs after allogeneic hematopoietic cell transplantation (HCT) (1–3). Although cGVHD often follows acute GVHD (aGVHD), it has distinguishing clinical features and wider range of target organs. While aGVHD features acute inflammatory infiltration in the gut, liver, lung, and skin, cGVHD appears to be an autoimmune-like disorder similar to scleroderma and systemic lupus erythematosus (SLE) (1–6). Besides the gut, liver, lung, and skin, other organs such as salivary glands, mucus membranes, and eyes also become the target of cGVHD (1, 7, 8). It has been proposed that autoreactive donor-type CD4⁺ T cells contribute to the pathogenesis of cGVHD (9–17), but it is not yet clear how alloimmune responses lead to the development of autoreactive donor-type CD4⁺ T cells.

T cell reconstitution following allogeneic HCT results from both thymus-dependent and independent pathways(18), and both pathways has been proposed to contribute to the generation of autoreactive CD4⁺ T cells that can mediate cGVHD. For example, a randomized trial comparing GVHD severity in patients given T cell-depleted (TCD) and non-TCD BM grafts from unrelated donors showed that T cell depletion markedly reduced the rate of aGVHD but not cGVHD(19); transplantation of TCD-BM cells from MHC-mismatched MHC II^−/− donor mice resulted in defective negative selection and generation of autoimmune-like cGVHD(13); and protection of thymus by administration of keratinocyte growth factor(KGF) or anti–IL-7Rα antibody ameliorated cGVHD(20, 21). All these reports indicate that thymus-derived autoreactive donor-type CD4⁺ T cells can mediate cGVHD.

On the other hand, the thymus-dependent pathway is not the only source of pathogenic CD4⁺ T cells that mediated cGVHD. For instance, elder patients that had little thymocyte generation showed severe cGVHD (1); increase of donor T cells in G-CSF mobilized transplants was associated with more severe cGVHD but not aGVHD (15, 22, 23); a report showed that allogeneic cGVHD recipients did not have a defect in thymic negative selection (24); transplantation of thymic tissues did not reduce the incidence or severity of cGVHD(25, 26). All these reports indicate that the thymus-independent pathway can give rise to autoreactive donor-type CD4⁺ T cells that mediate cGVHD. Our studies with the mouse model of DBA/2 donor to BALB/c recipient showed that autoimmune-like cGVHD can be induced in euthymic, T cell-deficient athymic, and thymectomized recipients, using donor spleen cells. In addition, depletion of donor CD4⁺ T cells in the spleen can prevent the disease induction (12). These results indicate that mature donor CD4⁺ T cells in transplants are required but de novo thymus-derived donor-type T cells, previously described extra-thymic differentiated donor-type T cells (27), or residual host-type T cells are not required for the disease induction. However, the mechanisms wherein donor CD4⁺ T cells become autoreactive in allogeneic recipients are still unclear.

In the current studies, using the MHC-matched mouse model of DBA/2 donor and thymectomized BALB/c host, we found that donor-type autoreactive CD4⁺ T cells in transplants were expanded following the alloimmune response and contributed to cGVHD pathogenesis; furthermore, cGVHD can be mediated by a population of donor CD4⁺ T cells in transplants that possess TCRs that can subsequently interact with host- and donor-type APCs.

Materials and Methods

Mice

Thymectomized DBA/2(H-2^d) and BALB/c (H-2^d) mice were purchased from The Jackson Laboratory (Bar Harbor, ME). MHC II^−/− DBA/2 or BALB/c mice were generated by back-crossing with MHC II^−/− C57BL/6 that has whole MHC II segment deleted (28) for 8 generations. All the mice were maintained in a pathogen-free room at City of Hope Research Animal Facilities (Duarte, CA). All animal protocols were approved by City of Hope Research Animal Care committee.

Monoclonal antibodies, flow-cytometric analysis and cell sorting

The FITC-, phycoerythrin-, APC-, Cy7-APC-, or bitotin conjugated mAb’s to mouse CD5.1, CD4, CD8, CD25, CD11c, TCRVβ panel (Vβ2, 3, 4, 5.1, 6, 7, 8.1, 8.3, 9, 10^b, 11, 12, 13, 14, 17^a), and TCR Vα panel (Vα2, 3, 8,11) were all purchased from BD Pharmigen (San Diego, CA). Enrichment or depletion of cell subsets with magnetic purification system from Miltenyi Biotec (Auburn, CA) and Multiple-color FACS analysis and sorting were performed as previously described (12, 29).

Proliferation assays

Sorted CD4⁺ T cells together with irradiated allogeneic or syngeneic CD11c⁺ DCs (1 × 10⁵ each) from spleen cells were cultured in a U-bottom 96-well plate for 5 days and ³H-TdR (1 μCi/ml) was added 18 hours before harvest. Background counts in the wells of responder T cell alone was <1000 cpm. The stimulating index was calculated using this formula: $[\text{CPM}\text{of}\text{culture}\text{of}\text{Responder}\times \text{Stimulator}-\text{CPM}\text{of}\text{Culture}\text{of}\text{Responder}\text{alone}]÷[\text{CPM}\text{of}\text{culture}\text{of}\text{Responder}\text{alone}]$.

Measurement of autoantibodies in serum

Anti-dsDNA IgG 2a was measured with ELISA as previously described (12, 29). Anti-dsDNA titers are expressed in units per milliliter, using a reference-positive standard of pooled serum from 6- to 7-month-old NZB/W mice. A 1:100 dilution of this standard serum was arbitrarily assigned a value of 100 U/ml.

TCR spectratyping

The primers for TCR Vβs and TCR spectratyping were previously described(30). Briefly, CD4⁺ T cells (0.5 – 5×10⁶) were collected and lyzed in Trizol for RNA extraction. RNA (1 μg) was reverse transcribed to cDNA using M-MLV. The cDNA was divided into separate wells and amplified using PCR primers for TCR-Vβ genes, with the resulting product run for 3–5 additional cycles using primers labeled with 6-FAM. The fluorescently labeled amplification products were then run on a sequencing gel and analyzed using a Hitachi AB model 3730 capillary DNA analyzer (Applied Biosystems, Los Angeles, CA). Analysis was performed with the aid of GeneMapper software (Applied Biosystems, Los Angeles, CA). Healthy DBA/2 CD4⁺ splenocytes were used to form a baseline for each Vβ spectrum, and samples from CD4⁺ recipient cells or cultured cells were compared to the baseline. A Vβ peak was considered to be skewed if a majority of samples tested contained at least one peak which met the following criteria, based on the work of Friedman and colleagues (31): 1) the peak accounts for a minimum of 10% of the total peak area for the given Vβ, which excludes minor clones, and 2) the relative use of the particular peak increased by 50% or more as compared to the baseline, indicating significant clonal expansion.

Establishment of hybridoma T clones

Thymoma BW5147.G.1.4 cell line was purchased from the American Type Culture Collection (ATCC; catalog No.TIB 48) and cultured in 10% FBS RPMI 1640 complete medium (Invitrogen, Carlsbad, CA). Autoreactive hybridoma CD4⁺ T cells were generated by fusing the donor-reactive CD4⁺ T cells with BW5147 cell line as previously described (32). After fusion, the cells were selected with HAT (ATCC, catalog No.69-X) and HT (ATCC, catalog No.71-X). Two weeks latter, hybridomas were expanded and TCR Vβ8.1 was screened by flow cytomettry with anti-TCR Vβ8.1 antibody. The hybridoma TCR Vβ8.1 positive cells were selected and expanded. Then, hybridoma Vβ8.1⁺ CD4⁺ T cell were cloned with single cell flow cytometry sorting.

Measurement of IL-2 in supernatant

To test the donor- or host-reactivity of hybridoma Vβ8.1⁺ CD4⁺ T clones, cloned hybridoma T cells (0.5 × 10⁶/well) were stimulated by DBA/2 or BALB/c-DCs (0.05 ×10⁶/well) sorted from spleen cells for 48 hours. Then, supernatants were harvested, and IL-2 concentration was quantified by a Mouse IL-2 ELISA kid (BD Biosciences, catalog No.555148).

Histopathology

Tissue specimens were fixed in formalin before embedding in paraffin blocks. Tissue sections were stained with hematoxylin and eosin. Slides were examined at 200x magnification and visualized with an Olympus and a Pixera (600CL) cooled CCD camera (Pixera, Los Gatos, CA). Tissue damage was blindly assessed on scoring system previously described (29). In brief, skin GVHD was scored on the basis of tissue damage in epidermis, dermis, and loss of subcutaneous fat; the maximum score is 10. Jejunum GVHD was scored based on the damage in intestine including villous blunting, crypt regeneration, crypt epithelial cell apoptosis, crypt loss, luminal sloughing of cellular debris, lamina propria inflammatory cell infiltrate, and mucosal ulceration and the maximum is 14, as previously described (15). Liver GVHD was scored on the number of involved tracts and the severity of disease in each tract, and the maximum score is 8. Lung GVHD was scored on the periluminal infiltrates, pneumonitis, and the severity of lung tissues involved; the maximum score is 9, as previously described (33). Salivary GVHD was scored on mononuclear cell infiltration and follicular destruction, as previously described (34). The degree of inflammatory infiltrates was graded as follows. Grade 1: 1–5 foci of mononuclear cells were seen (more than 20 cells per focus). Grade 2: more than 5 foci of mononuclear cells were seen, but without significant parenchymal destruction. Grade 3: multiple confluent foci were seen, with moderate degeneration of parenchymal tissue. Grade 4: extensive infiltration of the gland with mononuclear cells and extensive parenchymal destruction were seen. Kidney GVHD was graded by the glomerular inflammation, proliferation, crescent formation, and necrosis, and maximum score is 12. The mean± SE of the score of six recipients in each group was calculated.

Statistical analysis

Body weight changes, proteinuria incidences and survival in different groups were compared using the log-rank test (Prism Version 4.0, Graph Pad Software, San Diego, CA). Comparison of two means was analyzed using unpaired two-tail Student t-test.

Results

Donor CD4⁺ T cells in transplants mediated autoimmune-like manifestations of cGVHD via interaction with host-type APCs in allogeneic recipients

We previously reported that transplantation of DBA/2 donor spleen cells into sublethally irradiated MHC-matched euthymic or athymic BALB/c recipients induced severe autoimmune-like cGVHD (12, 29). However, the origin of the autorective T cells in the recipients is still unclear. In the current study, we explored the mechanisms how the mature donor T cells in transplants led to the development of autoimmune-like disease in allogeneic recipients. First, as shown in Fig. 1A of the experimental scheme, sublethally irradiated (800R) thymectomized allogeneic BALB/c and thymectomized syngeneic DBA/2 recipients were transplanted with donor CD25⁺ T-depleted spleen (CD25⁻-SPL) cells (50×10⁶). Majority of the donor CD4⁺ T cells in the CD25⁻-SPL had CD62L^hiCD44^lo naïve phenotype (Sup. Fig. 1). We found that, although none (0/12) of syngeneic DBA/2 recipients(1⁰ Syn-Rec) showed any signs of GVHD, all (12/12) of allogeneic BALB/c recipients(1⁰ Allo-Rec) showed high serum levels of IgG2a anti-dsDNA, hair loss, severe proteinuria, and died by 30 days after HCT (P<0.01, Fig. 2A–C). In addition, we observed that the cGVHD BALB/c recipients had severe tissue damage in the skin, jejunum, liver, lung, salivary gland, and kidney, although little tissue damage was observed in the syngeneic DBA/2 recipients (P<0.01, Fig. 2D–E). The identification of tissue damage in the jejunum and salivary gland of the cGVHD mouse recipients is of interest, because recent reports show that tissue damage in jejunum and salivary gland is often associated with cGVHD in patients (1, 7).

Figure 1. Experimental Scheme.

(A) DBA/2 donor spleen cells were transplanted into sublethally irradiated thymectomized MHC-matched allogeneic BALB/c recipients or syngeneic DBA/2 recipients. (B) Sorted donor-type CD4⁺ T cells with T cell-depleted spleen cells (TCD-SPL) from the primary thymectomized BALB/c recipients (1⁰ Allo-BALB/c) were transferred into secondary thymectomized BALB/c recipients (2⁰ Allo-BALB/c). (C) Sorted donor-type CD4⁺ T cells and TCD-SPL from the primary thymectomized recipients (1⁰ Allo-BALB/c) or control primary DBA/2 recipients (1⁰ Syn-DBA/2) were transferred into the secondary syngeneic thymectomized DBA/2 recipients (2⁰ Syn-DBA/2). (D) Sorted donor-type CD4⁺ T cells from secondary allogeneic BALB/c (2⁰-Allo) or syngeneic DBA/2 (2⁰-Syn) recipients were further expanded by host-type BALB/c-DCs or donor-type DBA/2 DCs to generate host-reactive or donor-reactive CD4⁺ T cells. Thereafter, the host-reactive or donor-reactive CD4⁺ T cells were transferred into tertiary syngeneic DBA/2 (3⁰-Syn) or allogeneic BALB/c (3⁰-Allo) recipients.

Figure 2. Donor CD4⁺ T cells in transplants induced autoimmune-like manifestations in allogeneic BALB/c but not syngeneic DBA/2 recipients.

Sublethally irradiated thymectomized BALB/c or DBA/2 mice were transplanted with CD25⁺-depleted spleen (CD25⁻-SPL) cells (50×10⁶) from DBA/2 donors. The recipients were monitored for survival daily and clinical GVHD and proteinuria twice a week. 15 days after HCT, serum was harvested for measuring IgG2a autoantibody levels. There were 12 mice in each group. Additional recipients were euthanized 25 days after HCT for proliferation of donor-type CD4⁺ T cells in the spleens of recipients and for histopathology of GVHD target tissues. (A) Percentage of recipients without proteinuria after HCT. (B) Survival percentage of recipients after HCT. (C) Serum levels of anti-dsDNA IgG2a (n=12). (D) A representative photo of histopathology of skin, jejunum, liver, lung, salivary gland, and kidney. (E) Mean± SE of histopathology scores (n=6).

It was proposed that autoimmunity in cGVHD recipients was mediated by thymus-derived donor-type T cells that interact with donor-type APCs (13, 20). Next, we tested whether donor-type CD4⁺ T cells derived from mature T cells in transplants and their interaction with host-type APCs could mediate autoimmune-like manifestations. Accordingly, as shown in Fig. 1B of the experimental scheme, 15 days after HCT, sorted CD4⁺ T cells (5 ×10⁶) plus T cell-depleted spleen cells (TCD-SPL, 45×10⁶) or TCD-SPL(45×10⁶) alone from the primary allogeneic BALB/c recipients were injected into sublethally irradiated thymectomized secondary BALB/c recipients. Donor DBA/2 T cells express CD5.1 and host BABL/c T cells express CD5.2; donor CD5.1⁺CD4⁺ T cells were sorted by flow cytometry after CD4⁺ cell enrichment with magnetic microbes as described in our previous publications(12, 29) and Sup. Fig 2. Majority of the donor-type CD4⁺ T cells from the primary allogeneic recipients had CD62L^loCD44^hi effector memory phenotype (Sup. Fig. 1). We found that all (12/12) of the secondary BALB/c recipients given the CD4⁺ T and TCD-SPL cells developed weight-loss, proteinuria, high serum levels of IgG2a anti-dsDNA, hair-loss, and died by 40 days after cell transfer. In contrast, the recipients given TCD-SPL cells only showed no signs of cGVHD (P<0.01, Fig. 3A–E). Histopathology showed that, while recipients given TCD-SPL cells only showed minimum signs of tissue damage, the recipients given CD4⁺ T cells had severe tissue damage in skin, jejunum, liver, lung, salivary gland, and kidney (P<0.01, Fig. 3F & G). These results indicate that mature donor-type CD4⁺ T cells from transplants and their interaction with host-type APCs can result in autoimmune-like manifestations in the allogeneic recipients.

Figure 3. Donor-type CD4⁺ T cells from cGVHD BALB/c recipients induced severe autoimmune-like manifestations in secondary BALB/c recipients.

Donor-type CD4⁺ T cells (5 ×10⁶) and/or T cell-depleted spleen cells (TCD-SPL, 45×10⁶) from the primary allogeneic BALB/c recipients (1⁰ Allo) 15 days after HCT were injected into sublethally irradiated thymectomized secondary BALB/c recipients (2⁰ Allo). The recipients were monitored for survival daily, proteinuria, body weight, and clinical GVHD twice a week for 50 days. 15 days after cell transfer, the serum levels of anti-dsDNA IgG2a was measured. In additional experiments, 30 days after cell transfer, the recipients were euthanized for histopathology study. (A) Percentage of body weight changes after cell transfer (n=12). (B) Percentage of recipients without proteinuria after cell transfer. (C) Survival percentage of the recipients after cell transfer. (D) Serum levels of anti-dsDNA IgG2a 15 days after cell transfer (n=12). (E) A representative photo of recipients given TCD-SPL or TCD-SPL and CD4⁺ T cell from the primary allogeneic recipients. (F) Mean± SE of histopathology scores (n=6). (G) A representative histopathology photo of the skin, jejunum, liver, lung, salivary gland, and kidney tissues from the secondary BALB/c recipients 30 days after cell transfer.

Donor-type autoreactive CD4⁺ T cells in transplants were expanded in allogeneic BALB/c but not in syngeneic DBA/2 recipients

Since we observed that DBA/2 donor spleen cells induced autoimmune-like manifestations in thymectomized allogeneic but not syngeneic recipients, we tested whether autoreactive donor-type T cells were expanded in the allogeneic recipients. Accordingly, as shown in Fig. 1C of the experimental scheme, sorted CD5.1⁺CD4⁺ T (5 ×10⁶) and/or TCD-SPL (45×10⁶) from primary allogeneic (1⁰ Allo-BALB/c) recipients or primary syngeneic (1⁰ Syn-DBA/2) recipients were transplanted into secondary DBA/2 recipients (2⁰ Syn-DBA/2), respectively. As mentioned above, majority of the donor-type CD4⁺ T cells from primary allogeneic recipients had CD62L^loCD44^hi effector memory phenotype (Sup. Fig. 1). Consistent with a previous report that T cells that go through homeostatic expansion upregulate CD44 and down-regulate CD62L (35), majority of the donor-type CD4⁺ T cells in the primary syngeneic recipients appeared to have CD62L^loCD44^hi phenotype (Sup. Fig. 1). We found that the recipients given the CD4⁺ T cells from the primary allogeneic recipients(1⁰ Allo-BALB/c) showed weight-loss, fur dyspigmentation, and increased serum anti-dsDNA IgG, as compared to recipients given CD4⁺ T cells from the primary syngeneic recipients(1⁰ Syn- DBA/2) (Fig. 4A–C). In addition, histopathology showed that, compared to recipients given CD4⁺ T cells from the 1⁰ Syn-DBA/2 recipients, the recipients given CD4⁺ T cells from the 1⁰ Allo-BALB/c recipients had much more severe tissue damage, including increased cell infiltration in the dermis, increased apoptotic cells in hair-follicle, increased regeneration in the small intestine [a sign of cGVHD as previously described(15)], and increased infiltration in the liver, salivary gland, and kidney glomeruli (P<0.01, Fig. 4D and E). The DBA/2 secondary recipients given TCD-SPL cells alone from either allogeneic or syngeneic recipients showed no signs of disease (data not shown).

Figure 4. Donor-type CD4⁺ T cells from cGVHD BALB/c recipients induced autoimmune manifestations in syngeneic thymectomized DBA/2 recipients.

Sorted CD4⁺ T (5 ×10⁶) and/or TCD-SPL (45×10⁶) from the primary allogeneic (1⁰ Allo) recipients or sorted CD4⁺ T (5 ×10⁶) and/or TCD-SPL (45×10⁶) from the primary syngeneic (1⁰ Syn) recipients were transplanted into sublethally irradiated thymectomized secondary DBA/2 recipients (2⁰ Syn), respectively. The recipients were monitored for survival daily, and proteinuria, body weight, and clinical GVHD twice a week for up to 50 days. 15 days after cell transfer, the serum levels anti-dsDNA IgG2a were measured. 50 days after cell transfer, the recipients were euthanized for histopathology study. (A) Percentage of body weight changes after cell transfer (n=12). (B) Serum levels of anti-dsDNA IgG2a (n=12). (C) A representative photo of mice given donor-type CD4⁺ T cells from primary syngeneic recipients (1⁰ Syn→Syn), which had a healthy appearance or mice given donor-type CD4⁺ T cells from primary allogeneic recipients (1⁰ Allo→Syn), which had depigmented fur. (D) A representative histopathology photo of the skin, jejunum, liver, lung, salivary, and kidney tissues in secondary recipients 50 days after cell transfer. (E) Mean± SE of histopathology scores (n=6). (F, G) Proliferation of CD4⁺ T cells. Sorted donor-type CD4⁺ T cells (0.2×10⁶/well) from the primary allogeneic (1⁰ Allo) recipients or the primary syngeneic (1⁰ Syn) recipients were stimulated with DCs (0.1×10⁶/well) from wild-type (WT) or MHC II^−/− donor DBA/2 or host BALB/c mice for 5 days, and ³H-TdR incorporation was measured. Mean± SE of Stimulating index (SI) the SI of 6 replicated experiments is shown.

In addition, 15 days after HCT, donor-type CD4⁺ T cells from syngeneic DBA/2 and allogeneic BALB/c recipients were sorted and stimulated in vitro with CD11c⁺ DCs from DBA/2 or BALB/c mice. We found that the proliferation of CD4⁺ T cells from allogeneic BALB/c recipients with cGVHD was 4-fold stronger than the CD4⁺ T cells from syngeneic recipients without cGVHD when they were stimulated with syngeneic DBA/2 DCs (P<0.01, Fig. 4F). The proliferation was DBA/2 MHC II restricted, because they did not proliferate in response to stimulation by MHC II^−/− DBA/2 DCs (Fig. 4F). In contrast, their proliferation was not significantly different, when they were stimulated with allogeneic BALB/c DCs: both proliferated vigorously in a BALB/c MHC II-dependent manner (Fig. 4G). Taken together, these results indicate that autoreactive CD4⁺ T cells in transplants are expanded in allogeneic but not in syngeneic recipients, and the autoreactive CD4⁺ T cells also contribute to the development of autoimmune-like manifestations in the allogeneic recipients with cGVHD.

Both donor-reactive and host-reactive CD4⁺ T cells derived from mature T cells in transplants induced autoimmune-like manifestations in allogeneic BALB/c as well as in syngeneic DBA/2 recipients

To further dissect the role of donor-reactive and host-reactive CD4⁺ T cells derived from mature T cells in transplants in mediating autoimmune-like manifestations in allogeneic hosts, we used serial in vivo transfer and in vitro culture to further expand the “donor-reactive” and “host-reactive” CD4⁺ T cells. Accordingly, as shown in Fig. 1D of the experimental scheme, 15 days after HCT, the CD4⁺ T cells from the primary allogeneic (1⁰ Allo) BALB/c recipients with GVHD were transferred into secondary (2⁰ Allo) BALB/c recipients or (2⁰ Syn) DBA/2 recipients.15 days after secondary transfer, the donor-type CD4⁺ T from the (2⁰ Allo) BALB/c recipients were sorted, and then repeatedly stimulated with host-type BALB/c CD11c⁺ DCs in complete medium with 100 u/ml of IL-2 for 15 days. The expanded CD4⁺ T cells were referred as host-reactive CD4⁺ T cells. Similarly, the donor-type CD4⁺ T cells from the (2⁰ Syn) DBA/2 recipients were sorted and in vitro expanded by donor-type DBA/2 CD11c⁺ DCs, and the expanded CD4⁺ T cells were referred as donor-reactive CD4⁺ T cells. The CD4⁺ T cells in both cultures were expanded about 20 fold.

The donor-reactive or host-reactive CD4⁺ T cells were measured for their donor- or host-reactivity in a proliferation assay with donor-type DBA/2 or host-type BALB/c DC stimulation. Surprisingly, we found that the expanded donor-reactive or host-reactive CD4⁺ T cells proliferated similarly in response to donor- or host-type DC stimulation (Fig. 5A), although the stimulation index was 2–3 fold higher when stimulated with host-type DCs (P<0.01, Fig. 5A).

Figure 5. Donor-reactive and Host-reactive CD4⁺ T cells from primary cGVHD recipients induced autoimmune-like manifestations in allogeneic BALB/c as well as syngeneic DBA/2 recipients.

Donor-reactive and host-reactive donor-type CD4⁺ T cells were generated by serial in vivo transfer and in vitro expansion. For generating donor-reactive CD4⁺ T cells, donor-type CD4⁺ T cells were sorted from spleens of secondary DBA/2 recipients (1⁰ Allo Rec→Syn Rec) on day 15 after cell transfer. The cells were cultured with 5% FBS RPMI1640 medium and 100U/ml rh-IL2 and stimulated by DBA/2-DCs weekly for two weeks. Similarly, host-reactive CD4⁺ T cells were generated by culturing the sorted donor-type CD4⁺ T cells from the spleen of secondary BALB/c recipients (1⁰ Allo Rec→Allo Rec). (A) Expanded donor- and host-reactive donor-type CD4⁺ T cells (0.1×10⁶/well) were re-stimulated with DBA/2-DCs or BALB/c-DCs (0.05×10⁶/well) for 5 days, and ³H-TdR incorporation was measured. Stimulating index (SI) was calculated. Mean± SE of the SI of four replicate experiments is shown. (B–I) The expanded donor-reactive or host-reactive donor-type CD4⁺ T cells (10×10⁶) were co-injected with T-cell depleted spleen (TCD-SPL) cells from primary allogeneic BALB/c recipients 15 days after HCT into sublethally irradiated thymectomized tertiary BALB/c (B–E) or DBA/2 (F–I) recipients. The control mice were injected with TCD-SPL cells alone. The recipients were monitored for survival daily, proteinuria, body weight, and clinical GVHD twice a week. 30 days after cell transfer, the serum levels of anti-dsDNA IgG2a was measured. 80 days after cell transfer, the recipients were euthanized for histopathology study. (B and F) Percentage of bodyweight changes after cell transfer (n=12); (C and G) Percentage of recipients without proteinuria after cell transfer (n=12); (D and H) Serum levels of anti-dsDNAIgG2a (n=12); (E and I) Mean± SE of histopathology scores of the skin, jejunum, liver, lung, salivary gland, and kidney tissues. (n=6).

Next, we tested whether the expanded donor-reactive and host-reactive CD4⁺ T cells could induce autoimmune-like manifestations in allogeneic BALB/c recipients. Accordingly, the in vitro expanded donor-reactive or host-reactive CD4⁺ T cells (10×10⁶) were injected with TCD-SPL (50×10⁶) from the primary allogeneic BALB/c recipients into sublethally irradiated thymectomized BALB/c recipients. The control recipients were injected with TCD-SPL cells alone. We found that, compared with TCD-SPL cells, the donor-reactive and the host-reactive CD4⁺ T cells induced the weight-loss in all (12/12), proteinuria in 50% (6/12), higher serum levels of anti-dsDNA IgG in all, and hair-loss in most of the recipients (P<0.01, Fig. 5B–D). Furthermore, histopathology showed that, as compared with TCD-SPL cells, both donor-reactive and host-reactive donor-type CD4⁺ T cells induced severe tissue damages in skin, jejunum, liver, lung, salivary gland, and kidney (P<0.01, Fig. 5E), and there was no significant difference in the severity of tissue damage between the two groups (Fig. 5E). These results indicate that both donor-reactive and host-reactive donor-type CD4⁺ T cells derived from the mature T cells in transplants can mediate autoimmune-like manifestations in allogeneic cGVHD recipients.

Similarly, we found that both donor-reactive and host-reactive CD4⁺ T cells induced autoimmune manifestations in the syngeneic DBA/2 recipients (Fig. 5F-I). We should point out that, consistent with the in vitro proliferation result, the donor-reactive and the host-reactive CD4⁺ T cells induced more severe disease in host-type BALB/c than in donor-type DBA/2 recipients (Fig. 5). Taken together, both donor-reactive and host-reactive CD4⁺ T cells derived from mature T cells in transplants could proliferate to stimulation by donor-type as well as host-type APCs, and they mediate autoimmune-like manifestations in both syngeneic and allogeneic recipients. These results indicate that the donor-reactive and host-reactive CD4⁺ T cells may recognize antigens presented by both donor and host APCs.

The donor-reactive and the host-reactive CD4⁺ T cells in transplants belong to the same population, and their expansion was triggered by alloimmune response

TCR-spectratyping has been used to measure tissue-specific and antigen-specific alloreactive T cell clonal expansion in GVHD recipients (36–41). Next, we used this technique to find out whether the alloreactive and autoreactive T cells in donor transplants express differential TCRs, by comparing the heavy chain (Vβ) TCR-spectra of donor-type CD4⁺ T cells before and after HCT in syngeneic and allogeneic primary and secondary recipients, using TCR spectratyping methods described by Friedman and colleagues (31). We found that DBA/2 donor-type CD4⁺ T cells showed no skewed TCR-CDR3 lengths in primary syngeneic recipients (1⁰ Syn), as compared to before HCT, but they showed skewed TCR-CDR3 lengths in primary allogeneic recipients (1⁰ Allo) (0/22 vs 13/22, P<0.01, Fig. 6A and B). This result indicates that clonal expansion of donor-type CD4⁺ T cells in transplants is initiated by alloimmune response after allogeneic HCT.

Figure 6. Expansion of donor-reactive CD4⁺ T cells in transplants was initiated by alloimmune response.

Day 15 after HCT, donor-type CD4⁺ T cells from the primary syngeneic DBA/2 recipients(1⁰ Syn Rec) or allogeneic recipients (1⁰ Allo Rec) given CD25⁻-DBA/2 donor spleen cells were sorted for TCR spectratyping analysis. Naïve donor DBA/2 CD4⁺ T cells were used as control. At the same time, CD4⁺-SPL cells from the primary allogeneic BALB/c recipients (1⁰ Allo Rec) were enriched and transferred into sublethally irradiated secondary syngeneic DBA/2 recipients(1⁰ Allo→Syn) or allogeneic recipients(1⁰ Allo→Allo). 15 days after the cell transfer, donor-type CD4⁺ T cells from the (1⁰ Allo→Syn) or (1⁰ Allo→Allo) recipients were sorted for TCR spectratyping analysis again. In addition, the TCR-spectrum of donor-reactive and host-reactive donor-type CD4⁺ T cells used in Fig. 5 was also measured. (A) Representative TCR spectra: Skewed TCR-CDR3 length was shaded. A representative of at least 3 repetitions is shown. (B) Summary of TCR spectratyping: A check mark indicates that tested samples contained at least one peak exhibiting significant skewing, while a dash indicates a lack of skewing. A representative of at least three replicate experiments is shown. (C) Left panel shows the percentage of Vβ subsets among total donor-type CD4⁺ T cells from spleens of DBA/2 donor, (1⁰ Allo→Syn) recipient, or (1⁰ Allo→Allo) recipient, as measured with flow cytometry; right panel shows the fold of expansion of percentage of the Vβ subsets as compared to DBA/2 donors. Mean ± SE of 4 mice in each group is shown. (D) Percentage of Vβ subsets among donor-reactive or host-reactive CD4⁺ T cells and fold of expansion of percentage of Vβ subsets as compared to percentage of CD4⁺ Vβ subsets in spleens of DBA/2 donors. Mean ± SE of three replicate experiments is shown.

Furthermore, transfer of the donor-type CD4⁺ T cells from the primary syngeneic recipients (1⁰ Syn) into the secondary syngeneic recipients (1⁰ Syn→Syn) still did not cause the skewing of TCR-CDR3 lengths (data not shown), and transfer of donor-type CD4⁺ T cells from primary thymectomized allogeneic recipients (1⁰ Allo) into the secondary thymectomized allogeneic recipients (1⁰ Allo→Allo) or syngeneic recipients (1⁰ Allo→Syn) did not lead to further increase in the frequencies of skewed CDR3 length, either. In addition, the CD4⁺ T cells with skewed TCR-CDR3 lengths in the secondary allogeneic (1⁰ Allo→Allo) or secondary syngeneic (1⁰ Allo→Syn) recipients were all originated from the primary allogeneic recipients (1⁰ Allo) (Fig. 6A and B). Interestingly, the expanded donor- or host-reactive CD4⁺ T cells both had similar TCR spectra to the donor-type CD4⁺ T cells in the primary allogeneic BALB/c recipients (Fig. 6B). These results indicate that the donor-reactive and host-reactive CD4⁺ T cells derived from the mature T cells in transplants may have similar antigen specificity.

Although the frequencies of TCRs with skewed CDR3-lengths was not further increased by further in vivo or in vitro stimulation, as compared to that in the primary allogeneic recipients, the percentage of some Vβ subsets was increased not only in the secondary allogeneic recipients (1⁰ Allo→Allo), but also in the secondary syngeneic recipients (1⁰ Allo→Syn) (P<0.01, Fig. 6C). Additional in vitro stimulation resulted in further expansion of some Vβ subsets. For example, after the serial in vivo and in vitro stimulation, Vβ8.1⁺ subset became the dominant subset in both donor-reactive and host-reactive CD4⁺ T cells, although the percentage of the Vβ8.1⁺ subset was higher among donor-reactive than that among host-reactive (35% versus 12%) (Fig. 6D). These results indicate that, although the donor- and host-reactive donor-type CD4⁺ T cells have similar TCR spectra and recognize antigens presented by both donor- and host-type APCs, different clones may differ in their expansion in response to donor- or host-type APC stimulation. Therefore, it remains unknown whether or not the donor- and host-reactivity can be mediated by a single TCR.

Single donor-type Vβ8.1⁺ CD4⁺ T clones possessed both donor-reactivity and host-reactivity

Since we observed that Vβ8.1⁺ subset became dominant among the donor-reactive and host-reactive donor-type CD4⁺ T cells after several rounds of in vitro expansion by syngeneic DBA/2 or allogeneic BALB/c DCs, we used the Vβ8.1⁺CD4⁺ T subset as an example for comparing their donor- and host-reactivity. First, we tested whether the reactivity of the Vβ8.1⁺CD4⁺ T cells were similar to the whole population. Accordingly, Vβ8.1⁺CD4⁺ T cells from the expanded donor-reactive donor-type CD4⁺ T cells were sorted and stimulated with DCs from syngeneic donor DBA/2 or allogeneic host BALB/c mice. We found that Vβ8.1⁺ CD4⁺ T cells proliferated vigorously to either type of DC stimulation, and there was no difference between the two groups (Fig. 7A). This is consistent with the donor- and host-reactivity of whole CD4⁺ T population (Fig. 5A).

Figure 7. Donor-reactive Vβ8.1⁺CD4⁺ T cells derived from primary cGVHD recipients expressed different Vα, and their clones proliferated to both syngeneic donor and allogeneic host DC stimulation.

(A) In vitro expanded donor-reactive CD4⁺ T cells were stained with anti-Vβ8.1 vs anti-CD4, and the Vβ8.1⁺CD4⁺ T cells were sorted for proliferation in response to re-stimulation by DBA/2 or BALB/c DCs for 5 days. ³H-TdR incorporation was measured and mean± SE of CPM of triplicate culture was calculated. One representative of two replicate experiments is shown. (B) The donor-reactive CD4⁺ T cells were stained with anti-Vβ8.1, CD4, and different anti-Vα (Vα2, Vα3, Vα8, or Vα11). The gated Vβ8.1⁺CD4⁺ T cells were shown in Vα vs Vβ8.1. (C) The gated Vβ8.1⁺CD4⁺ T cells are shown in Vα2 vs Vα8, Vα2 vs Vα11, Vα8 vs Vα11. One representative of three replicate experiments is shown. (D) TIB-48 Thymoma cell line and hybridoma T cells were stained for Vβ8.1 and CD4. Hybridoma Vβ8.1⁺CD4⁺ T cells were cloned by single cell sorting. (E) IL-2 production of hybridoma Vβ8.1⁺ T clones in response to stimulation by syngeneic donor DBA/2 or allogeneic host BALB/c DCs. Ten clones with different Vα from each plate were randomly picked up and expanded. Each clone cells (0.1×10⁶/well) were stimulated with DBA/2 or BALB/c DCs for 48 hours. Then, IL-2 concentration in the culture supernatant was measured by ELISA.

Second, we tested whether the donor-reactivity and the host-reactivity was mediated by T cells with dual TCRs, because it was reported that alloreactive CD4⁺ T cells could have dual TCR alpha chain (42). We checked the Vα expression by a panel of available anti-Vα antibodies. We found that, although the Vβ8.1⁺ cells expressed different Vα subsets (including Vα2, 8, and 11), no Vβ8.1⁺ cells expressed dual Vα (Fig. 7B and C). In addition, we used the panel of anti-Vβ antibodies used in Fig. 6C to check whether Vβ8.1⁺ CD4⁺ T cells expressed second Vβ, and we found no expression of second Vβ by the Vβ8.1⁺ T cells from primary allogeneic recipients or T hybridoma cells (Sup. Fig. 3). Thus, the Vβ8.1⁺CD4⁺ T cells that possess both donor- and host-reactivity are unlikely to express dual TCRs.

Third, we tested whether a single Vβ8.1⁺CD4⁺ T cell clone possessed both donor- and host-reactivity, using Vβ8.1⁺CD4⁺ hybridoma T cell clones. Accordingly, the donor-reactive CD4⁺ T cells were fused with TIB-48 thymoma cell line as previously described (32). After selective culture for 14 days, Vβ8.1⁺CD4⁺ T hybridomas were cloned with single cell flow cytometry sorting, and two plates of 96 clones each were set up (Fig. 7D). After clonal expansion, 10 clones with different Vα from each plate were stimulated with syngeneic donor-type DBA/2 or allogeneic host-type BALB/c DCs, and their production of IL-2 was measured. We found that, although the hybridoma T clones produced little IL-2 without DC stimulation (data not shown), they produced large mounts of IL-2 in response to syngeneic donor-type or allogeneic host-type DC stimulation, and there was no significant difference in response to syngeneic DBA/2 or allogeneic BALB/c DC stimulation (Fig. 7E). These results indicate that a single Vβ8.1⁺CD4⁺ T cell clone can possess both donor-reactivity and host-reactivity.

Discussion

We have demonstrated that, in a cGVHD model of DBA/2 donor to MHC-matched but minor antigen mismatched BALB/c recipient, autoimmune-like manifestations in cGVHD recipients can be mediated by a population of donor-type CD4⁺ T cells derived from mature T cells in transplants that possess both donor- and host-reactivity.

First, we observed that donor-type CD4⁺ T cells from thymectomized cGVHD recipients proliferated in response to stimulation by donor-type as well as by host-type DCs, and they induced autoimmune-like manifestations in both syngeneic as well as allogeneic secondary recipients. Second, donor-reactive and host-reactive donor-type CD4⁺ T cells established by serial in vivo transfer and in vitro culture from cGVHD recipients still proliferated to stimulation by donor- and host-type DCs as well as induced autoimmune-like manifestations in donor-type syngeneic and host-type allogeneic recipients. Third, TCR spectratyping analysis revealed that TCR-CDR3 skewing of donor CD4⁺ T cells took place in primary allogeneic recipients, and the TCR spectra of the donor-reactive and host-reactive CD4⁺ T cells from the cGVHD recipients was similar. Finally, hybridoma T cell clones derived from the donor-reactive CD4⁺ T cells also proliferated to stimulation by syngeneic donor-type and allogeneic host-type DCs. Taken together, donor CD4⁺ T cells that possess both donor- and host-reactivity in transplants can mediate autoimmune-like manifestations in allogeneic recipients.

We have also observed that the activation and expansion of the donor CD4⁺ T cells with both donor- and host-reactivity is initiated by the alloimmune response. However, the mechanisms are not yet clear. Since we have reported that depletion of donor B cells can prevent induction of autoimmune-like manifestations in allogeneic recipients (12), and this is associated with reduction of autoreactive CD4⁺ T cells in the recipients (Zhao and Zeng, unpublished data), we hypothesize that the donor CD4⁺ T cells in transplants include both alloreactive and autoreactive T cells, and the latter recognize non-polymorphic antigens presented by both donor and host APCs. After allogeneic HCT, alloreactive donor T cells are activated by interaction with host-APCs presenting alloantigens, then the activated alloreactive T cells interact with donor B cells that present both allo- and autoantigens including non-polymorphic antigens from both donor and host. Thereafter, the activated donor B cells subsequently activate and expand the autoreactive donor CD4⁺ T cells in transplants that contribute to the development of autoimmune-like manifestations in the recipients. We should also point out that the donor CD4⁺ T cells with both donor- and host-reactivity have high percentage of Vb8.1⁺ subsets after expansion with either donor- or host-type APCs. This is consistent with a previous report that majority of autoreactive CD4⁺ T cell clones from EAE mice were Vβ8.1⁺ (43). Therefore, the donor CD4⁺ T cells with both donor- and host-reactivity are most likely to be autoreactive T cells that recognize non-polymorphic antigens presented by MHC-matched donor and host APCs, as recently proposed by Shlomchik and Pavletic (44). The role of donor B cells in expansion of the donor-type autoreactive CD4⁺ T cells in cGVHD recipients is under investigation.

However, we can not exclude the possibility that some of those donor CD4⁺ T cells have both allo- and autoreactivity, and their expansion was initiated by their alloreactvity and continued by their autoreactivity, because it was reported by Allen and colleagues that an alloreactive CD4⁺ T cell could express a poly-specific TCR that recognize multiple distinct ligands, each in a highly specific way(45). Thus, some donor CD4⁺ T cells may recognize both allo- and autoantigens. In addition, it was reported that many alloreactive T cells possess dual TCRs (42), but we did not find dual TCRs on the donor CD4⁺ T cells with donor- and host-reactivity.

To our best knowledge, this is the first demonstration that autoimmune-like manifestations in cGVHD recipients can be mediated by donor CD4⁺ T cells in transplants that possess both donor- and host-reactivity. Our current studies have also made several new novel observations in contrast to a recent publication by Tivol and colleagues (11): 1) Our study is based on a MHC-matched chronic GVHD model in which the primary recipients showed signs of autoimmunity including high levels of serum autoantibodies and tissue damages in small intestine and salivary glands that represent characteristic futures of human chronic GVHD(1, 7). On the other hand, Tovol’s observation was based on a MHC-mismatched acute GVHD model in which the primary recipients did not show clear signs of chronic GVHD or autoimmunity(11). 2) The autoreactive CD4⁺ T cells in our studies induced similar autoimmune syndrome in both allogeneic and syngeneic secondary recipients with variety organ tissue damages; however, in Tivol’s studies, the autoreactive CD4⁺ T cells induced only autoimmune tissue damage in colon but not in other GVHD target tissues in the secondary syngeneic recipients, and they did not induce tissue damage in the secondary allogeneic recipients, either. Colitis is the typical sign of acute GVHD but not for chronic GVHD in humans (1, 7). 3) We have demonstrated that the autoreactive CD4⁺ T cells belong to a population of donor CD4⁺ T cells that express a single TCR that can interact with both allogeneic host-type and syngeneic donor-type APCs, indicating that they may recognize the non-polymorphic antigens presented by both donor- and host-type APCs. In contrast, in Tivol’s study, the origin of the autoreactive donor CD4⁺ T cells were not clear; they could be donor CD4⁺ T cells in transplants or those from de novo thymic development. It was also not clear in their study whether the donor-APC reactive and the host-APC reactive donor CD4⁺ T cells belong to the same population; and why the donor CD4⁺ T cells induced GVHD in the primary but not in the secondary allogeneic recipients. Therefore, our report is significantly different from Tivol’s report in many aspects.

Our observations may provide new insights into some clinical observations. First, although aGVHD damage of the thymus is proposed to be the cause of cGVHD (13, 46–50), it has been reported that some patients develop severe autoimmune-like cGVHD without obvious aGVHD that causes thymus damage (51, 52); and that old patients with little thymocyte production often develop more severe autoimmune-like cGVHD(1). Therefore, we speculate that, in those cases, the autoreactive T cells that mediate the cGVHD may be derived from donor T cells in transplants. Second, G-CSF-mobilized blood transplants causes reduction of aGVHD but augmentation of autoimmune-like cGVHD (22, 23). It is possible that the G-CSF-mobilized blood transplants contain higher numbers of the autoreactive CD4⁺ T precursors. Third, we have observed that, although the activation and expansion of the donor-type autoreactive CD4⁺ T cells in transplants are initiated by host-type APCs, those T cells can also interact with donor-type APCs and continue to expand and mediate GVHD after the elimination of host-type APCs. This observation provides a new explanation about why majority of cGVHD in patents are following acute GVHD and autoimmune-like chronic GVHD can occur in complete chimeric patients. This observation can also explain why cGVHD is often associated with better GVL activity (53).

On the other hand, we should still be cautious about the relevance of this mouse model to human cGVHD. Although this animal model has important autoimmune features as well as GVHD target organs (i.e. small intestine and salivary gland) that is similar to human chronic GVHD(1, 7, 54), the high serum levels of anti-dsDNA, high frequency of glomerulonephritis, and the temporal time course in this mouse model does not mirror in human cGVHD. Interestingly, recent reports showed a significant increase of glomerulonephritis in allogeneic recipients with non-myeloablative conditioning (55–58). Therefore, further study about the clinical relevance of this animal model is warranted.

We should also point out that, although both host-reactive and donor-reactive CD4⁺ T cells induced serve autoimmune-like cGVHD in allogeneic host-type BALB/c recipients, both T cells induced weaker autoimmune-like syndrome in syngeneic donor-type DBA/2 recipients. The disease severity difference in BALB/c and DBA/2 mice may be due to different genetic background in which APC cells present different amount of non-polymorphic autoantigens that influence the T cell differentiation and expansion, as previously reported (59, 60).

In summary, our studies have provided an alternative paradigm for the origin of pathogenic T cells that mediate cGVHD. Donor CD4⁺ T cells with both donor- and host-reactivity in transplants are initially activated and expanded by alloimmune response, and then they contribute to the pathogenesis of autoimmune-like manifestations in the cGVHD recipients. Targeting the donor CD4⁺ T cells possessing both donor- and host-reactivity in transplants before and after HCT may be an effective approach for preventing and treating cGVHD.

Abbreviations used in this paper

Allo

allogeneic

GVHD

graft versus host disease

HAT

hypoxanthine-Aminoptein-Thymidine

HCT

hematopoietic cell transplantation

HT

hypoxanthine- Thymidine

Rec

recipient

SPL

spleen

Syn

syngeneic

TBI

total body irradiation

Treg

regulatory T