Major Histocompatibility Complex Class I Matching is Necessary but not Sufficient for Tolerance Induction to Vascularized Composite Allografts After Donor Hematopoietic Cell Transplantation

Abstract

Background.

We previously reported a porcine model of vascularized composite allograft (VCA) transplantation with concurrent hematopoietic cell transplantation (HCT). Tolerance developed when major histocompatibility complex (MHC) class I was matched in the absence of class II matching but not when class II was matched without class I matching. We investigated whether rejection resulted from a mismatch between class I and class II.

Methods.

Recipient pigs were either mismatched from donors by a single class I haplotype or matched by a single class II haplotype. They received CD3 immunotoxin and total-body irradiation followed by HCT and VCA transplantation from the same donor. Cyclosporine immunosuppression was discontinued after 45 d.

Results.

Pigs mismatched for only 1 MHC class I haplotype maintained peripheral chimerism and accepted the VCA graft, except for 1 animal that rejected the epidermis. Pigs matched for only 1 MHC class II haplotype lost bone marrow engraftment and rejected the VCA graft.

Conclusions.

When combined with our prior experiments, our results suggest that some degree of MHC class I matching is necessary for tolerance induction to HCT/VCA, but this is dependent on the degree of class II matching. Full class I matching reliably generates VCA tolerance, regardless of the degree of class II matching. VCA tolerance fails to develop reliably when MHC class I is fully mismatched, regardless of the degree of class II matching. When a single class I haplotype is matched, VCA tolerance depends on the degree of class II matching.

INTRODUCTION

During the past decade, our laboratory has developed a swine model of hematopoietic cell transplantation (HCT) with nonmyeloablative conditioning that has produced tolerance and durable mixed chimerism across major histocompatibility complex (MHC) barriers, including haploidentical mismatch,¹ full class I mismatch,² and full class II mismatch.² In one variation of this model, HCT was combined with concomitant vascularized composite allograft (VCA) transplantation from the same donor to evaluate whether tolerance would extend to the allograft. The full VCA grafts were accepted in recipients haploidentical to the donor, even after cessation of immunosuppression. They were also accepted by recipients who were fully class I matched to the donor, even when class II was fully mismatched. However, they were not universally accepted in recipients who were fully class I mismatched to the donor, even when class II was fully matched. In the latter situation, 2 of 3 recipients rejected the epidermis, whereas the underlying dermis/fascia was additionally rejected in 1 recipient (see summary of findings in Figure 1). In all cases, the animals remained chimeric in peripheral blood.

FIGURE 1.

Summary of vascularized composite allograft (VCA) and split-thickness skin graft (STSG) survival in previous and current work from our laboratory. Studies include Leonard et al,¹ Shanmugarajah et al,² and the current experiment. MHC, major histocompatibility complex.

This observation generated an immunological question that has remained unanswered, which we aimed to resolve in the present study: did rejection of VCAs occur because class I was mismatched or because class II was matched? These findings presented a novel opportunity to study a fundamental question about the relative importance of class I versus class II matching in the induction of tolerance to VCA in hematopoietic chimeric animals, which is knowledge that could inform donor selection in clinical models of combined bone marrow/solid organ transplantation, several of which are ongoing. Therefore, we addressed that question by comparing the historical groups with 2 new experimental groups: (1) mixed chimeric animals differing in only a single class I haplotype (to test whether rejection is due to class I mismatching) or (2) mixed chimeric animals sharing only a single class II haplotype (to test whether rejection is due to class II matching; Figure 2).

FIGURE 2.

Donor and recipient MHC class I and II antigens, experimental timeline, and T-cell depletion. A, Donor and recipient MHC class I and II antigens in groups 1 and 2. B, Experimental timeline for nonmyeloablative conditioning including pCD3-IT given intravenous twice daily at 50 µg/ kg bodyweight at day –4, –3, –2, and –1, TBI of 100 cGy at day –2 and cyclosporine infusion from day –2 until day 30 with target through levels of 400–800 ng/mL and cyclosporine tapering day 30–45, HSCT with 15 × 10⁹ cells/kg bodyweight, VCA transplantation and STSG transplantation. C, Absolute T-cell count as measured by number of CD3⁺ cells/µL of blood before and after T-cell depletion in group 1 (closed circles) and group 2 (open circles). HSCT, hematopoietic stem cell transplant; MHC, major histocompatibility complex; pCD3-IT, porcine CD3 immunotoxin; STSG, split-thickness skin graft; TBI, total-body irradiation; VCA, vascularized composite allograft.

MATERIALS AND METHODS

Animals

All procedures were approved by the local Institutional Animal Care and Use Committee (protocol AC-AABI0555). The procedures and animal care practices complied with the guidelines approved by the Columbia Institutional Animal Care and Use Committee, the National Society of Medical Research’s “Principles of Laboratory Animal Care,” and the “Guide for the Care and Use of Laboratory Animals” by the Institute of Laboratory Animal Resources, National Research Council. We used Sachs miniature swine, which have been inbred such that swine leukocyte antigen (SLA; porcine MHC) class I and II have become fixed and defined.³ Recipients were either mismatched from donors by a single class I haplotype or matched by a single class II haplotype. Recipient animals were 2 mo old at the start of T-cell depletion. The mean age of HSC/VCA donors was 15 mo (range, 9–21). Different lines of Sachs miniature swine have been inbred for MHC haplotypes while maintaining minor antigen variation. In group 1, SLA^gd (class I^cd/II^dd) pigs were selected as donors, and either SLA^gg (class I^cc/II^dd) or SLA^dd (class I^dd/II^dd) pigs were selected as recipients to achieve a 1-haplotype class I mismatch (Figure 2). In group 2, SLA^gc (class I^cc/II^dc) pigs were selected as donors and SLA^gg (class I^cc/II^dd) pigs were selected as recipients to match only 1 class II haplotype (Figure 2). Animal characteristics and housing are described in Supplemental Materials and Methods (SDC, https://links.lww.com/TXD/A817).

Nonmyeloablative Conditioning and Hematopoietic Stem Cell Transplant

Donor and recipient pigs were prepared for apheresis and HCT as previously described^1,2 (see Supplemental Materials and Methods, SDC, https://links.lww.com/TXD/A817). Recipients were treated with CD3 immunotoxin⁴ and 100 cGy total-body irradiation. The leukapheresis product was administered intravenously to the recipients, with a total dose of 15 × 10⁹ cells/kg bodyweight. Cyclosporine (goal trough level 400–800 ng/m) was given intravenously for 45 d.

VCA and Split-thickness Skin Graft Transplantation and Rejection Monitoring

VCA donor pigs were placed in dorsal recumbency and the bilateral groins were prepared and draped for sterile surgery. The flap pedicle (saphenous artery) was identified, and its course was marked with Doppler ultrasound. A 10-cm diameter ovoid skin island was designed centered over the pedicle, with the proximal extent of the flap being marked 3 cm distal to the inguinal crease and the lateral extent of the flap being marked 1 cm medial to the patella as previously described.⁵ The distal flap was incised and dissection carried deep to the fascia whereupon the pedicle vessels were identified and ligated. The remainder of the flap markings were incised, and a fasciocutaneous flap was raised with careful attention paid to keep the pedicle vessels with the flap. The pedicle was then dissected back to its takeoff from the femoral vessels, and a 1.5 cm segment of the proximal femoral artery and vein was dissected and separated. The same procedure was then performed on the contralateral limb. When the recipient pigs were prepared and ready for flap transfer, the donor swine was systemically heparinized with 100 IU/kg (Heparin Sodium, McKesson, Irving, TX) and the flap was flushed with heparinized saline (10 IU/mL).

VCA recipient pigs were placed in dorsal recumbency, and the neck was prepared and draped for sterile surgery. A 5 cm incision was made lateral to the right sternocleidomastoid muscle. Dissection was carried down to the strap muscles, which were bluntly dissected to expose the carotid artery and internal jugular vein. These vessels were freed circumferentially, and side branches were clipped. Rizzuti clamps were placed on the carotid artery and the internal jugular vein, and each was clipped distally in preparation for anastomosis. The flap from the donor pig was then secured, and the pedicle vessels were separated and cleaned of adventitia. Under the operating microscope, the segment of femoral artery from the flap pedicle was sutured to the carotid artery in an end-to-end manner using interrupted 9-0 nylon sutures. The segment of the femoral vein from the flap pedicle was coupled to the internal jugular vein using an appropriately sized coupler. The vessel clamps were removed and reperfusion of the flap with good arterial and venous flow was confirmed. Finally, the flap was inset to the right neck of the swine, and skin was excised from the neck to allow for tension-free inset. The flap was sutured to the defect with 3-0 Vicryl deep dermal sutures and a subcuticular running 3-0 Vicryl. A Doppler signal was found on the flap and marked with a 3-0 nylon suture. After recovery of the recipient animals, the flaps were monitored closely for the first 48 h with assessment of color, turgor, and capillary refill.

For split-thickness skin graft (STSG) transplantation, the skin was sterilely prepared, and 0.6-mm-thick pieces were procured with a Zimmer dermatome and briefly stored in cold saline before transplantation to the recipients. To prepare the graft bed on the recipients, two 0.6-mm dermatome passes were used to deepen the wound. The grafts were protected with an occlusive dressing for the first 3 d postoperatively.

Both VCA skin and STSGs were inspected daily for evidence of rejection. VCA skin rejection was scored 0–4 according to Etra et al.⁶ The day of rejection for VCA skin and STSGs was defined as the day on which <10% of the skin showed signs of viability as judged by color, texture, and warmth to touch. Unless the STSGs had already been rejected, biopsies were collected in formalin on postoperative day 0, 4, 7, 9, 11, 14, 28, and 35. Biopsies from VCA grafts were collected in formalin during suspected rejection crises and at endpoint. Fixed tissues were moved to alcohol (RGA70, Azer Scientific, Morgantown, PA) and thereafter embedded in paraffin, and 3-µm sections were stained with hematoxylin and eosin before scoring. Sections were scored 0–4 according to Banff 2007 Working Classification of Skin-Containing Composite Tissue allografts.⁷

Sample and Tissue Collection and Processing

Peripheral blood was collected at day –4, –3, –2, –1, 0, 2, 5, and thereafter weekly for tracking of T-cell depletion and reconstitution, and/or chimerism monitoring (CD45, CD3, CD4, CD8α, CD25, CD45RA, CD21, CD14, MHC class IC, and MHC class ID). Blood was also collected for isolation of peripheral blood mononuclear cells (PBMCs) to be used in assays. After euthanasia, thymic and bone marrow tissues were collected for chimerism analysis. Details of sampling, tissue processing, and cell staining can be found in Supplemental Materials and Methods (SDC, https://links.lww.com/TXD/A817).

Mixed Lymphocyte Reaction and Lymphocyte Versus Skin Reaction Assay

Mixed lymphocyte reaction (MLR) assays were carried out pretransplant and at regular intervals after transplantation to assess recipient responses to donor and third-party MHC antigens. For pig 26542, a lymphocyte versus skin reaction (LSR) assay was plated at the time of skin rejection to assess in vitro responses against skin antigens. For comparison with a pig that did not reject any part of the VCA graft, an LSR assay was plated for pig 26442 at endpoint day 204. Details of the assays can be found in Supplemental Materials and Methods (SDC, https://links.lww.com/TXD/A817).

Assessment of Alloantibodies

The presence of antidonor antibodies (IgM and IgG) in serum was examined by flow cytometry for the animals that rejected any part of the VCA. Details of the assay can be found in Supplemental Materials and Methods (SDC, https://links.lww.com/TXD/A817).

Data Analysis

Flow cytometry data were analyzed using FCS (Flow Cytometry Software) Express (De Novo Software, Pasadena, CA).

RESULTS

Persistence of Peripheral Blood Chimerism Generated by Hematopoietic Stem Cell Transplantation was Associated With Class I Sharing

All animals tolerated the induction regimen without toxicity, and the degree of T-cell depletion was similar in both groups (Figure 2). No animals developed signs of graft-versus-host disease. Multilineage mixed chimerism was initially established in both groups (group 1, with 1-haplotype MHC class I mismatch and MHC class II fully matched, and group 2, with MHC class I fully mismatched and MHC class II 1-haplotype mismatched) after nonmyeloablative conditioning and HCT (Figure 3). In group 1, multilineage chimerism levels in peripheral blood remained stable throughout the experiment, even after immunosuppression was discontinued. In group 2, chimerism levels remained high during the initial period when blood cyclosporine levels were elevated (trough levels 400–800 ng/mL). However, chimerism levels gradually declined for both animals in this group when immunosuppression was weaned. The percent of lymphoid lineage donor cells decreased <20%, whereas CD14 (myeloid) and CD21 (B cell) chimerism eventually fell <1% in both animals, indicating loss of bone marrow engraftment (Figure 3).

FIGURE 3.

Chimerism in peripheral blood. A, Percentage of donor cells in peripheral blood for pigs in group 1 (No. 26347, 26442, 26541, 26542) and group 2 (No. 26667 and 26668). B, Gating strategy for chimerism data analysis in FCS (Flow Cytometry Software) Express. MHC, major histocompatibility complex; SSC, side scatter.

Persistence of Bone Marrow and Thymic Chimerism was Associated With Class I Sharing

Postmortem analysis of bone marrow demonstrated CD21 (B cell) and CD14 (myeloid) chimerism in group 1 but not in group 2 (Figure 4). Donor-derived CD3⁻CD4⁺CD8⁺ double-positive thymocytes were present in all pigs in group 1 but not in group 2 (Figure 4). These data demonstrate that engraftment was lost in group 2 but maintained in group 1.

FIGURE 4.

Bone marrow and thymus chimerism postmortem for groups 1 and 2. MHC, major histocompatibility complex.

HCT Generated Tolerance of Donor MHC in all Groups, but Tolerance of Donor Tissue-specific Antigens was Associated Only With Class I Sharing

A summary of clinical outcomes is presented in Figure 1, together with results from previous studies by Leonard et al¹ and Shanmugarajah et al,² and the clinical presentation and histology of VCA grafts at endpoint and/or the day of rejection are presented in Figure 5. Of the 4 swine in group 1, 2 (No. 26442 and 26541) accepted the VCA graft without any rejection episodes, whereas one-third (No. 26542) rejected the epidermis of the graft on postoperative day 96. The underlying VCA tissue remained intact, as judged by clinical examination, bleeding at the time of biopsy, and histopathology. The fourth pig (26347) lost the VCA graft 24 h after surgery due to a surgical complication (venous thrombosis).

FIGURE 5.

Clinical outcomes of VCA grafts. Pictures of VCA grafts with corresponding images of hematoxylin and eosin–stained sections at or at time of. The classification for each histology section was graded according to the Banff scoring system for skin-containing composite tissue allografts.⁷ Group 1: (A) pig 26442 at endpoint day 204, no clinical rejection, histology grade 0; (B) pig 26541 at endpoint day 142, no clinical rejection, histology grade 0; (C) pig 26542 at time of epidermal rejection day 96, histology grade III; and (D) pig 26542 at endpoint with no rejection of underlying VCA tissue day 152, histology grade 0. Group 2: (E) pig 26667 at endpoint day 203, clinical signs of chronic rejection, histology grade II; and (F) pig 26668 at time of full-thickness rejection day 99 (clinical picture) and 104 (histology slide), histology grade III. MHC, major histocompatibility complex; VCA, vascularized composite allograft.

In contrast, rejection occurred in both swine in group 2. In the first pig (26668), the VCA graft became hyperemic on day 54, progressing to full rejection of the epidermis on day 99 and subsequently full-thickness rejection within 4 wk, as demonstrated by no bleeding at the time of biopsy on day 127, histological examination, and the VCA graft separating from the underlying tissues on the following day. In the second pig (26667), the VCA graft became hyperemic on day 96 and developed crusting and ulceration. After recovering, the graft became hyperemic again on day 134 and remained discolored throughout the rest of the experiment. Histology confirmed low-grade chronic rejection.

We tested the nature of the VCA tolerance both in vivo and in vitro. In vivo, we challenged the swine approximately 100 d after HCT/VCA with STSGs from (1) the HCT/VCA donor, (2) a swine MHC-matched to the donor (differing only in minor antigens), and (3) self (autologous graft). All autografts appeared healthy indefinitely after transplantation. The pigs that had rejected the epidermis of the VCA graft rapidly rejected the STSGs from both the donor and donor-matched pigs (Figures 1 and 5). All pigs that accepted their VCA graft subsequently accepted a donor STSG without rejection and demonstrated delayed rejection of the donor-matched STSG, demonstrating tolerance of the donor’s MHC and additionally of the specific donor’s skin-specific antigens. Notably, this was also true for pig 26347, which had lost its VCA graft in the early postoperative period due to venous congestion, suggesting that 24 h of exposure to the skin-specific antigens in the vascularized graft was enough to tolerize the pig against those antigens, which are presented only in the tissue and not on the transplanted donor HSC.⁸

To assess tolerance in vitro, MLR assays were performed before transplant and at various time points after transplant. Representative MLR data for 1 pig from each group are presented in Figure 6. Regardless of whether epidermis or VCA grafts were rejected or whether engraftment was lost, all pigs maintained donor-specific hyporesponsiveness after HCT, consistent with rejection not being directed against MHC. To investigate the potential contribution of regulatory T cells (Treg) to the maintenance of tolerance, responder cells were depleted of CD25hi cells, approximately 75% of which are FOXP3⁺ (Figure 6). Donor-specific hyporesponsiveness was not affected by depletion of these cells, suggesting that Treg in peripheral blood did not play a significant role in maintaining tolerance against MHC (see Discussion). We also assessed the development of donor-specific antibodies in recipient serum collected at various time points from animals that had rejected the epidermis or the dermis/fascia of the VCA graft. None of these animals developed IgM or IgG donor-specific antibodies.

FIGURE 6.

MLR assays and the leukocyte vs skin reaction assay. Representative results from MLR assays for group 1 (A) and group 2 (B). PBMCs from pig 26542 (group 1) and pig 26668 (group 2) as responders were plated against PBMC stimulators from self, HSC, and VCA donors, donor MHC-matched pig, and third-party pig. Negative control was media and positive control was PHA. For MLR assays plated with and without CD25 depletion, flow cytometry data showing CD25 depletion are presented to the right of the MLR data. C, Representative data showing the proportion of CD25⁺ cells that are FOXP3⁺. D, Results from LSR assay, PBMCs from pig 26542 as responder, plated against skin cells from self skin and VCA skin at day 100. Negative control media and positive control PHA. E, Results from LSR assay, PBMCs from pig 26442 as responder, plated against skin cells from self skin, VCA skin, thawed donor skin, and third-party pig skin. Negative control media and positive control PHA. LSR, lymphocyte vs skin reaction; MHC, major histocompatibility complex; MLR, mixed lymphocyte reaction; PBMC, peripheral blood mononuclear cell; PHA, phytohemagglutinin; VCA, vascularized composite allograft.

Rejection was Associated With Sensitization to Donor Skin-specific Antigens but not to Donor MHC

Because rejecting animals did not lose tolerance of the donor’s MHC as assessed by MLR, we investigated whether the rejecting recipients had developed sensitization toward donor skin-specific antigens using a novel lymphocyte versus skin reaction (LSR) assay (article detailing the development of this assay is under review elsewhere, but all details of materials and methods are available in the “Supporting information for the web”). After pig 26542 rejected the epidermis of the graft, we plated its PBMCs with skin cells from the VCA donor and the self. The rejecting recipient did not respond to native skin but did respond to donor skin, whereas MLR responses showed no response against donor PBMCs, suggesting that rejection was caused by sensitization to skin-specific antigens not presented in the donor bone marrow (Figure 6).

DISCUSSION

We aimed to address a remaining knowledge gap raised by prior HCT/VCA experiments from our group regarding the contributions of MHC class I versus class II matching to tolerance. By assessing additional MHC class I and II mismatch combinations, this study provides new insight into the relative importance of MHC class I and II matching for transplantation tolerance after combined HST/VCA transplantation. Combining HCT with VCA, rather than another organ, provided the opportunity to assess tolerance not only to MHC antigens but also to minor antigens, because the skin contains multiple tissue-specific (ie, skin-specific) antigens that can precipitate rejection, even when MHC is fully matched.^9,10 All animals maintained at least some degree of chimerism in peripheral blood throughout the study, despite normal white blood cell counts, no immunosuppression, and normal T-cell response to third-party antigens in vitro. This is consistent with the tolerance of donor MHC. Because tissue-specific antigens are not presented by donor HSC, HCT alone is not sufficient to induce tolerance to skin from the same donor.^11,12 Therefore, rejection of the donor skin but not the dermis, further suggests that tolerance developed to the donor’s MHC, which could be a deletional or regulatory process, whereas rejection of the donor’s tissue-specific minor antigens, which would require a Treg-mediated regulatory process.¹⁰

In our previous work, tolerance of both VCA and skin was maintained when there was a full class II mismatch between donor and recipient, with class I fully matched but not when there was a full class I mismatch with class II fully matched (summarized in Figure 1). The unanswered question was whether rejection in the second scenario was due to a lack of shared class I antigens or, alternatively, due to the complete sharing of class II antigens. In the latter case, full class II matching would render all donor cells directly alloreactive, rather than the 1%–10% that are more typically alloreactive, promoting rejection of all MHC-matched cells presenting mismatched donor minor antigens.¹³ We addressed this question using HCT/VCA pairs in which either only a single class I haplotype was mismatched or a single class II haplotype was matched, permitting us to assess whether full class I matching was necessary for tolerance and whether full class II matching was associated with rejection.

Combining the data in this report with the prior data from our laboratory supports the following 3 principles:

1. When MHC class I is fully matched, combined HCT/VCA generates durable multilineage mixed chimerism and tolerance of both donor MHC and tissue-specific antigens, regardless of the degree of class II matching.

2. When MHC class I is fully mismatched, combined HCT/VCA does not reliably lead to tolerance of the VCA, and the induction of tolerance and persistence of engraftment appears to depend on the extent of class II matching. When class II is fully mismatched, durable lymphocyte mixed chimerism does not form, and VCA grafts are rejected. When 1 haplotype of class II is matched, durable lymphocyte mixed chimerism persists at a low level, but engraftment is lost, and skin ± dermis/fascia of the VCA is rejected. When class II is fully matched, durable mixed chimerism is maintained, but some animals nonetheless reject the skin ± dermis/fascia of the VCA.

3. When a single class I haplotype is matched, combined HCT/VCA generates durable multilineage mixed chimerism, but consistent tolerance of both donor MHC and tissue-specific antigens occurs only when a single class II haplotype is also matched (haplotype mismatch). When MHC class II is fully matched, durable mixed chimerism is maintained, but some animals reject the skin, suggesting that, in the presence of a partial class I match, some level of class II matching is required for regulatory tolerance to occur but that (counterintuitively) full class II matching is less tolerogenic than partial matching, likely due to a mechanism that we propose below.

In summary, class I mismatching rather than class II matching appears to have been responsible for rejection in our prior experiments. However, in the presence of class I mismatching, the degree of class II matching modulates the occurrence of tolerance and whether it develops to MHC alone or to all donor antigens, including the tissue-specific antigens of the skin.

Durable chimerism and tolerance occurred in all cases when class I was fully matched, but even a single class I haplotype mismatch could precipitate rejection of donor skin. The role of class II matching appears to be more nuanced. In the absence of any class I matching, a greater degree of class II matching is associated with greater engraftment and chimerism and less rejection, through a mechanism yet to be defined. We hypothesize that full matching of class II (group 1) limits allopresentation to a combined direct/indirect pathway, whereas mismatching 1 class II haplotype (group 2) permits separate indirect and direct allopresentation of donor antigen, potentially increasing immunogenicity. For this reason, the delayed direct allopresentation of tissue-specific antigens is likely the reason rejection of the donor-matched skin graft precipitated the loss of donor engraftment and the rejection of VCA skin in both group 2 animals. In contrast, when only a single class I haplotype is shared, epidermal rejection sometimes occurs when class II is fully matched but not when only a single class II haplotype is matched (haplotype match). In the latter case, the immunologic advantage of having one matched and one mismatched class II haplotype is likely due to linked suppression, or “infectious tolerance,” which is a Treg-dependent mechanism of tolerance. We have previously demonstrated this mechanism to be operative when disparate antigens are co-presented by antigen-presenting cells that also present a tolerated (or self) class II haplotype.¹⁴ Such a Treg-mediated mechanism could explain not only the tolerance of the foreign class II MHC but also of any minor antigen (in this case a tissue-specific antigen from the skin) that it presents. We hypothesize that this is why this combination is the only group without full class I matching in which rejection of the epidermis does not occur. We will investigate this hypothesis in future work.

Our study has at least 3 limitations. First, although our observations hold true for this particular model of tolerance (the HCT/VCA model), they are likely not as relevant to tolerance models involving solid organs, the tissue-specific antigens of which are generally less immunogenic.⁸ Second, although the use of inbred lines generally decreases the number of animals required to achieve meaningful results,¹⁴ our conclusions would be strengthened by the addition of more animals in group 2, which was not possible due to limitations in breeding additional swine of the required genotypes. Although both swine in group 2 had similar clinical and immunological courses, future studies will confirm the findings from this initial study. Third, 2 scenarios were not tested. One is full class I matching with a single class II haplotype mismatch. We think it is likely that this scenario would also lead to durable mixed chimerism and VCA tolerance, because this was the case in experiments in which class II was fully matched or fully mismatched. The second scenario is single haplotype class I matching in the absence of class II matching. We expect that this scenario would lead to rejection because swine were able to mount rejection of the VCA skin even when a single class I haplotype was matched in the presence of full class II matching. Although the scenarios tested were those we thought most likely to provide mechanistic insights, both untested scenarios should also be evaluated in future experiments. However, the results provided for these experiments nonetheless provide further insight into the relative importance of MHC class I and II matching in this model. We did not consider it necessary to test a third scenario, that of the full MHC mismatch, because no VCA grafts were accepted after HCT across a full MHC mismatch in unpublished prior attempts (K. Shanmugarajah, personal communication, May 2024).

Our findings may have clinical relevance because combined bone marrow/solid organ transplantation tolerance protocols have been the subject of clinical trials for the past 20 y, with varying success, which may depend on the degree of MHC mismatching.^15–20 In addition, tolerance mediated by bone marrow transplantation may have clinical relevance for VCA grafts that contain bone marrow (eg, limb transplants). Indeed, it has been shown in a preclinical model in nonhuman primates that VCA grafts co-transplanted with vascularized bone marrow have significantly prolonged graft survival compared with VCA grafts that do not contain bone marrow.²¹ Achieving tolerance without complications has been more readily achievable in MHC-matched pairs than in haploidentical pairs, and it is most difficult in fully MHC-mismatched pairs.²² Because almost all current tolerance protocols use living donors, who are often relatives, the ability to choose donors with more or less class I similarity may be possible and could allow selection of the most appropriate donors. Moreover, even in conventional allotransplantation without tolerance, our findings raise the possibility that prioritization of donors with greater similarity in class I could decrease rejection rates. Our work thus provides incentive for further evaluation of the potential immunological benefits of selective MHC matching.

ACKNOWLEDGMENTS

The authors thank Brittany Bhola and Fei Huang for their veterinary and technical support. They thank the Institute of Comparative Medicine at Columbia University for animal housing and care, and a special thanks to Olena Shevchenko for skillful animal handling and care.