Cultured thymus tissue implementation promotes donor-specific tolerance to allogeneic heart transplants

. In-Press


Introduction
Ray Owen first observed immune tolerance to other individual's blood in Freemartin calves (1), and later Peter Medawar was the first to induce transplant tolerance by injecting neonatal mice with adult donor tissue/cell suspension (2).These examples of tolerance induction were related to the immature nature of the neonatal immune system (3,4).Later, the central role of T cells in transplant rejection was identified (5,6), and deletion of developing alloreactive T cells in the host thymus was thought to be the primary mechanism of tolerance induction (7,8).There were many attempts to promote tolerance to allogeneic organs by introducing thymus composite tissues or vascularized thymus in which the thymus expressed donor MHC in miniature swine (9)(10)(11)(12).The thymoheart was formed by injecting finely minced thymus tissue of the donor type into the atrial appendages of miniature swine.
Those hearts were able to survive in an allogeneic recipient for up to approximately 194 days (10).A similar procedure was performed for the thymokidney (9).Tolerance was observed in the miniature swine that received a thymokidney and were followed for up to 60 days.Another study transplanted allogeneic vascularized donor thymic lobes and donor hearts at the same time (12).Allograft survival in the miniature swine was followed up to 301 days when the last animal was sacrificed.Despite success for up to 300 days, data have not been published showing the long-term success needed for human transplantation.
Experimental allogeneic cultured thymus tissue implantation used in 86 athymic infants with complete DiGeorge anomaly resulted in development of naïve T cells in the 61 survivors (71%) (13,14).Almost 30% of the infants died from preexisting conditions and infections.Studies of tolerance induction in survivors after CTTI from 1993 to 2018 using mixed lymphocyte reactions showed development of tolerance to the alloantigens of the thymus donor in the athymic infants (14).One example of tolerance induction to a solid organ occurred in an infant with complete DiGeorge anomaly who had profound hypoparathroidism.The infant was treated with CTTI plus parental parathyroid.The CTTI expressed the Class II antigens of the parent that were not inherited by the infant (15).The CTTI did not express the 3 mismatched parental Class I alleles.The parental parathyroid functioned for 10 years with not immune suppression until the patient was given a measles/mumps/rubella (MMR) vaccine.The parental parathyroid was rejected within 2 weeks and the patient had to return to calcium replacement.The mechanism for the loss of parathyroid function was likely rejection of the parental parathyroid by recipient CD8 T cells, one third of which are inherently alloreactive and would have recognized the parental unmatched Class I antigens as foreign.If the CTTI had expressed the uninherited Class I antigens, the CD8 T cells developing in the thymus would have been deleted if they reacted against the uninherited Class I antigens, and the patient would almost certainly still have parathyroid function.
Several other findings in recipients of CTTI should be mentioned.Recipients of CTTI were able to control viral infections such as Epstein-Barr virus that would have been fatal prior to CTTI (16).In addition, these infants, who had essentially no naïve T cells prior to CTTI, developed naïve T cells approximately 6 months after CTTI (13,17).
Biopsies of the transplanted cultured thymus tissue have demonstrated thymopoiesis on immunohistochemistry (18).

In vitro thymic cultures, in vivo engraftment
In order to perform allogeneic CTTI in a rat model with procedures used for the treatment of athymic infants with complete DiGeorge anomaly, thymus was harvested from 3-day-old F1(Lewis x Dark Agouti, LWxDA) rat pups.
Each postnatal thymus was cut into 4 pieces and cultured for 5 to 7 days in thymus organ medium (Fig. S1; see also methods).Histologic analysis (Fig. S2) showed reduced Ki67 + , CD3 + cells in the thymus after culture, similar to the change seen after culture of thymus tissue used for patients (18).As in cultured human thymus, the network of thymic epithelial cells (TEC) was preserved in the rat cultured thymus tissue based on cytokeratin (CK) staining (Fig. S2).The thymus, after being in culture for 5 to 7 days, was transplanted under the kidney capsule of a Lewis rat.As shown in Fig. 1A, recipients were thymectomized and treated with anti-CD5 mAb (OX19; 1 mg, IP) every 5 days for three doses.The recipients also received cyclosporine (CsA), approximately 2.5 mg/kg/day after thymus transplantation using osmotic pumps.CsA was discontinued 4 months after thymus transplantation when the test group had naïve T cells over 10%.At 6 to 7 months, the 3 rd party BN heart was transplanted in the neck.The thymus graft and all hearts were evaluated at necroscopy 8 months after CTTI when the test group rejected the cervical BN heart.As predicted, recipient-derived LW T cells, not expressing DA MHC, appeared in the peripheral blood of thymus transplant recipients (Fig. S3, panel B).We evaluated T cell and its subpopulations as well as non-

T cell populations (B cell and NK cell) via flow cytometry (Fig 1B and Fig S4).
Gating strategies for each subpopulation is shown in supplemental figure 5. We predicted that the LW T cells would be tolerant to the DA antigen of the DA heart because the LW T cells had developed in an LWxDA thymus (Fig. 1A).Total circulating CD3 T cell numbers were not significantly different in between groups prior to transplantation.As expected, LW recipients with CTTI showed significantly increased numbers of circulating CD4 and CD8 T cells compared to control animals without CTTI after transplantation (Fig. 1B).Furthermore, animals with CTTI showed significantly increased repopulation of naïve (CD62L + CD45RC + ) CD4 and CD8 T cells as well as recent thymic emigrant (RTE, CD90 + CD45RC -) T cells in the peripheral blood while control groups without thymus transplantation showed low level of circulating naïve CD4 and CD8 T cells and did not show circulating RTE CD4 and CD8 T cells (Fig. 1B).
In addition, immunohistologic analysis of transplanted cultured thymus tissue explanted on day 180 showed normal thymus histology, viable T cells (CD3), T cell proliferation (Ki67), and a lacy pattern of CK with Hassall body formation (arrow) on TECs from the CTTI (Fig. 1C).These observations confirm the viability and function (thymopoiesis) of the transplanted thymus in the animals receiving allogeneic heart transplantation.Taken together, rats that received cardiac allografts with CTTI demonstrated thymopoiesis with naïve T cell development.

No graft rejection with or without CTTI in thymectomized recipients
It was expected that T cells reactive to the DA donor would not develop since the T cells developed in CTTI expressing DA as well as LW.We evaluated the DA heart for evidence of rejection.As shown in Fig. 2A, LW rats with DA heart transplants without any immunosuppressive treatment rejected the DA heart grafts within 10 days (the DA control).However, even after developing RTE (CD90 + CD45RC -) T cells, LW recipients with CTTI did not reject (no cessation of beating) the DA cardiac allografts (n=8).Unexpectedly, LW control animals without CTTI also did not reject the DA cardiac graft (n=9).Both groups showed good beating quality for the entire study period (day 180).Since continuous graft beating does not necessarily imply absence of rejection, we sacrificed two recipients two months after cessation of immunosuppression (prior to 3 rd party BN cervical heart transplantation) to confirm that there was no rejection.The explanted cardiac allografts (DA hearts) from both animals showed minimal mononuclear cell infiltration (Fig. 2B) with no signs of rejection by 2004 International Society for Heart &Lung Transplantation (ISHLT) grading (Fig. 2C).Based on the reconstitution of naïve T cells after CTTI, we believe that animals with CTTI lost their donor-reactive T cell repertoire while animals without CTTI did not fully reconstitute their T cell populations (general hyporesponsiveness).

Alloreactivity against third-party vascularized heart transplantation
In order to confirm the donor-specific unresponsiveness (tolerance) versus general hyporesponsiveness, we performed additional fully MHC mismatched BN heart transplantation in both groups of animals at 6 to 7 months (day 180 to 210) after DA heart transplantation (Fig. 1A).As shown in Fig. 3A, LW rats with CTTI rapidly rejected (cessation of graft beating) the third-party BN heart (n=5, median survival time (MST) =10±1.0days).However, the control LW animals without CTTI did not reject the third-party hearts (Fig. 3A) (n=6, MST³38.5±8.9 days), possibly due to the lack of any allo reactive T cells.In accordance with the rejection of 3 rd party heart by animals with CTTI and the lack of rejection of the third party heart in animals without CTTI (Fig. 3A and 3B), histological analysis (Fig. 3C) confirmed that animals with CTTI showed increased mononuclear cell infiltration in the heart allograft (Fig. 3C, 3 rd panel) while animals without CTTI showed a pristine BN heart allograft (Fig. 3C, 4

th panel).
It is also notable that the BN hearts in the recipients with CTTI were greatly enlarged (Fig. 3C,

Selective T cell infiltration in third-party hearts but not in DA hearts that shared the DA MHC of the CTTI
We used two conventional ways to define graft rejection in this rat heart transplantation model: heart beating/cessation measurements and the ISHLT human grading system.The former is insensitive with respect to low-grade rejection, while the latter is insensitive with respect to high-grade rejection.Therefore, we evaluated inflammatory cell infiltration in DA hearts from 3 rats at day 180 and in BN hearts at the time of sacrifice at 7 to 8 months in 5 rats.As shown in Fig. 4A, rats that were treated with T cell depletion, CTTI, and four months of CsA did not show an increased level of inflammatory cell infiltration in the DA hearts after T cell repopulation.On the other hand, the animals given CTTI showed massive inflammatory cell infiltration in the third-party cardiac allograft (BN heart) (Fig. 4B).Rats without CTTI showed no infiltrates in the BN heart (Fig. 4B).Finally, we evaluated T cell infiltration with immunohistochemistry and confirmed a selective T cell infiltration in the BN (Fig. 4C, 3 rd panel), but not DA (Figure 4C, 2 nd panel), hearts of the animals with CTTI and a lack of T cell infiltration in both hearts of animals without CTTI (Fig. S7).These data confirm that the T cell infiltration occurs only in the thirdparty BN graft, but not in the graft sharing MHC (DA) with the transplanted thymus, possibly due to lack of T cell repertoire (by negative selection) against (DA heart) donor antigens.

Humoral response against donor antigens after thymus transplantation
After native LW thymectomy followed by LWxDA CTTI, we hypothesized that there would be a lack of T cell help for cognate B cell and downstream humoral response against the DA donor antigen in recipients.Therefore, we evaluated anti-donor antibody responses to determine whether the allogeneic T cell unresponsiveness was associated with humoral tolerance against donor DA MHC.We tested serially collected recipient serum samples and performed flow crossmatch with PBMCs from DA and BN rats.As expected, unmanipulated LW rats that received DA or BN heart transplants without immunosuppression developed antibody against their donors (DA or BN, respectively) (Fig. 5A, first panel in each row).Animals with a syngeneic cardiac allograft did not produce antibody against either DA or BN MHC (Data not shown).Interestingly, similar to T cell hyporesponsiveness, no anti-DA Ab was detected in animals with or without CTTI (5A, 2 nd and 3 rd panels of top row and Fig. 5B).Anti-BN Ab was readily detected in animals with LWxDA CTTI but not in animals without CTTI, p<0.01 (5A, 2 nd and 3 rd panels of bottom row and Fig. 5C).Taken together, thymus co-transplantation resulted in specific tolerance to the allogeneic DA MHC expressed in the donor thymus, and thus long-term survival of the DA heart transplant via preventing development of both the donor-specific anti-DA T cell repertoire as well as preventing the donor (DA)-specific humoral response.Immunocompetence was demonstrated in these rats by the rapid rejection of third-party BN hearts as well as alloantibody response against BN donor cells.

Discussion
Achieving donor-specific immune tolerance remains the ultimate immunologic goal in transplantation.Most of the current approaches focus on controlling peripheral mature donor-reactive T cells by depletion (e.g.alemtuzumab, thymoglobulin, etc.) or suppression (e.g.calcineurin inhibitors, basiliximab, etc.) without targeting the production of alloreactive T cells in thymus.However, even with the dramatic advancement of immunosuppressive drugs and new immunomodulatory regimens, transplant tolerance has not yet been consistently achieved.The use of thymus tissue to promote or transfer immunologic tolerance such as by intrathymic injection of antigens has been investigated in many animal models (19)(20)(21)(22)(23)(24).Over all, these studies have not been convincing due to the lack of histological evidence, proper controls, long-lasting efficacy, or translation into large animal models or humans.Our approach to achieve tolerance to donor antigens is to use donor CTTI to induce immune tolerance to allogeneic antigens.We hypothesize that tolerance to self is achieved because recipient dendritic cells induce apoptosis in thymocytes expressing T cell receptors that bind with high avidity/affinity to self-peptide:self-MHC on recipient dendritic cells (25)(26)(27).In addition, tolerance to donor is achieved because TEC (28) present donor antigens directly, donor-peptide:donor-MHC (29)(30)(31)(32)(33) or indirectly via recipient DC, donor-peptide:recipient-MHC to thymocytes.Thus, TECs are a key component promoting donor-specific tolerance by deletion of thymocytes that bind strongly to recipient or donor MHC during allogeneic thymus transplantation.
Tolerance induction by CTTI is similar to tolerance induction via donor DC in hematopoietic stem cell transplantation (34,35).A series of studies from Transplantation Biology Research Center (TBRC, Boston, MA) showed the crucial role of thymus in tolerance induction (36) and tested thymus transplantation with tolerance induction in a large animal model (9,37,38).In their series of studies, this group successfully used Class II matched/Class I mismatched donor (thymus and kidney or heart) as thymus composite tissues (thymokidney and thymoheart) with 12 days of CsA for transplant tolerance induction.This elegant concept of generating vascularized thymus prior to transplantation to induce tolerance, would be difficult to translate to the clinic without using xenotransplantation.The authors stated that non-vascularized thymus did not induce tolerance in their model.More precisely, however, non-vascularized thymus that was not cultured did not engraft long-term.We believe that the reason for lack of function of non-cultured thymus is that there is so much cell death in freshly harvested thymus.
The dying thymocytes release deoxyadenosine which diffuses into neighboring thymocytes.The deoxyadenosine is then phosphorylated to deoxyadenosine monophosphate (dAMP) which is trapped in the cell.The dAMP is further phosphorylated into dATP which inhibits ribonucleotide reductase which prevents DNA synthesis (39,40).
Thymocytes die in this environment.The culture system used in this report likely prevents buildup of deoxyadenosine.In particular media is dripped on the thin thymus slices that are floating on sponges in a tissue culture dish.Every day the old medium is removed and new medium is dripped onto the tissue.The deoxyadenosine from the thymocytes is washed off the slices and thus does not lead to the toxic pathway described above.
The limitation of non-vascularized thymus transplantation can be overcome with a culture system as well as T cell depletion as shown in the present study.Experimental transplantation of allogeneic CTTI that retains TECs has been successfully applied to treat pediatric patients with complete DiGeorge anomaly who are born without thymus or thymus function (17,41).Thymopoiesis has been documented by allograft biopsies and the presence of recipient naive T cells in the periphery (13,18,42).Studies of children treated with investigational CTTI show tolerance to donor MHC by mixed lymphocyte reactions (14).In addition, the infants with complete DiGeorge anomaly, after CTTI, are able to control infections such as Epstein Barr virus (16,19).Based on these data in human infants, we hypothesize that implantation of allogeneic thymus expressing the MHC of a solid organ donor will result in donorspecific tolerance to both self and to the donor and also will retain functional T cells that will protect the recipient from infection.
In the present study, we largely adapted techniques from the clinical setting of CTTI.Rat thymus was cultured in a similar manner to that used for human thymus (42,43).Cultured thymus tissue maintained the normal thymus structure but was partially thymocyte-depleted consistent with human data (Fig. S2).As shown in Fig. 1, panel C, and Fig. S1, panel D, the CTTI was well-engrafted under the kidney capsule with normal thymus structure.To test our tolerance hypothesis, we performed heterotopic DA heart transplantation together with F1(LWxDA) CTTI in thymectomized, T cell-depleted LW rats (Fig. 1A).The control group did not receive CTTI.Both groups of thymectomized LW rats developed circulating T cells after T cell depletion with anti-CD5 mAb followed by DA heart transplantation (Fig 1B).In the control group, the return of circulating T cells was likely due to T cell repopulation from the periphery (memory T cells) whereas the group with CTTI also had, in addition, development of naive T cells in the thymus (Fig. 1B).The CTTI in LW rats that were given DA heart transplants showed normal thymus structure with thymopoiesis (Fig. 1C).There was no DA heart graft rejection in either group (with or without a F1(LWxDA) CTTI) when the heart beating was assessed by palpation (Fig. 2A).We hypothesized that the T cells in the control group, although present in the circulation, were not functional.To test this, we used a thirdparty BN heart transplant to assess if either group could reject the third-party graft.The third-party BN heart allografts were promptly rejected from animals with CTTI but were not rejected by animals without CTTI.Thus, only the animals with CTTI had specific tolerance to donor MHC with immunocompetence demonstrated by the ability to reject allogeneic BN hearts.
Currently, long-term graft survival is often hindered by donor-specific antibodies even with successful T cell control (44).We tested whether the donor-specific tolerance induced by cultured donor thymus tissue transplantation would have an impact on post-transplantation humoral responses.Our hypothesis was correct in that thymectomized, T cell-depleted LW rats given F1(LWxDA) CTTI did not develop DSA against the DA antigens in the abdominal DA heterotopic heart transplant (Fig. 5A, top row, middle panel and Fig. 5B) but did make DSA against the BN heart as detected at the time of graft rejection (Fig. 5A, bottom row, middle panel and Fig. 5C).This suggests that allogeneic, thymus-induced, donor-specific tolerance might control anti-donor humoral response as well.Lastly, it is known that the major population of Foxp3 + Treg is generated in the thymus (45).Therefore, it is possible that Treg cells generated from CTTI could also provide additional regulation against donor Ag reactive T cells.
We believe that the present study provides proof-of-concept for donor thymus co-transplantation with solid organs for tolerance induction.The patient group that would most benefit from the procedure is infants needing heart transplants.Since postnatal thymic tissue is present and could be removed from the deceased donor infant, and the recipient thymus is routinely removed from infants undergoing heart transplantation, no additional procedure aside from CTTI would be needed to transfer this approach to the clinic.Overall, these results support the use of CTTI for the tolerance induction in organ transplantation, however, it remains to be tested as to how efficiently transplanted cultured thymus tissue will develop donor-specific tolerance in an immunocompetent large animal model with clinically relevant immunosuppressive regimens before moving into the clinic.

Animal Models
Lewis (RT-1 l ) and BN (RT-1 n ) rats were purchased from Charles River.DA (RT-1 av1 ) rats were purchased from Envigo.F1(LEW/DA; RT-1 l/av1 ) were bred by protocol staff at the Duke Breeding Core Division of Laboratory Animal Resources facility.Lewis recipients received thymectomies as previously described (46).Briefly, the submandibular glands and sternohyoid muscle were separated with blunt forceps to expose the overlying the trachea.
A 1-to 1.5-cm incision was made in the sternal manubrium.A 7 cm alms-type retractor was used to retract the manubrium and the two halves of the sternohyoid muscles to expose the thymus.The thymus was grasped with blunt forceps and extracted.The cut ends of the sternum were closed with a single 3 to 4-0 silk suture.Two drops of 2.5mg/ml bupivacaine were applied on the incision and the outer layer of skin was closed with three or four 9-mm wound clips.All thymectomized rats were maintained on a diet containing Septra (PMI Nutrition International, LLC).To induce T cell depletion in vivo, 1mg anti-CD5 mAb (OX19; BioXCell, NH) was intraperitoneally administered on day 0, 5, and 10 after thymectomy and suppression with 0.25 mg/kg/d cyclosporine pump from day 0 (heart transplant & CTTI time point) to 4 months with respect to heart transplantation.All rats were used and maintained in accordance with the guidelines and compliance of the Duke Institutional Animal Research Ethics Committee.

In vitro thymus culture and CTTI
Thymus from three-day old neonatal F1 (LEW/DA) rat pups were harvested sterilely, cut into four pieces along the longitude natural seam, and transferred onto sterile nitrocellulose filters (MF-Millipore, Millipore Sigma) in a tissue culture dish with TOM medium.Thymus tissue was cultured in a CO2 incubator with 5% CO2 at 37 o C for the desired length of time (5 to 7 days).The medium was changed daily.The thymus organ medium (TOM) was composed of HAMS F12 (Life Technologies) at 86.5%; Hepes (Life Technologies) at 25mM; L-Glutamine Life Technologies) at 2mM; Fetal Bovine Serum (Life Technologies) at 10%; and Pen-strep (Life Technologies) used at 1x.On the day of transplantation, the thymus pieces were rinsed with fresh medium and transplanted under the kidney capsule of a Lewis rat with one secure suture (10-0 monofilament).All manipulations took place under sterile conditions in a biological safety cabinet.
Abdominal heart transplantation was performed using a modified technique of the methods described by Schmid et al. (47).Briefly, the donor heart was transplanted into the abdominal cavity of the recipient after a short period of cold ischemia in Euro-Collins solution.The donor pulmonary artery and aorta were anastomosed to the recipient inferior vena cava and descending aorta with an end-side fashion as the inflow and outflow vessels for circulation, using running 9/0 non-absorbable monofilament sutures.Cyclosporine A (CsA) was given via the osmotic pump (Model 2ML4, Alzet).The pump was loaded sterilely and surgically inserted subcutaneously to mid-dorsal area of recipients.The osmotic pump was replaced every month for 4 months.For full MHC mismatched BN (RT-1 n ) thirdparty heart transplantation to the DA heart bearing Lewis recipients, we used cervical vascularized heart transplantation described by Heron in a modified fashion (48).Briefly, the third-party heart was transplanted into the right side of cervical area via a longitudinal incision from submaxilla to the xiphoid.The donor pulmonary artery and external jugular vein were anastomosed end to end and the aorta was anastomosed to the right common carotid artery by cuffing technique.The grafts were monitored by daily palpation and later confirmed by laparotomy at the time of sacrifice.Animals were sacrificed on the day of rejection of the graft (cessation of beating) or a designated time point.

Histology, Immunohistochemistry (IHC), and morphological analysis
All cultured thymus and CTTI samples from under the kidney capsule were frozen in OCT (Optimal Cutting Compound; Tissue Tek).Control thymus tissue was obtained from newborn to 5-day-old rat pups.Four to five mm sections were stained for CD3 (polyclonal; Dako), Ki-67 (clone: SP6; Thermo), CK, (polyclonal; Invitrogen).IHC images were obtained using an Olympus Vanox AH-3 Microscope of the Olympus DP-70 Digital Camera System.The explanted hearts underwent serial sectioning (5µm) from the midventricular level to the base.H&E stains were performed for routine examination and grading of rejection.Graft infiltrating T cells were evaluated with polyclonal anti-CD3 (Dako) staining.Whole slides of grafts were scanned with an Aperio ScanScope XT (Aperio Technologies, Inc., Vista, CA).

Statistical Analysis
Experimental results were analyzed by a GraphPad Prism (GraphPad Software 7.0, San Diego, CA) and SAS version 9.4 (SAS Institute, Cary, NC).The log-rank test for differences in graft survival, student t-test for two group comparison and Turkey test was used for multiple group comparison.All the data were presented as mean ± SD.
Values of p which were less than 0.05 were considered as statistically significant.

Flow
cytometry and spectratyping have shown development of a diverse T cell repertoire.Based on the human data showing tolerance to unmatched thymus MHC antigens, we evaluated CTTI, with the same methods used clinically, for its ability to induce donor-specific tolerance in a rat heart transplantation model.Our studies show that transplanting unmatched hearts along with donor CTTI expressing the heart donor's MHC class I and class II antigens induces tolerance to the antigens of the donor heart while preserving alloreactivity toward other MHC antigens.
3 rd panel, Fig. S6 panel w/CTTI) while the DA hearts were smaller than the native hearts (Fig. S6 panel w/o CTTI).BN hearts from recipients without CTTI did not show any increase in size (Fig. S6).Histological analysis (ISHLT grading) of explanted BN hearts from rats with CTTI (Fig. 3C, 3 rd panel) showed grade 3R rejection with significantly increased inflammatory cell infiltration compared to syngeneic controls (Fig. 3C first panel) or rats without CTTI (Fig. 3C, 4 th panel).