Correction of T-Cell Repertoire and Autoimmune Diabetes in NOD Mice by Non-myeloablative T-Cell Depleted Allogeneic HSCT

Abstract

The induction of partial tolerance toward pancreatic autoantigens in the treatment of type 1 diabetes mellitus (T1DM) can be attained by autologous hematopoietic stem cell transplantation (HSCT). However, most patients treated by autologous HSCT eventually relapse. Furthermore, allogeneic HSCT which could potentially provide a durable non-autoimmune T-cell receptor (TCR) repertoire is associated with a substantial risk for transplant-related mortality. We have previously demonstrated an effective approach for attaining engraftment without graft versus host disease (GVHD) of allogeneic T-cell depleted HSCT, following non-myeloablative conditioning, using donor-derived anti-3rd party central memory CD8 veto T cells (Tcm). In the present study, we investigated the ability of this relatively safe transplant modality to eliminate autoimmune T-cell clones in the NOD mouse model which spontaneously develop T1DM. Our results demonstrate that using this approach, marked durable chimerism is attained, without any transplant-related mortality, and with a very high rate of diabetes prevention. TCR sequencing of transplanted mice showed profound changes in the T-cell repertoire and decrease in the prevalence of specific autoimmune T-cell clones directed against pancreatic antigens. This approach could be considered as strategy to treat people destined to develop T1DM but with residual beta cell function, or as a platform for prevention of beta cell destruction after transplantation of allogenic beta cells.

Graphical Abstract

Open in new tabDownload slide

Introduction

The incidence of autoimmune diseases in general and T1DM in particular has been increasing worldwide over the last decade, affecting a significant proportion of the world population.¹ The most common therapeutic options, such as insulin replacement therapy in the case of T1DM, or chronic use of immunosuppression and biological drugs in the treatment of other autoimmune diseases, can alleviate symptoms but do not modify the underlying mechanism causing the disease. Exogenous insulin replacement therapies and techniques are advancing continuously but maintaining blood sugar level in normal ranges for long periods is still a struggle.^2-5 In other autoimmune diseases, the chronic use of immunosuppressive drugs exposes the patient to opportunistic infections, long-term risk of malignancies, and reduction in quality of life.⁶

Bone marrow transplantation, which could potentially “reset” the immune system to eliminate autoimmune clones and to create tolerance against the eliciting autoantigens, has been suggested as a potential treatment for these diseases. Considering the transplant-related mortality associated with allogeneic hematopoietic stem cell transplantation (HSCT), clinical applications of this approach were largely limited to autologous HSCT. However, such transplants can result in the remission of disease but most treated patients eventually relapse.^7-14 Thus, the development of relatively safe approaches for allogeneic HSCT, which potentially offer a more durable non-autoimmune TCR repertoire, is warranted.

To address this challenge, it is critical to eliminate the risk of GVHD, lethal infections, and other adverse effects associated with the use of myeloablation as part of the conditioning protocol.¹⁵ It is well established that GVHD associated with allogeneic HSCT can be prevented if T-cell contamination in the graft is kept under a threshold of 5 × 10⁵ cells per kg.^16,17 Such T-cell depleted HSCT (TD-HSCT) transplants are associated with high risk for lethal infections, due to the prolonged time required for immune reconstitution after the myeloablative conditioning.^18,19 This problem can be addressed by using mild non-myeloablative conditioning, which spares a substantial level of host immunity, but TD-HSCT following such mild conditioning is associated with high risk for allograft rejection.

We previously demonstrated in wild-type mice, that this challenge can be addressed by combining megadose TD-HSCT with donor-derived veto cells.²⁰ To that end, we have used naïve or memory CD8 veto T cells depleted of graft versus host (GVH) reactivity by selective expansion against 3rd party major histocompatibility (MHC), or against viral antigens, under culture conditions favoring expression of the central memory phenotype.²¹

The resulting central memory CD8 T cells are endowed with marked veto activity, able to enhance engraftment of TD-HSCT while effectively depleted of GVH reactivity in fully mismatched recipients.^21,22

However, whether such mild non-myeloablative T-cell depleted protocol can effectively eliminate the autoimmune T-cell clones in T1DM remained unknown.

Here, we treated T1DM-prone NOD mice with non-myeloablative megadose allogeneic TD-HSCT from C57BL/6 donors, in conjunction with donor-derived veto CD8 T cells. This transplantation modality led to marked chimerism without any transplant-related mortality and with a very high rate of diabetes prevention.

Using T-cell receptor sequencing, we show that the new chimeric immune system formed in NOD recipients exhibits marked changes in the entire T-cell repertoire and in the prevalence of specific autoimmune T-cell clones. Taken together, our results demonstrate a relatively safe allogeneic HSCT scheme able to prevent development of T1DM in pre-diabetic NOD mice.

Materials and Methods

Mice

Mice of the following strains were used at 6-12 weeks of age: C57BL/6 (H2K^b/H2D^b) and FVB (H2^q) were purchased from Harlan Israel (Rehovot, Israel). Progeny of Nude-C57BL/6 (Foxn1nu/J, H-2^b) was purchased from Jackson Laboratories (Bar Harbor, ME; Sacramento, CA, USA) and bred in the Weizmann Institute Animal Facility. NOD/LtJ (H2K^d/H2D^b) and β-actin GFP mice on C57BL/6J background were bred and maintained at the Weizmann Institute Animal Facility. Mice were kept in small cages and fed sterile food and acid water in a specific pathogen-free (SPF) environment under conditions approved by the Institutional Animal Care and Use Committee at the Weizmann Institute.

Diabetes Evaluation

NOD female mice were transplanted at the age of 8 weeks, before development of diabetes. Baseline measurements of weight and blood glucose were obtained and documented prior to conditioning administration. None of the mice in 4 different experiments developed diabetes prior to the conditioning at age of 8 weeks. After transplantation, mice were evaluated weekly for appearance and weight. Blood glucose was measured in tail-vein with a glucometer (Elite Diabetes Care System; Bayer, Leverkusen, Germany). Diabetes was defined as 3 consecutive glucose concentrations over 200 mg/dL over 2-6 days intervals.

In accordance with IACUC rules, mice that exhibited 2 blood glucose concentrations >400 mg/dL or loss of >20% percent body weight were euthanized. Deaths for other reasons were scored as transplant-related.

Preparation of Host Non-Reactive Donor Anti 3rd Party Veto Cells

Anti-3rd party T cells with a central memory phenotype (CD8+CD44+CD62L+; Tcm) were prepared as described.²⁰ Briefly, splenocytes from donor mice (C57BL/6J, H2^b) were cultured against irradiated 3rd-party splenocytes (FVB, H2^q) for 60 h without cytokines and CD8-positive cells positively selected using magnetic particles (BD Pharmingen, San Diego, CA, USA) and cultured in an antigen-free environment. Recombinant human interleukin-15 (rhIL-15; 20 ng/mL; R&D Systems. Minneapolis, MN, USA) was added every 2nd day. At day 16, Tcm veto cells were positively selected for CD62L expression using the Magnetic-Activated Cell Sorting (MACS Cell Separation, Miltenyi Biotec, Bergisch Gladbach, Germany), and cells retrieved for fluorescence-activated cell sorting (FACS) analysis (BD FACSCANTO II system running BD Diva software).

BM Preparation

Long bones were harvested from C57BL/6 nude (Foxn1nu/J, H-2^b) mice (7-12 weeks of age). Bone marrow was extracted by grinding the bones in a sterile bone crusher (Omni-Mixer Homogenizer, Omni International, Kennesaw, Georgia, USA).²³ To attain a single-cell suspension, the extract was strained and washed several times. Cells were counted using Türk’s solution (Sigma-Aldrich, Cat #1092770100), and a hematocytometer. Solution was brought to the correct concentration and was then injected into the tail vein. Fresh BM cells were used without any culturing or subsequent cell manipulation. No fat nor mesenchymal stroma cells were added.

Pre-Transplant Conditioning

Eight-week-old female NOD/LtJ (H-2K^d/H-2D^b) mice received 4.5 Gy total body irradiation (TBI) from X-ray irradiator (X-Rad320; Precision X Ray Inc., Madison, CT, USA) on day −1 (Fig. 1A). On day 0 mice received 25 × 10E+6 C57BL/6 nude (Foxn1nu/J, H-2^b) bone marrow cells mixed with 5 × 10E+6 anti-3rd party C57BL/6 (H-2K^b/H2D^b) veto Tcm by tail vein injection followed by subcutaneous injections of rapamycin (Rapamune, Pfizer Inc., New York, NY, USA), 12.5 μg/mouse/d, on days −1 to +4.

Chimerism Analysis

Blood chimerism was analyzed on day 30, and then periodically, on days 90, 180, and 260. Chimerism was determined by Multiparametric flow cytometry (MPFC). Blood was collected by retro-orbital bleeding using heparin-coated glass capillaries. Lymphocyte chimerism was determined in mononuclear cells fractionated on Ficoll-Paque Plus (Amersham Pharmacia Biotech, AB, Uppsala, Sweden) and double-stained for direct immune-fluorescence using labeled antibodies against donor and host MHC (H2K^b and H2K^d, respectively). FACS analyses were performed using a modified Becton Dickinson FACSCanto II. Multi-lineage chimerism was determined by processing spleen cells with 0.15 M potassium-ammonium chloride (ACK) lysing buffer as described followed by FACS analyses.²⁴ Cells were stained with labeled antibodies specific for CD8a, CD4, CD3, CD44, CD62L, CD11c, MHCII, CD80, CD86, CD11^b, B220, DX5, F4/F8, H2K^d, and H2K^b (Biolegend, San Diego, CA, USA; BD Bioscience, San Jose, CA, USA; Miltenyi Biotec, Bergisch Gladbach, Germany). Data were analyzed using FACSDiva 8.0. software and FlowJo v10.2 software (Tree Star, Ashland, OR, USA).

Immune Histochemistry

Mice were euthanized when they developed diabetes (defined above), loss of >20% body weight, or at day 260 post-transplant. Control pancreata were obtained from female NOD mice euthanized at ages 30 or 60 days. Pancreata were harvested and fixed with 4% paraformaldehyde (PFA) solution for 24 h, transferred into 70% ethanol, and paraffin embedded. Blocks were sectioned at 4 µ intervals, hydrated in H₂O, and antigen retrieved at pH 9.0. Slides were then incubated for 20 minutes with a recombinant anti-insulin antibody (ab181547, Abcam, Cambridge, UK), or anti-Glucagon antibody (NBP2-66869, Novus, Englewood, CO, USA), or anti-Somatostatin antibody (PAS-85759, Invitrogen, Waltham, MA, USA), or anti-pancreatic polypeptide (PPY) (NB100-1793, Novus), or incubated for 30 minutes with anti-CD45 antibody (ab10558, Abcam) followed by incubation in anti-rabbit HRP (DS9800, Leica, Wetzlar, Germany) for 15 minutes (for CD45, Insulin, Glucagon, and Somatostatin staining) or incubation in donkey anti-goat HRP (ab97110, Abcam) for 15 min (for PPY staining) visualized with diaminobenzidine, counter-stained with hematoxylin, dehydrated, and cover slipped. We used an inverted Olympus IX73 fluorescent microscope with X10, X20 air objectives, and Olympus megapixel digital camera (DP74-21) (Olympus Life Science, Waltham, MA, USA). Images were acquired by CellSens standard software (Olympus Life Science).

Image Acquisition by 2-Photon Laser Scanning Microscopy

Experiment was performed according to the scheme in Fig. 1A. Tcm were generated from GFP+ C57BL/6 donors. Hosts were sacrificed 1-year post-transplant and lymph nodes and spleens were harvested. Zeiss LSM 880 upright microscope fitted with Coherent Chameleon Vision laser was used to scan images acquired with a femtosecond-pulsed 2-photon laser tuned to 940 nm. The microscope was fitted with 505 LPXR mirror to split the emission to 2 GaAsp detectors (with a 500-550 nm filter for GFP fluorescence). Tile images were acquired as Z stacks. The zoom was set to 1.5 and pictures were acquired at 512 × 512 x-y resolution (Zeiss, Oberkochen, Germany).

Library Preparation for TCR-seq

All libraries in this study were prepared as described with minor modifications.²⁵ Briefly, total RNA was extracted from CD4+/CD8+ spleen cells using RNeasy Micro Kit (Qiagen, Germantown, MD, USA), and contaminating DNA was removed with DNAse 1 (Promega, Madison, WI, USA). RNA samples were reverse transcribed into cDNA and an encore region at the variable part of the TCR was added using single strand ligation. Ligation products were amplified by PCR in 3 reactions based on a nested primer approach. Our modified protocol for mice included specific primers for the constant region of the TCR α or β chains and used in the reverse transcription (RT) and the 1st 2 PCR (see below). Primers in the 2nd round PCR (PCR2) included constant region annealing part (“index”-constant sequence), 6 base pair Illumina (Nextseq 550 system, Illumina, San Diego, CA, USA) index (“index”), 6 random base pairs (“index”-random sequence), and the Illumina SP1 sequence. In the 3rd round PCR (PCR3), the reverse primer included SP1 and P5 sequences. In all PCR, the KAPA HiFi was used (KAPA Biosystems, Wilmington, MA, USA). Libraries were sequenced using NexsSeq 550 for 300 cycles (Illumina).

Pre-processing and Error Correction for Raw Reads

Data were processed using a tailor-made pipeline. Trimmomatic tool was used to filter out the raw reads containing bases with Q-value ≤ 20 and to trim reads containing adaptor sequences.²⁶ Remaining reads were separated according to their barcodes, and reads containing the constant region for α or β chain primer sequences were filtered

(CAGCAGGTTCTGGGTTCTGGATG/TGGGTGGAGTCACATTTCTCAGATCCT α and β chain) allowing up to 3 mismatches. Bowtie 2 was used to align reads to the germline V/J gene segments found in IMGT germline (using sensitive local alignment parameters).²⁷ Nucleotide sequences were translated in silico to amino acid sequences according to IMGT convention. Sequences were clustered according to their UMI, to correct for possible sequencing errors. The UMI of each CDR3 sequence was counted and UMI count reads with one copy number filtered. Sequences were used only if they were fully annotated (V, J segments assigned), in-frame (encode for a functional peptide without stop codons), and with copy number greater than one. We removed the invariant α chain of the iNKT CDR3 sequence (“CVVGDRGSALGRLHF”) which accounted for 0.001% of all sequences in our dataset and which are irrelevant to our TCRseq analysis.

Statistical Analysis

Statistical analyses were performed using R Statistical Software (version 4.4.0). For the pre-processing pipeline, we used the “ShortRead” package (version 1.48.0).²⁸ The package “vegan” (version 2.5-7)²⁹ was used to measure the Jaccard index,²⁹ and to project the Nonmetric Multidimensional Scaling.³⁰ The “ggplot2” (version 3.3.5)³¹ package was used for generating figures. Statistical significance was tested by the Wilcoxon rank test with FDR correction.³²

Results

Prevention of Diabetes After Reduced Intensity Mismatched Allogenic Bone Marrow Transplantation

To evaluate the ability of our reduced intensity protocol to prevent development of T1DM, we used 8-week-old NOD (H2^d) female mice, an age that precedes their development of diabetes. Recipients were irradiated with 4.5 Gy TBI (total body irradiation) (day −1) and transplanted with a mega dose of C57BL/6 (H2^b) “nude” mouse bone marrow (BM) combined with anti-3rd-party veto central memory T cells (Tcm) from the same donors (day 0). As indicated schematically in Fig. 1A, the recipient mice were also treated with rapamycin between day −1 and day +4, in accordance with our previous finding that such brief treatment with rapamycin enables chimerism induction if combined with veto cells following 4.5 Gy TBI.³³

Chimerism was evaluated after 30 days, and then periodically, at 90 days, 180 days, and 260 days.

In 4 repeated experiments, high levels of peripheral blood chimerism, ranging from 83.5% to 99.6% were found at different time points, between 30 and 260 days post-transplant in the group receiving bone marrow transplantation (BMT) plus veto cells (31 out of 35). Notably, the 4 mice which failed to exhibit chimerism have developed diabetes. In contrast, no chimerism could be detected, up to 150 days post-transplant, in the group receiving BMT without veto cells (10 mice out of 23 were evaluated as other mice developed diabetes earlier and were sacrificed). Further multi-linage analysis revealed a similarly high level of chimerism in various cell types in blood and spleen, including CD4 and CD8 T cells, B cells, dendritic cells, and macrophages (Fig. 2A–2D).

We next followed up to 260 days of age the survival and diabetic state of the different groups of mice. As shown in Table 1; Supplementary Fig. S1A, depicting the rate of diabetes development and the actual glucose level at different time points, respectively, the treatment group in the 4 experiments comprised a total of 35 mice, and only 4 mice (11.4%) developed diabetes, compared to 72.4% (21/29) of untreated NOD mice (P < .0001) or 95% (22/23) of NOD mice that received conditioning and BMT without veto Tcm cells (P < .0001). Furthermore, the actual cumulative development of diabetes or transplant-related mortality over time is shown in Fig. 1B, 1C, respectively. Thus, no GVHD as measured by body weight (Supplementary Fig. S1B) nor transplant-related mortality (Fig. 1C; Table 1) (0/35) was observed in the mice transplanted with BM and veto Tcm (Fig. 1C), strongly confirming the safety of this treatment modality.

To evaluate the transplantation effect on the pancreas tissue, we harvested the pancreata of control NOD mice early (30 and 60 days of age) and after outbreak of diabetes (beyond day 100), as well as of mice transplanted with BM and veto Tcm at 260 days of age.

As shown in Fig. 3, immuno-histological CD45 staining revealed marked infiltration of leukocytes into the islets at 30 (Fig. 3A) and 60 (Fig. 3B) days of age, and only remnants of islets could be detected at 125 days of age (Fig. 3C). In contrast, no infiltration of CD45+ leukocytes into islets could be detected in transplanted mice at the end of the follow-up period at 260 days of age (Fig. 3D).

Likewise, staining of insulin in the pancreata of untreated NOD mice harvested at 125 days of age could not detect insulin expressing islets (Fig. 4A, 4B) while in pancreata of treated mice harvested at 260 days of age, insulin staining revealed robust insulin expressing islets (Fig. 4C, 4D).

Staining for glucagon (Supplementary Fig. S2A), somatostatin (Supplementary Fig. S2B), and PPY (Supplementary Fig. S2C) was also positive at day 260 in the treated mice.

Veto Tcm Persistence in Transplanted NOD Mice

The degree of Tcm survival in recipient NOD mice is of major interest for evaluation of tolerance persistence as well as for future applications of these veto cells. To evaluate cell survival, we generated Tcm veto cells from GFP mice (C57BL/6 background). Cells were transplanted according to the protocol shown schematically in Fig. 1A, mice were sacrificed 1-year post-transplantation, and lymphoid organs were evaluated by 2-photon laser scanning microscopy. As shown in Supplementary Fig. S3, in 3 evaluated mice, GFP+ Tcm cells were present in the spleen and in lymph nodes.

T-Cell Repertoire Analysis in Chimeric NOD Mice Supports Modification of the T-Cell Repertoire

To interrogate potential shifts in the T-cell receptor repertoire of the transplanted mice, TCRseq analysis was performed on T cells from spleens of untreated NOD mice, treated NOD mice, and untreated C57BL/6 mice 8 weeks of age, used as donors in our transplantation experiments. NOD mice treated according to the protocol elaborated in Fig. 1A. Based on the time frame of diabetes progression in the untreated NOD mice described above, a TCR sequencing experiment was performed, comparing diabetic untreated control NOD mice at 22 weeks of age to transplanted mice at 22 weeks of age (14 weeks after transplantation). Thus, mice were sacrificed at this time point and tissue was obtained for T-cell repertoire analysis (detailed in the Methods section). Evaluation of the Vα and Vβ usage by principal component analysis (PCA) revealed significant differences between T cells classes (CD4 and CD8) and between the recipient (NOD) and donor (C57BL/6) mouse strains, as expected. Notably, the transplanted NOD chimeric mice show unique V-usage, which is localized between NOD and C57BL/6 cells in the PCA space (Fig. 5A). Comparison between the amino acid sequences of the variable complementary-determining region 3 (CDR3AA) across mice, revealed a unique Jaccard (clonal overlap ratio³⁴) sharing pattern. Transplanted groups show high similarity of CDR3AA across mice, which is distinct from the NOD or C57BL/6 repertoire. Notably, the α chain CDR3AA shows a greater separation between the treated NOD group and untreated NOD mice. This phenomenon can be explained by the lower diversity of the α vs the β chain CDR3AAs (mean Jaccard across all samples in α = 0.114/β = 0.03), which leads to higher Jaccard scores across mice (Fig. 5B).

Finally, we characterized the convergent recombination levels among mice and treatments. Convergent recombination (CR) is defined as the number of CDR3 nucleotide sequences that encode for the same CDR3AA sequence. High CR and frequency levels are associated with clonal expansion.^36,37 We focused on the β chain repertoire since it has more published diabetes-related sequences,³⁸ and it is intrinsically more diverse,³⁹ allowing us to use it as an identifier of annotated clones. CDR3AAβ clones from prediabetic and diabetic untreated NOD mouse groups showed significantly higher levels of CR than the treated NOD mice (Fig. 6A, 6B). We also searched for known CDR3 sequences connected to type I diabetes mellitus^40-42 and evaluated their presence in the high CR clone fraction. Notably, sequences related to GAD 65, nrpv7, or HSP60 were found only in untreated NOD mice, before and after development of diabetes (Fig. 6C). Furthermore, the annotated clones were found in the pre-T1DM untreated NOD mice only in the CD4 cells and in the sick untreated NOD mice only in the CD8+ cells, suggesting differential involvement of those clones in the disease process.

Discussion

It is well established that allogeneic HSCT can potentially “reset” the immune system in autoimmunity, destroying autoimmune T-cell clones, and replacing them with a normal donor-derived immune system.^43,44 However, HSCT is rarely used as a treatment, due to the difficulties in finding a matched donor, as well as the high-risk and toxicity involved in this treatment.⁴⁵ Nevertheless, there are reports of the curative potential of allogeneic HSCT in T1DM, in cases of allogeneic HSCT performed due to other indications, with cure of diabetes as an incidental outcome.^46,47

Thus, methodologies for allogeneic HSCT, which could potentially provide a durable non-autoimmune TCR repertoire, are needed. Accordingly, transplantation of TD-HSCT under mild conditioning, associated with minimal toxicity, and reduced risk of GVHD, offers an attractive therapeutic option in T1DM. However, a major basic question posed by this approach for “resetting” of the immune system in autoimmune diseases in general, and in autoimmune diabetes in particular, relates to the ability of such non-myeloablative transplant modalities to induce a significant change in the T-cell repertoire.

To address this questions, we used the well-established NOD murine model for autoimmune diabetes, and based on previous studies in wild-type mice,³³ we tested the efficacy and safety of a veto-based non-myeloablative TD-HSCT in T1DM prone NOD mice. Notably, there was no sign of GVHD and no transplant-related mortality in the transplanted mice.

Although minimal conditioning was used, chimerism analysis in the transplanted mice showed multi-lineage chimerism, in various immune cell types, including myeloid cells, dendritic cells, and lymphoid cells, suggesting a major alteration in the host immune system. Furthermore, analysis of lymphoid organs of treated NOD mice, using GFP-allogeneic Tcm, showed survival of Tcm in spleen and various lymph nodes up to one-year post-transplant, supporting the presence of durable immune tolerance.

In accordance with the normal glucose levels in 31 of 35 mice transplanted with BM and veto Tcm during a follow-up period of 260 days (compared to 8 of 27 in untreated NOD mice), evaluation of pancreatic tissue of treated mice showed the normal presence of Langerhans islets expressing insulin, glucagon, PPY, and somatostatin and without any islet infiltration of CD45+ leukocytes.

Notably, TCR sequencing revealed that the transplanted chimeric NOD mice showed unique V usage, which localized between NOD and C57BL/6 cells in the PCA space, reflecting the establishment of a new T-cell repertoire.

Considering that the V usage is known to correlate with the MHC type,⁴⁸ these results suggest that the transplant V usage is shaped by thymic negative (BM-derived antigen-presenting cells—donor) and positive selection (thymic cortex epithelial cells—host).⁴⁹ Even more strikingly, when we focused on the full CDR3AA sequences which are rarely shared across individuals, we found higher clonal overlap among mice in the same treated group than across conditions. This exclusive sharing pattern was more prominent in CDR3AA’s α vs. the β chain, consequently due to the low diversity and higher clonal sharing levels in the α chain.

Furthermore, and most relevant to our study, is the finding that CDR3AAβ clones from prediabetic and diabetic untreated NOD mice showed significantly higher levels of CR, consistent with expansion of autoimmune clones, compared to the chimeric treated NOD mice. From these high convergent clones, sequences related to GAD 65, nrpv7, or HSP60 were found only in untreated NOD mice, before and after development of diabetes, and were undetectable in the transplanted mice.

Overall, the general significant changes in TCR repertoire after transplantation and, in particular, the reduction of T1DM-related clones in the treated group, strongly support the hypothesis that our non-myeloablative allogeneic BMT procedure can completely and durably correct the pathogenicity of the T-cell repertoire in NOD mice.

Thus, as expected, when allogenic BM hematopoietic progenitor cells were “educated” in the thymus of NOD recipients, a different T-cell repertoire was generated, and autoimmunity was prevented even when using a mild non-myeloablative transplantation protocol with minimal risk for transplant-related mortality.

Conclusion

We developed an allo-transplant protocol that prevented T1DM in a pre-diabetic mouse model with long-term donor chimerism and no transplant-related mortality.

Notably, this protocol is readily adaptable to humans. A phase I-2 clinical trial, testing the safety and efficacy of a similar veto-based protocol, is currently accruing patients with hematological malignancies, sickle cell disease, and severe aplastic anemia in MD Anderson.⁵⁰ Once the safety of this non-myeloablative protocol will be established for clinical use, and upon further evaluation of its curative potential in diabetic mice before complete destruction of their pancreas, it could be primarily considered as strategy to prevent T1DM in patients with sufficient residual beta cell function or in recipients of allogeneic beta cells to avoid late rejection by autoimmune T-cell clones. It may also be relevant to patients with other serious auto-immune disorders such as systemic sclerosis and systemic lupus erythematosus.

Acknowledgments

Graphical abstract was created using “BioRender.com”

RPG acknowledges support from the National Institute of Health Research (NIHR) Biomedical Research Centre funding scheme.

Funding

This publication was funded by Cell Source Inc. and was supported by MD Anderson’s Histopathology Core Lab (HCL), Award Number P30CA016672 from the NIH National Cancer Institute. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Cancer Institute or the National Institutes of Health.

Conflict of Interest

RPG is a consultant to BeiGene Ltd., Fusion Pharma LLC, LaJolla NanoMedical Inc., Mingsight Pharmaceuticals Inc., CStone Pharmaceuticals, NexImmune Inc. and Prolacta Bioscience; advisor to Antengene Biotech LLC, Medical Director, FFF Enterprises Inc.; partner, AZAC Inc.; Board of Directors, Russian Foundation for Cancer Research Support; and Scientific Advisory Board: StemRad Ltd. YR is a consultant and shareholder of Cell Source Inc. All of the other authors declared no potential conflicts of interest.

Authors Contributions

R.S.M., B.N-L.: conception and design, collection and assembly of data, data analysis and interpretation, manuscript writing. S.R-Z., M.M.: conception and design, data analysis and interpretation, manuscript writing. L.S-B., C.R., I.M-K.: collection and assembly of data. E.B.L., N.F.: conception and design. R.P.G.: manuscript writing. Y.R.: conception and design, data analysis and interpretation, manuscript writing, final approval of manuscript. All authors reviewed the final typescript, take responsibility for the content, and agree to submit for publication.

Data Availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

References

Author notes