Research ArticleTRANSPLANT

Mixed chimerism and acceptance of kidney transplants after immunosuppressive drug withdrawal

See allHide authors and affiliations

Science Translational Medicine  29 Jan 2020:
Vol. 12, Issue 528, eaax8863
DOI: 10.1126/scitranslmed.aax8863

Staving off immunosuppression

Kidney transplant recipients are continually at risk of rejection despite intense immunosuppressive regimens that can have detrimental side effects. Busque et al. conducted kidney transplants with a selected composition of donor hematopoietic cells to attempt to evade graft-versus-host disease. They monitored mixed chimerism, defined as at least 1% of circulating cells being of donor origin, and kidney graft health. Most patients in the fully human leukocyte antigen (HLA)–matched cohort experienced persistent chimerism and no rejection episodes after removal of immunosuppressive drugs. Persistent chimerism was less prevalent in a partially HLA-matched cohort, and long-term withdrawal of immunosuppressants was not feasible in these patients. These findings indicate that methods to promote mixed chimerism may improve kidney transplant outcomes.

Abstract

Preclinical studies have shown that persistent mixed chimerism is linked to acceptance of organ allografts without immunosuppressive (IS) drugs. Mixed chimerism refers to continued mixing of donor and recipient hematopoietic cells in recipient tissues after transplantation of donor cells. To determine whether persistent mixed chimerism and tolerance can be established in patients undergoing living donor kidney transplantation, we infused allograft recipients with donor T cells and hematopoietic progenitors after posttransplant lymphoid irradiation. In 24 of 29 fully human leukocyte antigen (HLA)–matched patients who had persistent mixed chimerism for at least 6 months, complete IS drug withdrawal was achieved without subsequent evidence of rejection for at least 2 years. In 10 of 22 HLA haplotype–matched patients with persistent mixed chimerism for at least 12 months, reduction of IS drugs to tacrolimus monotherapy was achieved. Withdrawal of tacrolimus during the second year resulted in loss of detectable chimerism and subsequent rejection episodes, unless tacrolimus therapy was reinstituted. Posttransplant immune reconstitution of naïve B cells and B cell precursors was more rapid than the reconstitution of naïve T cells and thymic T cell precursors. Robust chimerism was observed only among naïve T and B cells but not among memory T cells. No evidence of rejection was observed in all surveillance graft biopsies obtained from mixed chimeric patients withdrawn from IS drugs, and none developed graft-versus-host disease. In conclusion, persistent mixed chimerism established in fully HLA- or haplotype-matched patients allowed for complete or partial IS drug withdrawal without rejection.

INTRODUCTION

Seminal observations in cattle fetuses with shared placentas or in neonatal mice after injections of allogenic bone marrow cells have shown that after mixed chimerism is established in these species, it is self-sustaining. Mixed chimerism refers to continued mixing of donor and recipient hematopoietic cells in recipient tissues after transplantation of donor cells. This stable mixed chimerism results in long-term acceptance (tolerance) of organ grafts from the hematopoietic cell donors without the need for immunosuppressive (IS) drugs (13). Subsequent studies in adult laboratory animals showed that stable mixed chimerism could be established without graft-versus-host disease (GVHD) in major histocompatibility complex–mismatched recipients given lymphoid tissue irradiation (46) or total body irradiation (7, 8). These adult mixed chimeras developed tolerance to combined organ and bone marrow transplants from the same donors in the presence or absence of posttransplant IS drugs (49). We hypothesized that establishment of persistent mixed chimerism in patients given combined hematopoietic cell and kidney transplants from the same donor would result in tolerance and the ability to completely withdraw IS drugs without subsequent rejection.

Rejection episodes and graft loss in the first 3 years have been markedly reduced with the advent of potent new IS drug combinations that include steroids, tacrolimus, and mycophenolate mofetil (MMF) (1012). Conversely, long-term graft loss due to chronic rejection remains relatively unchanged with this combination. Side effects such as drug toxicity—including hypertension, diabetes, heart disease, hyperlipidemia, infection, and cancer—continue to be a major clinical problem (1315). Thus, there is an unmet medical need to translate preclinical studies of tolerance to human organ transplantation safely, with the goal of preventing graft loss and removing IS drug side effects, both of which are major drivers of morbidity and mortality.

Recently, three different approaches have been reported to achieve chimerism and tolerance to human kidney transplants (1624). The goals of each approach were to eliminate the lifelong need for IS drugs and their attendant side effects and to prevent acute and chronic rejection that results in graft loss (1624). In some of the studies of combined hematopoietic cell and kidney transplantation in humans, tolerance to the kidney graft was achieved with complete chimerism (2224), whereas in others, it was achieved with transient mixed chimerism (1618) or stable mixed chimerism (1921). Complete chimerism and kidney graft acceptance persisted without loss after discontinuation of IS drugs in human leukocyte antigen (HLA)–mismatched (haplotype mismatched) recipients (2224). However, some of these complete chimeras developed severe GVHD or chronic infections (25). Tolerance associated with stable mixed chimerism was reported in HLA-matched patients (1921), but stable mixed chimerism off IS drugs was not achieved in tolerance protocols with HLA-mismatched patients (1618, 21). When mixed chimerism was lost early (less than 3 weeks) after HLA-mismatched combined kidney and hematopoietic cell transplantation, patients developed an associated engraftment syndrome with vascular injury of the kidney graft and dysfunction while on standard doses of IS drugs (26). Although stable mixed chimerism off IS drugs was frequently achieved in HLA-matched patients after hematopoietic cell transplantation to treat sickle cell anemia, stable mixed chimerism occurred rarely in HLA-mismatched recipients and only in those who were maintained on IS drugs (2729). The current study was undertaken to test the hypothesis that the establishment of persistent mixed chimerism in patients given combined hematopoietic cell and kidney transplants from the same donor would result in tolerance and the ability to completely withdraw IS drugs without subsequent rejection.

RESULTS

Mixed chimerism, kidney graft acceptance, and IS drug withdrawal in fully HLA-matched transplant recipients

Twenty-nine patients received kidney and hematopoietic cell transplants from fully HLA-matched donors using radiation targeted to the lymph nodes, spleen, and thymus [total lymphoid irradiation (TLI)]. Ten doses of 120 centigrays (cGy) each were administered along with rabbit antithymocyte globulin (ATG; 5 daily doses of 1.5 mg/kg) (Table 1). ATG was started on the day of kidney transplantation and ended on day 4 after transplant. TLI was started on day 1 and ended on day 11 after transplant, with the last dose of TLI immediately followed by the intravenous infusion of donor hematopoietic cells. The donor cells were obtained at least 1 month before the kidney donation after a five-injection course of granulocyte colony-stimulating factor (G-CSF). CD34+ hematopoietic progenitor cells were enriched from one or two apheresis products using immunomagnetic bead columns, and a defined dose of T cells was also collected in the column flow through. Both cell products were cryopreserved and thawed just before infusion. Recipients were discharged from the hospital on postoperative day 4 or 5 and continued TLI in the outpatient clinics. IS drugs, cyclosporine (CsA) or tacrolimus, were started at standard doses on the day of kidney transplantation. MMF was started at the standard dose on the day of the donor cell infusion and was discontinued 30 days later. Thus, patients were maintained on CsA or tacrolimus monotherapy after day 30.

Table 1 HLA-matched patient characteristics, donor cells, and outcomes.

ESRD, end-stage renal disease; FSGS, focal segmental glomerulosclerosis; IgA, IgA nephropathy; DM, diabetes; SLE, systemic lupus erythematosus; PKD, polycystic kidney disease; CIN, chronic interstitial nephritis.

View this table:

The demographic characteristics, donor cell doses, and outcomes of the recipients are shown in Table 1. Observations were from 27 to 157 months (median, 80 months). All 29 recipients were injected with at least 4 × 106 CD34+ donor cells/kg, and 27 were injected with 1 × 106 CD3+ T cells/kg. Two patients with pretransplant immune hyperreactivity were injected with 5 × 106 and 10 × 106 T cells/kg to increase chimerism magnitude. Chimerism was detected by differences in lengths of short tandem repeats (STRs) in DNA extracted from donor and recipient white blood cells (30). The threshold of sensitivity of the assay is at about 1% of donor-type cells (30).

Patient survival in the observation period of up to 13 years is shown in the bar graph in Fig. 1A. Three patients died, and causes of death were unrelated to the treatment (Table 1). Two patients who died had functioning grafts and were off IS drugs at the time of sudden death. As shown in Fig. 1B, two patients lost grafts because of kidney disease without evidence of rejection. Figure 1C shows the duration off IS drugs in the 24 patients who were completely withdrawn. Four patients had IS drugs resumed because of development of kidney disease in three cases and a rejection episode in one case. The duration of chimerism, defined as at least 1% donor-type cells in whole blood, in all 29 patients is shown in Fig. 1D. Ten patients had persistent chimerism at the time of the last test, and 19 lost chimerism at the last test (Fig. 1D). Five patients had chimerism for less than 6 months and did not meet criteria for IS drug withdrawal.

Fig. 1 Patient survival, graft survival, duration of complete withdrawal of IS drugs, and duration of chimerism in HLA-matched patients.

(A) Patient survival. (B) Graft survival. (C) Duration off IS drugs. Patients without bars did not meet criteria to achieve IS drug withdrawal. (D) Duration of chimerism up to the last time point tested.

The mixed chimeras who met criteria had CsA or tacrolimus tapered to discontinuation between 6 and 18 months after transplant. Tapering and discontinuation of IS drugs continued as long as chimerism persisted for at least 6 months, and there was no evidence of GVHD or rejection on protocol biopsies just before discontinuation. Figure 2 shows the magnitude of chimerism, IS drugs, and serum creatinine in 10 of 24 patients who maintained chimerism at the last time point tested without evidence of kidney graft rejection. Fourteen patients withdrawn from IS drugs lost detectable chimerism at the last time point tested in the first or second year after transplant without evidence of subsequent rejection except in patient no. 15 who developed a rejection episode 3 years after discontinuation of IS drugs (fig. S1).

Fig. 2 Persistent mixed chimerism and kidney graft acceptance in HLA-matched recipients.

The percentages of donor cells in whole blood, daily doses of CsA or tacrolimus, and serum creatinine concentrations (mg/dL) at serial time points after kidney transplantation are shown in 10 recipients who had at least 2% donor cells in all chimerism tests. Orange lines show whole blood chimerism, and blue lines show the daily dose of calcineurin inhibitor. MMF was discontinued in all patients at 1 month after transplant. Calcineurin inhibitor used in patient nos. 1 to 18 was CsA and was tacrolimus (Tac) in patient nos. 21 to 29.

Delayed reconstitution of naïve T cells is associated with low T cell chimerism

Figure 3A shows the mean percentages of donor-type cells among T cells, B cells, natural killer (NK) cells, granulocytes, CD34+ cells, and whole blood cells at serial time points during the first year after transplant of all 29 HLA-matched patients. Figure 3B shows the values in patients with persistent chimerism, and Fig. 3C shows the values in patients who lost detectable chimerism in whole blood within the first 2 years. The mean percentages of T cell chimerism during the first year after transplant were significantly reduced (P < 0.001) as compared to all other subsets of white blood cells as judged by area under the curve analysis using the Wilcoxon signed-rank test in all patients regardless of the persistence or loss of chimerism (Fig. 3, A to C). The mean magnitude of chimerism of B cells was at least 30% higher than those of T cells during the first 9 months after transplant in all fully matched patients.

Fig. 3 Differences in magnitude T and B cell chimerism are due to differences in kinetics of reconstitution of naïve cells in HLA-matched transplant recipients.

(A) Mean percentages (±SEM) of donor-type cells among purified T cells, B cells, NK cells, CD34+ cells, granulocytes, and whole blood (WB) cells before and at serial time points after kidney transplantation (KTx) in all 29 patients. (B) Percentages in 10 patients with chimerism that persisted for at least 24 months. (C) Percentages in patients with chimerism that was lost within 24 months. The Wilcoxon signed-rank test was used to make statistical comparisons. (D) Mean percentages of naïve CD3+ T cells among PBMCs. (E) Percentages of naïve CD4+ T cells among PBMCs. (F) Percentages of naïve CD8+ T cells among PBMCs. (G) Percentages of thymic-derived CD3+ T cell precursors (cells containing TRECs) among PBMCs. (H) Percentages of thymic-derived CD4+ T precursors among PBMCs. (I) Percentages of thymic-derived CD8+ T precursors among PBMCs. (J) Percentages of B cell precursors (pro-B cells) among PBMCs. (K) Percentages of naïve B cells among PBMCs. The Mann-Whitney U test was used to make statistical comparisons. (L) compares the mean percentage of donor-type cells among sorted naïve B cells to mean percentages among sorted pro-B cells, naïve CD4+ T cells, memory RO+CD4+ T cells, naïve CD8+ T cells, and memory CD8+ T cells. There were 6 to 12 available patient samples analyzed at each time point for the naïve and precursor cell studies. The two-sided Dunnett’s test was used for multiple comparisons. ns, not significant (P > 0.05); *P < 0.05 and **P < 0.01.

We determined the changes in the percentages of the different T cell subsets before and after kidney transplantation and the magnitude of chimerism in each subset. The mean percentage of naïve CD4+ T cells among peripheral blood mononuclear cells (PBMCs) fell at least 10-fold as compared to pretransplant values (P < 0.05) with a nadir in the 1- to 3-month posttransplant interval (Mann-Whitney U test) (Fig. 3, D to F). The percentage of naïve CD4+ T cells remained significantly reduced (P < 0.05) during the 7- to 12-month interval also (Fig. 3E). In contrast, the percentage of naïve CD8+ T cells at all posttransplant time points was not statistically different from the pretransplant value (P > 0.05). There was also a significant reduction in the mean percentages of thymic T cell precursors among CD4+ and CD8+ T cells in the blood, as measured by cells containing T cell antigen receptor (TCR) gene excision circles (TRECs), in the 1- to 3-month (P < 0.01) and 4- to 6-month (P < 0.05) intervals, as compared to the pretransplant magnitudes (Fig. 3, G to I). The mean percentages of T cell precursors gradually increased to greater than pretransplant values at 18 months. Opposite to the early reduction in T cell precursors, there was an early increase in the mean percentage of B cell precursors (CD19+CD20lgDCD38+ pro-B cells) that exceeded pretransplant frequencies at 1 to 3 months (Fig. 3J). There was also a rapid return of naïve B cells to above pretransplant frequencies starting at 4 to 6 months and continuing beyond 18 months (Fig. 3K).

The chimerism frequencies of sorted T and B cell subsets during the 7- to 9-month posttransplant interval were compared. As shown in Fig. 3L, robust chimerism observed in purified naïve B cells (mean, 52%) was not significantly different (P > 0.05) from that of pro–B cells (36%) or purified naïve CD4+ T cells (35%) cells. Naïve B cell chimerism was increased as compared to memory CD8+ T cells (5%; P < 0.01), naïve CD8+ T cells (7%; P < 0.01), or memory CD4+ T cells (11%; P = 0.05), as judged by the Tamhane-Dunnett test. The prolonged and marked reduction in the percentage and absolute numbers of CD4+ naïve T cells made this subset a small minority of total T cells during the first year after transplantation, whereas naïve B cells with rapid recovery accounted for the large majority (at least 80%) of total B cells (figs. S2 and S3). Thus, changes in the posttransplant balance of naïve T and B cells with high chimerism versus memory T cells including central memory, effector memory, and effector memory RA cells with low chimerism affected the chimerism magnitudes of total T and B cells (fig. S4).

Mixed chimerism, kidney graft acceptance, and IS drug minimization in HLA haplotype–matched kidney transplant recipients

Twenty-two patients received kidney and hematopoietic cell transplants from related HLA haplotype–matched living donors using the same TLI and ATG conditioning regimen used for fully matched patients. The patient demographics, serum creatinine concentrations, and duration of follow-up are shown in Table 2. The HLA typing of donors and recipients are shown in table S1. The duration of follow-up was 14 to 96 months after transplant (median, 47 months). Two patients lost grafts (year 4 because of return of kidney disease and year 7 because of chronic pyelonephritis).

Table 2 HLA-haplotype–matched patient characteristics and outcomes.

GN, glomerulonephritis; MN, membranous nephropathy; HTN, hypertensive nephropathy.

View this table:

Whereas the target T cell dose was constant at 1 × 106 per kg for 27 fully matched patients, the target T cell dose was gradually increased from 3 × 106 cell/kg in patient no. 1 to 100 × 106 cells/kg starting with patient no. 16 in an effort to achieve persistent mixed chimerism (Table 3). A combination of three IS drugs (tacrolimus, MMF, and prednisone) was started at standard dose after transplantation. Prednisone was withdrawn during month 2 in all haplotype-matched patients. MMF was tapered starting from months 6 to 9 and tapered to discontinuation by months 12 to 13 if detectable chimerism persisted without evidence of rejection or GVHD. Patients with persistent chimerism were maintained on tacrolimus monotherapy (4 to 8 ng/ml trough blood concentrations) at the end of year 1.

Table 3 HLA-haplotype–matched kidney transplant patients—Donor whole-blood chimerism.

Shaded boxes show patients with persistent chimerism for at least 12 months.

View this table:

Influence of donor cell dose and HLA mismatches on chimerism

Table 3 shows the number of donor CD3+ T cells and CD34+ hematopoietic progenitor cells injected, number of HLA antigen mismatches, and percentage of donor-type cells in recipient whole blood samples at serial time points during the first year after transplant in the 22 patients followed for 12 months. The range of percentages of donor-type cells at 9 months was from 10 to 60% and from 1 to 46% at 12 months in the 10 patients with persistent chimerism. Persistence of chimerism at 1 year was correlated with the magnitude of chimerism at day 60. Whereas 8 of the 10 persistent chimeras had at least 27% donor-type cells at day 60, all patients who failed to achieve chimerism for at least 12 month had less than 23% donor-type cells at day 60.

We determined the number of donor HLA antigens that were mismatched with the recipient HLA antigens. Eleven of the 22 patients had three HLA antigen mismatches; seven had two mismatches, and four had one mismatch (Table 3). None of five patients (0% with 95% confidence interval; 0 to 43%) who had three HLA mismatches and who were given less than 10 × 106 donor CD34 cells/kg maintained chimerism for at least 12 months. In contrast, four of four patients (100% with 95% confidence interval; 50 to 100%) who had one HLA antigen mismatch and who were given at least 10 × 106 CD34 cells/kg maintained chimerism for at least 12 months (nos. 2, 10, 15, and 16). The confidence intervals of these two groups did not overlap.

The dose of donor CD34 cells was increased starting with patient no. 13 by administering a single injection of the hematopoietic progenitor mobilizing agent plerixafor (0.24 mg/kg) to donors at the end of the course of five injections of G-CSF. The number of donor CD34 cells injected thereafter was at least 10 × 106 cells/kg (Table 3). When three HLA antigen–mismatched patients were given at least 10 × 106 CD34 cells/kg, then three of six (50% with 95% confidence interval; 19 to 81%) patients maintained chimerism for at least 12 months. Three of seven patients (43% with 95% confidence interval; 16 to 75%) with two HLA antigen mismatches who were given at least 10 × 106 CD34 cells/kg maintained chimerism for at least 12 months. Thus, a low number of mismatches and high doses of CD34 cells favored the persistence of chimerism. There was no correlation between the development of persistent chimerism and the dose of T cells in the 3 × 106 to 100 × 106 cells/kg dose range tested. The 12 patients who never developed or lost chimerism during the first year were maintained on standard-of-care doses of IS drugs.

Delayed reconstitution of naïve T cells is associated with low T cell chimerism in HLA haplotype–matched patients

Figure 4A shows the mean chimerism in whole blood and among CD34+ cells, T cells, B cells, granulocytes, and NK cells during the first year after transplant in all 22 haplotype-matched patients. Figure 4B shows the means in the 10 patients with persistent chimerism, and Fig. 4C shows the means in 12 patients who lost chimerism. As in the case of the fully HLA-matched patients, the mean percentages of donor-type cells were significantly reduced among T cells as compared to all other cell types in all 22 patients (P < 0.001, Wilcoxon signed-rank test). The highest mean chimerism was among the B cells.

Fig. 4 Differences in magnitude of T and B cell chimerism are due to differences kinetics of reconstitution of naïve cells in HLA haplotype–matched transplant recipients.

(A) Mean percentages (±SE) of donor-type cells among T cells, B cells, NK cells, CD34+ cells, granulocytes, and whole blood cells before and at serial time points after kidney transplantation in all 22 patients. (B) Percentages in 10 patients with detectable chimerism for at least 12 months. (C) Percentages in patients in whom chimerism persisted for less than 12 months. The Wilcoxon signed-rank test was used to make statistical comparisons. (D) Mean percentages of naïve CD3+ T cells among PBMCs. (E) Percentages of naïve CD4+ T cells among PBMCs. (F) Percentages of naïve CD8+ T cells among PBMCs. (G) Mean Shannon index diversity scores of the TCRα chain genes. (H) Scores of the TCRβ chain genes. (I) Percentages of B cell precursors among PBMCs. (J) Percentages of naïve B cells among PBMCs. The Mann-Whitney U test was used to make statistical comparisons. (K) compares the mean percentage of donor-type cells among sorted naïve CD4+ T cells to memory RO+CD4+ T cells, naïve CD8+ T cells, and RO+ memory CD8+ T cells. There were 6 to 12 available patient samples analyzed at each time point for the naïve and precursor cell studies. The two-sided Dunnett’s test was used to make statistical comparisons of sorted cells. P > 0.05, *P < 0.05, **P < 0.01, and ***P < 0.001.

The low chimerism among T cells was associated with significant reductions (P < 0.01) in the mean percentages of naïve CD3+, CD4+, and CD8+ T cells among PBMCs during the first 7 months after transplant as compared to pretransplant values (Mann-Whitney U test) (Fig. 4, D to F). The low frequencies of naïve T cells were associated with significantly reduced diversity of the TCRβ genes even at 13 to 18 months (P < 0.01), as determined by RNA sequencing (RNA-seq) analysis (Fig. 4, G and H). TCRα gene diversity reduction was not statistically significant (P > 0.05). There was a marked increase in the percentage of pro-B cells as compared to pretransplant magnitudes during the 1- to 3-month interval (Fig. 4I) that was associated with the rapid recovery of naïve B cells to above pretransplant magnitudes at the 4- to 6-month interval (Fig. 4J).

The percentages and absolute numbers of CD4+ memory T cells were significantly reduced for at least 18 months, whereas all CD8+ memory T cells were significantly reduced for only 1 to 3 months (figs. S5, A to I, and S6, A to I). In summary, the haplotype-matched patients showed that the recovery of CD4+ T cells was considerably slower than that of the CD8+ T cells and B cells as observed in the fully matched patients.

Chimerism among sorted naïve CD4+ and CD8+ T cells and sorted CD45RO+ memory CD4+ and CD8+ T cells were determined in haplotype-matched patients during the 7- to 9-month time interval using the same approach described for the fully matched patients. As shown in Fig. 4K, robust chimerism was observed only among naïve CD4+ T cells (mean percentage of donor-type cells, 30%) versus 2% (P < 0.01) and 9% (P < 0.05), respectively, among sorted CD4+ and CD8+ memory T cells and 9% among CD8+ naïve T cells (P = 0.05) (Tamhane-Dunnett test). Thus, the low magnitudes of chimerism among total T cells during the first 12 months after transplant in both fully HLA-matched and haplotype-matched patients reflected the predominance of memory T cells with low magnitudes of chimerism versus the low proportion of CD4+ naïve T cells with robust chimerism.

Instability of mixed chimerism in HLA haplotype–matched recipients during tapering of tacrolimus monotherapy to discontinuation

None of the 10 patients with persistent chimerism for 360 days (Table 3) developed clinical rejection episodes while they were on tacrolimus monotherapy at therapeutic doses (4 to 8 ng/ml trough blood concentration) during or shortly after the first year after transplant. Six of the ten patients had protocol biopsies at 360 to 487 days after transplant while on tacrolimus monotherapy at therapeutic doses (nos. 2, 10, 15, 16, 17, and 18). There was no evidence of microscopic antibody-mediated rejection, T cell–mediated rejection, or borderline rejection, and all interstitial fibrosis/tubular atrophy scores were 1 or less except in a patient who developed an arteriovenous fistula after the implantation biopsy (table S2). Figure 5 shows the serial percentages of whole blood chimerism, the daily dose of MMF, serum concentrations of tacrolimus, and serial serum creatinine concentrations in the 10 HLA haplotype–matched patients who had persistent chimerism during the first year. Seven of the latter patients had stable magnitudes of chimerism, and three patients (nos. 9, 10, and 14) had declining chimerism while on standard doses of MMF and tacrolimus during that interval. Nine had IS drugs further tapered to achieve discontinuation of tacrolimus monotherapy during the second year. Patient no. 2 lost chimerism shortly after discontinuation of tacrolimus and developed a biopsy-proven rejection episode 3 months later. Patient no. 3 lost chimerism after tapering of tacrolimus to subtherapeutic concentrations and had evidence of acute rejection on a protocol biopsy obtained 5 months later (Fig. 5). The rejections were treated, creatinine magnitudes returned to baseline, and maintenance IS drugs were restarted with tacrolimus alone (no. 3) or in combination with MMF (no. 2). Patient no. 3 had recurrent urinary tract infections, and serum creatinine concentrations increased gradually after year 3. She returned to dialysis in year 6. Biopsy results showed evidence of pyelonephritis without cellular or antibody-mediated rejection.

Fig. 5 Loss of mixed chimerism during or after the withdrawal of IS drugs in 10 HLA-mismatched patients.

The percentage of donor cells in whole blood (red), daily doses of MMF (blue), trough blood concentrations of tacrolimus (green), and serum creatinine concentrations are shown at serial time points for each patient up to 762 days (24 months). Red and blue symbols are no longer shown after loss of chimerism or discontinuation of MMF, respectively. Arrows show acute rejection episodes (AR) or microvascular injury (MVI). Patient no. 9 died during tapering of tacrolimus. Due to pulmonary embolism (PE).

The loss of chimerism during or after tapering to subtherapeutic trough blood concentrations (<4 ng/ml) or discontinuation of tacrolimus monotherapy was also observed in four other patients (nos. 10, 14, 15, and 16) (Fig. 5). As soon as chimerism was lost in these patients, maintenance tacrolimus monotherapy (4 to 8 ng/ml trough blood concentrations) was reinstituted. No microscopic rejection was observed on protocol biopsies performed 6 to 12 months after reinstitution in patient nos. 10 and 16. However, patient no. 10 subsequently developed increased creatinine concentrations in the fourth year because of immunoglobulin A (lgA) nephropathy noted on biopsy. Patient no. 14 developed a subclinical microvascular inflammation pattern that resolved on return to maintenance MMF and tacrolimus (Fig. 5). Patient no. 16 with pretransplant Kikuchi’s syndrome was returned to MMF therapy because of evidence of hemolytic anemia. In summary, five of five HLA-mismatched patients with persistent chimerism during the first year lost chimerism after tacrolimus monotherapy was tapered below therapeutic concentrations with evidence of rejection in three. All of these patients had whole blood chimerism magnitudes of <30% just before tapering tacrolimus to subtherapeutic concentrations. Patient nos. 15, 17, and 22 continue to have persistent chimerism at the last observation while on therapeutic concentrations of tacrolimus monotherapy. Patient no. 18 with persistent chimerism continues on MMF monotherapy (Fig. 5). The latter patient was switched from tacrolimus to MMF to treat recurrent kidney disease. Patient no. 9 with persistent chimerism died after an acute pulmonary embolism while on tacrolimus monotherapy.

Link between mixed chimerism and specific unresponsiveness to donor alloantigens in the mixed leukocyte reaction

We previously reported that fully HLA-matched patients who had been completely withdrawn from IS drugs had specific unresponsiveness to donor alloantigens in the mixed leukocyte reaction (MLR) (20). In the current study, we tested the MLR using recipient responder cell donor and third-party stimulator cells in haplotype-matched patients before transplant, after they developed persistent mixed chimerism on MMF and tacrolimus at 6 to 9 months after transplant, and after tapering to therapeutic tacrolimus monotherapy at 13 months after transplant with continued mixed chimerism. Figure S7 (A and B) shows representative results from haplotype-matched patient nos. 10 and 15 who both developed persistent chimerism. In both patients, T cells from the pretransplant blood samples showed proliferative responses as measured by carboxyfluoroxy succinyl esterase (CFSE) dilution (12 to 26% CFSElo responder T cells after 7 days of culture) to both the donor and third-party stimulator cells. A marked reduction of the responses to donor cells(<5% CFSElo cells) was observed at the posttransplant time points in both patients as compared to the responses (9 to 20% CFSElo cells) to third-party cells (P = 0.05, Mann-Whitney U test). Thus, both patients developed specific unresponsiveness to donor cells in association with IS drug–dependent mixed chimerism. Most of the T cells in these mixed chimeric patients were of recipient type (Fig. 4).

Figure S7C shows the results of MLR tests in haplotype-matched patient no. 2 who had persistent mixed chimerism at month 7 and who lost mixed chimerism in month 13 after IS drug withdrawal. Whereas a marked reduction of the response to donor stimulator cells was observed at 7 months, the response to donor cells was above that of the pretransplant value at 13 months. Thus, the loss of mixed chimerism at the later time point was associated with the return of a robust response to donor cells, and both mixed chimerism and unresponsiveness to donor cells were IS drug dependent.

DISCUSSION

The main objective of the study was to test the hypothesis that mixed chimerism and tolerance can be established in fully HLA-matched and HLA haplotype–matched living donor kidney transplant recipients conditioned with posttransplant TLI and ATG and infused with a combination of donor T cells and hematopoietic progenitors. Twenty-four of 29 fully HLA-matched recipients were completely withdrawn from IS drugs after mixed chimerism persisted for at least 6 months. Twenty-three of 24 had no evidence of rejection thereafter during an observation period of up to 10 years. Ten had persistent chimerism at all time points, and 14 lost chimerism within the first or second year during withdrawal of IS drugs as judged by the reduction of the percentage of donor-type cells to less than 1% in whole blood. We were also able to establish persistent mixed chimerism for at least 1 year in 10 HLA haplotype–matched recipients. The barrier to persistent chimerism was considerably greater than in fully matched patients, and we increased the dose of donor T cells and the dose of CD34+ cells to overcome this barrier.

Persistent mixed chimerism was most easily achieved in patients with one HLA antigen mismatch who were injected with at least 10 × 106 CD34+ cells/kg and was not achieved in patients with three mismatches who were injected with less than 10 × 106 CD34+ cells/kg. Patients with persistent chimerism during the first year after transplant were reduced from three IS drugs to tacrolimus monotherapy at the end of the year. However, chimerism was not sustained during the further reduction of tacrolimus to subtherapeutic blood concentrations or to discontinuation in the second year after transplant. Three of five of these patients who lost detectable whole blood chimerism developed rejection episodes shortly thereafter. In contrast, none of 14 fully HLA-matched patients who lost detectable whole blood chimerism during IS drug withdrawal developed rejection episodes in the subsequent 2 years. The cellular and molecular basis of the difference in freedom from rejection of kidney transplants after IS drug withdrawal in the fully HLA-matched patients versus the haplotype-matched patients remains to be elucidated. It is possible that the loss of chimerism in the haplotype-matched patients was complete in the blood and solid lymphoid tissues, whereas the loss in the fully matched patients was incomplete.

The instability of chimerism in patients with HLA mismatches may be related to the low magnitudes (mean of about 10%) of T cell chimerism at the end of the first year, and achievement of higher magnitudes may allow for increased stability of chimerism during withdrawal of tacrolimus monotherapy. Surveillance biopsies at the end of the first year after transplant in fully matched or haplotype-matched patients with persistent chimerism and off IS drugs or on tacrolimus monotherapy showed no evidence of microscopic rejection. Thus, persistent mixed chimerism was associated with protection against rejection despite the reduction of IS drugs.

Mixed chimerism was lowest among T cells and highest among B cells in both the fully HLA- and haplotype-matched patients. Robust chimerism was confined to naïve CD4+ T cells and naïve B cells. Whereas naïve B cells accounted for more than 80% of total B cells, naïve CD4+ T cells accounted for a small minority of total T cells. Thus, the marked difference in chimerism in the T versus B lineage was explained by differences in the balance of naïve T versus B cells. The low magnitudes of chimerism in T cells as compared to other white blood cell lineages and marked reduction of naïve T cells were previously reported in patients given hematopoietic cell transplants for hematologic malignancies and in our preclinical models of mixed chimerism and tolerance using the TLI and ATG conditioning regimen (30, 31).

The delayed reconstitution of naïve T cells in our preclinical models was explained by the injury and loss of thymic epithelial cells in the medulla induced by the TLI because the thymus was included in the radiation fields (32). Recovery of these irradiated epithelial cells delayed reconstitution for a few months in the mouse model as compared to recipients with shielding of the thymus (32). We have also shown that memory T cells are considerably more resistant to radiation than naïve T cells (33). This may account for the predominance of recipient memory T cells rather than naïve donor cells.

Our previous studies showed that specific unresponsiveness to donor alloantigens in the MLR was maintained after IS drug withdrawal in the fully matched patients even after loss of chimerism in whole blood (20). This study revealed that this is not the case in haplotype-matched patients. Although tolerance and complete IS drug withdrawal were not achieved in haplotype-matched patients in the current study with the TLI and ATG conditioning regimen, tolerance and complete withdrawal without evidence of subsequent rejection were achieved in 4 of 10 haplotype-matched patients after transient mixed chimerism persisted for only a few weeks in the previously reported clinical trial performed at Massachusetts General Hospital (1618). The difference is likely related to the conditioning regimen of cyclophosphamide and anti-CD2 monoclonal antibodies used in that study. The latter regimen was not associated with severe neutropenia, severe thrombocytopenia, or GVHD (1618) but was associated with transient kidney graft injury and early chimerism loss in 9 of 10 patients (26) and a high rate of early graft loss (three kidney grafts lost within the first 3 years after transplant) (1618). Early graft loss did not occur in any of the 51 patients enrolled in our clinical trial, and transient kidney injury was observed in only two of these patients.

There were limitations in the present chimerism studies because of both the lack of sensitivity of the STR assay below 1% donor-type cells in the blood and the lack of testing of chimerism in solid lymphoid tissues. Low magnitudes of chimerism may have persisted in these tissues in the 14 HLA-matched patients off drugs. Additional limitations were the small number of samples of sorted lymphocyte subsets available for chimerism studies early after transplantation because of lymphopenia. Studies of the impact of the number of HLA mismatches and of the dose of donor CD34+ cells injected into recipients on the persistence of chimerism were also limited by the small number of HLA haplotype–matched patients in each subgroup.

Another clinical trial of combined kidney and hematopoietic cell transplantation performed at Northwestern University established durable complete chimerism and complete withdrawal of IS drugs in most of haplotype-matched patients (2224). However, the conditioning regimen of total body irradiation, pre- and posttransplant cyclophosphamide, and fludarabine was considerably more intensive than the regimens used here, and almost all patients developed severe neutropenia and thrombocytopenia in the first 2 weeks after transplant (2224). The intensive conditioning regimen was associated with early graft loss from infection, and complete chimerism was associated with chronic GVHD in one recipient and lethal GVHD in another (25). The side effects and risks of graft loss associated with the different tolerization regimens need to be balanced against the benefits of IS drug withdrawal, and in our view, patient and graft survival should be at least equal to standard-of-care therapy. An important additional goal is to ensure that acute and chronic side effects are no worse than those of standard of care. The protocol described here appears to meet all of these goals.

In conclusion, we established persistent mixed chimerism for at least 1 year in fully HLA-matched and haplotype-matched kidney transplant recipients. Partial or complete withdrawal of IS drugs was not associated with rejection or GVHD in persistent chimeric recipients, and persistence is likely to provide continued prevention of rejection.

MATERIALS AND METHODS

Study design

Patients given fully HLA-matched or haplotype-matched living donor kidney transplants were enrolled in a single-arm tolerance induction protocol in which enriched CD34+ hematopoietic progenitor cells and CD3+ T cells obtained from the donors were infused after kidney transplantation (NCT00319657). The recipients were conditioned with posttransplant TLI and ATG to achieve persistent mixed chimerism. Standard-of-care IS drugs were administered after transplantation, and the latter drugs were partially or completely withdrawn if recipients showed persistent chimerism for at least 6 to 12 months, lack of microscopic rejection, and lack of GVHD. The primary end point of this phase 1/2 safety and feasibility study was the withdrawal of IS drugs without subsequent evidence of rejection. Serial immune monitoring was performed on blood samples to assess immune reconstitution, magnitudes of chimerism in immune cell subsets, antidonor immune reactivity, and relationship between persistence of chimerism and successful IS drug withdrawal. In the matched patient study, the target dose of donor CD3 T cells was kept constant at 1 × 106 cells/kg. In the haplotype-matched study, the dose of donor T cells was escalated in the range of 10 × 106 to 100 × 106 cells/kg. Dose escalation was triggered starting with the second enrolled patient if there were fewer than four engraftment successes in groups of five consecutive patients. Engraftment successes were defined as at least 5% chimerism in whole blood. Clinical trials included in this paper are registered with ClinicalTtrials.gov (Kidney and blood stem cell transplantation that eliminates requirement for immunosuppressive drugs: NCT00319657; Combined blood stem cell and kidney transplant of one haplotype match living donor pairs: NCT01165762). Primary data are reported in data file S1.

Patients

Fifty-one patients with end-stage renal failure who were candidates for living donor kidney transplants at the Stanford Medical Center were given transplants using an experimental protocol (no. 18731) that was approved by the Stanford Institutional Review Board and the Federal Drug Administration. Twenty-nine patients were fully HLA-matched, and 22 patients were HLA haplotype–matched after typing for HLA A, B, C, DR, and DP antigens by high-resolution DNA typing. An additional haplotype-matched patient with polycystic kidney disease began the protocol but was withdrawn on day 4 after transplant because of calcineurin-associated microangiopathic hemolytic anemia. HLA-mismatched recipients with pretransplant donor-specific antibodies or panel-reactive antibody screen of more than 20% were excluded from the study. Observations were made up to 30 June 2019.

Conditioning of recipients

All fully HLA-matched and haplotype-matched recipients were given 10 doses of TLI of 120 cGy each targeted to the lymph nodes, spleen, and thymus starting on postoperative day 1 and completed on day 11, as described previously (1921). Rabbit ATG (Thymoglobulin, Genzyme) was intravenously administered in five daily doses of 1.5 mg/kg each starting intraoperatively. Steroids were given for 10 days in matched patients to reduce ATG side effects and extended to 30 to 90 days in mismatched patients.

Processing of donor hematopoietic cells

HLA-matched donors were given a course of five daily injections (16 μg/kg) of G-CSF, and mononuclear cells were collected by one or two aphereses about 4 to 6 weeks before kidney donation. CD34+ cells were enriched on immunomagnetic bead columns (CliniMACS, Miltenyi), cryopreserved, and then thawed just before intravenous infusion. The column flow through cells were collected to contain a defined number of T cells for infusion, cryopreserved, and thawed at the same time at the CD34+ cells, as described before (21). In some of the HLA-mismatched donors, a single injection of plerixafor (0.24 mg/kg; Mozobil, Sanofi) was given after the first apheresis on day 5, and a second apheresis was performed on day 6 to increase the yield of CD34+ cells. Unmanipulated mononuclear cells containing a defined dose of T cells were cryopreserved on day 5, and the remaining day 5 cells along with all the day 6 cells were enriched for CD34+ cells and cryopreserved.

Chimerism testing

Chimerism was determined by analysis of polymorphisms of STRs contained in DNA samples extracted from whole blood or from white blood cells purified for CD3+ T cells, CD19+ B cells, CD56+ NK cells, and CD15+ granulocytes on immunomagnetic bead columns (Miltenyi) (19, 20, 30). Chimerism testing was continued at monthly intervals during the first 2 years after transplant as long as donor-type cells were detected with a threshold of sensitivity of about 1% (30). In some instances, chimerism testing was repeated after a few to several years to assess long-term stability.

Kidney transplant biopsies

Biopsies were performed if there was an unanticipated increase in serum creatinine concentrations that suggested graft injury (“for cause” biopsies). Protocol biopsies were also performed just before discontinuation of IS drugs and at time intervals after discontinuation or minimization of IS drugs.

Mixed leukocyte reaction

The MLR assays compared cryopreserved recipient pre- and posttransplant PBMCs as responder cells using donor and HLA unmatched third-party stimulator cells, as described previously (20). The stimulator cells were PBMCs or monocytes that were incubated with granulocyte-macrophage CSF and interleukin 4 for 6 days to generate dendritic cells (20). Responder cells were labeled with CFSE, and the T cell proliferation after 7 days of culture was measured by staining the CFSE intensity of gated CD3+, CD4+, and/or CD8+ T cells. The percentage of CFSElo cells was determined by flow cytometric analysis, as described previously (31).

Antimicrobial prophylaxis

Nystatin was administered for 1 month, trimethoprim-sulfamethoxazole, for 12 months, and valganciclovir for 3 to 6 months, followed by acyclovir until 18 months after transplant.

Immunofluorescent staining and analysis of T and B cell subsets

Blood mononuclear cells were stained with fluorochrome-conjugated monoclonal antibodies against CD3, CD4, CD8, CD62L, CD45RO, CD19, CD20, CD38, and lgD (BD Pharmigen). Up to 10 color analyses were performed by flow cytometry using LSR and Aria cytometers (BD Biosciences) using FlowJo software, as described previously (19, 31).

TREC analysis

Sorted T cells obtained by flow cytometry were cryopreserved and shipped to the Human Vaccine Institute Immune Reconstitution Facility at Duke University for TREC analysis. The polymerase chain reaction (PCR)–based assay has been described before (34).

Analysis of diversity of TCR gene sequences by RNA-seq

5′RACE (rapid amplification of cDNA ends) methods for TCR repertoire sequences used modified gene-specific primers in the TCRα and TCRβ constant regions. The purified 5′RACE PCR products were processed to make sequencing libraries using the KAPA HyperPrep Kit (KAPA Biosystems). Sequencing was performed using an lllumina MiSeq reagent 500-cycle V2 kit by paired-end 250 × 2 cycles (lllumina Inc.). The paired-end reads from the MiSeq were submitted to MixCR for TCRα and TCRβ rearrangement analyses. The unique CDR3 amino acid sequences (CDR3 clones) for each sample were summarized on the basis of the MixCR results. Single-copy CDR3 clones were removed. The frequency of a clonotype was calculated by the copy number of the clonotype divided by the total number of copies of all clonotypes in a sample. Bioinformatics processing and further analytics were performed using the Stanford Sherlock supercomputing cluster and customized analysis platform.

Statistical analysis

Comparisons of the serial percentages and absolute numbers of T and B cell subsets were analyzed using the two-sided Student’s t test of independent means. Analyses of the percentage of chimerism in T cell versus non–T cell white blood cell lineages over time were performed by area under the curve calculations via the trapezoidal rule, followed by analysis using the two-sided one-sample Wilcoxon signed-rank test adjusting for multiple comparisons with the Holm method. Comparisons of the percentages of chimerism in groups of sorted T and B cell subsets at a given time point were made using the Tamhane-Dunnett test for a two-sided hypothesis in which multiple comparisons are made to a single control when variances are heterogeneous and the design is unbalanced (35). Computations of the 95% binomial confidence intervals of the percentage of patients with persistent chimerism who received different doses of CD34+ cells and who had different numbers of HLA mismatches were made using the method of Wilson. Comparisons of the changes in the percentages of proliferating cells in the MLR assays at different time points were performed using the Mann-Whitney U test.

SUPPLEMENTARY MATERIALS

stm.sciencemag.org/cgi/content/full/12/528/eaax8863/DC1

Fig. S1. Loss of mixed chimerism and kidney graft acceptance in HLA-matched recipients.

Fig. S2. Changes in the mean percentages of total, naïve, and memory T cells before and at serial time points after transplantation in HLA-matched patients.

Fig. S3. Changes in B cell subsets and absolute numbers of T and B cells in HLA-matched patients.

Fig. S4. Changes in subsets of memory T cells among PBMCs in HLA-matched patients.

Fig. S5. Changes in subsets of memory T cells among PBMCs in HLA haplotype–matched patients.

Fig. S6. Changes in the absolute numbers of total and naïve T and B cells in HLA haplotype–matched patients.

Fig. S7. Representative MLR in HLA haplotype–matched recipients.

Table S1. HLA typing and mismatch of donor/recipient pairs enrolled in haplotype-matched protocol.

Table S2. Surveillance biopsy findings in haplotype-matched patients with mixed chimerism.

Data file S1. Primary data.

References and Notes

Acknowledgments: We thank B. M. Zhang for reviewing the chimerism tests for the last year and W. Leong, J. Arulprakasam, and L. Bulanadi for excellent and timely contributions in the performance of chimerism tests. We thank J. Troiano for manuscript preparation. Funding: The research was supported by the NIH (grants P01 HL075462 and R01 1085024; prinicipal investigator: S.S.), the California Institute of Regenerative Medicine (CIRM grant CLIN2-09439; principal investigator: S.S.), and the Stanford University Hospital (project director: S.B.). The research was also supported by the J.M. Sobrato, R. and P. Moore, and M. and J. Goldman Foundations. Author contributions: S.B., J.D.S., R.L., T.P., and J.S. helped enroll patients, provided informed consent, managed treatment of donors and recipients, collected and analyzed data, and wrote the paper. K.J., O.C., E.G.E., J.W., H.-H.W., and E.M. performed immune monitoring of blood samples and collected and analyzed data. A.S. arranged all patient visits, blood sample collections, testing and distribution, and communications with regulatory agencies. M.A.F.V. supervised and analyzed HLA typing. R.H. supervised radiotherapy. J.T. and P.L. performed biostatistical analyses. K.S. supervised the processing and analysis of donor cells and analyzed data. S.S. supervised all aspects of the study and wrote the paper. Competing interests: S.S., R.L., and E.G.E. are cofounders of a cell therapy company with a focus on organ transplantation, which provided no funding for the study. A patent related to this work (Combined organ and hematopoietic cells for transplantation tolerance of grafts) has been registered under U.S. Patent no. US 9,561,253 B2. Data and materials availability: All data associated with this study are present in the paper or the Supplementary Materials.

Stay Connected to Science Translational Medicine

Navigate This Article