Belatacept and Sirolimus Prolong Nonhuman Primate Renal Allograft Survival Without a Requirement for Memory T Cell Depletion


Corresponding author: Allan D. Kirk,


Belatacept is an inhibitor of CD28/B7 costimulation that is clinically indicated as a calcineurin inhibitor (CNI) alternative in combination with mycophenolate mofetil and steroids after renal transplantation. We sought to develop a clinically translatable, nonlymphocyte depleting, belatacept-based regimen that could obviate the need for both CNIs and steroids. Thus, based on murine data showing synergy between costimulation blockade and mTOR inhibition, we studied rhesus monkeys undergoing MHC-mismatched renal allotransplants treated with belatacept and the mTOR inhibitor, sirolimus. To extend prior work on costimulation blockade-resistant rejection, some animals also received CD2 blockade with alefacept (LFA3-Ig). Belatacept and sirolimus therapy successfully prevented rejection in all animals. Tolerance was not induced, as animals rejected after withdrawal of therapy. The regimen did not deplete T cells. Alefecept did not add a survival benefit to the optimized belatacept and sirolimus regimen, despite causing an intended depletion of memory T cells, and caused a marked reduction in regulatory T cells. Furthermore, alefacept-treated animals had a significantly increased incidence of CMV reactivation, suggesting that this combination overly compromised protective immunity. These data support belatacept and sirolimus as a clinically translatable, nondepleting, CNI-free, steroid-sparing immunomodulatory regimen that promotes sustained rejection-free allograft survival after renal transplantation.


calcineurin inhibitor


cytotoxic T lymphocyte associated antigen 4-immunoglobulin


donor-specific transfusion; IM, intramuscular


leukocyte function associated antigen 3-Immunoglobulin


peripheral blood mononuclear cells


rhesus Cytomegalovirus


central memory T cell


effector memory T cell


naive T cell


regulatory T cell


Organ transplantation is the accepted standard therapy for many forms of end-stage organ failure, and advancements in immunosuppression have facilitated maintenance of transplanted organs without immune mediated rejection. All clinically approved immunosuppressive therapies, however, are associated with side effects that temper the substantial benefits of organ transplantation. Thus, more specific, less toxic therapies are needed to optimize transplantation.

Costimulation blockade is a promising immunomodulatory approach that specifically targets the costimulatory pathways required for full T cell activation, leaving other nonimmune cellular pathways unperturbed. Its specificity for defined costimulation pathways provides an immunosuppressive effect, but avoids the renal, cardiovascular and metabolic morbidities associated with currently approved immunosuppressive agents such as calcineurin inhibitors (CNIs) and steroids. Apart from its prevention of rejection, costimulation blockade approaches have been shown to induce long-term allograft acceptance without any chronic immunosuppressive treatment—tolerance—in several experimental models [1, 2]. Successful clinical translation of therapies with this type of durable antirejection effect would represent a substantial improvement in posttransplantation care.

Belatacept is a high-affinity cytotoxic T lymphocyte associated antigen 4 (CTLA4) fusion protein that blocks the CD28-B7 costimulatory pathway [3]. It is the only clinically approved costimulatory blockade agent in transplantation, and is indicated as a CNI alternative in renal transplantation [4, 5]. Its use has been shown to avoid the predominant side effects associated with CNIs, although the clinically approved regimen remains coupled with chronic steroid therapy. Belatacept's initial clinical approval has prompted interest in studies exploring its optimal pairing with other available adjuvant immunosuppressive agents so as to foster its durable antirejection effects, and perhaps avoid maintenance immunosuppression altogether. Indeed, belatacept-related molecules (e.g. CTLA4-Ig) have been shown to induce tolerance in naive animal models, particularly when paired with the mammalian target of rapamycin (mTOR) inhibitor, sirolimus [6]. These approaches have not induced tolerance in outbred animal models expressing established memory T cell repertoires—models that are likely more clinically relevant [7]. Memory T cells are a known barrier to costimulation blockade as they exert effector functions without reliance on full costimulatory signaling [8-10]. They are, however, sensitive to CNI-based inhibition, as CNIs target signals required by both naive and memory T cells. Consistent with this biology, patients receiving belatacept in the pivotal phase 3 trials had a higher incidence of early acute rejection compared to the CNI-treated group [11]. It is thus plausible that, in clinical settings involving nonnaive recipients, adjuvant therapies may be needed specifically to control this cell population.

Alefacept (leukocyte function-associated 3-Ig; LFA3-Ig) is a fusion protein that targets CD2 [12, 13]. CD2 is upregulated on memory T cells [14], and alefacept has been shown to preferentially eliminate effector memory T cells in vivo [15, 16]. We have shown previously that alefacept neutralizes costimulation blockade (CTLA4-Ig)-resistant alloreactive memory T cells [16, 17]. However, alefacept previously has not been paired with belatacept in vivo, and thus, it is unclear if its effect is required when using the higher affinity agent, particularly when combined with other optimally delivered adjuvant agents.

In this study, we have tested belatacept and sirolimus as a CNI-free, steroid-sparing maintenance regimen after renal transplantation in nonhuman primates. Belatacept has been chosen to investigate the optimized, clinically available costimulation blockade agent, and sirolimus dosing has been optimized compared to prior studies [16] to rigorously test the combination's effect. We find that this dual therapy successfully prevents renal allograft rejection, and that additional therapy with alefacept is not required to address costimulation blockade-resistant cell populations.

Materials and Methods

Donor–recipient pair selection and kidney transplantation

All experiments described in this study were performed in compliance with the principles set forth in The Guide for the Care and Use of Laboratory Animals, Institute of Laboratory Animal Resources, National Research Council, DHHS. Outbred rhesus monkeys (Macaca mulatta) ranging between 3 and 5 years old were obtained from AlphaGenesis, Inc. (Yemassee, SC, USA) and Yerkes National Primate Research Center (Lawrenceville, GA, USA). Donor–recipient pairs were chosen to maximize genetic nonidentity at both MHC class I and class II alleles. Prior to transplantation, alloreactivity was confirmed by mixed lymphocyte reaction as previously described [2]. Left native nephrectomy was performed at least 3 weeks prior to transplantation, and kidney transplantation was performed using standard microvascular techniques [2]. A completion right native nephrectomy was performed at the time of transplantation. Animals were heparinized (100 U/kg) during organ procurement and implantation. All transplant recipients were monitored with daily clinical assessment and weekly labs including complete blood count, serum chemistry and sirolimus trough levels.

Experimental groups and immunomodulation

There were four experimental groups in this study. Three animals received maintenance therapy consisting of belatacept (Nulojix; Bristol–Myers Squibb, New York, NY, USA) and sirolimus (Rapamune; LC Laboratories, Woburn, MA, USA). Belatacept (10 mg/kg IV) was given on days −1, 3 and 7 before tapering to a maintenance dose (5 mg/kg IV) weekly for 24 weeks. This regimen was designed to approximate the drug exposure achieved with the clinically indicated dose of belatacept. Sirolimus was started on day −1 and given once daily via intramuscular (IM) injection for 16 weeks. The IM dosing differs from prior experiments by our group in which oral dosing was used, and was chosen to avoid the inconsistent levels achieved when relying on oral absorption in rhesus monkeys. Doses were adjusted to maintain a serum sirolimus trough level of 8–10 ng/mL.

To determine the additional effect of CD2 blockade, eight animals received extended induction therapy with alefacept (Amevive; Astellas Pharma, Deerfield, IL, USA) in addition to belatacept and sirolimus maintenance therapy. Alefacept (1 mg/kg IV) was administered on days −1, 3 and 7, and weekly for 8 weeks. A subgroup of three animals receiving alefacept also were given a single whole blood transfusion (7 mL/kg) from the donor animal on day 7 (Figure 1A). This cohort was included to specifically expand upon prior work by our group [16, 18] and others [19, 20] suggesting that donor-specific transfusion (DST) synergizes with costimulation blockade and mTOR combination therapies to induce tolerance. Two animals were maintained with IM sirolimus alone. These four experimental groups were compared to previously completed control cohorts that received no immunosuppression after transplantation [2]. All animals received a rapid 3-day methylprednisolone taper beginning the day of surgery to mimic the use of induction steroids likely to be preferred clinically. Animals were weaned sequentially off all immunomodulatory agents by posttransplant day 168 (Figure 1A). Belatacept was provided by Bristol–Myers Squibb, and alefacept was provided by Astellas.

Figure 1.

Belatacept and sirolimus promote rejection-free renal allograft survival. (A) The immunomodulation regimen is shown. Rhesus macaques received belatacept and sirolimus maintenance therapy beginning the day prior to transplantation. Eight animals received adjuvant therapy with alefacept, and three of these animals were given a donor-specific whole blood transfusion (DST) 7 days after transplantation. Immunomodulation was discontinued sequentially, and animals were weaned off all therapies by posttransplant day 168. (B) Animals receiving belatacept and sirolimus maintained stable renal allograft function well past the discontinuation of all immunomodulation. Alefacept-treated animals demonstrated decreased rejection-free renal allograft survival. Two-tailed p-values comparing treatment groups to the belatacept and sirolimus cohort are shown.

Histologic analysis

All histology specimens were prepared by a histopathologist (MS) and graded by a nephropathologist (ABF). Biopsies were performed using a 20-gauge percutaneous core needle on day 56 after transplantation (the time of alefacept withdrawal) and upon clinical suspicion of rejection. Standard hematoxylin and eosin staining was performed on all allograft biopsy and nephrectomy specimens. C4d staining was performed on nephrectomy specimens. Samples were graded for rejection based on modified 1997 Banff scoring criteria [21].

Antibodies and flow cytometric analysis

Peripheral blood T cell phenotypes were characterized pretransplant and serially posttransplantation to assess changes in the CD3, CD4 and CD8 T cell subsets with particular attention to CD28+CD95 naive T (TN) cells, CD28+CD95+ central memory (TCM), CD28CD95+ effector memory (TEM) and CD4+CD25hiFoxP3+ regulatory T (TReg) cells [22]. T cells were quantified by complete blood cell count and flow cytometric analysis. Fresh peripheral blood mononuclear cells (PBMCs) were isolated by ficoll density gradient centrifugation (BD Biosciences, Franklin Lakes, NJ, USA) and stained with the following monoclonal antibodies: CD3 PacBlue, CD3 APC-Cy7, CD4 PerCP-Cy5.5, CD8 V500, CD8 PerCP, CD127 PE-Cy7, CD2 PE, CD28 PE-Cy7, CD95 FITC (all BD Biosciences), CD95 PacBlue (Invitrogen, Grand Island, NY, USA) and CD25 PE (Miltenyi Biotech, Germany). PBMCs (5 × 105) were incubated with appropriately titered antibodies for 15 min at 4°C and washed twice. Intracellular staining was performed using the BD Cytofix/Cytoperm kit (BD Biosciences) according to the manufacturer's instructions. Intracellular staining with FoxP3 Alexa 488 (Biolegend, San Diego, CA, USA) was done to detect TReg cells. Samples were acquired immediately on a BD LSR II multicolor flow cytometer, and data were analyzed using FlowJo software (Tree Star, San Carlos, CA, USA).

rhCMV viral monitoring using quantitative real-time PCR analysis

Animals were monitored at regular intervals to quantify rhesus Cytomegalovirus (rhCMV) viral load using quantitative real-time polymerase chain reaction (qPCR). DNA was extracted from whole blood samples using the Qiagen QiaAmp DNA Blood Mini Kit (Qiagen, Valencia, CA, USA). rhCMV primer and probe sets were designed using ABI Primer Express software (Applied Biosystems, Foster City, CA, USA) to detect the rhCMV Intermediate Early (IE) gene (NCBI, Accession #M93360). Primers and probes used included 5′-ATCCGCGTTCCAATGCA-3′, 5′-CGGAGGAGCACCATAGAAGGT-3′ and 5′-6FAM-CCTTCCCGGCTATGG-3′ (Sigma-Aldrich, St. Louis, MO, USA). The qPCR reaction was performed in a total volume of 50 μL using 7.5 μL (∼200–800 ng) extracted DNA, 0.7 μM primer, 0.05 μM TaqMan probe and 25 μL 2X TaqMan Universal PCR Master Mix (Applied Biosystems). Samples were run in triplicate on the ABI 7900HT (Applied Biosystems) with an initial denaturation at 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C for 60 s. Transcript copy numbers were calculated utilizing Sequence Detection Systems 2.3 software (Applied Biosystems) and compared to the standard curve, which was generated using the pRhIE 9.4.2 plasmid (donated by Dr. Peter Barry Lab, University of California, Davis, CA, USA) containing the rhCMV IE gene sequence. Based on our prior experience with rhCMV, we deemed titers above 10 000 copies/mL as significant and treated animals with ganciclovir (6 mg/kg IM twice daily) until the viral load was undetectable.


Survival statistics were calculated using a log-rank test. T cell frequencies were compared using an unpaired t-test. Data were analyzed using Prism 4 (GraphPad Software, La Jolla, CA, USA). A two-tailed p-value of <0.05 was considered statistically significant.


Belatacept and sirolimus promote sustained rejection-free renal allograft survival

Animals maintained with belatacept and IM sirolimus alone (n = 3) experienced prolonged renal allograft survival of 231, 245 and 378 days (Figure 1B, Table 1). Rejection was observed only after animals had been completely withdrawn from all immunomodulation. The addition of belatacept significantly increased allograft survival, as animals receiving IM sirolimus monotherapy experienced allograft rejection 55 and 64 days after transplantation (p = 0.039). As previously reported [2], control animals receiving no treatment all rejected allografts within 7 days (p = 0.013).

Table 1. Rejection-free survival
Treatmentsurvival (days)p-value*
  1. *p value compared to belatacept group;

  2. **Historical controls (2).

No treatment**5, 6, 7, 7, 80.013
Sirolimus alone55, 640.039
Belatacept, sirolimus231, 245, 378N/A
Belatacept, sirolimus, alefacept129, 154, 203, 206, 2170.014
Belatacept, sirolimus, alefacept, DST231, 231, 2340.155

Immunohistochemical analysis was performed on all allograft specimens. Upon sacrifice all allografts demonstrated marked lymphocytic infiltrates, tubulitis and occasional vasculitis, consistent with acute cellular rejection (data not shown). There was no evidence of C4d deposition in any specimen (data not shown).

Belatacept and sirolimus promote the maintenance of naive T cells

Animals were followed with weekly polychromatic flow cytometry to detect changes in T cell populations. In the belatacept and sirolimus group, the absolute number of CD3, CD4 and CD8 T cells remained stable throughout therapy, with no evidence of T cell depletion (Figure 2A). After initiation of therapy, the percentage of CD4 naive T (TN) cells dramatically increased above baseline and remained elevated during belatacept treatment (Figure 2B). Correspondingly, the proportion of CD4 memory T cells was markedly lower than baseline during belatacept therapy. This phenotypic skew toward the naive phenotype may have been caused by a maturation arrest induced by belatacept. After belatacept was withdrawn there was a progressive phenotypic shift from predominantly CD4 TN to CD4 central memory T (TCM) cells, with TCM equal to or outnumbering TN 7 weeks after belatacept discontinuation. Two of three animals developed acute rejection within 4 weeks of this phenotypic shift. Similarly, CD8 TN cells comprised the majority of the CD8 T cell population early during belatacept therapy (Figure 2C). A CD8 T effector memory (TEM) phenotype predominated by posttransplant day 140. One animal demonstrated stable renal function for over 1 year after transplantation. Interestingly, this animal maintained high proportions of both CD4 and CD8 TN cells without converting to a more memory-laden phenotype until just prior to rejection (Figure 2B and C).

Figure 2.

Absolute T cell counts and T cell subset distribution after transplantation. Polychromatic flow cytometry was used to follow T cell counts over time. (A–C) Animals treated with belatacept and sirolimus alone maintained stable CD3, CD4 and CD8 absolute T cell counts throughout therapy (A). (B) The percentages of TN, TCM and TEM were analyzed. These animals maintained a high percentage of CD4 TN cells above baseline throughout belatacept therapy, with a shift to a TCM phenotype after belatacept was withdrawn. (C) CD8 TN cells predominated early after transplantation with a slow increase in TEM over time. (D–F) Animals receiving alefacept in addition to belatacept and sirolimus showed reduced absolute T cell counts during alefacept therapy (D). T cell counts began to rebound soon after alefacept was discontinued. (E) Alefacept-treated animals also experienced a more rapid conversion from CD4 TN to TCM and (F) CD8 TN to TEM compared to animals not receiving alefacept. Depletion of CD8 TEM was transient and repopulated more rapidly relative to other T cell subsets.

Addition of alefacept depletes CD3 T cells and accelerates progression to a memory phenotype

Alefacept was tested as an 8-week adjuvant therapy to the optimized belatacept and sirolimus regimen. Based on our previous work suggesting that DST might add some additional survival benefit [16], a single dose of DST in addition to alefacept was given to a subset of animals. The addition of alefacept alone (n = 5) did not improve survival over the optimized belatacept and sirolimus regimen. Surprisingly, animals receiving alefacept experienced decreased rejection-free survival (p = 0.014), with two of five animals rejecting during the period of belatacept monotherapy. The addition of alefacept and DST (n = 3) did not improve allograft survival compared to the optimized maintenance regimen alone (p = 0.155).

The addition of alefacept resulted in a substantial decrease in the absolute number of CD3, CD4 and CD8 T cells (Figure 2D). Absolute T cell counts gradually recovered after alefacept was discontinued on day 56, reaching pretransplant levels after both alefacept and sirolimus were withdrawn. Animals treated with alefacept demonstrated the same phenotypic shift from CD4 TN to TCM as previously described; however, this change occurred during belatacept therapy, which was much earlier than observed in the belatacept and sirolimus group (Figure 2E). Alefacept animals also more rapidly converted from CD8 TN to TEM compared to animals that did not receive alefacept (Figure 2F).

Alefacept targets CD3+CD2hi T cells

We have previously shown that T cells can be segregated based on CD2 expression into CD2hi and CD2lo subsets, with memory T cells expressing the highest levels of CD2 [16, 17]. Since alefacept targets cell surface CD2 polychromatic flow cytometry was used to evaluate T cell subsets for CD2hi and CD2lo phenotypes. Animals treated with alefacept showed a sharp decrease in both CD4 and CD8 CD2hi populations for the duration of alefacept treatment (Figure 3A and B). Corresponding reductions in absolute CD4 TCM and CD8 TEM cell counts were noted, suggesting that alefacept effectively targeted CD2hi T cells. Animals receiving belatacept and sirolimus alone also demonstrated a decrease in CD2hi populations, though not nearly as profound as in the alefacept group. This change reflected an enrichment of the CD2lo TN population rather than a selective elimination of CD2hi T cells or a decline in overall T cell counts.

Figure 3.

Alefacept targets CD3+CD2hi T cells. CD4 (A) and CD8 (B) T cells were segregated based on CD2 expression and evaluated for the proportion of CD2hi T cells over time. Alefacept efficiently targeted CD4 and CD8 CD2hi T cells (solid line). CD2hi T cells were detected again after discontinuation of alefacept therapy. Animals receiving belatacept and sirolimus alone (dashed line) demonstrated a modest decline in the proportion of CD4 and CD8 CD2hi T cells from baseline, which corresponded to an enrichment of both CD4 and CD8 TN cells.

FoxP3+ regulatory T cells were decreased throughout therapy

Regulatory T (TReg) cells are critical protective immune mediators that have been shown to be reliant on costimulation for persistence and function [23]. Indeed, some studies have suggested that both costimulation blockade and mTOR inhibitor-based therapies can potentially promote the expansion or function of TReg cells [24, 25]. Therefore, we used flow cytometry to follow circulating CD4+CD25hiFoxP3+ TReg cells over time. A reliable TReg staining panel was developed during the course of the study, and results are available for the belatacept and sirolimus group and five animals that received alefacept, with or without DST. For this analysis animals that received alefacept, with or without DST, were combined into one treatment group.

Treatment with belatacept and sirolimus decreased the absolute number of circulating TReg cells to approximately half of pretransplant levels, with this number remaining stable throughout therapy (Figure 4A). Discontinuation of belatacept was associated with a rebound in circulating TReg cells to near pretransplant levels. Treatment with alefacept compounded the reduction of circulating TReg cells caused by belatacept and sirolimus alone, resulting in profoundly low levels that did not recover to baseline even after therapy was discontinued. Alefacept animals had a significantly decreased number of circulating TReg cells compared to the belatacept and sirolimus group at all time points analyzed (Figure 4A), even after withdrawal of all therapy. To determine the basis for this marked reduction in TReg cells associated with alefacept we specifically analyzed this population for CD2 expression. TReg cells consistently expressed higher levels of CD2 compared to all CD4 T cells, regardless of the treatment group (Figure 4B). The increased CD2 density on TReg cells may have rendered them more susceptible to deletion by alefacept.

Figure 4.

Belatacept, sirolimus and alefacept decrease the number of peripheral TReg cells. (A) The absolute number of circulating CD4+CD25hiFoxP3+ TReg cells was followed over time. Belatacept and sirolimus caused a twofold reduction in circulating TReg cells that was maintained throughout belatacept therapy. The number of peripheral TReg cells rebounded and approached baseline once all therapies were discontinued. The addition of alefacept caused a profound sixfold decrease in circulating TReg cells compared to baseline, which marginally recovered after all therapies were weaned. *p<0.01, **p<0.001. (B) The ratio of CD2 mean fluorescence intensity (MFI) on TReg cells relative to all CD4 T cells was analyzed. TReg cells consistently expressed higher levels of CD2 compared to total CD4 T cells, regardless of the treatment group.

Alefacept increases the incidence and rapidity of rhCMV reactivation

Selective targeting of T cells may unintentionally compromise antiviral protective immunity. Therefore, transplanted animals were monitored serially for rhCMV viral reactivation by PCR quantification of whole blood rhCMV viral DNA. All animals had undetectable rhCMV viral titers at the time of transplantation. Six of eight animals receiving alefacept developed rhCMV reactivation requiring antiviral treatment (Figure 5A). Five animals developed severe rhCMV viremia of greater than 50 000 copies/mL, and four animals had multiple rhCMV reactivations despite antiviral therapy. Initial rhCMV reactivation in all six animals occurred within the first 56 days after transplantation during alefacept therapy (Figure 5B). In contrast, one of three animals in the belatacept and sirolimus group experienced a delayed reactivation on posttransplant day 175. All rhCMV reactivations were clinically asymptomatic.

Figure 5.

Alefacept increases the frequency and rapidity of rhCMV reactivation. Animals were serially monitored for rhCMV reactivation via quantitative real-time polymerase chain reaction. Animals with rhCMV viral loads above 10 000 copies/mL were treated with ganciclovir. (A) rhCMV viral loads over time from five alefacept-treated animals with at least one severe rhCMV viremia (>50 000 copies/mL) are shown. Four animals had recurrent rhCMV reactivation despite treatment. (B) The time to first rhCMV reactivation is shown. Animals receiving alefacept demonstrated a strikingly earlier time to rhCMV reactivation, with the first episode occurring during alefacept treatment in all six animals that experienced reactivation. One animal in the belatacept and sirolimus group experienced delayed CMV reactivation on posttransplant day 175.


Belatacept is a clinically available, costimulation blockade agent approved for use as maintenance therapy after renal transplantation. Randomized clinical trials in kidney transplantation have demonstrated the efficacy of belatacept when paired with steroids and the antiproliferative agent, mycophenolate mofetil [4, 5]. In this study we investigated alternative belatacept-based immunomodulation regimens that avoid chronic steroid use and could be clinically translatable. We demonstrate in a rigorous preclinical model that belatacept and sirolimus is a CNI-free, steroid-sparing immunomodulation regimen that promotes sustained rejection-free renal allograft survival. These data are consistent with those observed by Lowe et al. (manuscript accepted) in that optimally achieved mTOR inhibition avoids belatacept-resistant rejection in this model. Our group has previously published data on belatacept monotherapy in this experimental model, indicating that the typical onset of rejection with belatacept monotherapy is 20–42 days [3]. Thus, it is likely that the addition of sirolimus provided a survival benefit in comparison. Although the belatacept dosing was not identical to previous studies, the need for a contemporaneous control group to make this point was not viewed as sufficiently compelling to warrant additional primate surgery.

In this study, treatment with belatacept and sirolimus skewed the T cell population toward a high percentage of TN cells without causing depletion. Belatacept may have induced maturation arrest by blocking costimulatory signals required by TN cells to differentiate into effector or memory phenotypes, therefore allowing for the accumulation of TN cells. This maturation arrest may be critical to maintaining the antirejection effect of belatacept. Consistent with this is one animal that maintained excellent graft function for over 1 year after transplantation, and did not reject until after the naive predominance was replaced by a memory dominant phenotype. Monitoring the overall ratio of naive to memory T cells over time may be a useful index to gauge the efficacy of costimulation blockade in prophylaxis against rejection. Changes in this ratio toward memory could be an early harbinger of potential rejection prior to clinical signs of renal dysfunction. Similarly, highly memory dominant individuals may be at a disadvantage early after transplant.

One important concern regarding belatacept use is the higher incidence of acute cellular rejection compared to CNIs found in clinical trials [11]. While belatacept may slow the differentiation of TN cells into effectors, preexisting alloreactive memory T cells have been implicated as central mediators of costimulation blockade-resistant rejection [10]. Our previous work suggested that CD2 blockade with alefacept preferentially depletes memory T cells in vivo, and initial trials pairing alefacept with CTLA4-Ig yielded encouraging results [16]. However, unlike our previous experience, alefacept did not provide an additional survival benefit when added to our belatacept and sirolimus regimen. We surmise this difference to be related to the optimized nature of the costimulation blockade (belatacept as opposed to CTLA4-Ig) and the more consistent mTOR inhibition (IM rather than oral administration), leaving less opportunity for early rejection.

Of concern, the addition of alefacept to the optimized belatacept/sirolimus combination significantly diminished renal allograft survival. Since it powerfully targets both alloreactive and nonalloreactive CD2hi T cells indiscriminately, the addition of alefacept to an already effective immunomodulation regimen may have been overly immunosuppressive, compromising antiviral protective immunity. rhCMV reactivation in animals that received alefacept was more frequent, more severe and occurred soon after transplantation during alefacept treatment. Viral reactivation may have contributed in part to early rejection by creating a proinflammatory cytokine milieu, triggering alloreactive T cell activation through a bystander effect [26]. Viral infections have also been associated with acute rejection through heterologous immunity, or cellular immunity arising from T cell cross-reactivity to both allo- and viral antigen [27, 28]. Other latent viruses, such as Epstein–Barr virus homologs and Simian Virus-40, were not monitored in this study, but protection against these viruses also could have been compromised, similarly predisposing to acute rejection.

Alefacept therapy resulted in moderate T cell depletion, with CD3 T cell counts reaching the lowest levels 5 weeks after alefacept initiation and rebounding in the weeks after alefacept cessation. CD8 TEM cells repopulated more rapidly compared to other T cell subsets, a consistently described feature of homeostatic repopulation [29, 30]. The combination of rhCMV reactivation during alefacept treatment along with moderate T cell depletion induced by alefacept resulted in T cell homeostatic repopulation often occurring in the setting of recent or ongoing rhCMV viremia. The presence of antigen, either viral or alloantigen, in the setting of homeostatic proliferation has been suggested to influence the character of the repopulating immune repertoire, either in favor of enhanced reactivity or toward exhaustion/anergy [31, 32]. In this instance, the presence of rhCMV antigen during the postalefacept repopulation phase may have resulted in relative exhaustion against rhCMV antigen and heightened susceptibility to repeated viral reactivation. Alternatively, T cell priming with viral antigen at the time of transplantation has been shown to prevent tolerance induction [33]. T cell priming during an acute episode of rhCMV viremia may have synergized with the homeostatic proliferative drive to further heighten alloreactivity.

The CD28/B7 pathway has been shown to be critical in the thymic development and peripheral homeostasis of TReg cells [23, 34, 35], and the effect of costimulation blockade with belatacept on TReg cells requires further investigation. While some have suggested that costimulation blockade and sirolimus may promote TReg cell number or function [24], although ex vivo effects have suggested a more nuanced effect dependent on exposure duration [25]. While the reasons for the discrepancy between various models and regimens remains speculative, numerous factors, including the difference in memory T cell repertoires between relatively naive mice and immunologically experienced large animals, differences in the adjuvant agents and dosing regimens could be implicated. In this study, we found a twofold decrease in the absolute number of peripheral CD4+CD25hiFoxP3+ TReg cells during belatacept treatment; with TReg cells reemerging after belatacept was discontinued. We hypothesize that the deprivation of CD28 costimulatory signaling was detrimental to peripheral TReg survival [35], though this effect was not significant enough to precipitate rejection. We did not assess changes in TReg cell function. It is not well established whether Tregs behave like effector memory T cells in regard to their sensitivity to immunosuppressive drugs, but functional Tregs are generally thought to be antigen experienced, and thus one could expect them to be targeted in a way similar to that of antigen-experienced effectors. The data from this series of experiments suggest this to be the case. The TReg depletion in alefacept-treated animals was significantly greater than animals receiving belatacept and sirolimus alone, suggesting that alefacept had a deleterious effect on these cells. The addition of alefacept may have altered the TReg to T effector cell ratio sufficiently to impair the ability of this regulatory population to control alloreactive effectors in this CNI-free regimen.

In summary, our study shows that belatacept and sirolimus are a CNI-free, steroid-sparing maintenance regimen that effectively prevents renal allograft rejection in nonhuman primates. Our companion investigations (Lowe et al., manuscript submitted) extend these observations to islet allotransplantation as well. This nondepleting regimen arrested T cell maturation and promoted accumulation of TN cells, but did adversely affect the TReg population without overt negative consequences to the allograft. Alefacept therapy was associated with worse outcomes, and may have been overly immunosuppressive given the increase in rhCMV viremia and TReg depletion. Though durable tolerance was not achieved, our studies support the clinical translation of belatacept and sirolimus as a maintenance regimen after kidney transplantation and illustrate the need for robust preclinical evaluation when considering new biologic therapies for transplantation.


This study was funded by support from the NIH (5 U01 AI079223), Yerkes National Primate Research Center (P51RR-00065) and the Georgia Research Alliance. The authors would like to thank Bristol-Myers Squibb for providing belatacept and Astellas, Inc. for providing alefacept. Additionally, the authors are grateful to the Yerkes National Primate Research Center staff and the Emory Transplant Center Biorepository staff for their contributions to this work.


The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.