Calcineurin Inhibitor Withdrawal After Renal Transplantation with Alemtuzumab: Clinical Outcomes and Effect on T-Regulatory Cells

Authors


Corresponding author: Stuart J. Knechtle, stuart@surgery.wisc.edu

Abstract

To address the results of calcineurin inhibitor (CNI) withdrawal after alemtuzumab induction relative to CNI continuation, we performed a pilot randomized clinical trial in renal allograft recipients on CNI, a mycophenolic acid derivative and steroids after the first 2 months posttransplantation. Forty patients were randomized to taper off CNI or to maintain it, and followed for at least 1 year. Four patients in the withdrawal group were treated for acute rejection while no patient received antirejection treatment in the control group. Two control patients withdrew CNI due to nephrotoxicity. Estimated GFR was similar in both groups after 1 year. Flow cytometry of CD4+CD25+CTLA-4+FoxP3+ regulatory T cells (Treg) demonstrated a significant increase in Treg percentages in the peripheral blood of alemtuzumab-treated patients on CNI early postransplant. Furthermore, the increased Treg percentages in the withdrawal cohort were unchanged at month 6 postenrollment, whereas they decreased significantly in those patients maintained on CNI. Patients withdrawn from CNI after alemtuzumab trend toward a higher rejection rate, but most patients can be weaned from a CNI using this regimen. With the exception of maintaining increased Treg levels, the benefits are not appreciable in this short follow-up, and a larger randomized trial is justified.

Introduction

The development of new immunosuppressive agents for use in organ transplantation has resulted in substantial improvements in the survival of grafts and patients over the past decades. Nevertheless, success remains relatively short term, with survival of renal transplants approximating 40–50% at 10 years (1). The most significant problems in the field are accelerated cardiovascular disease leading to patient death, chronic rejection and side effects of long-term immunosuppressive therapy. Each of these problems is related directly to the immunosuppressive strategies that have been employed. Lymphocyte depletion at the time of renal transplantation with induction with alemtuzumab, has allowed allografts to be maintained with reduced immunosuppression (reviewed in 2). A pilot study by our group demonstrated that a majority of renal allograft recipients treated with alemtuzumab induction therapy maintained good graft function while on low-dose sirolimus monotherapy, but 28% of them developed rejection, with strong humoral component in some of them (3). However, Calne et al. reported a low-rejection rate over a 5-year period in patients receiving alemtuzumab perioperatively, with subsequent low-dose cyclosporine monotherapy (4). Other relevant experiences, including ours show great efficacy and adequate safety of CNI treatment after alemtuzumab induction (5–10). Taken these studies together, the impression is that after depletion with alemtuzumab, CNI treatment is adequate for preventing acute rejection. In contrast with this apparent unavoidable reliance, a number of potential advantages of a calcineurin inhibitor-free regimen are evident, including improved renal allograft function, lower incidence of hypertension and diabetes post transplant and less drug side effects due to calcineurin inhibitors.

We have tested in a pilot randomized trial the previously unexplored way of obtaining a CNI-free maintenance therapy in alemtuzumab-treated patients after short period of CNI therapy, in comparison with CNI maintenance. In addition, we analyzed CD4+CD25+FoxP3+ T-regulatory cells in the peripheral blood to assess how CNIs affect the increase and maintenance of these cells after alemtuzumab induction.

Methods

Study design, objective and outcomes

This IRB-approved, randomized trial enrolled adult recipients of kidney transplants who had received induction therapy with alemtuzumab, cyclosporine or tacrolimus, mycophenolate mofetil (MMF) or enteric-coated sodium mycophenolic acid (ECSM) and low-dose steroids. Subjects were randomized between 2 and 16 months posttransplant to either continue CNI or to taper off of CNI. Patients with PRA > 10%, estimated Nankivell GFR < 40 mL/min (11) or prerandomization antibody-mediated or Banff IA acute rejection were excluded. Planned follow-up was set up for 3 years, and in the present report we will analyze 1-year results.

Successful CNI withdrawal was assessed by the incidence of biopsy-proven acute renal allograft rejection at 12 months after enrollment, comparing patients who stopped CNI at 3 months and the control group on CNI therapy. Additional end-points were comparisons between the two groups regarding patient and graft survival, renal function as measured by serum creatinine and calculated GFR, incidence of posttransplant diabetes, hypertension, hypercholesterolemia and infection and CD4+CD25+FoxP3+ T-regulatory cell percentages in the peripheral blood.

Immunosuppressive protocol

Alemtuzumab (Campath-1H, Ilex, Inc., San Antonio, TX), was administered intraoperatively on the day of transplant (30 mg) and a second dose of 30 mg was given the day following transplantation. Steroid treatment began minutes prior to the first infusion (500 mg of i.v. methylprednisolone), 250 mg i.v. on day 1, 10 mg orally on day 3 and until month sixth and 5–7.5 mg thereafter. Cyclosporine or tacrolimus was started on day 1 or when serum creatinine < 3.0 mg/dL in case of delayed graft function, and doses were adjusted to achieve trough levels of 100–200 ng/mL or 5–10 ng/mL respectively. Subjects randomized to the control group continued receiving CNI dosing at the specified target levels, unless the subject develops CNI drug toxicity, necessitating a reduction or withdrawal of the CNI therapy. In subjects randomized to CNI withdrawal, the CNI daily dosing was reduced by 25–50% on the day of randomization. This reduced CNI dosing continued for 7–14 days, at which time the CNI dosing was discontinued. Subjects were on maintenance doses of MMF/ECMS (minimum of 500/360 mg every 12 h) at the time of study enrollment. Post enrollment, subjects continued MMF/ECMS dosing, as tolerated, up to a maximum of 1000/720 mg every 12 h.

Protocol biopsy

Renal transplant biopsies were performed after well-informed consent at 12 months and if clinically indicated by graft dysfunction. Kidney biopsies were graded following Banff 97 (with updated 2005 report) criteria (12,13).

Flow cytometry and treg populations

Six-color flow cytometry was performed on Ficoll-purified PBMCs isolated at the time of enrollment and again 6 months later. Cell surface and intracellular fixing and staining was performed according to manufacturer instructions (eBioscience, San Diego, CA). Live/dead cells were determined using the Hoechst stain (BD Bioscience). The following labeled antibodies were also used: anti-human FoxP3 PCH101 FITC (eBioscience), anti-human CD25 PE, anti-human/NHP CD4 PerCp and anti-human CTLA-4 APC (all BD Bioscience). All flow was performed on a FACSLaserII (BD Bioscience) and the machine calibrated with Rainbow fluorescent particles (Spherotech, Inc.) for more consistent measurements between runs. Cells that expressed these markers were deemed Tregs and the percentage of all CD4+ lymphocytes that were Tregs calculated. Data were analyzed using FlowJo software (TreeStar, Inc).

Statistics

The primary efficacy variable was the occurrence of a first biopsy-proven acute rejection. We tested the hypothesis of a unilateral equivalence for an expected incidence of acute rejection of 10% in the control group. With a noninferiority limit of 15%, the number of patients to include in the study was 98 (49 in each group), for α= 0.05 and β= 0.80. This one-center study had recruitment limitations, so the final patient sample was smaller than planned. Consequently, the study had only moderate statistical power for detecting a difference of a clinically meaningful magnitude, and was finally conceived as a proof of concept or pilot trial. The patients were assigned to the groups by computer generated random numbers, supplied in sealed envelopes.

Values are shown as means ± standard deviation. Patient and graft survival were estimated with the Kaplan-Meier method. Statistical comparisons were performed with the log-rank test. Chi-square and Fisher's exact tests, and t-test and Wilcoxon test were used as needed for parametric and nonparametric variables. All data management and statistical analyses were performed with SPSS software, v. 14.0 (SPSS, Chicago, IL).

Results

Patient disposition and characteristics

A total of 267 patients were screened for study inclusion, and finally 40 were randomized (Figure 1). No significant difference was observed between included and excluded patients (mean age of those excluded 52 years, 73% males, 98% caucasians, 61% deceased donor, 6% retransplantations). Twenty patients were randomized to CNI withdrawal and 20 to the control group. General characteristics are summarized in Table 1, showing no significant differences between both groups.

Figure 1.

Patient disposition during the study.

Table 1.  Demographic data
 Withdrawal (n = 20)Controls (n = 20)p
  1. SD = standard deviation; BMI = body mass index; PRA = panel reactive antibodies; MMF = mycophenolate mofetil; ECMS = enteric-coated mycophenolic sodium.

Age at transplantation (years ± SD)55.2 ± 9.5 53.6 ± 9.2 0.59
Gender (% Male)85750.69
Caucasian race (%)1001001.00
Pretransplant diabetes mellitus [n(%)] 4 (20) 5 (25)1.00
Pretransplant hypertension [n(%)]17 (85)18 (90)1.00
Pretransplant coronary artery disease [n(%)] 6 (30) 5 (25)0.72
Pretransplant BMI > 34 kg/m2[n(%)] 6 (30) 5 (25)0.72
Deceased donor [n(%)]11 (55)12 (60)0.46
Living-related donor [n(%)] 7 (35) 4 (20) 
Living-unrelated donor [n(%)] 2 (10) 4 (20) 
Retransplantation [n(%)]1 (5)1 (5)1.00
HLA A mismatches1.00 ± 0.861.15 ± 0.810.57
HLA B mismatches1.40 ± 0.681.15 ± 0.750.27
HLA DR mismatches1.00 ± 0.791.20 ± 0.620.38
Peak PRA2.1 ± 3.14.1 ± 8.80.34
Delayed graft function (%)20201.00
Posttransplant time at inclusion (days) 0.09
Mean (± SD)141 ± 105212 ± 150 
Median (range)103 (61–466)176 (61–477) 
Calcineurin inhibitor at inclusion 1.00
Receiving cyclosporine16 (80)17 (85) 
Receiving tacrolimus 4 (20) 3 (15) 
Micophenolic derivative at inclusion 0.52
Receiving MMF 8 (40)10 (50) 
Receiving ECMS12 (60)10 (50) 

Efficacy

After 1 year of follow-up, all patients were alive and with functioning grafts (Table 2).

Table 2.  Efficacy results
 Withdrawal (n = 20)Controls (n = 20)p
Patient survival (%)1001001.00
Graft survival (%)1001001.00
Acute rejection [n(%)] 4 (20)0 (0)0.11
Biopsy proven and clinical treated 2 (10)0 (0) 
Nonbiopsy-proven clinical treated1 (5)0 (0) 
Subclinical treated1 (5)0 (0) 
Progressive graft dysfunction due to CNI toxicity leading to CNI stop02 (10%)
Serum creatinine (mg/dL)
  At randomization1.47 ± 0.321.55 ± 0.370.45
  3 months postramdomization1.46 ± 0.291.54 ± 0.300.45
  12 months postrandomization1.52 ± 0.641.45 ± 0.300.62
Estimated GFR (mL/min)
  At randomization69.9 ± 11.964.1 ± 12.50.14
  3 months postramdomization70.6 ± 11.665.3 ± 11.40.16
  12 months postrandomization72.1 ± 11.668.0 ± 12.10.28
Proteinuria (mg/L)
  At randomization185 ± 171254 ± 2890.37
  3 months postramdomization158 ± 190223 ± 2540.39
  12 months postrandomization271 ± 303343 ± 5010.59

Acute rejection:  Biopsy-proven acute rejection was diagnosed in two patients in the withdrawal group, 3 months after CNI withdrawal. One case was graded as IIA and initially responded to steroids and thymoglobulin therapy. During the subsequent months, kidney function remained suboptimal (SCr at 1 year after enrollment, 4 mg/dL), and cyclosporine was not reintroduced. The other case was diagnosed after a mild SCr increase and a biopsy showing acute tubular injury, diffusely positive C4d staining in peritubular capillaries, minimal inflammation and no tubulitis. Donor-specific antibodies were not tested at that time, and the patient was diagnosed as having suspected acute humoral rejection. Kidney function was stabilized after treatment with steroids, i.v. immunoglobulins and rituximab. Low-dose tacrolimus was reintroduced. Two more patients in the withdrawal group showed signs of acute rejection but they did not fulfill criteria for biopsy-proven acute rejection. One of them showed a mild increase in SCr from 1.6 to 1.9 mg/dL 5 days after randomization to withdrawal group, and after only a 25% CsA decrease. The patient completely recovered after steroid boluses and finally was not considered for CsA withdrawal. Another patient showed subclinical acute rejection grade IA on the 1-year protocol biopsy (9 months after CNI withdrawal). This patient received steroid boluses, tacrolimus was reintroduced and kidney function remained stable. No acute rejection was diagnosed in the CNI maintenance group.

No differences were detected between rejectors and nonrejectors regarding the demographic parameters detailed in Table 1, except for HLA-B matching. Complete HLA-B mismatch was observed in all four patients with postrandomization acute clinical or subclinical rejection, and these 2-B mismatches were only observed in six out of the other 16 withdrawal patients (p = 0.04).

Kidney function:  No difference was detected during follow-up regarding kidney graft function determined by SCr or estimated GFR (Table 2). Overall, both groups minimally increased their estimated GFR at 12 months of randomization when compared with baseline GFR (withdrawal group +3.9 ± 9.7 mL/min vs. +4.3 ± 11.5, p = 0.89). Proteinuria was nonsignificantly higher in controls at baseline, and after 12 months it had mildly increased in both groups (+86.1 ± 303 vs. +114 ± 457, p = 0.82).

Protocol kidney biopsies:  Protocol biopsies were performed 1 year after randomization in only 14 of the 40 patients (Table 3). The remaining 26 patients either refused (n = 21) or the biopsy was considered to be contraindicated due to oral anticoagulants (n = 2) or other reasons (n = 3). Among the 14 biopsied patients, 4 out of the 8 withdrawal patients and 3 out of 6 controls showed normal histology. The remaining controls (n = 3) showed grade I interstitial fibrosis and tubular atrophy, and the findings detected in the four withdrawal patients with abnormalities were: grade I interstitial fibrosis and tubular atrophy, borderline rejection lesions, grade IA acute rejection and C4d positive without cellular rejection, respectively (Table 3).

Table 3.  Protocol kidney allograft biopsies
 Withdrawal (n = 20)Controls (n = 20)
Not performed [n(%)]12 (60)14 (70)
Normal [n(%)] 4 (20) 3 (15)
Rejection [n(%)]
 Borderline lesions1 (5)0 (0)
 Subclinical IA1 (5)0 (0)
 C4d+ without cellular lesions1 (5)0 (0)
 Interstitial fibrosis and tubular atrophy [n(%)]1 (5) 3 (15)

Peripheral Treg levels in the first 6 months of the trial

Alemtuzumab-mediated depletion promotes a significant increase in peripheral Tregs that can subsequently be affected by the type of maintenance immunosuppression (14,15). This clinical trial gave us the opportunity to assess directly how CNI withdrawal alone affected Treg levels. We therefore performed six-color flow cytometry using the Treg markers CD4, CD25, CTLA-4 and FoxP3. We did this on samples from 12 patients in the control group and eight patients in the withdrawal group obtained at the time of enrollment and at the 6-month time point. At 6 months, the withdrawal group had been off, of CNI, for over 5 months. The gating strategy is shown in Figure 2A, and consisted of first gating on live cells, then on lymphocytes, then on CD4+ lymphocytes and finally on FoxP3 and CTLA-4. All double positives in the latter gate were CD25+. As shown in Figure 2B, the vast majority of Campath-1H-treated patients had increased Tregs at the time of enrollment with an average of 9.4% Tregs within the CD4+ population compared with levels of healthy individuals (3.8%) and compared to 4% Tregs pretransplant (15). Interestingly, those patients maintained on CNI had a significant decrease in peripheral Treg percentages over the 6-month time frame (p = 0.0031, Figure 2C). However, those patients weaned from CNIs maintained their Treg levels, as the differences in the 6-month time frame were not significant (p = 0.46).

Figure 2.

Figure 2.

T-regulatory cells in peripheral blood. CD4+CD25+FoxP3+ T-regulatory cells were identified by six-color flow cytometry at the time of enrollment and 6 months thereafter for 8 withdrawal and 12 control patients. (A) Gating strategy for Tregs from frozen PBMC preparations. Cells were first gated on the live cells using the Hoechst stain, then gated on the lymphocyte gate by forward and side scatter, then gated on the CD4+ population and then on the CTLA-4+FOXP3+ cells. 100% of double positives in this latter gate were CD25+. Tregs are measured as a percentage of the total CD3+CD4+ population. (B) Comparison of 20 patients at the time of enrollment in the study (before CNI withdrawal) versus 5 healthy individuals. Statistical analysis was performed using an unpaired two-tailed t-test. (C) Comparison of Treg percentages in the CD4+ gate between eight patients in the withdrawal group versus 12 patients in the maintenance (control) group. Statistical analysis was performed using a paired two-tailed t-test.

Figure 2.

Figure 2.

T-regulatory cells in peripheral blood. CD4+CD25+FoxP3+ T-regulatory cells were identified by six-color flow cytometry at the time of enrollment and 6 months thereafter for 8 withdrawal and 12 control patients. (A) Gating strategy for Tregs from frozen PBMC preparations. Cells were first gated on the live cells using the Hoechst stain, then gated on the lymphocyte gate by forward and side scatter, then gated on the CD4+ population and then on the CTLA-4+FOXP3+ cells. 100% of double positives in this latter gate were CD25+. Tregs are measured as a percentage of the total CD3+CD4+ population. (B) Comparison of 20 patients at the time of enrollment in the study (before CNI withdrawal) versus 5 healthy individuals. Statistical analysis was performed using an unpaired two-tailed t-test. (C) Comparison of Treg percentages in the CD4+ gate between eight patients in the withdrawal group versus 12 patients in the maintenance (control) group. Statistical analysis was performed using a paired two-tailed t-test.

Safety

Blood pressure, serum total, LDL and HDL-cholesterol, serum triglycerides and glycemia were similar in both groups during the year of study period (data not shown). After 1 year, three patients in the withdrawal group decreased antihypertensive drug needs while no control patient showed this decrease (p = 0.10). Similarly, at 1 year, two withdrawal patients and two controls had stopped the statin treatment prescribed before randomization. No patient had posttransplant diabetes mellitus before randomization. Two control patients developed posttransplant diabetes mellitus 5 and 8 months after randomization. No withdrawal patient developed this complication.

Six withdrawal patients developed CMV infection (n = 3), urinary tract infection (n = 2) and sinusitis (n = 1), while six controls developed CMV infection (n = 2), herpes- zoster infections (n = 2), gastroenteritis (n = 1) and pneumonia (n = 1), without significant differences between the groups. The number of hospitalizations per patient during the study period was identical between the groups.

Discussion

We have described for the first time that CNI withdrawal is feasible and safe in most patients after induction with alemtuzumab and maintenance with CNI, MMF/ECMS and steroids. However 20% patients assigned to CNI stop after a median time of 4 months posttransplantation needed antirejection treatment, whereas no control patient assigned to CNI maintenance needed such treatment. In particular, 3 out of the 8 withdrawal patients showed acute alloimmune activity on the protocol biopsy, whereas no control patient did. Although some nephrotoxicity was evident in control patients maintained on CNI, with CNI withdrawal in two of them, no relevant benefit in estimated GFR at 1 year was observed, thus questioning the real benefit of CNI discontinuation. Estimated GFR was higher at 1 year in withdrawal patients only when excluding the four patients maintained on CNI despite the intention to stop CNI. This result is similar to that observed by Smak-Gregoor et al. in the first controlled trial of cyclosporine stop in MMF-treated patients: estimated GFR did not improve unless withdrawal patients who were actually on CNI treatment were excluded (16).

The prediction of acute rejection after CNI withdrawal remains a challenge. A recent clinical trial has shown that borderline acute rejection changes at randomization protocol biopsy and mycophenolic acid area under the curve were the only good predictors (17). Regrettably, we did not perform prewithdrawal biopsies or mycophenolic acid levels. However, we have partially confirmed the previous findings by Anjum et al., showing an increased risk for acute rejection after CNI stop in patients with HLA-B complete mismatch (18).

Considerable interest has developed in applying immunosuppressive regimens that allow elimination of CNIs and their toxicity while maintaining adequate immunosuppression. After classical trials based on azathioprine (19,20), the advent of MMF-based therapy has been very promising in achieving these goals. Five previous studies have examined the possibility of CNI withdrawal in MMF-treated kidney allograft recipients (16,17,21–25). The designs, populations and primary endpoints were variable, but in general, our results are in agreement with previous and recent experiences of CNI stop with MMF-based therapy. In particular, a modest improvement in estimated GFR and an increase in acute rejection rates after CNI withdrawal are common findings. In our study, given the low number of patients, this increase in acute rejection did not reach significance, but the percentage is similar to other trials. When no antibody induction treatment was used (16,21), the increase in acute rejection rate and the severity of the episodes were greater, suggesting that either anti-IL2 receptor antibodies (25), antithymocyte globulin (17,24) or alemtuzumab may significantly limit this complication.

Alemtuzumab induction has been previously used to permit further maintenance minimization strategies (3,4,6,7,9, 26–29). After two initial CNI-free pilot trials including ours (3,26), only two other small trials have tried to maintain patients without CNI and nonnephrotoxic maintenance sirolimus-MMF (27) or deoxyspergualin (28). High acute rejection incidence and suboptimal safety characterized those approaches. All the other attempts of alemtuzumab induction were associated with prolonged cyclosporine or tacrolimus maintenance therapy (4,6–9,29).

Data obtained from this laboratory has shown that alemtuzumab enhances the expression of FoxP3+ Tregs in vitro, and in vivo, inducing an increase in the percentage of Tregs in the peripheral blood. CNIs, however, have been shown to inhibit the generation of Tregs (30,31). Therefore, using flow cytometry, we assessed the percentage of Tregs within the CD4+ population in the peripheral blood of kidney transplant patients who had received alemtuzumab induction therapy and either long-term CNI therapy or withdrawal from short-term CNI therapy. We have previously reported that Campath-1H patients maintained on sirolimus monotherapy had a significant increase in Tregs by 6 months posttransplant (15). However, we could not definitively determine whether Campath-1H-mediated depletion or sirolimus was fundamentally responsible for the increase. The data presented in this study, in which all depleted patients were treated with CNI, demonstrate that depletion alone is responsible for the increase in Treg percentages, albeit rapamycin likely further augments these levels (14). Alemtuzumab induction therapy followed by long-term CNI therapy in the control group resulted in a significant decrease in Treg percentages. This decrease was not found after CNI withdrawal and in fact increased Treg levels were maintained. Since CNIs interfere with IL-2 production and its signal transduction, an important cytokine in the generation and maintenance of Tregs (32) it is conceivable that the inhibition of IL-2 by CNIs blocks Treg production. CNI withdrawal, therefore, would be associated with IL-2 generation and subsequent Treg generation. This is in line with recent findings showing increased Treg populations in rapamycin-treated patients versus those treated with CNI (31). We previously showed that Treg populations are better preserved using sirolimus-based therapy than with CNI-treated patients (15). However, this is the first time that better Treg preservation is suggested after CNI withdrawal in comparison with CNI maintenance.

The main limitation in this study is the low number of patients finally randomized. We had difficulty obtaining informed consent for inclusion in the study, mainly because these were stable patients and we were requesting extra clinic visits and protocol kidney biopsies. Even after patients consented to protocol biopsies, many of them declined consent when approached again to consent. In addition, 1 year is a short period of time to adequately address protocol-related changes in kidney graft function and metabolic profile.

In conclusion, alemtuzumab induction allows short- to medium-term CNI withdrawal with successful evolution in the majority of kidney transplant recipients, and some potential benefits in Treg populations. However, CNI maintenance was even more successful in terms of efficacy. Future development in this area will include standardization of some immune monitoring tools, particularly cytokine kinetics assays, trying to predict possible rejections after CNI withdrawal thus individualizing drug minimization.

Acknowledgment

This work was supported by a grant from ILEX, Inc., San Antonio, TX, USA. JP is supported by a grant from the Institute Carlos III-Spanish Health Department (BA06/90020).

Ancillary