Effect of the Proteasome Inhibitor Bortezomib on Humoral Immunity in Two Presensitized Renal Transplant Candidates

Abstract

Background. 

Recipient presensitization represents a major hurdle to successful renal transplantation. Previous case series have suggested that the proteasome inhibitor bortezomib directly affects the alloantibody-secreting plasma cells in rejecting allograft recipients. However, the ability of this agent to desensitize nonimmunosuppressed transplant candidates before transplantation is currently unknown.

Methods. 

In this analysis, two sensitized hemodialysis patients were selected to receive two subsequent bortezomib cycles. Bortezomib was given at 1.3 mg/m² on days 1, 4, 8, and 11. Dexamethasone was added to the second cycle to enhance treatment efficiency. Serial immune monitoring included cytotoxic panel reactive antibody testing, Luminex single antigen testing for anti-human leukocyte antigen (HLA) IgG with or without C4d-fixing capability, and ABO antibody detection.

Results. 

During a half-year follow-up period, cytotoxic panel reactive antibody decreased from 87% to 80% (patient 1) and 37% to 13% (patient 2). Patient 1 showed a 40% reduction in binding intensities of identified Luminex HLA single antigen reactivities and, in parallel, slight reductions in ABO blood group antibody and total immunoglobulin levels. In patient 2, bortezomib did not affect circulating antibody levels in a meaningful way. Both patients showed a more than 50% reduction in the levels of anti-HLA antibody-triggered C4d deposition to Luminex beads.

Conclusion. 

Our initial experience suggests that, without additional immunosuppressive measures, bortezomib has modest effects on circulating antibodies against HLA or blood group antigens. The reduced levels of antibody-triggered complement fixation, however, imply potential clinical relevance of proteasome inhibition for recipient desensitization.

Humoral allosensitization represents a major barrier to successful transplantation (^{1, 2}). A variety of immunosuppressive regimens, such as intravenous immunoglobulin (IVIG), apheresis, and CD20 antibody rituximab, have been proposed for recipient desensitization (^{3, 4}). However, down-regulation of antibody levels is often incomplete and transient, presumably because such strategies may not considerably affect the integrity and function of plasma cells as a major source of secreted immunoglobulin (^5–7).

There is now accumulating evidence that bortezomib, a selective inhibitor of the 26S proteasome, which has proven effective in the treatment of malignant plasma cell disorders (^{8, 9}), could be effective as “anti-humoral” treatment strategy in allogeneic transplantation (^{6, 10–12}). Recent case series have suggested that bortezomib, together with other strategies, such as plasmapheresis, IVIG, and rituximab, could be effective in the treatment of antibody-mediated rejection (^{6, 10–12}). In an elegant analysis of two rejecting transplant recipients subjected to a single bortezomib cycle, Perry et al. (⁶) observed a transient decrease in the number of bone marrow-derived plasma cells, which was paralleled by a marked down-regulation of anti-human leukocyte antigen (HLA) antibody production.

The ability of bortezomib to desensitize transplant candidates before transplantation remains to be established. Herein, we selected two sensitized dialysis patients to receive treatment with bortezomib and, after 3 to 4 months, bortezomib plus dexamethasone, a combination proven to enhance efficiency in myeloma treatment (¹³). The results of serial immune monitoring are provided in detail.

PATIENTS AND METHODS

Patients

Baseline characteristics and immunologic data are detailed in Table 1. Patient 1 has been sensitized by a first allograft lost because of not further specified interstitial fibrosis and tubular atrophy. A second transplant had been performed under desensitization with IVIG according to a previously published protocol (¹⁴). On the basis of severe acute antibody-mediated rejection, he had been included in a trial designed to assess the treatment efficiency of immunoadsorption and randomized to the control group (no apheresis treatment) (¹⁵). Despite rescue treatment 3 weeks after inclusion, the patient remained dialysis dependent. In 2005, he was wait listed for a third kidney transplant and, on the basis of an extensive level of sensitization (complement-dependent cytotoxicity [CDC]-panel reactive antibody [PRA] reactivity 80% to 100%), included in the Eurotransplant acceptable mismatch program (¹⁶), however, without a suitable transplant offer over a period of 3 years. Bortezomib treatment was initiated in October 2008.

TABLE 1: Baseline characteristics and immunological data

Patient 2 has lost her first graft because of chronic allograft injury associated with morphologic features suggestive of chronic calcineurin inhibitor toxicity. A second graft was lost because of de novo immune complex glomerulonephritis. In 2004, the patient was again put on the waiting list. Over time, CDC-PRA levels steadily increased to 30% to 40%. Because of an extended waiting time (4.5 years) and two subsequent transplant offers that could not be realized because of positive crossmatches, she was offered bortezomib treatment in November 2008.

Treatment

Patients received two cycles of bortezomib at an interval of 4 (patient 1) and 3 (patient 2) months. Each cycle consisted of 1.3 mg/m² of bortezomib administered twice weekly on days 1, 4, 8, and 11. Dexamethasone (second treatment cycle) was added at 20 mg on the day of bortezomib and the day after. Treatment was initiated after obtaining written informed consent.

Antibody Detection

Sera were sampled over a period of 7 (patient 1) and 6 (patient 2) months. Samples were obtained immediately before hemodialysis treatment and stored at −80°C before testing. To avoid day-by-day test variations, for each applied assay principle, all the samples of a single patient were tested on the same day. CDC-PRA reactivity was assessed applying the standard microcytotoxicity technique originally described by Terasaki and McClelland (¹⁷).

Anti-HLA single antigen (SA) reactivity was detected on a Luminex platform (LABScan 100 flow analyzer; Luminex Corporation, Austin, TX) according to the manufacturer's protocol using LABScreen Single Antigen assays (HLA class I: LABScreen Single Antigen HLA Class I Antibody Detection Test—Combi; HLA class II: LABScreen Single Antigen HLA Class II Antibody Detection Test—Group 1; One Lambda, Canoga Park, CA). To avoid lot-to-lot variation, for all experiments, the same lots of beads were used. The results were recorded as mean fluorescence intensity (MFI). An MFI value more than 500 was reported as positive.

Anti-HLA antibody-triggered C4d-fixation to Luminex beads was assessed using a DyLight 549-labeled anti-C4d polyclonal antibody (Biomedica, Vienna, Austria) as described previously in detail (¹⁸). Thresholds for C4d test results were determined according to the results obtained with five negative control sera and binding to no-antigen control beads (¹⁸).

Anti-blood group A₁ and anti-B reactivities were determined by indirect antiglobulin test (IAT) and direct agglutination (DA) using gel cards containing rabbit anti-human IgG and neutral gel cards (DiaMed; Cressier, Switzerland), respectively. Serial 2-fold dilutions of serum samples were incubated with commercially available reagent red cells (DiaMed) for 15 min at 37°C for IAT and 10 min at room temperature for DA. After gel card centrifugation, the titer was recorded as the inverted value of the highest plasma dilution that gave a weak agglutination reaction (1+). In parallel, standardized sera were tested for control.

RESULTS

Bortezomib treatment was well tolerated without adverse events. On administration of the second treatment cycle (bortezomib plus dexamethasone), patient 1 showed a moderate (<50%) reduction of total immunoglobulin levels (IgG, IgA, and IgM) (Fig. 1A). No such effect was observed in patient 2 (Fig. 2A). Flow cytometric analysis did not reveal major changes in the counts of peripheral B cells, T cells, monocytes, granulocytes, or natural killer cells (data not shown).

FIGURE 1.: Effect of bortezomib (Bz) and bortezomib plus dexamethasone (Bz/D) on immunoglobulin levels and allosensitization in patient 1. Serial serum samples obtained before and after treatment were tested for total IgG, IgM, and IgA levels (A) and complement-dependent cytotoxicity (CDC)-panel reactive antibody (PRA) reactivity (B). Levels of Luminex single antigen (SA) reactivities are shown as mean of 3 to 6 subsequent measurements per month (mean±SD). For five anti-human leukoctye antigen (HLA) class I (C) and five anti-HLA class II SA reactivities (D), which were selected on the basis of highest initial mean fluorescence intensity (MFI) levels, the individual course of binding intensity is illustrated. The average of all 74 detectable IgG SA reactivities was calculated for IgG (E) and C4d binding (F). For statistical comparison of detected binding levels with MFI recorded before treatment nonparametric testing (Mann Whitney U test) was applied. Asterisks indicate P values below 0.05.

FIGURE 2.: Effect of bortezomib (Bz) and bortezomib plus dexamethasone (Bz/D) on humoral alloreactivity in patient 2. Analysis included serial detection of total IgG, IgM, and IgA (A) and complement-dependent cytotoxicity (CDC)-panel reactive antibody (PRA) reactivity (B). Levels of Luminex single antigen (SA) reactivities are shown as the mean of 3 to 6 subsequent measurements at 1-month intervals (mean±SD). The course of five anti-human leukocyte antigen (HLA) class I (C) and five anti-HLA class II single reactivities (D) with highest levels of mean fluorescence intensity (MFI) is illustrated. For a total 22 SA reactivities identified before treatment, the course of average MFI levels is shown for IgG (E) and C4d binding (F). For statistical comparisons between pre- and posttreatment MFI levels nonparametric testing was applied. Asterisks indicate P values less than 0.05. For C4d MFI, there was a trend toward lower MFI at month 4 (P=0.071).

Effect of Bortezomib on Anti-HLA Sensitization

In patient 1, administration of the second treatment cycle was associated with a slight decrease of panel reactivity to 80% (Fig. 1B). Patient 2 showed a more pronounced decrease to 23% and 13% after the first and second cycles, respectively (Fig. 2B).

Applying Luminex SA testing, individual anti-HLA alloreactivity patterns were analyzed at close intervals. A remarkable observation was a substantial variation in MFI values within short sampling intervals. For simplified presentation of the kinetics of Luminex reactivities, we calculated the average of three to six subsequent measurements within monthly intervals. The course of 10 individual anti-HLA SA reactivities (five anti-HLA class I and five anti-HLA class II reactivities, respectively), which were selected according to highest initial MFI, is illustrated in Figures 1 and 2. In patient 1, all these reactivities showed a moderate decrease in MFI (Fig. 1C and D). In contrast, for patient 2, this effect was observed only for anti-HLA class I antibodies (Fig. 2C and D). To assess the effect of bortezomib treatment on overall levels of allosensitization, we then calculated the average of the MFI values of all identified anti-HLA reactivities (patient 1: n=74; patient 2: n=22) in monthly intervals. In patient 1, average pretreatment MFI was 1907±707 (mean±SD) (Fig. 1E). There was a slight decrease at month 3 (MFI: 1506±730) and after the second treatment cycle, a more pronounced, even though not significant decrease at months 5 (1330±535) and 7 (1145±455). Patient 2 (initial MFI 895±392) showed virtually no change in binding intensities (month 6: 706±348) (Fig. 2E).

Effect of Bortezomib on the C4d-Fixing Capability of HLA Antibodies

For the 74 IgG SA reactivities identified in patient 1, (C4d) Luminex SA testing revealed average pretreatment MFI levels (three consecutive measurements) of 352±48, and for the 22 IgG SA reactivities identified in patient 2, levels of 141±32 were revealed. As shown in Figures 1F and 2F, we observed a significant more than 50% decrease in MFI values during follow-up (patient 1: 132±92 at month 7; patient 2: 61±22 at month 6; four consecutive measurements). Nevertheless, although there was a clear decrease in MFI levels, we did not observe a considerable reduction of the number of C4d-fixing IgG SA reactivities. As shown in Table 1, before treatment, 59 of 74 (patient 1) and 10 of 22 individual SA reactivities (patient 2) showed significant levels of C4d fixation. At the end of follow-up, this was the case for 57 of 77 and 9 of 22 detectable IgG reactivities. Moreover, the proportions of HLA class I versus class II reactivities with or without C4d-fixing capability did not change on treatment (data not shown).

Effect of Bortezomib on ABO Blood Group Antibody Reactivity

As shown in Figure 3, bortezomib did not affect ABO blood group antibody titers in a meaningful way. In patient 1 (blood group O), after the second treatment cycle, a transient reduction of anti-A₁ or anti-B IgG (IAT) blood group antibody titers by one titer step was observed. Anti-A₁ or anti-B IgM (DA) decreased by one titer step (Fig. 3A). In patient 2 (blood group B), there was no apparent effect on anti-A₁ titers (Fig. 3B).

FIGURE 3.: Effect of bortezomib (Bz) and bortezomib plus dexamethasone (Bz/D) on ABO blood group antibody titers. Patient 1 (blood group 0) was serially tested for anti- A₁ and anti-B antibody titers (A), patient 2 (blood group B) for anti-A₁ titers (B). The indirect antiglobulin test (IAT) method was applied for preferential detection of IgG, the direct agglutination (DA) method for preferential detection of IgM type reactivities.

DISCUSSION

This analysis of two sensitized transplant candidates revealed that two subsequent treatment cycles with bortezomib, without additional extracorporeal therapy or immunosuppression, did not or only moderately affect the levels of circulating HLA antigen-specific IgG or ABO blood group antibodies.

At first glance, our findings may be in contrast to recent studies demonstrating a virtually complete down-regulation of alloantibody levels by proteasome inhibition. Perry et al. (⁶) demonstrated a permanent depletion of anti-HLA antibody levels in two rejecting kidney allograft recipients after a single cycle of bortezomib. In their study, which included investigations of bone marrow aspirates, they demonstrated apoptosis of a majority of isolated antibody-secreting cells associated with a substantial decrease of in vitro alloantibody production (⁶). A similar reduction of anti-HLA antibody levels was also reported in two other posttransplant series (^{11, 19}).

A limitation of available studies, including ours, is the small number of included patients. Accordingly, interindividual variability in responsiveness to treatment could have contributed to some of the observed discrepancies. However, there could be some other explanations for the higher efficiency reported for rejecting patients.

First, a particular sensitivity of a recipient immune system challenged with alloantigen could at least in part have contributed to the higher efficiency of bortezomib posttransplantation. There is experimental evidence that contact with antigen may substantially increase the sensitivity of the B cells or plasma cells to proteasome inhibition-mediated apoptosis (²⁰). This remarkable effect has been extended also to the T-cell compartment (²¹).

Second, in earlier reported kidney transplant recipients, concomitant basal immunosuppression and additional measures including high dose steroids, CD20 antibody rituximab, and plasmapheresis could have considerably enhanced the impact of bortezomib treatment on humoral alloreactivities (^{6, 10–12}). For particular immunosuppressants, including tacrolimus or mycophenolate mofetil, anti-humoral efficiency has been suggested earlier (^{22, 23}). Moreover, steroids are well established to enhance the effect of proteasome inhibition on (malignant) plasma cells (¹³). Of notice, in our study, some effects, such as the reduction of total immunoglobulin or blood group antibodies in patient 1 were observed only after the second cycle including dexamethasone.

Our finding of an incomplete reduction of total immunoglobulin levels may be in line with previous experimental and clinical observations (^{6, 24}). In our patients, allo- and blood group-specific IgG and IgM essentially appeared to parallel the course of total immunoglobulin levels. This observation argues against a selective or preferential depletion of the respective clones in a pretransplant context. Notably, there was no considerable difference in the impact of proteasome inhibition on HLA versus ABO reactivities, suggesting that B-cell compartments to T-dependent or T-independent antigens may not differ in their susceptibility to the effects of proteasome inhibition.

An interesting observation was the more pronounced reduction of the ability of detected anti-HLA antibodies to fix complement split product C4d in vitro. However, the observed down-regulation of C4d binding intensities, which was paralleled by some reduction in CDC-PRA levels, did not result in a decreased number of SA reactivities with detectable C4d fixing ability. Our results suggest that bortezomib exerts particular effects on the quality of alloreactive antibody patterns, which can be speculated to result from subtle changes in the distribution of IgG subclasses. Assuming a causative role of complement in the process of rejection (²⁵), this effect could be beneficial in a clinical context.

Similar to previous studies evaluating bortezomib as anti-humoral treatment, this study is limited by its uncontrolled design and a small sample size. A randomized controlled trial including a nontreated control group to assess the spontaneous course of reactivity patterns will be necessary to definitely clarify the actual impact of bortezomib either before or after transplantation. Nevertheless, we suggest that our preliminary observations could provide a valuable basis for the design of future studies. Such analyses will have to elucidate whether the changes in dosage, introduction of additional antihumoral measures, or intensification or prolongation of treatment could improve the efficiency of bortezomib-based recipient desensitization in advance of transplantation.

ACKNOWLEDGMENTS

The authors thank Romana Raab and the team members of the Department of Clinical Cell Biology and FACS Core Unit at the Children's Cancer Research Institute for excellent technical assistance.