Bortezomib in Late Antibody-mediated Kidney Transplant Rejection

Background

Recent studies have underscored a dominant role of alloimmune injury as a leading cause of long-term graft loss in kidney transplantation. In this respect, the formation of antibodies against specific polymorphic donor antigens, commonly human leukocyte antigens (HLA), has turned out to be an important trigger of graft rejection (Colvin, RB et al. 2007).

Humoral rejection (antibody-mediated rejection; AMR) of organ transplants has been established to constitute a separate rejection entity and in recent years accurate biopsy-based (C4d) and serological criteria for this rejection type have been defined to provide a valuable basis for targeted treatment (Sis, B et al. 2010). It has become evident that a considerable proportion of recipients develop features of AMR late after transplantation, a process culminating in chronic irreversible tissue damage, graft dysfunction and loss (Mauiyyedi, S et al. 2001). Indeed, there are studies suggesting that newly formed donor-specific antibodies (DSA) represent the primary cause of late graft loss (Einecke, G et al. 2009).

In contrast to early acute AMR, where various treatment protocols including immunoadsorption, intravenous immunoglobulin (IVIG) or CD20 antibody (Rituximab) exist, appropriate targeted treatment for late AMR still remains to be established (Vo, A et al. 2008).

In an earlier uncontrolled study of 11 kidney transplant recipients with advanced C4d-positive chronic AMR we found that conversion of Cyclosporine A and Azathioprine-based basal immunosuppression to Tacrolimus and Mycophenolate-Mofetil (MMF) did not have any impact on DSA levels and graft performance (Schwarz, C et al. 2006). Our observations, which are in some contrast to a previous small study (Theruvath, TP et al. 2001), suggested that simple conversion strategies may not suffice to ameliorate ongoing late AMR.

Rationale

More recently, we reported on the treatment with Bortezomib, a proteasome inhibitor, in a patient diagnosed with advanced chronic active C4d-positive AMR (Schwaiger, E et al. 2011). In this patient, a single cycle of Bortezomib led to an impressive decrease in DSA levels and a complete disappearance of capillary C4d deposits in a follow-up biopsy. Also a remarkable observation was the stable decrease in proteinuria. As expected, advanced chronic lesions (severe transplant glomerulopathy) in this patient remained unchanged. Nevertheless, 16 months after treatment, kidney function stabilized and the protein/creatinine ratio was still half of the initial value of 4,000 mg/g.

Objective

The primary objective of this investigator-driven, randomized prospective double-blind, placebo- controlled trial is to assess the efficiency of the innovative concept of proteasome inhibition with Bortezomib in the treatment of late AMR. Our primary hypothesis is that, by inhibiting the DSA production of plasma cells, Bortezomib can halt the progression of ongoing graft injury and dysfunction.

Study Design

The study (anticipated duration time is 36 months) will be performed in two major steps, Part A and B.

Part A:

This part consists of a cross-sectional screening analysis of a large cohort of approximately 1000 kidney transplant recipients from our outpatient clinic for the presence of late AMR.

Part A: Objective

We will screen kidney transplant recipients for the presence of circulating DSA and biopsy features of late AMR. Key inclusion criteria are a functioning graft at > 6 months post-transplantation and an eGFR above 20 ml/min/1.73m2. The GFR threshold was chosen to avoid inclusion of transplants showing a high degree of irreversible chronic damage. For patients with very advanced graft injury a sustainable treatment benefit can no longer be expected.

Part A: Sample size considerations

Approximately 90% of 1000 recipients available for a pre-screening will be eligible for HLA antibody testing. In at least 10% (n≥90) of tested patients DSA detection can be expected (Lachmann, N et al. 2009). We estimate biopsy-based late AMR features in at least 60% of the DSA-positive patients (Hidalgo, LG et al. 2009). Hence, at least 50 patients will be eligible for inclusion into Part B, the randomized controlled trial.

Part A: Methodology

Estimation of GFR eGFR will be calculated using the Mayo equation. For kidney transplants, this equation was reported to be superior with respect to estimations of GFR slopes (Rule, AD et al. 2004).

DSA screening:

Serum obtained in the context of a routine outpatient laboratory control (10 ml blood; no additional venipuncture) will be pre-screened for anti-HLA immunoglobulin G (IgG) alloantibodies using LabScreenMixed assays (One Lambda, Canoga Park, CA, USA). For identification of HLA antigen/allele specificities, pre-screen-positive samples will be subjected to single-antigen flow-bead (SAFB) testing (LABscreen Single Antigen assays, One Lambda). SAFB results will be documented as mean fluorescence intensities (MFI), and MFI levels >500 will be considered to be positive. Donor-specificity will be defined according to donor and recipient HLA typing results. Virtual panel-reactive antibody (PRA) levels will be calculated using specific software tools (http://www.eurotransplant.eu).

Renal biopsies:

All biopsies will be performed under local anesthesia (lidocaine) using ultrasound-guided percutaneous techniques (two cores per biopsy). After biopsy, patients will be monitored closely for at least 6 hours for any complications.

Biopsies will be evaluated on standard paraffin-embedded sections and by electron microscopy.

(i) Histomorphological, (ii) electron-microscopical and (iii) immunohistochemical criteria will be applied to assess a score for late AMR.

Part B:

This is a phase IIa, proof of concept, randomized prospective, double-blind, placebo-controlled trial.

It will examine whether treatment with Bortezomib is capable to halt the progression of late AMR. We plan to include 44 patients into this trial.

Part B: Objective

To evaluate the efficiency and safety of Bortezomib on allograft outcome in recipients with late AMR.

Part B: Randomization

44 patients will be centrally randomized by computer assignment between two study arms (Bortezomib versus Placebo) using a 1:1 randomization scheme. To avoid the bias of unbalanced baseline variables that potentially affect treatment responses, stratification will be performed for eGFR (> versus <50 ml/min/1.73 m2) and the presence or absence of Banff I T-cell-mediated rejection (index biopsy).

For each patient a study ID will be assigned. The participating investigators and the patients will be blinded to group allocation until the completion of the study.

Part B: Intervention

Treatment with Bortezomib or Placebo. Patients allocated to the intervention group will receive two cycles of Bortezomib at an interval of 3 months. Each cycle will consist of Bortezomib 1.3 mg/m2 administered twice weekly on days 1, 4 and days 8, 11. Bortezomib will be given by injection intravenously (within 3-5 seconds). Bortezomib-treated patients will receive oral antiviral prophylaxis to prevent the development of herpes zoster infection: Valacyclovir 500 mg per day (eGFR <30 ml/min/1.73 m2: 250 mg per day) for 3 weeks after initiation of each cycle. Patients allocated to the control group will receive oral Placebo instead of Valacyclovir prophylaxis.

According to our center standard, upon diagnosis of late AMR, all recipients (both study groups) on therapy with a calcineurin inhibitor (Tacrolimus or Cyclosporine A) or a mammalian Target of Rapamycin (mTOR) inhibitor (Everolimus or Sirolimus), without Azathioprine or MMF, will receive MMF (initially 2 x 500 mg per day; in absence of gastrointestinal side effects and significant leukopenia or thrombocytopenia stepwise increase of dose to 2 x 1000 mg per day) to avoid underimmunosuppression. Recipients weaned off steroids will receive low dose Prednisolone (initiation with 10 mg/day, tapered to 5 mg/day within 4 weeks).

Part B: Sample size calculations, power calculation, statistical methodology and interim analyses

Pilot studies using data from the OEDTR (OEsterreichische Dialysis and Transplant Registry) were conducted to estimate the eGFR decline and variance of sequential eGFR determinations. When using only one eGFR determination at two years as primary endpoint, analyses showed an impracticable high sample size number under the assumption of a median treatment effect of Bortezomib (0.5 SD). Therefore, the difference in slope of half yearly determined eGFR between the two treatment groups will be used as quantitative outcomes measure. Mixed linear models for longitudinal data were used for analyses. Power calculation using an autoregressive covariance matrix of the first order using a correlation of 0.9, an alpha of 0.05 for the interaction term of treatment and time and an attrition of 8% per year showed that 2 x 22 subjects would be required to uncover a difference in GFR slopes of 5ml/min/year with a power of 80%.

To provide a solid background for our power analysis, we re-analyzed a large retrospective transplant cohort (transplantation and follow-up at the Medical University Vienna. Evaluating the impact of late (> 6 months) (C4d-positive) AMR on the clinical performance of kidney allografts a GFR slope of -8.2 ml/year (over a follow-up of 6 years), as compared to a slope of -1.8 ml/min/year in non-biopsied and -2.8 ml/min patients in biopsied C4d-negative subjects was observed.

All analyses will be conducted according to the intention-to-treat principle. Continuous data will be analyzed by t-test, categorical data by a chi-square test or Fisher's exact test when appropriate. The analysis of the creatinine trajectories stratified by treatment (and eventually by baseline GFR and T-cell-mediated rejection) will be performed by a mixed linear model with time and therapy as the independent parameters. The most applicable covariance matrices will be determined by graphical analysis and evaluated by the log likely hood ratio.

This study will be monitored by an independent data and safety monitoring board (DSMB) of the Medical University of Vienna (MUW). Interim analyses will be performed by the DSMB after completion of 10 and 20 cases in the treatment group. The Lan & DeMets extension of the O'Brian-Fleming stopping rules will be applied (DeMets, LD et al. 1994). The trial will be stopped if the observed p-value is <0.00001 (first interim analysis after 10 patients) or <0.00305 (second interim analysis after 20 patients).

Part B: Adverse events (AEs)

The most probable AEs caused by Bortezomib (usually transient) are mild to moderate thrombopenia and leukopenia, decreased appetite, gastrointestinal side effects (vomiting, nausea, diarrhea), fatigue, peripheral neuropathy and other neurological symptoms such as headache, dizziness, tremor and hypotension. Intensified immunosuppression may be associated with an increased risk of infection (common: herpes zoster, herpes simplex, pneumonia, bronchitis, sinusitis, nasopharyngitis). Hence, a careful patient follow-up will include a close monitoring for infectious complications (bacterial, viral and fungal infections). Valacyclovir adapted to kidney function will be added to prevent the occurrence of herpes virus infections (herpes zoster).

The most probable AEs caused by Valacyclovir are headache, nausea, gastrointestinal side effects (vomiting, diarrhea, abdominal pain), dizziness, hallucinations, confusion, changes in blood cell counts (Leukocytopenia, thrombocytopenia, anemia), increased liver and kidney parameters and fever.