Belimumab in kidney transplantation: an experimental medicine, randomised, placebo-controlled phase 2 trial

Abstract

Background: B cells produce alloantibodies and activate alloreactive T cells, negatively affecting kidney transplant survival. By contrast, regulatory B cells are associated with transplant tolerance. Immunotherapies are needed that inhibit B-cell effector function, including antibody secretion, while sparing regulators and minimising infection risk. B lymphocyte stimulator (BLyS) is a cytokine that promotes B-cell activation and has not previously been targeted in kidney transplant recipients. We aimed to determine the safety and activity of an anti-BLyS antibody, belimumab, in addition to standard-of-care immunosuppression in adult kidney transplant recipients. We used an experimental medicine study design with multiple secondary and exploratory endpoints to gain further insight into the effect of belimumab on the generation of de-novo IgG and on the regulatory B-cell compartment.

Methods: We undertook a double-blind, randomised, placebo-controlled phase 2 trial of belimumab, in addition to standard-of-care immunosuppression (basiliximab, mycophenolate mofetil, tacrolimus, and prednisolone) at two centres, Addenbrooke's Hospital, Cambridge, UK, and Guy's and St Thomas' Hospital, London, UK. Participants were eligible if they were aged 18-75 years and receiving a kidney transplant and were planned to receive standard-of-care immunosuppression. Participants were randomly assigned (1:1) to receive either intravenous belimumab 10 mg per kg bodyweight or placebo, given at day 0, 14, and 28, and then every 4 weeks for a total of seven infusions. The co-primary endpoints were safety and change in the concentration of naive B cells from baseline to week 24, both of which were analysed in all patients who received a transplant and at least one dose of drug or placebo (the modified intention-to-treat [mITT] population). This trial has been completed and is registered with ClinicalTrials.gov, NCT01536379, and EudraCT, 2011-006215-56.

Findings: Between Sept 13, 2013, and Feb 8, 2015, of 303 patients assessed for eligibility, 28 kidney transplant recipients were randomly assigned to receive belimumab (n=14) or placebo (n=14). 25 patients (12 [86%] patients assigned to the belimumab group and 13 [93%] patients assigned to the placebo group) received a transplant and were included in the mITT population. We observed similar proportions of adverse events in the belimumab and placebo groups, including serious infections (one [8%] of 12 in the belimumab group and five [38%] of 13 in the placebo group during the 6-month on-treatment phase; and none in the belimumab group and two [15%] in the placebo group during the 6-month follow-up). In the on-treatment phase, one patient in the placebo group died because of fatal myocardial infarction and acute cardiac failure. The co-primary endpoint of a reduction in naive B cells from baseline to week 24 was not met. Treatment with belimumab did not significantly reduce the number of naive B cells from baseline to week 24 (adjusted mean difference between the belimumab and placebo treatment groups -34·4 cells per μL, 95% CI -109·5 to 40·7).

Interpretation: Belimumab might be a useful adjunct to standard-of-care immunosuppression in renal transplantation, with no major increased risk of infection and potential beneficial effects on humoral alloimmunity.

Funding: GlaxoSmithKline.

Figure 1. Enrolment, Randomisation and Follow-up.

The randomised treatment was given in addition to standard immunosuppression consisting of basiliximab, mycophenolate mofetil, tacrolimus and corticosteroids (see methods for full details). ¹260 at Addenbrooke’s Hospital, Cambridge (28 randomised); 43 at Guy's and St Thomas’ NHS Foundation Trust, London (0 randomised). ²Reasons for not meeting inclusion criteria shown in Supplementary Table 1. Logistical reasons included unavailability of study staff or randomised study drug, insufficient time for consent and recipient resident out of region. ³The pre-specified per protocol (PP) population used for biomarker and clinical endpoint analysis consisted of all subjects in the modified intention to treat (MITT) population who received at least five infusions of belimumab/placebo, no prohibited medications or plasma exchange and at least 24 weeks follow-up. FSGS Focal Segmental Glomerulosclerosis, SAE Serious Adverse Event

Figure 2. Belimumab Affects Naïve B Cells, Activated Memory B Cells and Circulating Plasmablasts, and Reduces Allo/Auto-antibody Generation.

Panel A shows the adjusted mean (95% CI) change from baseline in naïve (CD20⁺CD27^-) B cell count (cells/mm³), by visit and treatment. Panel B is a radar plot summarising the median difference from baseline to week 24 compared to baseline (i.e. percentage at week 24 – percentage at week 0) for the labelled B cell populations by belimumab or placebo, where each B-cell population is expressed as a percentage of total B cells. Panel C shows memory (CD20⁺CD27⁺) B cells expressed as a percentage of B cells by visit and treatment group. Panel D shows the relationship between memory B cell count (cells/mm³) and activated memory B cell (CD95⁺CD27⁺) percentage at the post treatment week 12 timepoint for individual subjects labelled by treatment group. Panel E shows expression (arbitrary units) of a B cell whole blood gene expression module by treatment group at baseline, end of on-treatment phase (Week 24) and end of follow-up (Week 52). Horizontal lines correspond to median and interquartile ranges. Panel F shows a heat map of differential gene expression (belimumab versus placebo groups) for genes in the B cell module at baseline, week 24 and week 52. Genes are ordered by fold-change at week 24 and immunoglobulin-coding transcripts are highlighted. Colour corresponds to log₂-fold change, with blue indicating higher expression in the placebo group (Placebo) relative to the belimumab group (Belimumab). Panel G shows the number of unique ProtoArray antigen specificities with significant antibody binding at Week 24 but not at baseline. Horizontal lines on boxplot (left) correspond to median and interquartile ranges. Ordered individual participant data are also shown (right), with participants coloured by treatment group. The Wilcoxon rank-sum test was used to compare the belimumab treatment group (BEL) and placebo treatment group (PBO); * denotes p<0.05. Panel H shows the number of kidney-specific unique ProtoArray antigen specificities (as defined in 37) with significant antibody binding at Week 24. Horizontal lines on boxplot (left) correspond to median and interquartile ranges. Ordered individual participant data are also shown (right), with participants coloured by treatment group. The Wilcoxon rank-sum test was used to compare the belimumab treatment group (BEL) and placebo treatment group (PBO); exact p value displayed. Panel A uses the MITT population. Panels B, C, D, G and H use the PP population. Panels E and F use the MITT population for baseline and the PP population thereafter. C shows raw values at baseline for comparison and adjusted mean estimate with 95% confidence intervals at subsequent timepoints. Adjusted mean estimates and 95% confidence intervals are obtained from MMRM model, with fixed categorical effects of treatment, visit, donor type and treatment-by-visit interaction and fixed continuous covariates of baseline and baseline-by-visit interaction. A compound symmetry variance structure was used to model the within-patient errors, shared across treatments. # indicates that the 95% confidence interval of the treatment difference does not include zero. D shows data for individual PP subjects labelled by treatment group. Panels B, C, D, E, F, G and H represent analyses performed post hoc.

Figure 3. Memory B Cell IL-10/IL-6 Ratio is BLyS Dependent. Belimumab Treatment Increases Regulatory B Cells and is Associated with Reduced T cell Proliferation.

Peripheral blood mononuclear cells (PBMC) were stimulated ex vivo for 5 hours and intracellular cytokine production quantified by flow cytometry (Panel A-D). Individual data points represent individual subjects with horizontal lines signifying the median for each treatment group. Wilcoxon rank-sum tests were performed to compare samples by treatment and visit. Panel A shows representative flow cytometry plots for IL-6 (upper) and the percentage of CD19+ B cells expressing IL-6 (lower). Panel B shows representative flow cytometry plots for IL-10 (upper) and the percentage of CD19+ B cells expressing IL-10 by treatment group and timepoint (lower). Panel C shows the calculated ratio of IL-10/IL-6 with higher values indicating a more anti-inflammatory cytokine milieu. Panel D shows the IL-10/IL-6 ratio for individual subsets of naïve (CD27-), transitional (CD24^hiCD38^hiIgD⁺), CD24⁺CD27⁺ memory, switched memory (CD27⁺IgD^-), and non switched memory (CD27⁺IgD⁺) B cells at post treatment Wk 12. Panel E shows the mean IL-10/IL-6 ratio (relative to 0ng/mL BLyS) with 95% confidence intervals for PBMC from healthy volunteers enriched for CD27+ memory B cells and stimulated for 48 hours with increasing quantities of BLyS. In the presence of increasing quantities of BLyS a more inflammatory cytokine milieu was observed. This change was blocked by the addition of belimumab. N, number of healthy volunteers tested for each experimental condition. T-tests were performed to determine whether the mean percentage changes from baseline differed significantly from 0 for each experimental condition. Panel F shows a heatmap of differential gene expression in circulating CD4+ T-cells for the most downregulated genes at week 24 in the belimumab group relative to the placebo group, with genes ordered by unsupervised hierarchical clustering. Genes annotated to the cell-cycle (GO:0007049) are highlighted. Colour corresponds to log₂-fold change, with blue indicating higher expression in the placebo group (Placebo) relative to the belimumab group (Belimumab). Panels A-D use the PP population and Panel F uses the MITT population for baseline and PP population thereafter. P-values in Panels A-D were calculated using Wilcoxon rank-sum tests and indicated where significant; in all other cases the values are not statistically significantly different. P-values in Panel E were calculated using Student’s t-test. * p<0.05, ** p<0.01, *** p<0.001, NS not significant (p≥0.05). All analyses were performed post hoc.