Purpose: Identify subgroups of patients with relapsed/refractory follicular lymphoma deriving substantial progression-free survival (PFS) benefit with bortezomib–rituximab versus rituximab in the phase III LYM-3001 study.

Experimental Design: A total of 676 patients were randomized to five 5-week cycles of bortezomib–rituximab or rituximab. The primary end point was PFS; this prespecified analysis of candidate protein biomarkers and genes was an exploratory objective. Archived tumor tissue and whole blood samples were collected at baseline. Immunohistochemistry and genetic analyses were completed for 4 proteins and 8 genes.

Results: In initial pairwise analyses, using individual single-nucleotide polymorphism genotypes, one biomarker pair (PSMB1 P11A C/G heterozygote, low CD68 expression) was associated with a significant PFS benefit with bortezomib–rituximab versus rituximab, controlling for multiple comparison corrections. The pair was analyzed under dominant, recessive, and additive genetic models, with significant association with PFS seen under the dominant model (G/G+C/G). In patients carrying this biomarker pair [PSMB1 P11A G allele, low CD68 expression (≤50 CD68-positive cells), population frequency: 43.6%], median PFS was 14.2 months with bortezomib–rituximab versus 9.1 months with rituximab (HR 0.47, P < 0.0001), and there was a significant overall survival benefit (HR 0.49, P = 0.0461). Response rates were higher and time to next antilymphoma therapy was longer in the bortezomib–rituximab group. In biomarker-negative patients, no significant efficacy differences were seen between treatment groups. Similar proportions of patients had high-risk features in the biomarker-positive and biomarker-negative subsets.

Conclusions: Patients with PSMB1 P11A (G allele) and low CD68 expression seemed to have significantly longer PFS and greater clinical benefit with bortezomib–rituximab versus rituximab. Clin Cancer Res; 19(9); 2551–61. ©2013 AACR.

Translational Relevance

Given the heterogeneity of follicular lymphoma and range of possible therapeutic options for patients with relapsed/refractory disease, the use of selected biomarkers based on mechanistic rationales to identify patient subgroups most likely to benefit from specific therapies is important. Candidate proteins and target gene polymorphisms were prespecified as potential prognostic markers for exploratory biomarker analyses in the phase III LYM-3001 study of bortezomib–rituximab versus rituximab in patients with relapsed/refractory follicular lymphoma. Comprehensive pairwise analyses, with genetic model testing, identified a biomarker pair, PSMB1 P11A (C/G+G/G) plus low CD68 expression, associated with a significant progression-free survival benefit, improved response rates, and longer overall survival with bortezomib–rituximab versus rituximab. The benefit with bortezomib–rituximab was substantially greater than seen in the overall, unselected study population. The two biomarkers identified, for which a mechanistic hypothesis is provided for the reported efficacy benefit, would be feasible and practical to screen for in the clinical setting.

The goal of treatment for patients with follicular lymphoma, a generally incurable, common indolent subtype of non-Hodgkin lymphoma (NHL; refs. 1, 2), is to prolong progression-free survival (PFS) and improve overall survival (OS). Follicular lymphoma is a highly heterogeneous disease, as evidenced by variability in disease course, responsiveness to treatment, and outcomes (3–5). Therefore, to optimize treatment for individual patients, identification of subgroups that are most likely to benefit from a specific therapy is important.

The anti-CD20 monoclonal antibody rituximab is the mainstay of treatment for follicular lymphoma (1). Additional treatment options may enhance the activity of rituximab-based therapy in the relapsed setting. The proteasome inhibitor bortezomib has shown single-agent activity in follicular lymphoma and other indolent NHL subtypes (6–8), as well as promising activity in combination with rituximab, with or without other agents, in patients with relapsed/refractory follicular lymphoma (9–12).

Results from the international, multicenter, randomized, phase III LYM-3001 study (13) showed improved PFS with bortezomib–rituximab versus rituximab alone in patients with relapsed/refractory rituximab-naïve or rituximab-sensitive follicular lymphoma (median 12.8 vs. 11.0 months, HR 0.822, P = 0.039). The combination resulted in a significantly greater overall response rate (ORR; 63% vs. 49%, P = 0.0004), complete response (CR) rate (25% vs.18%, P = 0.035), and durable (≥6 months) response rate (50% vs. 38%, P = 0.002; ref. 13). After a median follow-up of 33.9 months, there was no difference in OS between arms (13).

The LYM-3001 protocol included as an exploratory end point a biomarker analysis aimed at identifying patient subgroups that derived a longer PFS benefit with, and were more likely to respond to, bortezomib–rituximab or rituximab alone. Candidate proteins and target gene polymorphisms were specified on the basis of prior studies of prognostic molecular markers for lymphoma and of target or response markers for bortezomib or rituximab. Here, we present the findings of this exploratory biomarker analysis.

Patients and clinical study design

The LYM-3001 (ClinicalTrials.gov trial registration ID: NCT00312845) study design has been reported previously (13). Eligible patients were aged 18 years or more with relapsed/refractory, rituximab-naïve, or rituximab-sensitive (13) grade 1/2 follicular lymphoma. Patients with grade ≥2 peripheral neuropathy or clinical evidence of transformation to aggressive lymphoma were excluded. Patients were randomized (1:1) to receive five 5-week cycles comprising bortezomib (1.6 mg/m2, days 1, 8, 15, and 22; all cycles) plus rituximab (375 mg/m2, days 1, 8, 15, and 22, cycle 1, and day 1, cycles 2–5), or rituximab alone. Randomization was stratified according to Follicular Lymphoma International Prognostic Index (FLIPI; ref. 14) score, previous rituximab treatment, time since last dose of antilymphoma treatment, and region.

The primary end point was PFS. Secondary efficacy end points included ORR, CR rate, time to progression, and 1 year OS. Response was assessed using modified International Working Group response criteria (15). Time to next antilymphoma treatment (TTNT; time from randomization to first dose of next treatment) was an additional predefined efficacy end point.

All patients provided written informed consent. Review boards at all participating Institutions approved the study, which was conducted according to the provisions of the Declaration of Helsinki, the International Conference on Harmonization, and the Guidelines for Good Clinical Practice.

Biomarker analysis study design

This prespecified focused analysis of potential biomarkers of sensitivity to bortezomib–rituximab or rituximab was an exploratory objective. All patients were required to provide consent for biomarker testing, and were included in the biomarker study if they had evaluable biomarker data and data for at least one clinical end point. Archived tumor tissue (samples requested from diagnosis) was collected at baseline, and samples were forwarded to a central laboratory as paraffin-embedded, formalin-fixed blocks or 6-micron slides. Whole blood samples for DNA analysis were collected at baseline. Serum samples were collected at multiple time points and stored for optional exploratory protein analysis.

Prespecified candidate biomarkers

Protein candidates were NF-κB p65, proteasome subunit α-5 (PSMA5), p27, and CD68. These were chosen based on their attenuation by bortezomib (NF-κB, PSMA5, p27; refs. 16–20), and associations with poor prognosis in lymphoma (CD68; refs. 21, 22) and rituximab activity (CD68; ref. 22).

Drug target candidate genes were included for both bortezomib and rituximab. The proteasome subunit β (PSMB) genes 1, 2, and 5 interact with bortezomib and variation in these and in PSMB 6, 7, and 8 may be important for interindividual variation in responses (23–25). Variation in FCGR2A (26, 27) and FCGR3A influences affinity of immunoglobulin G antibodies for the Fcγ immune cell surface receptor (28, 29); and an association has already been shown between FCGR3A genotype and response to rituximab-based therapy (30, 31).

Assay methods

Validated immunohistochemistry tests were used for protein analysis at a central laboratory; a single pathologist evaluated all samples, averaging counts over 3 high-powered fields for each marker. Assays used included: CD68, DakoCytomation (#M0184; average macrophage counts overall and for the follicular and perifollicular spaces were calculated; CD68 correlated directly with the number of infiltrating macrophages); NF-κB, Cell Signaling (#C22B4); PSMA5, Biomol International (#PW8125); and p27, Transduction Laboratories (BD Biosciences, #K25020). Cut points for protein markers used in the analyses are summarized in Supplementary Table S1. When insufficient sample was available for testing all biomarkers, immunohistochemistry analyses were prioritized; CD68 had the lowest priority.

TaqMan single-nucleotide polymorphism (SNP) assays (Applied Biosystems) and custom PCR/ligase detection reaction were used for genotyping (see Supplementary Materials and Methods). Alleles from PSMB subunit and FCGR2A/3A genes with sufficient variation (>10%) included PSMB1 P11A (rs12717), PSMB5 R24C (rs11543947), PSMB8 G8R (rs114772012), PSMB9 R60H (rs17587) and V32I (rs241419), FCGR2A H131R (rs1801274), Q62R (rs9427398), and Q62X (rs9427397) and FCGR3A V212F (rs396991).

Statistical analysis

The primary single-marker association analysis was aimed at identifying differentially expressed proteins or genotypes associated with clinical study end points (PFS, ORR, CR, TTNT, OS). For single-marker association analyses and pairwise comparisons, the log-rank test and Cox proportional hazard model were used for assessments of PFS, TTNT, and OS with bortezomib–rituximab and rituximab groups, including calculation of HRs and 95% confidence intervals (CI).

For pairwise comparison analyses, biomarker pairs were formed by exhaustive combinations of 2 markers. Genotypes of SNPs were used individually to combine with other biomarkers. Multiple testing corrections were conducted using the false discovery rate (FDR) method for pairwise comparisons (32). Biomarker pairs containing a SNP marker that were associated with a significant PFS benefit with bortezomib–rituximab versus rituximab were evaluated under dominant, recessive, and additive genetic models to determine association with PFS and OS, with the aim of identifying an association with improved outcome with combination treatment. Analyses were conducted in all evaluable patients, stratified by expression level of the non-SNP member of the biomarker pair. Effects of genotype, treatment, and interaction of treatment and genotype were evaluated using a Cox regression model, including calculation of HRs and 95% CIs. The log-rank test was used for comparisons of outcomes with bortezomib–rituximab and rituximab subgroups according to genotype.

Kaplan–Meier methodology was used to estimate PFS, TTNT, and OS distributions. Response rates were compared between groups using the Fisher exact test. Clinical covariates were FLIPI score (low, intermediate, high); prior rituximab therapy (yes, no); time since last dose of antilymphoma therapy (≤1 year, >1 year); region (United States/Canada, European Union, Rest of World), age (≤65, >65 years), sex, race, Ann Arbor stage (I, II, III, IV), number of prior lines of therapy (1, ≥2), and high tumor burden (yes, no). Treatment exposure and safety were summarized using descriptive statistics in bortezomib–rituximab and rituximab subgroups for biomarker pairs associated with a significant PFS benefit, with no statistical comparison between subgroups or between biomarker-positive and biomarker-negative patients.

In the absence of independent datasets to validate findings, all LYM-3001 patients with no missing biomarker values were assigned (7:3 ratio) to discovery and confirmation test sets using simple randomization. The discovery set was used for identification of biomarkers significantly associated with a PFS benefit. The confirmation set was used for independent validation; patients with missing data were included in the confirmation set provided data were available for significant biomarkers identified in the discovery phase. All markers and covariates were treated as categorical variables; protein biomarkers were dichotomized by staining pattern and frequencies of patients within score groups (Supplementary Table S1). PFS after treatment with bortezomib–rituximab and rituximab in biomarker subpopulations was compared using the log-rank test, with 5-fold cross-validation for the biomarker pair, PSMB1 P11A (dominant C/G + G/G) and CD68 follicular expression ≤50, in the discovery set, and subsequently tested in the confirmation set.

Patient characteristics

In total, 336 and 340 patients were randomized to treatment with bortezomib–rituximab and rituximab, respectively, and 334 and 339 received treatment (13). Patients providing evaluable samples and included in the biomarker analyses are summarized in Supplementary Fig. S1.

Single-marker associations and pairwise comparisons

Initial analyses focused on single-marker associations. Significant (P < 0.05) differences in PFS in patients treated with bortezomib–rituximab versus rituximab were seen in patient subgroups defined by biomarkers including CD68 positivity, NF-κB p65 cytoplasmic signal, and PSMB1 P11A, FCGR2A/H166R, and FCGR3A/V212F genotypes; however, the PFS benefit in these patient subgroups (the difference in median PFS in patients treated with bortezomib–rituximab versus rituximab) was generally less than 5 months or population frequencies were low (Supplementary Fig. S2).

In pairwise analyses (1,140 comparisons), in which genotypes of SNP markers were used individually in combination with other biomarkers, 102 biomarker pairs were identified as defining a patient subgroup in which there was a significant (P < 0.05) difference in PFS in patients treated with bortezomib–rituximab versus rituximab. Of these, the difference in median PFS in patients treated with bortezomib–rituximab versus rituximab was 6 months or more for 14 biomarker pairs (Table 1). Following FDR correction to control for multiple comparisons (32), this PFS difference remained significant for one biomarker pair (PSMB1 P11A C/G heterozygote, ≤50 CD68-positive cells). The population frequency of this biomarker pair was 33%. This biomarker pair was subsequently tested under different genetic models.

Table 1.

Significant biomarker pairs identified by pairwise analysisa

Bortezomib–rituximabRituximab
Marker AMarker BNPFS, moNPFS, moΔ PFS, moLog-rank PFDR
PSMB5 R24C C/T NF-κB p65 cytoplasmic signal intensity ≤1+ 27.0 10.4 16.6 0.0439 0.489 
PSMB1 P11A C/G PSMA5-positive cytoplasmic signal >90% 50 18.9 50 9.5 9.4 0.0145 0.447 
PSMB1 P11A C/G CD68 positive (follicular) ≤50 57 16.6 61 9.1 7.5 0.0001 0.051 
PSMB1 P11A C/G Time since last treatment >1 year 72 18.2 74 10.7 7.5 0.0198 0.447 
PSMB1 P11A C/G CD68 positive (perifollicular) >50 24 16.6 28 9.2 7.4 0.0365 0.471 
PSMB9 R60H G/G NF-κB p65 nuclear positive >0% 35 16.2 28 9.5 6.7 0.0303 0.455 
PSMB5 R24C C/T CD68 positive (follicular) ≤50 18 13.7 21 7.2 6.5 0.0220 0.447 
PSMB1 P11A C/G Age ≤65 years 86 15.3 96 9.2 6.1 0.0071 0.437 
1 prior regimen CD68 positive (follicular) ≤50 63 18.2 69 9.3 8.9 0.0129 0.447 
Race group ‘other’ PSMA5 nuclear staining >20% 11 11.4 3.8 7.6 0.0320 0.455 
High tumor burden: No CD68 positive (overall) ≤50 64 22.8 68 16.0 6.8 0.0177 0.447 
High tumor burden: No CD68 positive (follicular) ≤50 64 20.5 66 13.8 6.7 0.0310 0.455 
No prior rituximab CD68 positive (follicular) ≤50 73 15.9 86 9.2 6.7 0.0066 0.437 
Male sex PSMA5 nuclear staining >20% 63 13.7 48 7.7 6.0 0.0050 0.437 
Bortezomib–rituximabRituximab
Marker AMarker BNPFS, moNPFS, moΔ PFS, moLog-rank PFDR
PSMB5 R24C C/T NF-κB p65 cytoplasmic signal intensity ≤1+ 27.0 10.4 16.6 0.0439 0.489 
PSMB1 P11A C/G PSMA5-positive cytoplasmic signal >90% 50 18.9 50 9.5 9.4 0.0145 0.447 
PSMB1 P11A C/G CD68 positive (follicular) ≤50 57 16.6 61 9.1 7.5 0.0001 0.051 
PSMB1 P11A C/G Time since last treatment >1 year 72 18.2 74 10.7 7.5 0.0198 0.447 
PSMB1 P11A C/G CD68 positive (perifollicular) >50 24 16.6 28 9.2 7.4 0.0365 0.471 
PSMB9 R60H G/G NF-κB p65 nuclear positive >0% 35 16.2 28 9.5 6.7 0.0303 0.455 
PSMB5 R24C C/T CD68 positive (follicular) ≤50 18 13.7 21 7.2 6.5 0.0220 0.447 
PSMB1 P11A C/G Age ≤65 years 86 15.3 96 9.2 6.1 0.0071 0.437 
1 prior regimen CD68 positive (follicular) ≤50 63 18.2 69 9.3 8.9 0.0129 0.447 
Race group ‘other’ PSMA5 nuclear staining >20% 11 11.4 3.8 7.6 0.0320 0.455 
High tumor burden: No CD68 positive (overall) ≤50 64 22.8 68 16.0 6.8 0.0177 0.447 
High tumor burden: No CD68 positive (follicular) ≤50 64 20.5 66 13.8 6.7 0.0310 0.455 
No prior rituximab CD68 positive (follicular) ≤50 73 15.9 86 9.2 6.7 0.0066 0.437 
Male sex PSMA5 nuclear staining >20% 63 13.7 48 7.7 6.0 0.0050 0.437 

NOTE: The shaded biomarker pair was determined to remain significant following FDR correction for multiple testing (32).

aSamples from China were not included in these initial pairwise biomarker assessments but were included in the subsequent genetic model analyses.

The biomarker pair of PSMB1 P11A and CD68 follicular expression was evaluated for association with PFS and OS under dominant (PSMB1 P11A C/C vs. C/G + G/G), recessive (C/C + C/G vs. G/G), and additive (CC vs. C/G vs. G/G) genetic models, stratified by CD68 follicular expression (0–50 vs. >50 CD68-positive cells). By Cox regression analysis, among all evaluable patients, significant association with PFS was only seen under the dominant genetic model in patients with low CD68 expression. PSMB1 P11A C/G + G/G was significantly associated with PFS in the rituximab arm (P = 0.0238), and there was a trend of significant interaction between PSMB1 P11A genotype and treatment (P = 0.0647; Supplementary Table S2 and Supplementary Fig. S3), indicating that the association of the marker pair with PFS may be different between treatment arms. No significant associations with OS were seen, although there was a trend for association with OS in the bortezomib–rituximab arm with low CD68 expression and PSMB1 P11A under both the dominant (P = 0.0801; Supplementary Fig. S3) and additive (P = 0.0888) models. In subsequent analyses, the dominant model (G allele: C/G + G/G) was adopted for PSMB1 P11A in the significant biomarker pair.

Clinical outcomes in patients with/without significant biomarker pair

A total of 376 patients (186 bortezomib–rituximab treated, 190 rituximab treated) were evaluable for both PSMB1 P11A and CD68. There were 164 (43.6%) patients positive for the biomarker pair (PSMB1 P11A G allele, low CD68 expression, hereafter referred to as “biomarker-positive” patients; 78 bortezomib–rituximab treated, 86 rituximab treated). The PSMB1/CD68-evaluable population (total, and by treatment arm) was representative of the overall population in terms of demographics and baseline characteristics (Table 2), with no significant differences (except region and race, likely due to lower collection rate of tumor blocks/slides in China compared with the rest of the world). In addition, there was a similar distribution in baseline characteristics between the bortezomib–rituximab and rituximab arms in the biomarker-evaluable and overall populations, including the asymmetry in male gender between the 2 treatment groups (13). Notably, similar proportions of patients with high-risk features were seen in biomarker-positive and biomarker-negative subsets; FLIPI score was high in 68 (41%) and 89 (42%) patients, respectively, intermediate in 54 (33%) and 76 (36%) patients, and low in 42 (26%) and 47 (22%) patients, respectively (P = 0.703). A total of 93 (57%) biomarker-positive and 110 (52%) biomarker-negative patients had high tumor burden (P = 0.404). Both high FLIPI score and high tumor burden were seen in 52 (32%) biomarker-positive and 59 (28%) biomarker-negative patients, respectively (P = 0.427).

Table 2.

Comparison of baseline demographics and disease characteristics between the overall study population and the PSMB1 P11A G allele/CD68 biomarker-evaluable population

Overall populationBiomarker-evaluable populationPa, overall vs. biomarker population
Btz-R (n = 336)R (n = 340)Total (N = 676)Btz-R (n = 186)R (n = 190)Total (N = 376)Btz-RRTotal
Median age, y (range) 57 (24–83) 57 (21–84) 57 (21–84) 57 (24–79) 57 (22–82) 57 (22–82) 0.9666 0.8824 0.8895 
Age >65 years, n (%) 81 (24) 87 (26) 168 (25) 45 (24) 49 (26) 94 (25) 0.9971 0.9747 0.9788 
Male, n (%) 172 (51) 137 (40) 309 (46) 96 (52) 71 (37) 167 (44) 0.9530 0.4916 0.6552 
White, n (%)b 249 (74) 257 (76) 506 (75) 158 (85) 158 (83) 316 (84) 0.0050 0.0487 0.0007 
Region, RoW, n (%)b 183 (55) 182 (54) 365 (54) 88 (47) 90 (47) 178 (47) 0.1093 0.1630 0.0341 
High FLIPI (≥3), n (%) 138 (41) 140 (41) 278 (41) 74 (40) 83 (44) 157 (42) 0.8974 0.8642 0.9715 
High tumor burden, n (%) 185 (55) 179 (53) 364 (54) 102 (55) 101 (53) 203 (54) 0.9325 0.9374 0.9959 
Prior rituximab, n (%) 147 (44) 147 (43) 294 (44) 83 (45) 85 (45) 168 (45) 0.8700 0.7600 0.7399 
>1 prior therapy, n (%) 192 (57) 202 (60) 394 (59) 104 (56) 109 (57) 213 (57) 0.9233 0.8776 0.8063 
Overall populationBiomarker-evaluable populationPa, overall vs. biomarker population
Btz-R (n = 336)R (n = 340)Total (N = 676)Btz-R (n = 186)R (n = 190)Total (N = 376)Btz-RRTotal
Median age, y (range) 57 (24–83) 57 (21–84) 57 (21–84) 57 (24–79) 57 (22–82) 57 (22–82) 0.9666 0.8824 0.8895 
Age >65 years, n (%) 81 (24) 87 (26) 168 (25) 45 (24) 49 (26) 94 (25) 0.9971 0.9747 0.9788 
Male, n (%) 172 (51) 137 (40) 309 (46) 96 (52) 71 (37) 167 (44) 0.9530 0.4916 0.6552 
White, n (%)b 249 (74) 257 (76) 506 (75) 158 (85) 158 (83) 316 (84) 0.0050 0.0487 0.0007 
Region, RoW, n (%)b 183 (55) 182 (54) 365 (54) 88 (47) 90 (47) 178 (47) 0.1093 0.1630 0.0341 
High FLIPI (≥3), n (%) 138 (41) 140 (41) 278 (41) 74 (40) 83 (44) 157 (42) 0.8974 0.8642 0.9715 
High tumor burden, n (%) 185 (55) 179 (53) 364 (54) 102 (55) 101 (53) 203 (54) 0.9325 0.9374 0.9959 
Prior rituximab, n (%) 147 (44) 147 (43) 294 (44) 83 (45) 85 (45) 168 (45) 0.8700 0.7600 0.7399 
>1 prior therapy, n (%) 192 (57) 202 (60) 394 (59) 104 (56) 109 (57) 213 (57) 0.9233 0.8776 0.8063 

Abbreviations: Btz, bortezomib; R, rituximab; RoW, rest of the World, excluding USA, Canada, and Europe.

at test/Mann–Whitney test.

bSignificant differences likely due to lower collection rate of tumor blocks/slides in China (30%) compared with the rest of the world (80%).

In biomarker-positive patients, PFS was significantly longer in patients treated with bortezomib–rituximab versus rituximab (median 14.2 vs. 9.1 months, HR 0.47, P < 0.0001; FDR P = 0.047; Table 3 and Fig. 1A). There was also a significantly higher ORR, a numerically higher CR rate, and a significantly longer TTNT. These patients also seemed to have longer OS (HR 0.49, uncorrected P = 0.0461; Fig. 1B), although this was not significant after correction for multiple testing. In biomarker-negative patients, no significant efficacy differences between treatment with bortezomib–rituximab and rituximab were seen. The PFS findings were reflected when stratified by clinical covariates (Fig. 2); the PFS benefit with bortezomib–rituximab versus rituximab was consistent and generally statistically significant in biomarker-positive patients, but no significant differences were seen in biomarker-negative patients. PFS findings were also similar in high-risk patients with both high FLIPI score and high tumor burden (data not shown).

Figure 1.

Kaplan–Meier distributions of (A) PFS, (B) OS, and (C) TTNT with bortezomib–rituximab and rituximab in biomarker-evaluable patients (N = 376) who were positive or negative for the biomarker pair PSMB1 P11A (G allele) and low CD68 expression (≤50 CD68-positive cells).

Figure 1.

Kaplan–Meier distributions of (A) PFS, (B) OS, and (C) TTNT with bortezomib–rituximab and rituximab in biomarker-evaluable patients (N = 376) who were positive or negative for the biomarker pair PSMB1 P11A (G allele) and low CD68 expression (≤50 CD68-positive cells).

Close modal
Figure 2.

PFS stratified by clinical covariates for biomarker-evaluable patients (N = 376) who were positive or negative for the biomarker pair PSMB1 P11A (G allele) and low CD68 expression (≤50 CD68-positive cells).

Figure 2.

PFS stratified by clinical covariates for biomarker-evaluable patients (N = 376) who were positive or negative for the biomarker pair PSMB1 P11A (G allele) and low CD68 expression (≤50 CD68-positive cells).

Close modal
Table 3.

Outcomes and response rates with bortezomib–rituximab versus rituximab in biomarker-positive [PSMB1 P11A G allele, and low CD68 expression (≤50 CD68-positive cells)] and biomarker-negative patients, in all biomarker evaluable patients (N = 376), and in the overall study population (N = 676; ref. 13)

Biomarker positiveBiomarker negativeAll biomarker evaluableTotal study population
Btz-R (n = 78)R (n = 86)Btz-R (n = 108)R (n = 104)Btz-R (n = 186)R (n = 190)Btz-R (n = 336)R (n = 340)
Median PFS, months 14.2 9.1 11.6 13.7 13.6 11.0 12.8 11.0 
HR (95% CI) 0.47 (0.32–0.68) 1.12 (0.80–1.58) 0.79 (0.62–1.01) 0.82 (0.68–0.99) 
P <0.0001 0.5017 0.0602 0.039 
Median OS, months NE NE NE NE NE NE NE NE 
HR (95% CI) 0.49 (0.24–1.00) 1.04 (0.62–1.74) 0.80 (0.53–1.20) 0.97 (0.71–1.33) 
P 0.0461 0.8732 0.2805 0.854 
ORRa,% 73.1 47.6 62.0 56.1 66.9 52.2 63.2 49.4 
P 0.0013 0.4701 0.0053 0.0004 
CR ratea,% 34.6 22.6 38.0 36.7 36.5 30.2 25.1 18.2 
P 0.1166 0.8840 0.2199 0.035 
Median TTNT, months 32.6 14.1 22.1 24.2 24.9 17.5 23.0 17.7 
HR (95% CI) 0.44 (0.29–0.66) 1.10 (0.77–1.57) 0.75 (0.57–0.97) 0.80 (0.66–0.97) 
P <0.0001 0.6062 0.0293 0.024 
Biomarker positiveBiomarker negativeAll biomarker evaluableTotal study population
Btz-R (n = 78)R (n = 86)Btz-R (n = 108)R (n = 104)Btz-R (n = 186)R (n = 190)Btz-R (n = 336)R (n = 340)
Median PFS, months 14.2 9.1 11.6 13.7 13.6 11.0 12.8 11.0 
HR (95% CI) 0.47 (0.32–0.68) 1.12 (0.80–1.58) 0.79 (0.62–1.01) 0.82 (0.68–0.99) 
P <0.0001 0.5017 0.0602 0.039 
Median OS, months NE NE NE NE NE NE NE NE 
HR (95% CI) 0.49 (0.24–1.00) 1.04 (0.62–1.74) 0.80 (0.53–1.20) 0.97 (0.71–1.33) 
P 0.0461 0.8732 0.2805 0.854 
ORRa,% 73.1 47.6 62.0 56.1 66.9 52.2 63.2 49.4 
P 0.0013 0.4701 0.0053 0.0004 
CR ratea,% 34.6 22.6 38.0 36.7 36.5 30.2 25.1 18.2 
P 0.1166 0.8840 0.2199 0.035 
Median TTNT, months 32.6 14.1 22.1 24.2 24.9 17.5 23.0 17.7 
HR (95% CI) 0.44 (0.29–0.66) 1.10 (0.77–1.57) 0.75 (0.57–0.97) 0.80 (0.66–0.97) 
P <0.0001 0.6062 0.0293 0.024 

Abbreviations: Btz, bortezomib; NE, not estimable; R, rituximab.

an = 57 for bortezomib–rituximab and n = 60 for rituximab biomarker-positive group; n = 111 and n = 113, respectively, for biomarker-negative group; n = 168 and n = 173 for all biomarker evaluable patients; n = 315 and n = 324 for the total study population.

Validation of these findings with the significant biomarker pair, using LYM-3001 patient discovery and confirmation test sets, is summarized in Supplementary Tables S3 and S4.

Treatment exposure and safety profiles of bortezomib–rituximab and rituximab in the overall safety population and in biomarker-positive and biomarker-negative patients are summarized in Table 4. Among patients treated with bortezomib–rituximab, the proportion receiving all 5 cycles of bortezomib appeared higher, and rates of grade ≥3 adverse events and serious adverse events appeared numerically lower, in biomarker-positive patients versus biomarker-negative patients and the overall safety population. Conversely, among rituximab-treated patients, rates of grade ≥3 adverse events and serious adverse events appeared numerically higher in biomarker-positive patients.

Table 4.

Treatment exposure and safety with bortezomib–rituximab and rituximab in biomarker-positive and biomarker-negative patients in the biomarker-evaluable population (N = 376) and in the overall safety population (N = 673)

Overall safety populationBiomarker positiveBiomarker negative
Btz-R (n = 334)R (n = 339)Btz-R (n = 78)R (n = 86)Btz-R (n = 108)R (n = 104)
Treatment exposure 
 Cycles received, median (range) 5 (1–5) 5 (1–5) 5 (1–5) 5 (1–5) 5 (1–5) 5 (1–5) 
 Patients receiving all 5 cycles, n (%) 232 (69)a 252 (74) 63 (81)a 62 (72) 74 (68.5)a 80 (77) 
 Rituximab relative dosing intensity, mean % 96 97 97 96 90 96 
 Bortezomib relative dosing intensity, mean % 88 – 92 – 90 – 
Safety profile 
 Any AE, n (%) 316 (95) 265 (78) 74 (95) 71 (83) 105 (97) 85 (82) 
 Any treatment-related AE, n (%) 290 (87) 156 (46) 69 (88.5) 48 (56) 93 (86) 39 (37.5) 
 Related to rituximab 206 (62) 156 (46) 45 (58) 48 (56) 69 (64) 39 (37.5) 
 Related to bortezomib 276 (83) – 67 (86) – 88 (81) – 
 Any grade ≥3 AE, n (%) 152 (46) 70 (21) 26 (33) 20 (23) 47 (43.5) 20 (19) 
 Any serious AE, n (%) 59 (18) 37 (11) 9 (11.5) 13 (15) 21 (19) 12 (11.5) 
 AE leading to treatment discontinuation, n (%) 19 (6) 5 (2) 3 (4) 2 (2) 6 (6) 2 (2) 
 Deaths due to AEs, n (%) 6 (2) 2 (<1) 1 (1) 2 (2) 1 (1) 
Overall safety populationBiomarker positiveBiomarker negative
Btz-R (n = 334)R (n = 339)Btz-R (n = 78)R (n = 86)Btz-R (n = 108)R (n = 104)
Treatment exposure 
 Cycles received, median (range) 5 (1–5) 5 (1–5) 5 (1–5) 5 (1–5) 5 (1–5) 5 (1–5) 
 Patients receiving all 5 cycles, n (%) 232 (69)a 252 (74) 63 (81)a 62 (72) 74 (68.5)a 80 (77) 
 Rituximab relative dosing intensity, mean % 96 97 97 96 90 96 
 Bortezomib relative dosing intensity, mean % 88 – 92 – 90 – 
Safety profile 
 Any AE, n (%) 316 (95) 265 (78) 74 (95) 71 (83) 105 (97) 85 (82) 
 Any treatment-related AE, n (%) 290 (87) 156 (46) 69 (88.5) 48 (56) 93 (86) 39 (37.5) 
 Related to rituximab 206 (62) 156 (46) 45 (58) 48 (56) 69 (64) 39 (37.5) 
 Related to bortezomib 276 (83) – 67 (86) – 88 (81) – 
 Any grade ≥3 AE, n (%) 152 (46) 70 (21) 26 (33) 20 (23) 47 (43.5) 20 (19) 
 Any serious AE, n (%) 59 (18) 37 (11) 9 (11.5) 13 (15) 21 (19) 12 (11.5) 
 AE leading to treatment discontinuation, n (%) 19 (6) 5 (2) 3 (4) 2 (2) 6 (6) 2 (2) 
 Deaths due to AEs, n (%) 6 (2) 2 (<1) 1 (1) 2 (2) 1 (1) 

Abbreviation: AE, adverse event; Btz-R, bortezomib–rituximab; R, rituximab.

aData shown for cycles of bortezomib therapy; an additional 12 patients in the overall safety population, and an additional 2 and 4 patients in the biomarker-positive and biomarker-negative populations, respectively, received all 5 cycles of R.

The findings of our analyses of protocol-specified biomarkers in the LYM-3001 study suggest that subgroups of patients with relapsed/refractory follicular lymphoma can be identified that experience significantly longer PFS benefit together with improved OS with addition of bortezomib to rituximab. Our results can be considered robust, being derived from one of the largest prospective randomized studies conducted in this setting. Per protocol, there was mandatory collection of archival tumor samples and whole blood samples, enabling substantial sample collections for these exploratory analyses of prespecified candidate biomarkers. The biomarker subgroup that we identified had a high population frequency (43.6%). There was also a substantially greater efficacy benefit with bortezomib–rituximab versus rituximab in biomarker-positive patients than that seen in the overall, unselected study population (13). These findings indicate the usefulness of such biomarker analyses for identifying specific patient subgroups that benefit from bortezomib–rituximab (33) and the potential of such analyses for optimizing treatment for individual patients with follicular lymphoma.

Using pairwise analyses, we identified one biomarker pair that showed a significant PFS benefit with bortezomib–rituximab versus rituximab in biomarker-positive patients after multiple comparison correction. This pair was tested under different genetic models, and the presence of PSMB1 P11A (G allele) and low CD68 expression (≤50 CD68-positive cells) was associated with a median PFS benefit of 5 months, improved response rates and TTNT, and an OS benefit in patients treated with bortezomib–rituximab versus rituximab. Notably, patients positive for this biomarker pair were representative of the overall LYM-3001 study population (13), with approximately half having high-risk disease and/or poor prognostic features. Furthermore, the safety profile of bortezomib–rituximab in biomarker-positive patients appeared comparable with that in the overall population. It should be noted that this was an exploratory analysis and, in the absence of an independent dataset with which to validate our results with the significant biomarker pair, we split our data from LYM-3001 patients into discovery and confirmation test sets. The data from these analyses supported our findings. Nevertheless, independent validation studies are required for confirmation.

Importantly, the two biomarkers in this pair would be feasible and practical to screen for in the clinical setting, if these findings are confirmed in independent studies. CD68 has established prognostic value in lymphoma and with rituximab-based treatment (21, 22, 34), and readily available assays enable measurement using methodology that is likely reproducible in laboratories with experience in lymphoma diagnosis. In addition, because we used a relatively straightforward genotyping assay for PSMB1 P11A, a validated assay could potentially be developed for clinical use.

There are currently no published data on the functional consequence of the PSMB1 P11A variant. The prognostic significance of variants in PSMB genes and in some proteasome α-subunits (PSMA) may be hypothesized to be associated with reduced cellular levels of functional proteasomes (35). Sequence changes such as a G allele in the PSMB1 P11A leader sequence may interfere with the assembly or function of proteasomes (36) and could translate into greater bortezomib activity. Such greater activity could be related to a reduction of active proteasome sites and a consequent requirement for fewer bortezomib molecules to sufficiently inhibit proteasome function, leading to cell death (37). This hypothesis is supported by previous findings suggesting that low PSMA5 levels are associated with longer PFS with bortezomib in mantle cell lymphoma (18). In addition, recent RNAi screens showed that silencing of individual proteasome genes including PSMA5 and PSMB2/3/7 sensitized multiple myeloma cells to bortezomib (37, 38).

The proteasome also regulates CD20, the target of rituximab; therefore, functional mutations in proteasome subunits such as PSMB1 P11A may influence the activity of rituximab in patients. In the presence of rituximab, there is upregulation of both the ubiquitin and proteasome systems (39). In patients with lower levels of functional proteasomes at baseline, alternative proteolytic systems (e.g., autophagy) may be used for degradation of ubiquitinated CD20 (40). Autophagy is less efficient than degradation through the proteasome (41), and this may translate into less efficacy for patients treated with rituximab when these proteasome subunit anomalies are present. For such patients, as shown in this report, treatment with bortezomib is more effective than in patients without these anomalies; this may be due to direct inhibition of the proteasome, which leads to control of other survival signaling pathways such as NF-κB.

It has been reported previously that patients with high levels of tumor-associated macrophages (TAM) have favorable outcomes when treated with rituximab but poorer outcomes when treated with chemotherapy alone (22, 42). Our findings (Supplementary Fig. S2A) confirm the relatively poorer outcome of patients with low CD68 expression treated with rituximab alone (22). The antitumoral activity of rituximab is dependent on Fc-receptor–mediated interactions with effector cells including neutrophils, natural killer cells, and macrophages. Macrophages can eliminate B-lymphocytes by direct Fc-receptor–mediated phagocytosis (43), or they may secrete cytolytic factors or release cytokines, thereby recruiting other effector cells to amplify the inflammatory response (44, 45), inferring a direct relationship between CD68 TAM content and efficacy of rituximab. Patients with PSMB1 P11A (G allele), which may indicate lower proteasome levels, and low CD68 TAM did much worse on rituximab alone compared with bortezomib–rituximab, presumably due to less elimination of B-lymphocytes by direct Fc-receptor–mediated phagocytosis and slower clearance of ubiquitinated CD20 protein through autophagy. For these patients, addition of bortezomib overcame this inefficiency, presumably by controlling other survival signaling pathways such as NF-κB.

In conclusion, these findings suggest that prespecified biomarker combinations can identify follicular lymphoma patient subgroups deriving substantial clinical benefit from bortezomib–rituximab versus rituximab, without a notable impact on safety. Further confirmation in independent cohorts of similar patients, and in patients treated with other bortezomib–rituximab–based combination regimens showing activity in relapsed/refractory follicular lymphoma (9, 11, 12), would be warranted.

W. Li is employed as a senior scientist in Janssen Research & Development. J. Karkera is employed as a prinicpal reasearch scientist in Janssen Research & Development. R. Favis is an employee of Johnson & Johnson. Y. Elsayed is employed as a vice president, hemeatological malignancy, and has ownership interest (including patents) in Johnson & Johnson. H. van de Velde is employed as a senior director oncology, Research & Development, in Janssen and has ownership interest (including patents) in Johnson & Johnson. M. Schaffer has ownership interest (including patents) in Johnson & Johnson. J. Mayer has commercial research grant from Janssen. S.A.J. Rule has commercial research support for a Cancer Research UK randomized trial and is a consultant/advisory board member of Johnson & Johnson. S. de Vos is a consultant/advisory board member of Millennium. O. Shpilberg has a commercial research grant from Janssen. A. Cakana is an employee of Janssen Pharmaceuticals. D.-L. Esseltine is employed as a vice president in Millennium Pharmaceuticals Inc. G. Mulligan is employed as a director in Millennium Pharmaceuticals. D. Ricci is employed as a director, Biomarker, in Janssen Pharmaceuticals. No potential conflicts of interest were disclosed by the other authors.

Conception and design: B. Coiffier, W. Li, E.D. Henitz, P. Theocharous, Y. Elsayed, G. Mulligan, D. Ricci

Development of methodology: W. Li, J. Karkera, R. Favis, D. Gaffney, A. Shapiro, Y. Elsayed, A. Cakana, G. Mulligan, D. Ricci

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): E.D. Henitz, J. Karkera, R. Favis, D. Gaffney, A. Shapiro, P. Theocharous, Y. Elsayed, X.-N. Hong, A. Scheliga, J. Mayer, F. Offner, S.A.J. Rule, A. Teixeira, J. Romejko-Jarosinska, S. de Vos, M. Crump, O. Shpilberg, P.L. Zinzani, A. Cakana, D. Ricci

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): B. Coiffier, W. Li, E.D. Henitz, A. Shapiro, P. Theocharous, Y. Elsayed, H. van de Velde, M. Schaffer, A. Scheliga, P.L. Zinzani, A. Cakana, D.-L. Esseltine, G. Mulligan, D. Ricci

Writing, review, and/or revision of the manuscript: B. Coiffier, W. Li, E.D. Henitz, J. Karkera, R. Favis, P. Theocharous, Y. Elsayed, H. van de Velde, M. Schaffer, E.A. Osmanov, A. Scheliga, J. Mayer, F. Offner, S.A.J. Rule, J. Romejko-Jarosinska, S. de Vos, M. Crump, O. Shpilberg, P.L. Zinzani, A. Cakana, D.-L. Esseltine, G. Mulligan, D. Ricci

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): W. Li, E.D. Henitz, A. Shapiro, D. Ricci

Study supervision: B. Coiffier, W. Li, P. Theocharous, Y. Elsayed, H. van de Velde, F. Offner, D. Ricci

The authors thank the patients for their participation in this study, Dr Eugene Zhu at Janssen R&D for provision of adverse event and drug exposure data, and Steve Hill of FireKite for writing assistance in the development of this manuscript, which was funded by Millennium Pharmaceuticals, Inc., and Janssen Global Services.

This work was funded by Janssen Research and Development, L.L.C., and Millennium Pharmaceuticals, Inc.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Rummel
M
. 
Reassessing the standard of care in indolent lymphoma: a clinical update to improve clinical practice
.
J Natl Compr Canc Netw
2010
;
8
Suppl 6
:
S1
14
.
2.
The Non-Hodgkin's Lymphoma Classification Project
. 
A clinical evaluation of the International Lymphoma Study Group classification of non-Hodgkin's lymphoma
.
Blood
1997
;
89
:
3909
18
.
3.
Buske
C
,
Hoster
E
,
Dreyling
M
,
Hasford
J
,
Unterhalt
M
,
Hiddemann
W
. 
The Follicular Lymphoma International Prognostic Index (FLIPI) separates high-risk from intermediate- or low-risk patients with advanced-stage follicular lymphoma treated front-line with rituximab and the combination of cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) with respect to treatment outcome
.
Blood
2006
;
108
:
1504
8
.
4.
Johnson
PW
,
Rohatiner
AZ
,
Whelan
JS
,
Price
CG
,
Love
S
,
Lim
J
, et al
Patterns of survival in patients with recurrent follicular lymphoma: a 20-year study from a single center
.
J Clin Oncol
1995
;
13
:
140
7
.
5.
Dave
SS
,
Wright
G
,
Tan
B
,
Rosenwald
A
,
Gascoyne
RD
,
Chan
WC
, et al
Prediction of survival in follicular lymphoma based on molecular features of tumor-infiltrating immune cells
.
N Engl J Med
2004
;
351
:
2159
69
.
6.
Di Bella
N
,
Taetle
R
,
Kolibaba
K
,
Boyd
T
,
Raju
R
,
Barrera
D
, et al
Results of a phase 2 study of bortezomib in patients with relapsed or refractory indolent lymphoma
.
Blood
2010
;
115
:
475
80
.
7.
O'Connor
OA
,
Portlock
C
,
Moskowitz
C
,
Hamlin
P
,
Straus
D
,
Gerecitano
J
, et al
Time to treatment response in patients with follicular lymphoma treated with bortezomib is longer compared with other histologic subtypes
.
Clin Cancer Res
2010
;
16
:
719
26
.
8.
Ribrag
V
,
Tilly
H
,
Casasnovas
O
,
Bosly
A
,
Bouabdullah
R
,
Delarue
R
, et al
Final results of a randomized phase 2 multicenter study of two bortezomib schedules in patients with recurrent or refractory follicular lymphoma. Groupe d'Etude Des Lymphomes De l'Adulte (GELA) study FL-05 [abstract nr 768]. Blood
2010
;
116
:
338a
.
9.
Fowler
N
,
Kahl
BS
,
Lee
P
,
Matous
JV
,
Cashen
AF
,
Jacobs
SA
, et al
Bortezomib, bendamustine, and rituximab in patients with relapsed or refractory follicular lymphoma: the phase II VERTICAL study
.
J Clin Oncol
2011
;
29
:
3389
95
.
10.
De Vos
S
,
Goy
A
,
Dakhil
SR
,
Saleh
MN
,
McLaughlin
P
,
Belt
R
, et al
Multicenter randomized phase II study of weekly or twice-weekly bortezomib plus rituximab in patients with relapsed or refractory follicular or marginal-zone B-cell lymphoma
.
J Clin Oncol
2009
;
27
:
5023
30
.
11.
Craig
M
,
Hanna
WT
,
Cabanillas
F
,
Chen
C-S
,
Parasuraman
S
,
Neuwirth
R
, et al
Bortezomib in combination with rituximab, cyclophosphamide, and prednisone with or without doxorubicin followed by rituximab maintenance in patients with relapsed or refractory follicular lymphoma: results of a phase 2 study. [abstract nr 2798].
Blood
2010
;
116
:
1153a
4a
.
12.
Friedberg
JW
,
Vose
JM
,
Kelly
JL
,
Young
F
,
Bernstein
SH
,
Peterson
D
, et al
The combination of bendamustine, bortezomib, and rituximab for patients with relapsed/refractory indolent and mantle cell non-Hodgkin lymphoma
.
Blood
2011
;
117
:
2807
12
.
13.
Coiffier
B
,
Osmanov
EA
,
Hong
X
,
Scheliga
A
,
Mayer
J
,
Offner
F
, et al
Bortezomib plus rituximab versus rituximab alone in patients with relapsed, rituximab-naive or rituximab-sensitive, follicular lymphoma: a randomised phase 3 trial
.
Lancet Oncol
2011
;
12
:
773
84
.
14.
Solal-Celigny
P
,
Roy
P
,
Colombat
P
,
White
J
,
Armitage
JO
,
Arranz-Saez
R
, et al
Follicular lymphoma international prognostic index
.
Blood
2004
;
104
:
1258
65
.
15.
Cheson
BD
,
Horning
SJ
,
Coiffier
B
,
Shipp
MA
,
Fisher
RI
,
Connors
JM
, et al
Report of an international workshop to standardize response criteria for non-Hodgkin's lymphomas. NCI Sponsored International Working Group.
J Clin Oncol
1999
;
17
:
1244
53
.
16.
Sunwoo
JB
,
Chen
Z
,
Dong
G
,
Yeh
N
,
Crowl
BC
,
Sausville
E
, et al
Novel proteasome inhibitor PS-341 inhibits activation of nuclear factor-kappa B, cell survival, tumor growth, and angiogenesis in squamous cell carcinoma
.
Clin Cancer Res
2001
;
7
:
1419
28
.
17.
Yin
D
,
Zhou
H
,
Kumagai
T
,
Liu
G
,
Ong
JM
,
Black
KL
, et al
Proteasome inhibitor PS-341 causes cell growth arrest and apoptosis in human glioblastoma multiforme (GBM)
.
Oncogene
2005
;
24
:
344
54
.
18.
Goy
A
,
Bernstein
SH
,
McDonald
A
,
Pickard
MD
,
Shi
H
,
Fleming
MD
, et al
Potential biomarkers of bortezomib activity in mantle cell lymphoma from the phase 2 PINNACLE trial
.
Leuk Lymphoma
2010
;
51
:
1269
77
.
19.
Keats
JJ
,
Fonseca
R
,
Chesi
M
,
Schop
R
,
Baker
A
,
Chng
WJ
, et al
Promiscuous mutations activate the noncanonical NF-kappaB pathway in multiple myeloma
.
Cancer Cell
2007
;
12
:
131
44
.
20.
Mulligan
G
,
Mitsiades
C
,
Bryant
B
,
Zhan
F
,
Chng
WJ
,
Roels
S
, et al
Gene expression profiling and correlation with outcome in clinical trials of the proteasome inhibitor bortezomib
.
Blood
2007
;
109
:
3177
88
.
21.
Kelley
T
,
Beck
R
,
Absi
A
,
Jin
T
,
Pohlman
B
,
Hsi
E
. 
Biologic predictors in follicular lymphoma: importance of markers of immune response
.
Leuk Lymphoma
2007
;
48
:
2403
11
.
22.
Taskinen
M
,
Karjalainen-Lindsberg
ML
,
Nyman
H
,
Eerola
LM
,
Leppa
S
. 
A high tumor-associated macrophage content predicts favorable outcome in follicular lymphoma patients treated with rituximab and cyclophosphamide-doxorubicin-vincristine-prednisone
.
Clin Cancer Res
2007
;
13
:
5784
9
.
23.
Kraus
M
,
Ruckrich
T
,
Reich
M
,
Gogel
J
,
Beck
A
,
Kammer
W
, et al
Activity patterns of proteasome subunits reflect bortezomib sensitivity of hematologic malignancies and are variable in primary human leukemia cells
.
Leukemia
2007
;
21
:
84
92
.
24.
Lichter
DI
,
Danaee
H
,
Pickard
MD
,
Tayber
O
,
Sintchak
M
,
Shi
H
, et al
Sequence analysis of β-subunit genes of the 20S proteasome in patients with relapsed multiple myeloma treated with bortezomib or dexamethasone
.
Blood
2012
;
120
:
4513
6
.
25.
Oerlemans
R
,
Franke
NE
,
Assaraf
YG
,
Cloos
J
,
van Zantwijk
I
,
Berkers
CR
, et al
Molecular basis of bortezomib resistance: proteasome subunit beta5 (PSMB5) gene mutation and overexpression of PSMB5 protein
.
Blood
2008
;
112
:
2489
99
.
26.
Wang
SS
,
Cerhan
JR
,
Hartge
P
,
Davis
S
,
Cozen
W
,
Severson
RK
, et al
Common genetic variants in proinflammatory and other immunoregulatory genes and risk for non-Hodgkin lymphoma
.
Cancer Res
2006
;
66
:
9771
80
.
27.
Hosgood
HD
 III
,
Purdue
MP
,
Wang
SS
,
Zheng
T
,
Morton
LM
,
Lan
Q
, et al
A pooled analysis of three studies evaluating genetic variation in innate immunity genes and non-Hodgkin lymphoma risk
.
Br J Haematol
2011
;
152
:
721
6
.
28.
Binstadt
BA
,
Geha
RS
,
Bonilla
FA
. 
IgG Fc receptor polymorphisms in human disease: implications for intravenous immunoglobulin therapy
.
J Allergy Clin Immunol
2003
;
111
:
697
703
.
29.
Wang
H
,
Liu
X
,
Xu
B
. 
[Proteasome inhibitor induces apoptosis and influences the expression of Notch1 and NF-kappaB in multiple myeloma RPMI8226 cells]
.
Zhongguo Shi Yan Xue Ye Xue Za Zhi
2008
;
16
:
531
7
.
30.
Cartron
G
,
Dacheux
L
,
Salles
G
,
Solal-Celigny
P
,
Bardos
P
,
Colombat
P
, et al
Therapeutic activity of humanized anti-CD20 monoclonal antibody and polymorphism in IgG Fc receptor FcgammaRIIIa gene
.
Blood
2002
;
99
:
754
8
.
31.
Kim
DH
,
Jung
HD
,
Kim
JG
,
Lee
JJ
,
Yang
DH
,
Park
YH
, et al
FCGR3A gene polymorphisms may correlate with response to frontline R-CHOP therapy for diffuse large B-cell lymphoma
.
Blood
2006
;
108
:
2720
5
.
32.
Benjamini
Y
,
Hochberg
Y
. 
Controlling the false discovery rate: a practical and powerful approach to multiple testing
.
J Roy Stat Soc Ser B (Method)
1995
;
57
:
289
300
.
33.
Salles
G
. 
Is there a role for bortezomib combinations in the management of patients with follicular lymphoma?
J Clin Oncol
2011
;
29
:
3349
50
.
34.
Wahlin
BE
,
Aggarwal
M
,
Montes-Moreno
S
,
Gonzalez
LF
,
Roncador
G
,
Sanchez-Verde
L
, et al
A unifying microenvironment model in follicular lymphoma: outcome is predicted by programmed death-1–positive, regulatory, cytotoxic, and helper T cells and macrophages
.
Clin Cancer Res
2010
;
16
:
637
50
.
35.
Schmidt
M
,
Zantopf
D
,
Kraft
R
,
Kostka
S
,
Preissner
R
,
Kloetzel
PM
. 
Sequence information within proteasomal prosequences mediates efficient integration of beta-subunits into the 20 S proteasome complex
.
J Mol Biol
1999
;
288
:
117
28
.
36.
Murata
S
,
Yashiroda
H
,
Tanaka
K
. 
Molecular mechanisms of proteasome assembly
.
Nat Rev Mol Cell Biol
2009
;
10
:
104
15
.
37.
Chen
S
,
Blank
JL
,
Peters
T
,
Liu
XJ
,
Rappoli
DM
,
Pickard
MD
, et al
Genome-wide siRNA screen for modulators of cell death induced by proteasome inhibitor bortezomib
.
Cancer Res
2010
;
70
:
4318
26
.
38.
Zhu
YX
,
Tiedemann
R
,
Shi
CX
,
Yin
H
,
Schmidt
JE
,
Bruins
LA
, et al
RNAi screen of the druggable genome identifies modulators of proteasome inhibitor sensitivity in myeloma including CDK5
.
Blood
2011
;
117
:
3847
57
.
39.
Czuczman
MS
,
Olejniczak
S
,
Gowda
A
,
Kotowski
A
,
Binder
A
,
Kaur
H
, et al
Acquirement of rituximab resistance in lymphoma cell lines is associated with both global CD20 gene and protein down-regulation regulated at the pretranscriptional and posttranscriptional levels
.
Clin Cancer Res
2008
;
14
:
1561
70
.
40.
Pandey
UB
,
Nie
Z
,
Batlevi
Y
,
McCray
BA
,
Ritson
GP
,
Nedelsky
NB
, et al
HDAC6 rescues neurodegeneration and provides an essential link between autophagy and the UPS
.
Nature
2007
;
447
:
859
63
.
41.
Kraft
C
,
Peter
M
,
Hofmann
K
. 
Selective autophagy: ubiquitin-mediated recognition and beyond
.
Nat Cell Biol
2010
;
12
:
836
41
.
42.
Canioni
D
,
Salles
G
,
Mounier
N
,
Brousse
N
,
Keuppens
M
,
Morchhauser
F
, et al
High numbers of tumor-associated macrophages have an adverse prognostic value that can be circumvented by rituximab in patients with follicular lymphoma enrolled onto the GELA-GOELAMS FL-2000 trial
.
J Clin Oncol
2008
;
26
:
440
6
.
43.
Lefebvre
ML
,
Krause
SW
,
Salcedo
M
,
Nardin
A
. 
Ex vivo-activated human macrophages kill chronic lymphocytic leukemia cells in the presence of rituximab: mechanism of antibody-dependent cellular cytotoxicity and impact of human serum
.
J Immunother
2006
;
29
:
388
97
.
44.
Gong
Q
,
Ou
Q
,
Ye
S
,
Lee
WP
,
Cornelius
J
,
Diehl
L
, et al
Importance of cellular microenvironment and circulatory dynamics in B cell immunotherapy
.
J Immunol
2005
;
174
:
817
26
.
45.
Uchida
J
,
Hamaguchi
Y
,
Oliver
JA
,
Ravetch
JV
,
Poe
JC
,
Haas
KM
, et al
The innate mononuclear phagocyte network depletes B-lymphocytes through Fc receptor-dependent mechanisms during anti-CD20 antibody immunotherapy
.
J Exp Med
2004
;
199
:
1659
69
.

Supplementary data