Bortezomib is a proteasome inhibitor with pleotropic antitumor effects, including putative activity targeting NF-κB. Elevated levels of NF-κB DNA binding have been documented in HL and NF-κB activity is thought to mediate resistance to standard therapeutic approaches. Despite this rationale, and significant activity in non-Hodgkin lymphoma (NHL), the experience to date with bortezomib in HL has been disappointing. Studies in solid tumor lines have suggested that bortezomib markedly enhances cytotoxicity of gemcitabine, a nucleoside analog with significant activity against HL. Thus, we combined bortezomib with gemcitabine in a prospective trial for patients (pts) with relapsed HL. Eligible pts had confirmed relapsed classical HL and normal organ function. Based upon phase I data in solid tumors (
Cancer 107:2482
), bortezomib (1mg/m2 days 1, 4, 8, 11) was administered prior to gemcitabine (800 mg/m2 days 1 and 8) on a 21 day cycle. Response was assessed following two cycles of therapy; patients could then proceed to stem cell collection and ASCT or continue with gemcitabine and bortezomib for up to 6 cycles. 18 pts (10 female) were enrolled. Median age was 36 (range 19–62). 10 pts had received more than one prior chemotherapy regimen. 4 patients had toxicity requiring gemcitabine to be held during the first two cycles of therapy: 1 for thrombocytopenia, and 3 for grade 3 transaminase elevation. 1 additional pt had grade II transaminase elevation; LFTs returned to baseline after cessation of therapy. Other grade 3 toxicities included neutropenia (n=4), thrombocytopenia (n=2), and 5 single events likely unrelated to therapy. 4 pts experienced transient rash. Therapy administered was 5 cycles (n=1); all other pts 2 cycles. 4 pts (22%; 95% CI 3–42%) responded (1CR; 3PR) after 2 cycles of therapy. 7 pts had PET/CT imaging performed before and after therapy; in 4, standardized uptake values (SUV) following treatment were all ≤ 3.0 and 1 additional pt had significantly decreased SUV following treatment suggesting metabolic response. All 12 pts planned for ASCT eventually received it after salvage; 1 pt required harvest due to inadequate mobilization. Biochemical assays were performed to measure activity of the 20S proteasome pre/post treatment on day 1 and day 8 in all patients. Following the first dose of bortezomib, 16 pts demonstrated significant proteasome inhibition in whole blood compared to baseline (median 50%, range −39% to 77%). Proteasome inhibition on day 8 before and after bortezomib was less, with a median inhibition of 36%. Additionally, EMSA analysis for NF-κB activity in peripheral blood mononuclear cells was performed before and after treatment; there was no correlation between results from this assay and clinical response. In this study, there was no obvious improvement in clinical outcome with the gemcitabine/bortezomib combination over what would be expected with gemcitabine alone (39% RR; JCO 18:2615
). Moreover, unexpected, severe but reversible transaminase elevations requiring modification of prescribed therapy were noted in 16% of pts. Despite this lack of clinical response, we demonstrated proteasome inhibition occurred in nearly all patients following bortezomib, and there was no correlation between degree of proteasome inhibition in blood and clinical response to therapy. We conclude the combination of gemcitabine and bortezomib should not be pursued in HL, and caution should be exercised when combining these agents in other diseases, including NHL, due to potential for severe liver toxicity.
