Promising activity of nelfinavir-bortezomib-dexamethasone in proteasome inhibitor–refractory multiple myeloma

TO THE EDITOR:

Treatment of proteasome inhibitor (PI)–refractory multiple myeloma (MM) is challenging, with response rates of only 15% to 35% being achieved with next-generation drugs such as pomalidomide, carfilzomib, and daratumumab, alone or in combination.^1-3 

The oral HIV protease inhibitor nelfinavir has multiple activities against malignant human cells, including phosphorylated AKT inhibition and radiosensitization.⁴  It has single-agent activity against non–small-cell lung cancer,⁵  MM,⁴  and breast cancer.⁶  Combination of nelfinavir with bortezomib or carfilzomib overcomes resistance of MM cells to these PIs in vitro, via induction of the endoplasmic reticulum–resident endonuclease IRE1 and its cytosolic target X-box protein 1 (XBP1), which control the unfolded protein response, a key modulator of protein production, folding, and destruction^7,8  that is downregulated in PI-refractory MM.^9,10  In the phase 1 Swiss Group for Clinical Cancer Research (SAKK) 65/08 dose-escalation trial, a nelfinavir-bortezomib-dexamethasone combination (NeVd) was active in 5 of 6 prospectively selected patients with heavily pretreated, bortezomib-lenalidomide double-refractory MM.¹¹ 

We conducted a multicenter, open-label phase 2 trial (SAKK 39/13) to evaluate the activity and safety of NeVd in patients refractory to their most recent PI-containing regimen (International Myeloma Working Group [IMWG] criteria¹² ) and previously exposed or intolerant to ≥1 immunomodulatory drug (thalidomide, lenalidomide, pomalidomide). Study treatment consisted of nelfinavir 2500 mg on days 1 to 14 twice daily; bortezomib 1.3 mg/m² IV/subcutaneously on days 1, 4, 8, and 11; and dexamethasone 20 mg orally on days 1 to 2, 4 to 5, 8 to 9, and 11 to 12 for up to 6 21-day cycles. No antibiotic or antithrombotic prophylaxis was mandated. Objective response was assessed at the start of each cycle and confirmed by a central committee, using IMWG criteria.¹²  All patients provided written informed consent. The trial (registered at www.clinicaltrials.gov as #NTC02188537) was conducted according to applicable national and international regulations and approved by regulatory authorities.

Between December 2014 and April 2016, 34 patients were treated at 9 Swiss centers. All had received prior lenalidomide, 16 (47%) had received prior pomalidomide, and 26 (76%) had received prior high-dose chemotherapy (median, 5 prior treatment lines per 2 prior PI-containing lines). All were bortezomib refractory (79% PI-lenalidomide double refractory; 44% PI-pomalidomide double refractory; 38% PI-lenalidomide-pomalidomide triple refractory; Table 1). They received a median of 4.5 cycles (range, 1-6 cycles) of study treatment. The objective response rate (ORR) was 65% (90% confidence interval [CI], 49%-76%; 17 partial responses [PRs] and 5 very good PRs; Table 2). ORRs >60% were observed in those with poor-risk cytogenetic features (77%) and in lenalidomide-bortezomib double-refractory (70%) as well as pomalidomide-lenalidomide-bortezomib triple-refractory patients (62%; Table 2). PR was achieved after a median of 2 therapy cycles. Four patients (12%) achieved stable disease (2 lasting ≥2 cycles), and 3 patients (9%) showed progressive disease after cycle 1. Overall, 25 (74%) of 34 patients experienced minimal response or better (25%-49% decrease in serum paraprotein).

Median progression-free survival (PFS; from start of trial therapy to disease progression/death resulting from any cause) was 12 weeks (95% CI, 9-21) overall and 16 weeks for patients reaching ≥PR (95% CI, 12-30). During follow-up, 27 of 29 patients received further antimyeloma treatment. Five patients received ≤8 additional cycles of NeVd on a compassionate-use basis. Seventeen patients had died as of November 2016, resulting in a median overall survival of 12 months (95% CI, 6 to not available; supplemental Figure 1).

We did not observe any unexpected safety signals for NeVd. The most common adverse events were anemia (97% of patients), thrombocytopenia (82%), hypertension (53%), diarrhea (47%), fatigue (38%), and dyspnea (35%; supplemental Table 1, available on the Blood Web site). There were 4 deaths (3 resulting from septicemia and 1 from heart failure) during trial treatment; 3 of these were associated with underlying pneumonia. Although this mortality rate is consistent with the background mortality among patients with heavily pretreated, refractory myeloma,¹³  these findings suggest that prophylactic antibiotic therapy should be considered in those with low neutrophil counts and/or advanced age undergoing NeVd treatment.

The 65% ORR achieved with NeVd in our study is similar to that observed with first-line bortezomib-dexamethasone in bortezomib-naïve patients.¹⁴  To our knowledge, a similar ORR has not been reported in a PI-refractory myeloma population with comparable prior drug exposure. Our trial cohort was very heavily pretreated, with no upper age limit or minimum granulocyte counts, and the minimally required blood counts were allowed to be reached by transfusion. Moreover, it seems likely that clinical benefit was underestimated in responding patients, because the maximum number of study cycles was limited to 6 (4.2 months) as a result of cost constraints (no external funding support). The time course of paraprotein levels also suggests that individual patients might potentially have experienced myeloma control if NeVd had been continued until progression. The median PFS of 12 weeks is similar to the 4 months achieved with pomalidomide-dexamethasone in pomalidomide-naïve, bortezomib-lenalidomide double-refractory patients in the MM-002 phase 2 study.¹⁵ 

Importantly, the ORR in those patients whose myeloma met IMWG criteria for bortezomib-refractory disease in their most recent treatment line before study inclusion was 63%. Because the median time between last prior treatment and study inclusion was <1 month (27 days; range, 3-402 days), this precludes the likelihood that emergence of bortezomib-sensitive myeloma subpopulations by clonal tiding over several lines of therapy might have contributed to the high ORR.

Among 462 bortezomib-lenalidomide double-refractory patients, only 12% of patients achieved ≥PR to bortezomib-containing therapy and 35% to carfilzomib- or pomalidomide-containing therapy,¹³  with a median treatment duration of 3 months. Pomalidomide-dexamethasone–based therapies are standard for heavily pretreated, PI-refractory patients and have resulted in an ORR of 32% and PFS of 5 months. Response rates were up to 50% to 65% with pomalidomide-based triplet combination therapies containing additional cyclophosphamide, daratumumab, or carfilzomib in study populations from which pomalidomide-refractory patients were excluded.^2,16-18  Daratumumab-dexamethasone with an ORR of 29% currently represents the only approved treatment with proven activity in triple-refractory myeloma patients.¹⁹  Of note, the ORR of NeVd in our trial was 62% in bortezomib-lenalidomide-pomalidomide triple-refractory MM patients.

Although the specific molecular target of nelfinavir remains to be identified, its safety and long-term toxicities are well established after decades of experience. Orphan drug designation as potential MM treatment was granted by Swissmedic in 2016 and by the US Food and Drug Administration in 2017. Both nelfinavir and bortezomib are entering their generic era, with monthly treatment costs only 10% to 15% of the prices of other therapies active in double-refractory myeloma. The unprecedented ORR of NeVd observed in this heavily pretreated, multirefractory setting warrants further investigation to explore the potential of nelfinavir in combination with PIs, not only in relapsed/refractory myeloma but also in other lines of therapy and other cancer types.