Reduced Occurrence of Tumor Flare with Flavopiridol Followed by Combined Flavopiridol and Lenalidomidein Patients with Relapsed Chronic Lymphocytic Leukemia (CLL)

Abstract

Flavopiridol and lenalidomide have activity in refractory CLL without immunosuppression or opportunistic infections seen with other therapies. We hypothesized that flavopiridol treatment could adequately de-bulk disease prior to lenalidomide therapy, decreasing the incidence of tumor flare with higher doses of lenalidomide. In this Phase I study, the maximum tolerated dose was not reached with treatment consisting of flavopiridol 30 mg/m² intravenous bolus (IVB) + 30 mg/m² continuous intravenous infusion (CIVI) cycle (C) 1 day (D) 1 and 30 mg/m² IVB + 50 mg/m² CIVI C1 D8,15 and C2-8 D3,10,17 with lenalidomide 15 mg orally daily C2-8 D1-21. There was no unexpected toxicity seen, including no increased tumor lysis, tumor flare (even at higher doses of lenalidomide) or opportunistic infection. Significant clinical activity was demonstrated, with a 51% response rate in this group of heavily pretreated patients. Biomarker testing confirmed association of mitochondrial priming of the BH3 only peptide Puma with response.

Introduction

Chronic Lymphocytic Leukemia (CLL) is the most common adult leukemia, characterized by leukemic cells accumulating secondary to survival and proliferation signals delivered to the cells. CLL has been increasingly recognized as a heterogeneous disease, with many patients being observed for several years after diagnosis prior to disease progression requiring therapy. Certain disease characteristics such as high-risk interphase cytogenetic abnormalities (del17p13 and del11q22) and unmutated immunoglobulin heavy chain-variable status are associated with a more rapid progression and inferior survival [1]. Despite the improvement in overall response rates with the current standard treatment using combination chemoimmunotherapy, patients with the mentioned high-risk characteristics continue to have shorter time to progression with conventional therapies and lower response rates with repeated therapy.

CLL cell survival has been attributed to a role of proliferation and also disordered apoptosis, with constitutive expression of different anti-apoptotic proteins including BCl2, MCL1 and XIAP. Contributions of the microenvironment and aberrations in innate immune response are recognized as CLL survival mechanisms [2]. Significant survival signals are derived from both direct contact of CLL cells with stromal cells and cytokines, making development of novel therapeutics targeting the molecular alterations preventing apoptosis and the immunoregulatory defects interesting therapeutic targets.

Two promising agents that either down-regulate MCL-1 or favorably enhance immune effector cell function include lenalidomide and flavopiridol [3–4]. Both drugs have shown single agent activity in phase I and II trials in this population of patients [3–6].

Lenalidomide is an oral immunomodulatory drug approved for the treatment of 5q-myelodysplastic syndrome [7–8], multiple myeloma [9], and mantle cell lymphoma [10] with activity in other hematologic malignancies including CLL. Lenalidomide does not promote direct cytotoxicity to CLL cells but does alter the microenvironment and immune regulation. Lenalidomide promotes cellular and innate immune activation, activates T-cells and natural killer cells altering the tumor microenvironment [11–13] and down-regulates critical pro-survival cytokines including vascular endothelial growth factor (VEGF), tumor necrosis factor α (TNF-α), interleukin 8 (IL-8) and interleukin 6 (IL-6) [14–15].

The optimal dose and schedule of lenalidomide in relapsed/refractory CLL patients is controversial due to the risks of tumor flare reaction and tumor lysis syndrome (TLS). Two phase II trials with single agent lenalidomide administered at doses of 10–25 mg daily in relapsed/refractory CLL reported response rates ranging from 32–47% with clinical activity in patients with high-risk cytogenetics and/or bulky disease [4–5]. Tumor flare was reported in 30–58% of these patients and two cases of TLS were reported at the 25 mg daily dose [16–17]. A phase 2/3 study was planned to address the question of optimal dosing comparing 10 mg versus 25 mg orally daily for 21 of 28 days; however, after four patients experienced sever tumor flare and /or TLS, the study was amended to a phase 1 trial with intra-patient-dose escalation with lenalidomide given at 2.5 mg initially, increasing to 5 mg after 28 days with subsequent 5 mg incremental increases to 10 mg, 15 mg, or 20 mg. In that study, one-third of the patients could not escalate beyond 2.5 mg, and the response rate was only 11.5%, much lower than previously described [6].

Flavopiridol is a broad cyclin dependent kinase (CDK) inhibitor, that directly antagonizes cell cycle CDK’s (CDK 1 and 2) [18–19] and those involved in transcription initiation (CDK 7 and 8) and elongation (CDK9) [20–23]. Flavopiridol induced apoptosis in CLL cell lines and primary human CLL cells, which was independent of p53 [24]. Flavopiridol also globally inhibits gene transcription, potentially leading to down-regulation of several important intracellular proteins in CLL including Mcl-1 and XIAP [23–25].

Despite promising pre-clinical data, initial clinical trials of flavopiridol using a 72-hour continuous infusion, 24-hour continuous infusion and 1-hour bolus infusion were disappointing with limited clinical efficacy, likely due to plasma protein binding and sub-optimal exposure of free drug [26–28]. Alterations in the dose scheduling to a 30-minute bolus (IVB) followed by 4-hour continuous infusion (CIV) weekly improved efficacy in a Phase I trial with 42 patients enrolled in 3 separate cohorts [3, 29]. Nineteen (45%) patients achieved a partial response to therapy with responses noted in the high risk patients, including 51% of patients with bulky lymphadenopathy, 42% of those with del(17p13.1) and 72% of those with del(11q22.3). The dose limiting toxicity was acute TLS requiring dialysis, occurring in 14% of patients, but 63% of those with WBC > 200×10⁹/L. A Phase II trial followed [30], restricting eligibility to patients with < 200×10⁹/L and using aggressive TLS prophylaxis with hydration, rasburicase, IV dexamethasone, frequent monitoring for and treatment of hyperkalemia. A total of 64 patients were enrolled with only 4 patients not being treated with dose escalation for severe TLS, 3 of those requiring hemodialysis. Thirty-four (53%) patients achieved a response to therapy with 12 (57%) of patients with del(17p13.1) and 14(50%) of patients with del(11q22.3) responding. Subsequent data analysis from these earlier trials evaluating flavopiridol in relapsed or refractory CLL patients identified female sex, bulky lymphadenopathy, high WBC, β-2-microglobulin, and high exposure to flavopiridol-glucuronide as risk factors for TLS [30–31]. Flavopiridol plasma exposure was only weakly associated with response [2, 32]. It is unclear if Mcl-1 down-regulation is solely responsible for the rapid tumor cell death observed in a subset of CLL patients or if other factors are important, such as priming for early mitochondrial damage [33–34].

Given the promising efficacy of both of these agents in high risk and refractory patients coupled with the fact that neither agent was immunosuppressive nor associated with opportunistic infections, we elected to perform a phase I trial evaluating the potential synergy of direct flavopiridol-induced CLL cell cytotoxicity with indirect lenalidomide immunomodulatory effects. Based on the apparent opposite mechanisms for severe toxicities (i.e. tumor flare for lenalidomide vs. TLS for flavopirodol), we hypothesized dosing with single agent flavopiridol during cycle 1 with the established phase II dose would sufficiently de-bulk patients thus reducing risk for tumor flare when combined with escalating doses of lenalidomide in later cycles. Finally, we retrospectively evaluated several factors as potential biomarkers for outcomes from this combined therapy, including mitochondrial profiling of BH3-domain proteins, a functional assay surrogate for cellular ability to respond to pro-apoptotic cues [35–38].

Methods

All patients were enrolled and treated on the National Cancer Institute (NCI)-sponsored (NCI 8046, U01CA076576, NCT00735930) and The Ohio State University Medical Center (OSU) institutional review board-approved phase I trial of flavopiridol in combination with lenalidomide in previously treated CLL. Written informed consent approved by The Ohio State University Human Studies committee was obtained for all patients prior to study entry.

Eligibility Criteria

The study enrolled adult patients (≥ 18 years of age) with relapsed and refractory CLL that required therapy by the IWCLL 2008 criteria [39]. Other eligibility included: at least one prior therapy that included fludarabine unless a contra-indication existed; Eastern Cooperative Oncology Group (ECOG) performance status ≤ 2; normal organ and marrow function; no active infection.

Study Design, Treatment Plan and Dose Escalation Schema

The study was conducted to determine the maximum tolerated dose (MTD) of the combination of flavopiridol and lenalidomide in this population of patients. Due to potential risks of coincident tumor flare and TLS from flavopiridol and lenalidomide administered concomitantly, flavopiridol was administered alone during cycle 1 to cytoreduce patients prior to combination therapy with the two agents during cycles 2–8 (Table 1). Combination therapy and dose escalation was started with the initiation of cycle 2 (Day 29). A standard 3+3 phase I design was used with 3–6 patients enrolled at each dose level with dose escalation starting during cycle 2 (Table 2). Patients were required to complete 2 cycles of therapy with < 2 patients experiencing dose limiting toxicity (DLT) prior to dose escalation to the next dose level. DLT was defined during cycle 2 of protocol treatment (Supplemental Table 1).

Table 1.

Treatment Plan

Agent	Dose	Route	Day	ReRx
Flavopiridol (C1 only)	30 mg/m2 IV bolus + 30 mg/m2 IV continuous infusion D1. 30 mg/m2 IV bolus+ 50 mg/m2 IV continuous infusion D8, 15	30 mg/m2 IV bolus over 30 minutes, followed by a 4-hour 30 mg/m2 IV continuous infusion D1 or 50 mg/m2 IV continuous infusion D8, 15	C1: D1, 8, 15	Initiate C2 on D29 (treatment may be delayed until D43)
Flavopiridol (C2–8)	Table 2. Dose Escalation	Immediately after lenalidomide dosing, begin IV flavopiridol bolus over 30 minutes, followed by a 4-hour IV continuous infusion	C2–8: D3, 10, 17	Initiate C3–8 on D36 (treatment may be delayed until D50)
Lenalidomide (C2–8)	Table 2. Dose Escalation	PO, supplied in 2.5, 5, and 25 mg capsules	C2–8: D1–21

Table 2.

Dose Escalation Schema

Dose Levels/ Cohorts	Lenalidomide (C2–8) D1–21	Flavopiridol (C2–8) D3, 10 and 17	Number of patients enrolled	Number of patients completing C2
−1	2.5 mg po qd	C2–8: 20 mg/m2 IV over 30 min, 20 mg/m2 IV over 4 hrs
1	2.5 mg po qd	C2–8: 30 mg/m2 IV over 30 min, 30 mg/m2 IV over 4 hrs	8	6
2	5 mg po qd	C2–8: 30 mg/m2 IV over 30 min, 30 mg/m2 IV over 4 hrs	9	7
3	7.5 mg po qd	C2–8: 30 mg/m2 IV over 30 min, 30 mg/m2 IV over 4 hrs	4	4
4	10 mg po qd	C2–8: 30 mg/m2 IV over 30 min, 30 mg/m2 IV over 4 hrs	4	3
5	10 mg po qd	C2–8: 30 mg/m2 IV over 30 min, 50 mg/m2 IV over 4 hrs	10	6
6	15 mg po qd	C2–8: 30 mg/m2 IV over 30 min, 50 mg/m2 IV over 4 hrs	4	4

To minimize risk of TLS, the first dose of flavopiridol on days 1 and 8 of cycle 1 and day 3 of cycle 2 was administered in the hospital with standard supportive care measures and monitoring for TLS. On days 15 of cycle 1, days 10 and 17 of cycle 2 and days 3, 10 and 17 of cycles 3–8, flavopiridol was administered in the outpatient clinic provided no serious adverse events were noted with the first dose. During cycles 2–8, lenalidomide was administered orally on days 1–21, with flavopiridol dosing immediately after lenalidomide on days 3, 10 and 17. Cycle 1 was defined as 28 days and starting with cycle 2, each cycle was defined as 35 days.

Patients who developed a DVT were allowed to continue protocol therapy, but were treated with a low molecular weight heparin. In patients who were unable to receive anti-coagulation due to contra-indication (i.e. the platelet count was < 50,000/mm³ or serious bleeding occurred), pulmonary embolism occurred, or a DVT recurred despite anti-coagulation, were removed from protocol therapy.

Supportive Care

Rasburicase (4.5 mg) was administered two hours prior to flavopiridol on day 1 of cycle 1 and day 3 of cycle 2 with remaining doses left to the discretion of the treating physician. Potassium monitoring was done every two hours prior to, during and one hour following treatment. Dexamethasone 20 mg orally was administered 30 minutes prior to each dose of flavopiridol and 4 mg 24 hours after each dose to prevent flavopiridol-related cytokine mediated infusion reactions. Additional corticosteroids were used at the discretion of the treating physician for patients with flavopiridol-related cytokine release or lenalidomide induced tumor flare. Anti-emetic therapy with ondansetron 8 mg orally was given 30 min prior to each flavopiridol treatment. GCSF (Neulasta) 6mg SQ was administered on day 16 of cycle 1 and day 18 of cycles 2–8. Prophylactic antiviral and antibiotic therapy including valacyclovir, ciprofloxacin, and trimethoprim/sulfmethoxazole DS were continued throughout protocol therapy.

Dose Limiting Toxicity

DLT was defined during cycle 2 of treatment. Those patients experiencing DLT during cycle 1, who had TLS, or who were unable or unwilling to proceed to cycle 2 were replaced. DLT was defined as the occurrence of any of the following events when considered possibly, probably or definitely related to the study treatment including 1) treatment delays > 14 days for any non-hematologic or hematologic toxicity or any grade 5 hematologic or non-hematologic toxicity; 2) hematologic toxicities including grade 4 febrile neutropenia or infection, grade 3 febrile neutropenia or infection with fever or infection that fails to resolve within 7 days, or ANC < 1000/mm³ or platelets < 30,000/mm³ > 14 days after the completion of cycle 1 or 2 (i.e. day 44 or later after cycle 1 or day 51 or later after cycle 2); and 3) any grade 3 or 4 non-hematologic toxicity. The following were not considered DLT: TLS, tumor flare reaction, cytokine release syndrome, alopecia, DVT responsive to anticoagulation, diarrhea responsive to medical treatment, and nausea and vomiting controllable with steroids and anti-emetic therapy, transient grade 3 and 4 electrolyte or liver function test (AST, ALT, total bilirubin) abnormalities that resolve ≤ grade 1 in ≤ 7 days. The National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE) version 3.0 was used for toxicity grading.

Response Assessment

Response was assessed by the revised National Cancer Institute-Sponsored Working Group Guidelines for CLL. Response was evaluated after cycles 2, 4, 6, and 8 (during days 22–35 of each cycle).

Pharmacokinetics

Pharmacokinetic evaluations were conducted in collaboration with the Pharmacoanalytical Shared Resource at OSU. Pharmacokinetic assessments of flavopiridol and its glucuronide metabolite were completed on cycle 1 day 1 and cycle 2 day 3. Plasma samples were collected pre-dose and at 0.5, 4.5, 8, 12, 24 and 48 hours after the start of flavopiridol infusion. Pharmacokinetic assessments of lenalidomide were completed during cycle 2 on days 2 and 3. Plasma samples were collected pre-dose, 0.5, 1, 1.5, 2, 3, 4.5, 6, 8 and 24 hours after administration. Plasma concentrations of lenalidomide, flavopiridol and flavopiridol glucuronide were simultaneously determined using previously published LC-MS/MS methods [31, 40–41].

Statistics

The primary endpoint of the study was to define the MTD of the combination of flavopiridol and lenalidomide using the standard 3+3 phase I design as described above and to define the specific toxicities and the dose limiting toxicities of this combination in this population of patients. Common toxicities are summarized with frequency by dose level and for all patients. Baseline characteristics are described using frequencies and medians with ranges for categorical and continuous variables, respectively. PFS was measured from the date on study to the date of disease progression or death from any cause. Patients who were still alive and didn’t have disease progression were censored at the last visit for PFS. OS was determined from the date of study to the date of death from any cause or date of observation. Patients who were still alive were censored at the last visit. Survival curves were estimated using the method of Kaplan-Meier. Pharmacokinetic evaluations were performed in Phoenix WinNonlin version 6.3.0.395 and resulting parameters summarized in graphic and tabular format.

BH3 Profiling

Thawed aliquots of pretreatment peripheral blood mononuclear cells containing CLL cells were purified for untouched B-cells by non-B-cell depletion using a cocktail of biotinylated monoclonal Abs (CD2, CD4, CD11b, CD16, CD36, Anti-IgE, and CD235a (glycophorin) and magnetic beads (Miltenyi Biotec, Auburn, CA). Of 32 patients treated with lenalidomide and flavopiridol, specimens were available and assessed for 26. The extent of cell purification was monitored by flow cytometry of stained cells before and after purification with anti-CD19-APC and anti-biotin antibody-FITC for the presence of non-B cells labeled with MAbs from depletion cocktail (Miltenyi Biotec, Auburn, CA). Specimens (2×10⁵ cells/assay) were permeabilized with digitonin and incubated with JC-1 mitochondrial dye and 100 µM BH3 peptides (Bim, Puma, Noxa, Bad, Bmf, Hrk; Bim and Puma were also assayed at 0.1 µM and 10 µM, respectively) or with dimethyl sulfoxide (DMSO [(1%]) or carbonyl cyanide m-chlorophenyl hydrazone (CCCP [10 µM]). Samples were run in triplicate and fluorescent traces of JC-1 dye monitored over 300 min of assay. Area under the curve was integrated relative to the positive control uncoupling reagent after normalization for DMSO background:

$$\%\text{priming}=(1-(\frac{\mathit{\text{Peptide}}-CCCP}{DMSO-CCCP}))\times 100$$

Association between biomarker status (% priming) and clinical response classification was by Mann-Whitney Test. We pre-determined a statistical analysis plan with significance of p<0.05. Marker predictive ability was assessed using the area under the receiver operator characteristic curve (AUC). Analyses utilized SAS software, version 9.2 (Cary, NC), R version 2.14.2 (Vienna, Austria), and/or Graphpad Prism version 5.04 (La Jolla, CA).

Results

Patient Characteristics

A total of 39 patients were treated on Cohorts 1 (n=8), Cohort 2 (n=9), Cohort 3 (n=4), Cohort 4 (n=4), Cohort 5 (n=10), Cohort 6 (n=4) (Table 2). Median age was 62 years (range 26–74) and there was a male predominance (26). Median number of prior therapies was 3 (range 1–10). Patients had the following high-risk features: del(17p13.1) in 22 patients; del(11q22.3) in 15 patients; complex karyotype in 31 patients; fludarabine refractory in 14 patients. Baseline patient characteristics are summarized in Table 3.

Table 3.

Patient Characteristics, N=39

Characteristic	N (%)
Age Median (range)	62 (26–74)
Gender	26 Male : 13 Female
Rai Stage:
1–2	11 (28)
3–4	28 (72)
Prior therapies
Median (range)	3 (1–10)
Prior Fludarabine	39 (100)
Fludarabine refractory	14 (36)
Bulky lymphadenopathy
≥ 5 cm	24 (62)
≥ 10 cm	6 (15)
Cytogenetics (FISH)
Del(17p13.1)	22 (56)
Del(11q22.3)	15 (38)
Complex Karyotype	31 (79)
Median number of cycles (range)	4 (1–8)

Dose escalation and toxicity

Dose limiting toxicities occurred in three patients and included Grade 3–4 transaminitis in two patients, one at dose level 1 and one at dose level two, and grade 3 absolute neutrophil count lasting greater than 14 days in one patient at dose level 5. The most common grade 3–4 adverse events were hematologic: neutropenia (n=32 or 82%), anemia (n=16 or 41%), thrombocytopenia (n=20 or 51%) and lymphopenia (n=9 or 23%); gastrointestinal: diarrhea (n=18 or 46%); metabolic: transaminitis (n=17 or 44%), hypophosphatemia (n=12 or 31%), hypocalcemia (n=6 or 15%); hypokalemia (n=11 or 28%), hyperkalemia (n=6 or 15%); hyperglycemia (n=17 or 44%) and TLS (n=10 or 26%). TLS occurred only during cycle 1, and three of the patients with TLS were replaced. There was only one patient with Grade 3 tumor flare. This patient had lenalidomide held for painful splenomegaly on day 10 of cycle 2 until cycle 3 and was treated with three days of steroids with lenalidomide dose reduced to 7.5 mg for cycle 3. The patient had recurrence of painful splenomegaly on day 10 of cycle 3, and lenalidomide was held again with two days of steroid treatment given. When lenalidomide was restarted at 7.5 mg, the patient again had recurrence of symptoms in four days and was removed from study. A total of fifteen patients required dose reductions. Grade 3–4 adverse events are summarized in Table 4.

Table 4.

Grade 3–4 Adverse Events

Adverse Events	No. of patients with at least one occurrence (%)
Hematologic
Neutropenia	32 (82)
Anemia	16 (41)
Thrombocytopenia	20 (51)
Lymphopenia	9 (23)
Infection
Febrile Neutropenia	3 (8)
Cellulitis	2 (5)
Pneumonia	5 (13)
Catheter related	1 (3)
URTI	2 (5)
Shingles	1 (3)
Dermatologic
Rash	1 (3)
Gastrointestinal
Diarrhea	18 (46)
Constipation	1 (3)
Metabolic
Transaminitis	17 (44)
Hyperbilirubinemia	1 (3)
Hypophosphatemia	12 (31)
Hypocalcemia	6 (15)
Hypokalemia	11 (28)
Hyperkalemia	6 (15)
Hyperglycemia	17 (44)
Elevated creatinine	2 (5)
Other
Tumor flare	1 (3)
Tumor lysis syndrome	10 (26)
Cytokine release	1 (3)
Fatigue	3 (8)

Response

Thirty-one of the thirty-nine patients were evaluable for response completing at least two cycles. Of the eight patients that were not evaluable for response 2 had progressive disease in cycle 1, 3 were replaced for TLS in cycle 1, 1 had thrombocytopenia after cycle 1 whose platelet count did not recover to > 30,000 prior to cycle 2, and 2 had DLT with transaminitis during cycle 2. Of the thirty-one evaluable patients there was a 51% overall response rate, all of which were partial responses.

Of the 22 patients with del(17p13.1), 11 (50%) of patients responded and of the 14 patients with del(11q22.3), 10 (71%) responded. These factors associated with inferior outcome to standard therapy did not show a statistically significant association with response (p value 0.62, 0.14 respectively).

Median PFS for all patients was 7.7 months (95% CI: 4.7–11.2). Median PFS for patients with del(17p13.1) was 7.7 months (95% CI: 4.5–13.0) versus 8.6 months (95% CI: 3.0–16.0) for patients without del(17p13.1), showing PFS is not significantly associated with del(17p13.1) (p-value=0.71). Seven patients went on to allogeneic stem cell transplant. Of these seven patients, one was lost to follow-up; one had graft failure with multi-organ failure due to complications and died at 2 months post-transplant; one progressed within 6 months of transplant and died of infectious complications at 11 months post-transplant; one relapsed at 12 months post-transplant and went on to receive three other therapies prior to dying of infectious complications 32 months post-transplant; and at last follow-up three were still alive and without evidence of disease at 35, 37, and 38 months post-transplant.

Pharmacokinetics

Thirty-six patients were evaluable for flavopiridol pharmacokinetics on cycle 1 day 1 and thirty-two on cycle 2 day 3 (Figure 1). Pharmacokinetics of lenalidomide were evaluable for thirty-one patients on cycle 2 day 2 and thirty-two patients on cycle 2 day 3. Average observed maximum concentrations (Cmax) of lenalidomide reached 0.97uM, and the average observed Cmax of flavopiridol was 3.50uM (Figure 1 and Supplemental Material Figure 1). Non-compartmental parameter estimates for clearance, Cmax and volume of distribution were consistent with previously reported values [28, 42] (Table 5). When flavopiridol and lenalidomide were administered in combination, PK plots and parameter estimates of lenalidomide were consistent with those for lenalidomide alone (Supplemental Table 2). At 60 mg/m², flavopiridol AUC (p=0.03) and Cmax (p=0.01) tended to increase when lenalidomide dose was increased, as shown in Supplemental Table 3. However, this trend was not observed at the higher 80 mg/m² flavopiridol dose. Flavopiridol-glucuronide data was available from 30 patients. Of the 10 patients with TLS, flavopiridol-glucuronide data was available for 8 patients (data not shown). No trends or significant associations were observed between plasma flavopiridol-glucuronide pharmacokinetics and TLS.

Figure 1.

Semi-logarithmic concentration-time plots for (A) flavopiridol and (B) lenalidomide by patient cohort. Data points represent the average of individual plasma concentrations with standard deviations represented by error bars.

Table 5.

Non-compartmental parameter estimates for lenalidomide and flavopiridol by treatment category, where C1D1 represents flavopiridol alone, C2D2 is lenalidomide alone, and C2D3 is combined lenalidomide and flavopiridol. Values represent the average of individual parameter estimates followed by standard deviations in parenthesis.

Treatment		N	AUCINF_pred (hr*µmol/L)	Cmax (µmol/L)	Cl_F_pred (L/hr)	Vz_F_pred (L)	HL_Lambda_z (hr)	Tmax (hr)
Flavopiridol	C1D1	36	19.23 (11.45)	2.53 (1.46)	12.66 (7.20)	270.9 (244.9)	15.78 (11.89)	1.34 (1.63)
	C2D3	32	20.51 (11.91)	2.57 (1.30)	11.71 (7.16)	264.9 (264.4)	16.28 (12.18)	1.31 (1.37)
Lenalidomide	C2D2	31	2.93 (2.04)	0.65 (0.46)	14.43 (9.05)	76.0 (48.5)	3.71 (1.23)	1.33 (0.77)
	C2D3	32	2.53 (1.90)	0.45 (0.37)	18.43 (13.97)	85.7 (64.5)	3.36 (1.08)	3.73 (6.69)

Biomarker Testing of Clinical Response via BH3 Profiling

From 32 study participants who would be treated with flavopiridol and lenalidomide, aliquots of pre-treatment specimens were available from 26 with defined evaluable outcomes (median age of 60 years [range: 26–74 years]). Specimens were thawed and exposed in vitro to individual BH3 peptides, including an activator (Bim) and several sensitizers (Noxa, Puma, Bad, Hrk) as surrogates for the function of Bcl-2 family proteins. One sample was eliminated from statistical analysis due to insufficient number of viable cells thus yielding 25 specimens with analyzable data. All specimens were analyzed in triplicate, with coefficients of variation (CV) for repeat samples from individual patients generally <5%.

The percent priming, i.e. quantifiable propensity of a given BH3 peptide to induce mitochondrial depolarization relative to an uncoupling control agent, for each peptide is summarized in Table 6 separately for patients who responded to study therapy (i.e. achieved partial response) and those who did not respond to treatment (stable disease/progressive disease). Among the peptides assayed, Puma(10) alone elicited a priming trend between responders (32.8 ± 16.5% [mean±SD]) and non-responders (22.3 ± 9.9%; p= .059); the percent priming with Puma(10) for individual patients is depicted in Figure 2. To test the ability of Puma to serve as a predictive biomarker, we employed the area under the receiver operator characteristic curve (AUC) to analyze the sensitivity and specificity of this biomarker, which yielded an AUC of 0.73 (95%CI: 0.53–0.94; p=0.027; Figure 2). In assessment of patient clinical information (Table 6,) bulky disease of >5 cm trended with patient response (Mann-Whitney p-value =0.088; AUC=0.68 (95% CI: 0.53–0.84), AUC p-value=0.022). We performed adjusted analyses of Puma priming in which we accounted for bulky disease >5 cm as a second covariate. As shown in Figure 2, adjustment for bulky disease (>5 cm) improved the AUC to 0.84 (95% CI: 0.69–0.99). The AUC p-value =0.0063 was significant for Puma combined with bulky disease (p-value <0.01; p<0.05/5 BH3 profiling biomarkers with Bonferroni correction for multiple analyses).

Table 6.

Relationship Between BH3 Peptide Priming Biomarkers and Clinical pathologic Information Relative to Patient Response to Study Therapy

A. BH3 Profiling Biomarkers Association with Response
	Mean %Priming ± SD		p-value (Mann-Whitney)	AUC [95% CI]	p-value (AUC)
	Non-Responders	Responders
Bad (100 µM)	54.8 ± 9.3	61.3 ± 20.0	0.13	0.69 [0.19, 0.84]	0.090
Bim (0.1 µM)	5.5 ± 6.1	7.7 ± 12.1	0.95	0.51 [0.29, 0.73]	0.92
Hrk (100 µM)	41.2 ± 20.1	48.6 ± 25.9	0.36	0.61 [0.38, 0.85]	0.35
Noxa (100 µM)	26.8 ± 12.0	34.4 ± 21.9	0.43	0.60 [0.38, 0.82]	0.38
Puma (100 µM)	72.4 ± 14.9	77.1 ± 12.8	0.42	0.60 [0.37, 0.83]	0.40
Puma (10 µM)	22.3 ± 9.9	32.8 ± 16.5	0.059	0.73 [0.53, 0.93]	0.027

B. Clinical Patient Information Association with Response
	p-value (Mann-Whitney)	AUC [95% CI]	p-value (AUC)
Age (median=60; range 26–74)	0.99	0.51[0.28,0.74]	0.93
Gender	1.00	0.51[0.30, 0.72]	0.90
RAI Stage	0.87	0.52[0.33, 0.70]	0.86
# Prior Therapies	0.61	0.57[0.35, 0.80]	0.52
Fludarabine Refractory	0.66	0.57[0.53, 0.84]	0.45
Bulky Disease (>5cm)	0.088	0.68[0.53, 0.84]	0.022
Bulky Disease (>10cm)	0.25	0.60[0.50, 0.70]	0.053
Dose Level	0.41	0.60[0.37, 0.84]	0.38
Tumor Lysis Syndrome (TLS)	0.54	0.57[0.43, 0.71]	0.35

Figure 2.

Dot-plot and ROC-plot depictions of Puma patient response discrimination

Discussion

The purpose of this study was to evaluate the safety and efficacy of combined flavopiridol and lenalidomide in relapsed and refractory CLL with the hypothesis that higher doses of lenalidomide would be tolerated with less tumor flare observed in patients who received flavopiridol for disease reduction prior to the initiation of lenalidomide.

In this Phase I dose-escalation study of combined flavopiridol and lenalidomide in relapsed CLL, the MTD was not reached with the highest dose being given at dose level 6 with treatment of flavopiridol given in cycle 1 at a dose of 30 mg/m² IVB + 30 mg/m² CIVI day 1 of cycle 1 and 30 mg/m² IVB + 50 mg/m² CIVI day 8 and 15 of cycle 1 and days 3, 10 and 17 of cycles 2–8 immediately after lenalidomide dosing and lenalidomide 15 mg orally daily days 1–21 of cycles 2–8. There was no unexpected toxicity seen, including no increased risk of TLS, tumor flare, or opportunistic infection.

Significant clinical activity was demonstrated, with a 51% response rate in this group of relapsed, heavily pretreated patients. The response rate was not affected by the presence of high-risk genomic features, with a 50% response rate in patients with del(17p13.1) and a 71% response rate in patients with del(11q22.3). Additionally, seven patients (18%) were able to proceed with allogeneic stem cell transplant, a potentially curative therapy in these poor-risk patients. Three patients remained without evidence of disease at 3 years post-transplant.

Ten (26%) of patients did experience grade 3–4 TLS, all occurring in cycle 1 of therapy with 3 of these patients being replaced. This was a slightly lower rate of TLS compared to previous studies [28–30]. Furthermore, unlike those previous studies, which revealed higher rates of TLS in female CLL patients, TLS in this study appeared evenly distributed between males (7/26, 27%) and females (3/13, 23%). Furthermore, unlike previous studies that revealed significant associations between TLS and flavopiridol glucuronide exposures, we found no such associations in this study.

Only one patient (3%) experienced grade 3 tumor flare and was unable to continue with lenalidomide therapy due to symptoms, despite being able to reach a dose of 15 mg orally of lenalidomide in Cycles 2–8. In prior studies using single agent lenalidomide at doses of 10–25 mg daily in relapsed CLL [4–5], 30–60% of patients experienced tumor flare. Two separate phase I studies had to modify the planned study to reduce the dose of lenalidomide from 10–15 mg daily initial dose down to a dose of 2.5 mg daily due to serious tumor flare [6–16]. It may be possible to further increase lenalidomide dosing in these patients that have received cytoreduction prior to lenalidomide initiation.

Biomarker testing yielded interesting results for association of mitochondrial priming (BH3 profiling) of the BH3 only peptide Puma with response. A pro-apoptotic modulator of anti-apoptotic proteins Mcl-1, Bcl-2 and Bcl-xL, Puma functions as a general sensitizer. As such, Puma may not be used to specifically address whether a malignancy is driven by specific anti-apoptotic BH3 proteins, but rather it displays a global propensity of cells to respond to intrinsic pro-apoptotic cues. Previous BH3 profiling studies established that flavopiridol response in CLL patients was Bcl-xL dependent [34]. It is perhaps not surprising that a more complex combination treatment comprising therapies with different mechanisms would be associated with a BH3 peptide that modulates multiple anti-apoptotic proteins. Further, while we would classify this as a trend between Puma profiling and response (Mann-Whitney p = 0.059; AUC p=0.027) the observation is interesting when placed in the context of the dose escalation component of the study design. Further, Puma priming when combined with bulky disease as a clinical adjustment variable displayed significant association with response (p=0.0063) indicating that additional observation may be warranted in follow-on CLL clinical studies with uniformly defined treatment regimens comprising lenalidomide and flavopiridol.

In summary, this phase I dose escalation study demonstrates that the combination of flavopiridol and oral lenalidominde is feasible and well tolerated in heavily pretreated CLL patients. By treating patients with flavopiridol prior to lenalidomide therapy to decrease tumor burden, patients were able to tolerate higher doses of lenalidomide and only one patient experienced tumor flare. This combination had significant clinical activity in patients with high-risk cytogenetic features and is a promising therapy for patients requiring disease de-bulking prior to allogeneic transplantation.