Phase 1 study of the histone deacetylase inhibitor entinostat plus clofarabine for poor-risk Philadelphia chromosome-negative (newly diagnosed older adults or adults with relapsed refractory disease) acute lymphoblastic leukemia or biphenotypic leukemia

Abstract

PURPOSE:

Despite advances in immunotherapies, the prognosis for adults with Philadelphia chromosome-negative, newly diagnosed (ND) or relapsed/refractory (R/R) acute lymphoblastic leukemia/acute biphenotypic leukemia (ALL/ABL) remains poor. The benzamide derivative entinostat inhibits histone deacetylase and induces histone hyperacetylation. The purine nucleoside analogue clofarabine is FDA-approved for R/R ALL in children 1–21 years of age. Low doses of clofarabine have been reported to induce DNA hypomethylation. We conducted a phase 1 study of low dose clofarabine with escalating doses of entinostat in adults with ND or R/R ALL/ABL.

EXPERIMENTAL DESIGN:

Adults ≥60 years with ND ALL/ABL or ≥21 years with R/R ALL/ABL received repeated cycles every 3 weeks of entinostat (4mg, 6mg or 8mg orally days 1 and 8) and clofarabine (10mg/m²/day IV for 5 days, days 3–7) (Arm A). Adults aged 40–59 years with ND ALL/ABL or age ≥21 years in first relapse received entinostat and clofarabine prior to traditional chemotherapy on day 11 (Arm B). Changes in DNA damage, global protein lysine acetylation, myeloid-derived suppressor cells and monocytes were measured in PBMCs before and during therapy.

RESULTS:

Twenty-eight patients were treated at three entinostat dose levels with the maximum administered dose being entinostat 8 mg. The regimen was well tolerated with infectious and metabolic derangements more common in the older population versus the younger cohort. There was no severe hyperglycemia and no peripheral neuropathy in this small study. There were 2 deaths (1 sepsis, 1 intracranial bleed). Overall response rate was 32%; it was 50% for ND ALL/ABL. Entinostat increased global protein acetylation and inhibited immunosuppressive monocyte subpopulations, while clofarabine induced DNA damage in all cell subsets examined.

CONCLUSION:

Entinostat plus clofarabine appears to be tolerable and active in older adults with ND ALL/ABL, but less active in R/R patients. Further evaluation of this regimen in ND ALL/ABL appears warranted.

INTRODUCTION

The prognosis for adults over age 60 with Philadelphia chromosome negative (Ph-neg) acute lymphoblastic leukemia (ALL) or acute bilineage leukemia (ABL) remains a challenge relative to the favorable outcomes for children and younger adults. This poor prognosis relates to multiple disease-related factors including higher incidence of Ph chromosome-like genetic aberrations, multidrug resistance, and inability to tolerate high dose cytotoxic chemotherapy due to comorbidities in this older population. Two analyses have shed light on the age-dependence of overall response. The international ECOG E2993/MRC UKALL XII trial ¹ of 1500 adults ages 15–59 demonstrated that 5-year overall survival (OS) for adults age < 30 was 45% with median of 3.5 years and for adults 30–39 was 34% with median of 2 years. In contrast, the 5-year OS for patients age 40–49 was 23% with median of 1 year and for patients ≥ 50 was 15%. In the SWOG 9400 trial ², complete remission (CR) rates for patients ≤ 30 years of age were 90% but only 60% for patients ages 50–65 years. In contrast, cure rates in children approach 80–90%. Moreover, once leukemia relapses in adults, the outcome historically has been dismal, with median survival of 5–6 months and only 22% alive at 1 year and 7% alive at 5 years, prior to recent advances.

While the outcomes for adults with relapsed/refractory (R/R) disease have improved with immune-based constructs such as blinatumomab (Blin) ^3–8 and inotuzumab ozogamicin (Ino) ^6,9, there still remains the opportunity to find alternative combination treatment paradigms for adults with R/R ALL/ABL and older adults with newly diagnosed (ND) ALL/ABL (especially if they are or become CD19 or CD22 negative). In the INOVATE trial of 218 adults with R/R ALL randomized to Ino vs standard of care (SOC), Ino improved survival outcomes as defined by CR/CRi (80.7% vs. 29.4%; P<0.001); measurable residual disease negativity (78.4% vs. 28.1%, P<0.001); prolonged progression-free survival (PFS, HR, 0.45, P<0.001); median survival (5.0 vs 1.8 months) and overall survival (OS) (median 7.7 vs. 6.7 months, HR, 0.77, P=0.04). Veno-occlusive disease (VOD) affected 11% of patients ⁶. Additionally, patients age ≥ 55 years had decreased OS (median, 5.6 vs 8.6 months; HR, 0.610) and an increased likelihood of VOD if given Ino versus SOC (41% vs 17%). The phase 3 randomized study evaluating Blin vs. SOC in R/R ALL in 405 Ph-neg ALL patients demonstrated improved OS in those receiving Blin (median OS 7.7 vs. 4.0 months; HR, 0.71, P=0.01) and remission rates within 12 weeks of treatment initiation (34% vs. 16%, P<0.001) as well as a longer median remission duration (7.3 vs. 4.6 months)⁶. Nonetheless, although Blin demonstrated previously unattainable single-agent activity in this high-risk disease setting, salutary long-term outcomes were achieved mainly in younger patients able to undergo successful hematopoietic stem cell transplant (HSCT).

Aberrant gene expression, due to intrinsic structural genetic abnormalities, epigenetic alterations, or a combination of genetic and epigenetic changes is a hallmark of carcinogenesis. Histone deacetylases (HDACs) decrease gene expression through direct deacetylation of core histones and association with proteins such as transcription factors. HDAC inhibitors (HDACi) can permit re-expression of abnormally silenced genes (such as tumor suppressor genes) which are deacetylated in the malignant state and have been shown to induce cell cycle arrest, and differentiation and apoptosis with clinical responses in leukemia and in solid tumors ^10–13. The orally available, Class I selective HDACi entinostat (MS-275) induces expression of the tumor suppressor p21^WAF1/CIP1 with associated anti-proliferative effect in vitro and in vivo regardless of the p53 status ^11,13–21. Phase 1 testing of single-agent entinostat by Gojo, et al. ²² in 39 adults with R/R acute leukemias demonstrated increases in global protein and histone H3/H4 acetylation, p21^WAF1/CIP1 expression and caspase-3 activation in bone marrow mononuclear cells, despite a lack of clinical responses by classical criteria. Tsapis, et al. ²³ showed that HDAC inhibition induces apoptosis in glucocorticoid-resistant ALL cells in vitro, suggesting that HDACi might reverse some element of the steroid resistance often seen in R/R adult ALL.

Lastly, entinostat also enhances the cytotoxicity of fludarabine in vitro ²⁴. Fludarabine is a purine analog that inhibits multiple enzymes involved in DNA synthesis and kills cells via apoptosis. Synergism of these agents was noted when leukemia cell lines were pre-treated with entinostat. Specifically, prior exposure of leukemia cell lines to entinostat enhanced fludarabine lethality, raising the possibility that HDACi could trigger time-dependent events that lower the threshold for fludarabine-mediated apoptosis. Similar synergistic interactions have been reported with other HDACi (i.e. sodium butyrate) and nucleoside analogs (i.e. gemcitabine) ²⁴.

Clofarabine is a second-generation deoxyadenosine analog with clinical activity in a variety of acute leukemias including R/R ALL when given as a single agent ²⁵ or in combination with other anti-leukemia agents such as cyclophosphamide (CY), with responses in at least 50% of adults with R/R ALL ^25–27. The cytotoxic effects of clofarabine are mediated by mechanisms that converge on DNA synthesis and repair of damaged DNA. Inhibition of ribonucleotide reductase subunit M1 results in depletion of intracellular pools of preexisting normal deoxyribonucleotides and their triphosphorylated forms which, in turn, leads to enhanced accumulation of the triphosphorylated deoxyadenosine analog and its incorporation into DNA. Incorporation of clofarabine-triphosphate into DNA leads to termination of chain elongation, inhibition of DNA polymerases, and inhibition of DNA synthesis. Additionally, low doses of clofarabine can reverse methylation, thereby potentially exerting complementary epigenetic modification when combined with an HDACi ²⁸.

To assess the potential for epigenetic synergy between clofarabine and entinostat, we conducted a Phase I trial combining entinostat and clofarabine in adults with poor-risk, ND Ph-neg ALL/ABL. In addition to determining safety and efficacy, we examined protein acetylation by flow cytometry, as well as DNA damage as assessed by determination of γH2AX, also by flow cytometry, and impact on selected aspects of peripheral immunity in an exploratory fashion in samples obtained at baseline and during treatment.

PATIENTS AND METHODS

Patient eligibility and selection

Between May 2010 and April 2014, 28 adults with ND or R/R Ph-negative ALL/ABL were enrolled (Figure 1). Adults ≥ 60 years with ND ALL/ABL and adults ≥ 21 years with R/R ALL/ABL were eligible to receive multiple cycles of entinostat plus clofarabine (Arm A). Adults aged 40–59 years with confirmed diagnoses of ND ALL/ABL and adults ≥ 21 in first relapse were eligible to receive a single cycle of entinostat plus clofarabine (prior to standard chemotherapy) as a window of opportunity to obtain disease control and correlative data. (Arm B). Inclusion/exclusion criteria are provided in Supplement 1.

Figure One: Consort Diagram and Treatment/Lab Schema for Entinostat Clofarabine in Adults with Acute Leukemia.

Patients started at DL1. DLMinus1 was a dose de-escalation arm in case of toxicity. No patients required DLMinus1.

Treatment plan and study design

Entinostat was administered orally once daily on days 1 and 8 (D1/D8) at escalating doses to three cohorts using a traditional “3+3” design (Figure 1). The first cohort (DL1) received a 4mg dose, the second cohort received a 6mg dose and the third cohort received an 8mg dose, with each cohort dosed on D1/D8. Clofarabine 10 mg/m²/day was administered intravenously for 5 days on days 3 through 7 (D3–7) for each cohort. Adults age ≥60 years with ND ALL/ABL or ≥ 21 years with R/R ALL/ABL (Arm A) could continue to receive up to a total of 4 cycles at their assigned dose level if there was evidence of response. There was no intracohort dose escalation. Adults ages 40–59 years with ND ALL/ABL or adults > 21 years of age in first relapse (Arm B) received a single cycle of entinostat plus clofarabine before initiating traditional multi-agent chemotherapy or HSCT on Day 11. For patients in Arm A who achieved CR or CR with incomplete recovery (CRi), partial response (PR), or stable disease (SD) following cycle 1, a repeat cycle of the same regimen was administered no sooner than day 21 and no later than day 63 from initiation of the previous cycle. Patients who achieved PR could continue to receive treatment as long as there was evidence for clinical benefit without limiting toxicities. The maximal tolerated dose (MTD) was determined as the highest dose level up to and including DL3, at which no more than 1 of the first 3 patients experienced a DLT. Once MTD was defined, 10 patients were enrolled in an expansion cohort to better define safety and efficacy.

Toxicity evaluation

Clinical and laboratory monitoring of the study participants was performed according to the standards of practice for adults with acute leukemia undergoing chemotherapy. Toxicities were described and graded using NCI-CTCAE version 4.0. DLT were assessed as described previously (27). Additional details provided in Supplement 1.

Evaluation of Response

BM aspirate and biopsy were obtained at baseline prior to D1 entinostat and again on D8 of the first cycle of chemotherapy, as well as up to 72 hours before subsequent cycles and end of study or at any time that leukemia re-growth was suspected. Responses were assessed according to the International Working Group criteria for CR, CRi, PR, or PD ²⁹. Statistical considerations are included in Supplement 1.

Laboratory correlates

Serial samples were obtained from peripheral blood (PB) and/or BM in 21 patients prior to and 2 hours after entinostat administration on D1, D3 (PB only pre- and post clofarabine), D8 (PB and BM following the completion of D3–7 clofarabine and prior to and 2 hours post D8 entinostat administration), and D11 (PB only).

The impact of (a) entinostat alone, (b) clofarabine, and (c) entinostat following clofarabine on the presence and magnitude of DNA damage as measured by γH2AX expression, global protein acetylation in PBMCs (T cells, B cells, monocytes and NK cells) and modulation of myeloid-derived suppressor cell (MDSC) and monocyte subsets was determined at four specific time points during cycle 1 (D1 pre- and 2hrs post-entinostat, D8 post clofarabine pre- and 2hrs post-entinostat) as described previously ^30–32. Details regarding isolation and preparation of PBMCs along with analyses are in Supplement 1.

RESULTS

Patient characteristics

As depicted in Table 1, 28 adults with ALL/ABL (3 bi-lineage) with median age of 55 years were enrolled between Jan 2010 and Dec 2014. This was a high-risk group of patients, with 50% having R/R ALL and 46% ≥ 65 years old (Supplemental Table 3). The median number of prior therapies for ALL was 2, including one who had received allogeneic HSCT. A total of 21 patients were treated in Arm A (10 ND, 11 R/R; median age 62 years) and 7 in Arm B (4 ND, 3 R/R; median age 48 years). Median follow-up was 25 days (range 7–690) for all patients, 27 days for Arm A patients, 65 days for all responders, and 92 days for those achieving CR/CRi. Per protocol, patients on Arm A continued with entinostat/clofarabine if a response was obtained (not allowed to move to other therapy: Figure 1). Four patients in Arm B proceeded to multi-agent chemotherapy with HyperCVAD, one received Hoelzer-regimen chemotherapy and another E2993-based therapy (1–3). One patient on Arm B received a HSCT.

Table 1.

Patient characteristics

		Dose level		Treatment phase		Treatment arm		Treatment history
Characteristics	All patients (n=28)	DL 1 & 2 (n=12)	DL3 (n=16)	Dose-escalation (n=18)	Dose-expansion (n=10)	Arm A (n=21)	Arm B (n=7)	Previously treated (n=14)	No prior treatment (n=14)
Age (median yrs, range)	55 (21–75)	45 (21–72)	58.5 (23–75)	48 (21–72)	63 (23–75)	62 (21–75)	48 (23–66)	41 (21–69)	65 (45–75)
Male, %	18 (64)	9 (75)	9 (56)	12 (67)	6 (60)	14 (67)	4 (57)	11 (79)	7 (50)
PS = 2, %	3 (11)	2 (17)	1 (6)	2 (11)	1 (10)	2 (10)	1 (14)	3 (21)	0 (0)
Lineage
- B cell, %	20 (71)	9 (75)	11 (69)	14 (78)	6 (60)	16 (76)	4 (57)	10 (71)	10 (71)
- T cell, %	5 (18)	2 (17)	3 (19)	3 (17)	2 (20)	3 (14)	2 (29)	3 (21)	2 (14)
- Bilineage, %	3 (11)	1 (8)	2 (13)	1 (6)	2 (20)	2 (10)	1 (14)	1 (7)	2 (14)
Previously untreated, n (%)	14 (50%)	5(42%)	9(56%)	7(39%)	7(70%)	10(48%)	4(57%)	-	-
Previously treated, n (%)	14 (50%)	7 (58%)	7 (38%)	11 (61%)	3 (30)	11 (52%)	3 (43%)	-	-
(a)1 prior tx	6	3	3	4	2	4	2
(b) ≥2 prior	8	4	4	7	1	7	1
(c) allo BMT	1	1	0	1	0	0	0

Toxicities

No DLT was observed at entinostat flat fixed doses of 4mg, 6mg or 8mg on D1 and D8 in combination with clofarabine 10 mg/m² daily on D3–7. Thus, the MTD was defined to be entinostat 8mg by mouth daily on D1/D8 based on previous demonstration that an 8mg dose was the MTD achieved in the Phase I trial and was sufficient to induce protein hyperacetylation, p21 expression, and caspase-3 activation in leukemia BM cell populations ²². The median number of cycles administered was 1 (range 1–4).

The most common non-infectious, non-hematologic adverse events of any grade were metabolic derangements (82%), nausea and vomiting (64%), elevated alkaline phosphatase/bilirubin (64%) and elevated AST/ALT (61%), with grade ≥3 toxicities in 54%, predominantly in hepatic enzymes (Supplemental Table 1). The rate of non-infectious, non-hematological grade ≥ 3 toxicities was similar among patients treated at DL1/DL2 and those treated at DL3 (50% and 56%, repectively) but may have been higher in those cohorts with older (age ≥ 60 years) adults, in particular Arm A compared with Arm B (57% vs. 43%), albeit small numbers in all groups.

The most common serious adverse events during cycle 1 were infections (57%). The rate of grade ≥3 infections tended to be somewhat higher in the dose expansion cohort compared to the dose escalation cohort (70% vs. 50%), and in the ND patients compared with those who received prior chemotherapy (69% vs. 47%) although patient numbers are small. Notably, the median age for the dose expansion cohort (as compared to the dose escalation cohort) was older (age 62 versus 47.8 years) and this was also true for the ND patients due to predefined criteria for this cohort. The majority of infections were grade 3, with only 1 grade 5 (during cycle 2). Importantly, there was no major hyperglycemia and no evidence of peripheral neuropathy in this small group of patients.

The 30- and 60-day all-cause mortality rates were 10.7% (3 patients) and 21.4% (6 patients) respectively, with equal numbers of ND (n=3) and R/R (n=3) patients. With regard to treatment-related deaths, one patient developed tumor lysis syndrome with acute kidney injury, respiratory distress, and cytokine release syndrome on day three in response to beginning clofarabine and died 2 weeks later from intracranial hemorrhage. A second patient, who had achieved CRi with his first cycle, succumbed to Pseudomonas cellulitis with sepsis during cycle two.

Responses

For all 28 patients, the overall response rate was 32% (9/28). CR/CRi was attained in 4 (14%) patients; three at DL-1 and one at DL3). All CR patients were ND, age ≥ 60 years, treated in Arm A (4/14, 29%) and received at least 2 cycles of therapy. An additional 4 (25%) patients achieved PR at DL3 (4/16, 25%; 2 each in Arm A and Arm B; 2 ND and 2 R/R). Responses were observed at all dose levels, across all subtypes, and primarily in those who were ND rather than R/R (Table 2).

Table 2:

Response to therapy as it relates to prior therapy, number of cycles and designation in either Arm A or B as well as dose level.

	ALL PTS (N=28)	DL 1&2 (N=12)	DL3 (N=16)	ARM A (N=21)	ARM B (N=7)	1 CYCLE (N=22)	2 CYCLES (N=4)	3 OR 4 CYCLES (N=2)	PRIOR TREATMENT (N=14)	NO PRIOR TX (N=14)	NO PRIOR TX AND ARM A (N=10)	NO PRIOR TX AND RECEIVED > 1 CYCLE (N=5)
NR	18	9	9	14	4	18	0	0	13	5	4	0
	64%	75%	56%	67%	57%	82%	0%	0%	93%	36%	40%	0%
SD	1	0	1	0	1	1	0	0	0	1	0	0
	4%	0%	6%	0%	14%	5%	0%	0%	0%	7%	0%	0%
PR	4	0	4	2	2	2	2	0	1	3	1	1
	14%	0%	25%	10%	29%	9%	50%	0%	7%	21%	10%	20%
CR/CRI	4	3	1	4	0	0	2	2	0	4	4	4
	14%	25%	6%	19%	0%	0%	50%	100%	0%	29%	40%	80%
SD, PR OR	9	3	6	6	3	3	4	2	1	8	5	5
CR/CRI	32%	25%	38%	29%	43%	14%	100%	100%	7%	57%	50%	100%
NOT EVALUABLE	1	0	1	1	0	1	0	0	0	1	1	0
	4%	0%	6%	5%	0%	5%	0%	0%	0%	7%	10%	0%

The median follow-up for the entire cohort was 1.8 months (range 0.67–87.3 months). In patients who achieved CR/CRi, mean OS was 31.7 months (range 1.8–84.1 months) and median OS was 41.6 months. Patients having responses of PR/SD had a median OS of 23.6 months (range 9–38.3 months). Patients in Arm A had a shorter median follow-up (1.57 months) compared to patients in Arm B (30.5 months).

Correlative Studies

As demonstrated in Supplemental Table 2, measurement of γH2AX by multiparametic flow cytometry demonstrated that neither the tumor cell population (defined by CD34 positivity and/or aberrant or biphenotypic marker expression) nor the T cells, B cells, NK cells, and monocyte subsets exhibited significant changes in the amount of DNA damage induced by the initial D1 dose of entinostat 2 hours post exposure (median 0.96 fold change, range 0.4–1.21 fold). However, after 5 days of clofarabine (as measured at C1D8 pre- and C1D8 2hrs-post entinostat), DNA damage was induced in all cell types, with 2.4-fold (range 0.3 – 14.1 fold; 12/18 ≥ 2.0 fold increase) median increase in γH2AX in the tumor cell population and 2.5-fold (range 0.9 – 6.8) in these immune cell subsets. Thus, robust DNA damage was induced after clofarabine administration despite the relatively low dose (10 mg/m² daily x 5 as compared to FDA-approved dosing of 40 mg/m² daily x5). Two hours after entinostat administration on D8 only modest increases in DNA damage (median 1.16-fold increases for all cell types) were detected beyond what was already induced by one dose of entinosat (C1D1) and 5 daily doses of clofarabine (C1D3-C1D7). These data suggest that entinostat does not antagonize ongoing DNA damage. Notably, among immune subsets DNA damage was highest in the monocyte population post exposure to combination of entinostat and clofarabine (median 1.42 fold increase, range 0.4 – 11.1, with >70% of monocytic populations exhibiting ≥ 1.2-fold increases in γH2AX mean fluorescence intensity). Clofarabine-induced DNA damage was demonstrated for all patients by increased γH2AX median fluorescence intensity (Supplemental Figure 1). As in previous studies ^26,27, there were no consistent differences observed between 13 NR and 5 responders (3 CR/CRi, 1PR, 1SD) in the patterns or extent of the DNA damage induced by clofarabine in any of the cell types. However, in the tumor cell population, responders tended to have higher baseline γH2AX levels. In these patients, D8 entinostat consistently induced additional DNA damage beyond what was present after clofarabine (Figure 2, A–B).

Figure 2.

Evaluation of DNA damage and acetylation status. γH2AX median fluorescence intensity (MFI) in the tumor cell population, (A) non-responders, (B) responders. Acetylated lysine median fluorescence intensity (MFI) in each PBMC immune subset; B cells (C), T cells (D), and NK cells (E). C1D1 pre=cycle 1 day 1 pre entinostat dosing, C1D1 2hr=cycle 1 day 1, 2 hours post entinostat dosing, C1D8 pre=cycle 1 day 8 pre entinostat dosing, C1D8 2hr=cycle 1 day 8, 2 hours post entinostat dosing.

Evaluation of acetylation was performed at similar time points for 18 patients and demonstrated increased acetylation for all immune subsets (T, B and NK cells) (Figure 2, C–E). An attempt was made to evaluate only the population that best represented the leukemia clone specific to each patient. This was performed in 7 patients (5 NR and 2 CR/CRi) and demonstrated increased acetylation in all. The amount and pattern of acetylation was independent of clinical response (data not shown).

MDSCs and monocytes represent two important immunoregulatory cell classes with multiple subsets based on lineage derivation and differentiation and cell surface marker expression. Both cell types infiltrate the tumor microenvironment and expand in response to chronic inflammation. All subsets of functional MDSCs are potently immunosuppressive and potentially directly tumorigenic as well. Surface expression of the co-stimulatory molecule CD40 is central to MDSC immunosuppressive activities and expansion of regulatory T cell populations ³³. Treatment with entinostat on D1 did not appear to have any early effects on MDSC numbers and CD40 expression. On D8, however, after 5 days of clofarabine, significant increases in monocytic MDSCs (Figure 3A) and in their CD40 expression (Figure 3B) were detected, without further increases induced by entinostat. The sole difference between responders and NR among the MDSC populations was a significant increase (p=0.027) in the fold change in monocytic MDSCs (CD11b+, CD33+, CD14+, HLA-DR-) on D8 from D1 in responders (Figure 3C). This was not accompanied by changes in CD40 expression (Figure 3D).

Figure 3.

Changes induced in monocytic MDSCs. (A) Changes in the frequency of monocytic MDSCs among viable PBMCs, (B) Changes in CD40 expression (MFI) on monocytic MDSCs, (C) Fold change of monocytic MDSCs at C1D8 from C1D1 pre, (D) Fold change of CD40 expression on monocytic MDSCs at C1D8 from C1D1 pre. C1D1 pre=cycle 1 day 1 pre entinostat dosing, C1D1 2hr=cycle 1 day 1, 2 hours post entinostat dosing, C1D8 pre=cycle 1 day 8 pre entinostat dosing, C1D8 2hr=cycle 1 day 8, 2 hours post entinostat dosing. NR=nonresponder, CR=Complete response, PR=partial response and SD=stable disease, MDSC=myeloid-derived suppressor cell.

A different pattern was observed for monocyte subsets and accompanying HLA-DR expression. Within the overall monocyte population, there are multiple subpopulations characterized by differing intensities of CD14 and CD16 surface expression. Classical monocytes exhibit strong CD14 expression (CD14++) without CD16 expression, while non-classical monocytes fall within a spectrum of CD14+ CD16++ expression ³⁴. These subpopulations have distinctive immunologic effects, with classical monocytes typically being immunosuppressive and non-classical monocytes promoting antitumor immunity. As with MDSCs, the initial administration of entinostat did not induce early changes in monocyte numbers. On D8, however, while there was no change in the frequency of total monocytes relative to D1 (Figure 4A), there was a decrease in the percentage of the classical subpopulation (p=0.02) (Figure 4B) and a concomitant increase in the percentage of the non-classical subpopulation (p=0.011) (Figure 4C). Furthermore, on D8 responders had a higher level of HLA-DR expression on non-classical monocytes than non-responders (Figure 4D).

Figure 4.

Changes induced in monocytes. (A,B,C) Changes in the frequency of total monocytes (A), classical monocytes (B), nonclassical monocytes (C) among viable PBMCs and (D) HLA-DR expression (MFI) on nonclassical monocytes at C1D8. C1D1 pre=cycle 1 day 1 pre entinostat dosing, C1D1 2hr=cycle 1 day 1, 2 hours post entinostat dosing, C1D8 pre=cycle 1 day 8 pre entinostat dosing, C1D8 2hr=cycle 1 day 8, 2 hours post entinostat dosing.

DISCUSSION

Our study demonstrates that the combination of entinostat and clofarabine has clinical activity in adult patients diagnosed with ALL/ABL with poor-risk features. Furthermore, the regimen is well tolerated and, for those who received multiple cycles, without cumulative toxicity. The majority of infectious and non-infectious, non-hematologic serious adverse events occurred during the initial cycle. Grade 3 toxicities appeared to be more frequent in the older patient cohort possibly due to greater tumor burden, greater number of pretreatment comorbidities and underlying organ dysfunction. Unlike treatment with streoids and vinca-alkyloids this regimen does not induce hyperglyclemia or peripheral neuropathy. The MTD for entinostat when given with clofarabine at 10 mg/m² daily for 5 days was 8 mg orally once daily on D1/D8. The clofarabine dose is one quarter of the approved single-agent dose, thus allowing the regimen to be managed on an outpatient basis, particularly in the subsequent cycles. Responses including CRs occurred at all entinostat dose levels and occurred more often in patients with newly diagnosed disease who were treatment naive. Responses were durable, and one patient remains alive in first complete remission.

Clofarabine induces DNA damage in tumor cell populations ^26,27. We assessed if HDAC inhibition with entinostat could directly cause DNA damage or impact the effects of clofarabine. Data from this study substantiate previous findings with regard to the ability of clofarabine to damage DNA significantly in all cell subpopulations including tumor cells in both responders and NR and support the notion that induction of DNA damage is necessary but not sufficient to achieve a clinically meaningful response. In contrast, entinostat given alone on D1 had negligible short-term effects on baseline levels of DNA damage and induced only modest increases in DNA damage in all cell types following clofarabine administration.

In the ENCORE 301 phase II trial we performed flow cytometric evaluation of acetylated lysine in of PBMCs in breast cancer patients treated with exemestane with or without entinostat and reported that a lysine acetylation signature was associated with response to entinostat therapy ³⁰. The antibody-based acetylation assay used in the ENCORE 301 study and the present study does not employ an antibody to a specific lysine residue, but rather is directed at all proteins with acetylated lysine (global acetylation), thus capturing both chromatin-associated proteins as well as the other nuclear proteins and proteins in other cell compartments including cytoplasmic proteins.Accordingly, we explored the current trial for an acetylation signatiure and response to therapy similar to that seen in the ENCORE 301. Importantly, increased acetylation was found in both responders and NR with no correlation to response (Figure 2). Numerous differences exist between these trials, including the drugs used in combination with entinostat, dissimilar tumor populations, and different time points analyzed. It is possible that a Class 1 selective HDACi may be optimal for antitumor activity in relation to specific clinical settings (e.g. co-administered with an aromatase inhibitor versus cytotoxic chemotherapy), as opposed to a pan-inhibitor.

Lastly, we explored the impact of entinostat and clofarabine on the viability of selected immune cell subsets including MDSC populations with analysis of the functional marker CD40, and three monocyte populations including analysis of HLA-DR expression. Our findings in peripheral blood MDSCs in R/R ALL demonstrated increases in monocytic MDSC frequency and level of CD40 expression on D8 following D7 clofarabine therapy in lineage-negative, immature, and monocytic subsets, without reversal following entinostat administration (D8, 2 hrs post-entinostat). An increase in the MDSC population was associated with response. . While the expectation would be that the MDSC population would fall in responders, Orillion et al ³⁵ have previously demonstrated functional neutralization of MDSCs by entinostat.

It should be noted that this study is examining MDSCs in peripheral blood and not in bone marrow. Although the details are not clear it appears that MDSCs primarily arise in bone marrow where they can be induced to undergo expansion, potentially in response to growth factors and proinflammatory mediators. They can then enter peripheral blood and accumulate within the tumor microenvironment, where activating signals from proinflammatory stimuli promote the enhancement of suppressor function. In addition to their prominent role in suppressing effector CD8 T cells, MDSCs impact other immune cells, remodel the tumor microenvironment and promote angiogenesis, tumor progression and metastasis in response to factors including VEGF, basic FGF, and matrix metalloproteinase-9. Additionally, intratumorally, MDSCs display metabolic plasticity in response to environmental stimuli. Thus, MDSCs in PB, in bone marrow and in various tumor sites are in very different regulatory environments and are likely to vary in phenotype and function. ^36,37 In the current study, while there was no apparent change in the total numbers of PB monocytes with therapy, a shift in monocyte subpopulations was noted from the classical (CD14++CD16-) subset which has immunosuppressive activities to the CD14+CD16++ nonclassical subset which promotes antitumor immunity. Similar to the ENCORE 301 study, increased expression of HLA-DR was detected in monocytes on D8 following clofarabine and also post-entinostat. Of note, HLA-DR expression on non-classical monocytes was higher in responders than in NR (for both D8 time points), consistent with the notion that response may relate at least in part to induction of an immunostimulatory milieu by entinostat and clofarabine.

Since the conception and conduct of this trial, diverse immunotherapeutic targeted drugs have been developed and approved or are under investigation and have shown exceptional single agent promise with tolerable and in some cases unique toxicites. These novel agents demonstrate important activity in both ND and R/R disease and, in the case of blinatumomab, the minimal residual disease setting. Nonetheless, development of resistance, including loss of target antigen expression has been associated with these targeted molecules, limiting their long-term applicability. Furthermore, traditional therapies, with which the novel agents are often combined, are accompanied by serious toxicity profiles, particularly in older adults including steroid-induced glucose intolerance, bone loss and myopathy and vincristine-induced neuropathies. ^38,39 Therefore, there is still a need to pursue alternative therapeutic modalities for adult ALL, particularly in subgroups for which outcomes remain poor.

In summary, the combination of entinostat and clofarabine exhibits clinical activity in older adults with newly diagnosed, previously untreated Ph-negative ALL/ABL. These results support development of a Phase II trial with intensive biomarker analysis. The absence of limiting toxicities including hyperglycemia and peripheral neuropthy and the ability to deliver this regimen in an outpatient setting suggests a role of combined entinostat plus clofarabine in the maintenance setting could be considered.

Highlights.

1. The clinical outcome for older adults with Acute Lymphoblastic Leukemia or Acute Biphenotypic Leukemia ineligible for multi-agent chemotherapy are poor and are a rare unmet need.

2. The regimen of entinostat plus clofarabine was well tolerated with infectious and metabolic derangements more commonly in the older population versus the younger cohort, but without hyperglycemia or peripheral neuropathy.

3. Entinostat plus clofarabine appears to be active in older adults with de novo ALL/ABL, and further study of this combination should be considered.