Initial Testing (Stage 1) of the histone deacetylase inhibitor, quisinostat (JNJ-26481585), by the Pediatric Preclinical Testing Program

Abstract

Background

Quisinostat (JNJ-26481585) is a second generation pyrimidyl-hydroxamic acid histone deacetylase (HDAC) inhibitor with high cellular potency towards class I and II HDACs. Quisinostat was selected for clinical development as it showed prolonged pharmacodynamic effects in vivo and demonstrated improved single agent antitumoral efficacy compared to other analogs.

Procedures

Quisinostat was tested against the PPTP in vitro panel at concentrations ranging from 1.0 nM to 10 μM and was tested against the PPTP in vivo panels at a dose of 5 mg/kg (solid tumors) or 2.5 mg/kg (ALL models) administered intraperitoneally daily x 21.

Results

In vitro quisinostat demonstrated potent cytotoxic activity, with T/C% values approaching 0% for all of the cell lines at the highest concentration tested. The median relative IC₅₀ value for the PPTP cell lines was 2.2 nM, (range <1 nM to 19 nM). quisinostat induced significant differences in EFS distribution compared to control in 21 of 33 (64%) of the evaluable solid tumor xenografts and in 4 of 8 (50%) of the evaluable ALL xenografts. An objective response was observed in 1 of 33 solid tumor xenografts while for the ALL panel, two xenografts achieved complete response (CR) or maintained CR, and a third ALL xenograft achieved stable disease.

Conclusions

Quisinostat demonstrated broad activity in vitro, and retarded growth in the majority of solid tumor xenografts studied. The most consistent in vivo activity signals observed were for the glioblastoma xenografts and T-cell ALL xenografts.

INTRODUCTION

Transcription of eukaryotic genes is tightly regulated by a complex interplay between DNA methylation and covalent histone modifications (acetylation, methylation, phosphorylation) [1]. The acetylation of lysine residues on histones H3 and H4 modifies the structure of nucleosomes leading to an open chromatin configuration, which facilitates gene transcription. The level of acetylation of histones is the result of dynamic interplay between histone acetyl transferases (HATs) and histone deacetylases (HDACs). Aberrant regulation of the acetylation/deacetylation equilibrium is considered to play a role in generating the abnormal gene expression patterns observed in many forms of cancer [2,3]. For example, the pathogenesis of acute promyelocytic leukemia (APL) and its responsiveness to treatment with all-trans retinoic acid is based on the recruitment by PML-RARα of an HDAC nuclear co-repressor complex to target genes [4].

There are 18 described HDACs distributed in four classes (including the sirtuins). Class I encompasses HDAC1, HDAC2, HDAC 3 and HDAC8. The activity of these HDACs is considered crucial for cell proliferation [5] and is frequently augmented in cancer [6]. Mice lacking Class I HDACs show a range of defects in embryogenesis, particularly of the vasculature, and are not viable [5]. Inhibition of HDAC6 leads to hyperacetylation of tubulin and HSP90, with the latter effect resulting in reduced activity and/or levels of HSP90 client proteins [7,8].

The effects of HDAC inhibition can be broadly grouped as follows: a) induction of apoptosis; b) modification of cell cycle (either G1 or G2/M phase arrest); c) differentiation induction; d) changes in angiogenesis and the tumor microenvironment; and e) production of reactive oxygen species [9]. HDAC inhibition induces a plethora of gene expression changes in cancer cells that may lead to these effects [10], with increases in levels of the cell cycle regulator p21^waf being commonly reported as a consequence of HDAC inhibition [11,12]. Reports on the changes induced on gene expression upon histone deacetylase inhibitor (HDI) exposure range from 2% to more than 20% of genes [13,14]. Based on the relaxation of nucleosomes and increased transcription due to histone hyperacetylation, increases in gene expression would be the expected response to HDAC inhibition. There are, however, numerous examples of reduced expression of genes [15], reflecting the complex factors affecting the equilibrium established by HATs and HDACs on their gene targets. Furthermore, besides the lysine residues on histones there are a number of other proteins which can undergo structural and functional changes as a consequence of acetylation/deacetylation (reviewed in [16]), including transcription factors such as p53, Hif1a, Foxo proteins, p65, and STAT proteins.

HDIs are diverse in chemical structure including short chain fatty acid moieties, hydroxamic acid derived compounds, and cyclic oligopeptides. A large number of HDIs have been or are currently under investigation as therapeutic options for cancer [17]. The hydroxamic acid family of drugs (i.e. trichostatin A, vorinostat (SAHA), panobinostat, oxamflatin) is the larger group and many of its members are pan-HDIs (reviewed in [11]). Two HDIs, vorinostat and romidepsin, have received FDA approval for treatment of cutaneous T cell lymphoma [18,19].

Quisinostat is an orally available, second generation pyrimidyl-hydroxamic acid derivative HDAC inhibitor that has completed its phase 1 evaluation in adults with cancer [20]. Quisinostat was selected for development from among 140 candidate HDAC inhibitors based on its prolonged in vivo pharmacodynamic response against an ovarian cancer model [21]. JNJ-26481858 shows relative selectivity for HDAC1 and HDAC2 in enzymatic assays, but in cellular assays it shows broad spectrum activity against both Class I (e.g., HDAC1-3) and Class II (e.g., HDAC6) isozymes [21]. It is approximately 500-fold more potent than vorinostat at inhibiting HDAC1 [21]. Quisinostat potently induced apoptosis against leukemia cell lines [22], and it induced tumor growth inhibition as well as regressions against multiple adult cancer xenografts models [21,23]. Quisinostat was selected for systematic testing by the PPTP based on its improved preclinical efficacy relative to other HDIs and on the potential relevance of HDAC inhibition to the treatment of childhood cancers. This report describes testing of quisinostat as a single agent against the PPTP’s in vitro panel as well as against the in vivo tumor panel.

MATERIALS AND METHODS

In vitro testing

In vitro testing was performed using DIMSCAN, as previously described in a characterized panel of 23 cell lines [24]. Cells were incubated in the presence of quisinostat for 96 hours at concentrations from 1 nM to 10 μM and analyzed as previously described [24].

In vivo tumor growth inhibition studies

CB17SC scid^−/− female mice (Taconic Farms, Germantown NY), were used to propagate subcutaneously implanted kidney/rhabdoid tumors, sarcomas (Ewing, osteosarcoma, rhabdomyosarcoma), neuroblastoma, and non-glioblastoma brain tumors, while BALB/c nu/nu mice were used for glioma models, as previously described [25,26]. Human leukemia cells were propagated by intravenous inoculation in female non-obese diabetic (NOD)/scid^−/− mice as described previously [27]. Female mice were used irrespective of the patient gender from which the original tumor was derived. All mice were maintained under barrier conditions and experiments were conducted using protocols and conditions approved by the institutional animal care and use committee of the appropriate consortium member. Eight to ten mice were used in each control or treatment group. Tumor volumes (cm³) [solid tumor xenografts] or percentages of human CD45-positive [hCD45] cells [ALL xenografts] were determined as previously described [25]. Responses were determined using three activity measures as previously described [25]. An in-depth description of the analysis methods is included in the Supplemental Response Definitions section.

Statistical Methods

The exact log-rank test, as implemented using Proc StatXact for SAS®, was used to compare event-free survival distributions between treatment and control groups. P-values were two-sided and were not adjusted for multiple comparisons given the exploratory nature of the studies.

Pharmacodynamic studies

Rhabdomyosarcoma xenografts were harvested prior to treatment, or at 4, 8 and 24 hr after dose 1 (5 mg/kg). Additional tumors were harvested 24 hr post day 3 dosing and at 4 hr post day 4 dosing. Samples were prepared for immunoblotting as previously described [28]. Membranes were probed for acetylated Histone H3 and H4, p21 PARP and cleaved PARP. ALL-8 engrafted mice were treated with quisinostat at 2.5 mg/kg daily, and spleen cells (>90% huCD45) were harvested at days 0 (no treatment), 1, 3 and 7 and were analyzed for acetylated histone H4. Loading was normalized to GAPDH (solid tumors) or actin (ALL-8).

Drugs and Formulation

Quisinostat was provided to the Pediatric Preclinical Testing Program by Janssen Pharmaceutica, through the Cancer Therapy Evaluation Program (NCI). Powder was stored at room temperature, protected from light. Drug was formulated in 10% hydroxy-propyl-β-cyclodextrin, 25 mg/mL mannitol, in sterile water for injection, and made fresh prior to administration. Quisinostat was administered intraperitoneally (IP) to mice using a daily schedule. Based upon toxicity testing (non-tumored mice) a dose of 5 mg/kg was selected for the solid tumors, and at a reduced dose of 2.5 mg/kg was used for the ALL panel, for 21 days. Quisinostat was provided to each consortium investigator in coded vials for blinded testing.

RESULTS

In vitro testing

Quisinostat demonstrated potent cytotoxic activity, with T/C% values approaching 0% for all of the cell lines at the highest concentration tested. The median relative (Rel) IC₅₀ value for the PPTP cell lines was 2.2 nM, with a range from <1 nM (MOLT-4, CHLA-9, and CHLA-258) to 19 nM (NB-EBc1), Table I. The ALL cell lines tended to be more sensitive to quisinostat (median Rel IC₅₀ of 1.9 nM) and the rhabdomyosarcoma and neuroblastoma cell lines less sensitive (median Rel IC₅₀ of 5.1 and 6.8 nM, respectively), but these differences were not significant.

Table I.

In vitro activity of JNJ-26481585 against PPTP cell lines.

Cell Line	Histotype	Rel IC50 (nM)	Panel Rel IC50/Line Rel IC50	Ymin% (Observed)	p53 Mutation Status
RD	Rhabdomyosarcoma	7.2	0.30	0.02	R248W
Rh41	Rhabdomyosarcoma	3.0	0.73	0.02	Del1001–1013
Rh18	Rhabdomyosarcoma	11.5	0.19	0.08	WT
Rh30	Rhabdomyosarcoma	2.1	1.02	0.00	R273C
BT-12	Rhabdoid	9.5	0.23	0.03	A88T
CHLA-266	Rhabdoid	1.9	1.18	0.10	A88T
TC-71	Ewing sarcoma	3.2	0.69	0.00	R213(Stop)
CHLA-9	Ewing sarcoma	<1.0	>2.21	0.00	??
CHLA-10	Ewing sarcoma	2.5	0.88	0.00	??
CHLA-258	Ewing sarcoma	<1.0	>2.21	0.00	WT
SJ-GBM2	Glioblastoma	1.6	1.40	0.11	R273C
NB-1643	Neuroblastoma	3.8	0.57	0.11	WT
NB-EBc1	Neuroblastoma	19.0	0.12	0.00	WT
CHLA-90	Neuroblastoma	1.1	2.01	1.46	E286K
CHLA-136	Neuroblastoma	9.8	0.22	0.56	WT
NALM-6	ALL	2.2	0.98	0.00	WT
COG-LL-317	ALL	1.3	1.75	0.00	WT
RS4;11	ALL	3.2	0.69	0.00	WT
MOLT-4	ALL	<1.0	>2.21	0.01	WT
CCRF-CEM (1)	ALL	3.1	0.71	0.00	R248Q, R175H
CCRF-CEM (2)	ALL	1.6	1.33	0.00	R248Q, R175H
Kasumi-1	AML	1.1	2.02	0.00	R248Q
Karpas-299	ALCL	1.9	1.16	0.09	R273C
Ramos-RA1	NHL	1.4	1.57	0.00	I254D
Median		2.2	1.00	0.00
Minimum		<1.0	0.12	0.00
Maximum		19.0	>2.21	1.46

In vivo testing

Quisinostat was tested in vivo using a 5 mg/kg dose for solid tumors and 2.5 mg/kg for ALL xenografts, administered intraperitoneally daily for 3 weeks. Seventeen of 779 mice died during the study (2.2%), with 7 of 390 in the control arms (1.8%) and 10 of 389 (2.6%) in the quisinostat treatment arms. All 41 xenograft models studied were considered evaluable for efficacy. A complete summary of results is provided in Supplemental Table I, including total numbers of mice, number of mice that died (or were otherwise excluded), numbers of mice with events and average times to event, tumor growth delay, as well as numbers of responses and T/C values.

Quisinostat induced significant differences in EFS distribution compared to control in 21 of 33 (64%) evaluable solid tumor xenografts, and in 4 of 8 (50%) of the evaluable ALL xenografts (Table II). Quisinostat induced tumor growth inhibition meeting criteria for intermediate EFS T/C activity (EFS T/C > 2) in 6 of 32 (19%) evaluable solid tumor xenografts. The only solid tumor panel with two xenografts with EFS T/C > 2 was the glioblastoma panel, with both D645 and D456 showing substantial delay in time to event (EFS T/C of 4.5 and >6.7, respectively). For the ALL panel, 3 of 8 (38%) xenografts met criteria for intermediate activity, while one T lineage xenograft (ALL-16) met criteria for high activity.

Table II.

Summary of in Vivo Activity of JNJ-26481585

Line	Tumor Type	Time to Event	P-value	EFS T/C	Median Final RTV	Tumor Volume T/C	T/C Volume Activity	EFS Activity	Response Activity
BT-29	Rhabdoid	31.4	<0.001	2.3	>4	0.47	Low	Int	Int
KT-11	Wilms	12.9	0.007	1.4	>4	0.64	Low	Low	Low
KT-13	Wilms	18.1	<0.001	2.1	>4	0.32	Int	Int	Int
SK-NEP-1	Ewing	12.9	<0.001	2.1	>4	0.55	Low	Int	Int
EW5	Ewing	14.5	0.004	1.2	>4	0.73	Low	Low	Low
TC-71	Ewing	6.3	0.994	1.1	>4	0.81	Low	Low	Low
CHLA258	Ewing	10.1	0.807	1.2	>4	0.91	Low	Low	Low
Rh28	ALV RMS	> EP	<0.001	> 1.3	0.0	0.22	Int	NE	High
Rh30	ALV RMS	12.7	0.232	1.3	>4	0.62	Low	Low	Low
Rh30R	ALV RMS	23.4	<0.001	1.8	>4	0.43	Int	Low	Int
Rh41	ALV RMS	13.8	0.071	0.8	>4	1.39	Low	Low	Low
Rh18	EMB RMS	15.9	0.008	2.1	>4	0.60	Low	Int	Int
BT-28	Medulloblastoma	11.6	0.021	1.6	>4	0.68	Low	Low	Int
BT-45	Medulloblastoma	6.9	0.207	1.1	>4	0.78	Low	Low	Low
BT-44	Ependymoma	11.2	<0.001	1.3	>4	0.26	Int	Low	Low
BT-35	Astrocytoma	6.1	<0.001	1.7	>4	0.40	Int	Low	Int
BT-40	Astrocytoma	6.0	0.281	1.0	>4	1.06	Low	Low	Low
GBM2	Glioblastoma	15.1	0.355	1.8	>4	0.84	Low	Low	Int
BT-39	Glioblastoma	10.8	0.632	1.1	>4	0.84	Low	Low	Low
D645	Glioblastoma	37.0	<0.001	4.5	>4	0.40	Int	Int	Int
D456	Glioblastoma	> EP	<0.001	> 6.7	1.7	0.40	Int	Int	Int
NB-SD	Neuroblastoma	10.4	0.031	1.2	>4	0.76	Low	Low	Low
NB-1771	Neuroblastoma	19.8	0.055	1.5	>4	0.78	Low	Low	Low
NB-1691	Neuroblastoma	10.9	0.017	1.4	>4	0.81	Low	Low	Low
NB-EBc1	Neuroblastoma	15.7	<0.001	1.3	>4	0.53	Low	Low	Low
CHLA-79	Neuroblastoma	9.6	0.801	1.0	>4	1.02	Low	Low	Low
NB-1643	Neuroblastoma	6.6	0.893	1.1	>4	1.01	Low	Low	Low
OS-1	Osteosarcoma	32.5	0.054	1.3	>4	0.88	Low	Low	Low
OS-2	Osteosarcoma	32.7	<0.001	1.3	>4	0.53	Low	Low	Low
OS-17	Osteosarcoma	20.2	0.001	1.1	>4	0.78	Low	Low	Low
OS-9	Osteosarcoma	34.4	<0.001	1.5	>4	0.56	Low	Low	Low
OS-33	Osteosarcoma	19.5	0.029	1.3	>4	0.83	Low	Low	Low
OS-31	Osteosarcoma	26.2	0.004	1.3	>4	0.65	Low	Low	Low
ALL-2	ALL B-precursor	32.9	0.004	1.9	>25	.		Low	Int
ALL-3	ALL B-precursor	38.5	<0.001	4.0	>25	.		Int	Int
ALL-4	ALL B-precursor	4.5	0.612	0.9	>25	.		Low	Low
ALL-7	ALL B-precursor	4.8	0.029	0.7	>25	.		Low	Low
ALL-8	ALL T-cell	> EP	0.019	> 4.6	23.6	.		Int	High
ALL-16	ALL T-cell	> EP	0.103	> 2.5	0.2	.		High*	High
ALL-17	ALL B-precursor	19.1	0.025	2.4	>25	.		Int	Int
ALL-19	ALL B-precursor	3.0	0.467	1.0	>25	.		Low	Low

^*High EFS T/C is called despite p>0.05 resulting from irregular growth in control animals.

An objective response was observed in 1 of 33 solid tumor xenografts: Rh28 in the rhabdomyosarcoma panel achieved a maintained complete remission (MCR). For the ALL panel, two xenografts achieved CR and MCR, respectively, and a third xenograft achieved stable disease. The two ALL xenografts achieving complete responses (ALL-8 and ALL-16) are both T-cell immunophenotype, and the B-precursor ALL xenograft achieving stable disease (ALL-3) has an MLL gene rearrangement. The in vivo testing results for the objective response measure of activity are presented in Figure 1 in a ‘heat-map’ format as well as a ‘COMPARE’-like format, based on the scoring criteria described in the Supplemental Response Definitions section. The latter analysis demonstrates relative tumor sensitivities around the midpoint score of 5 (stable disease). Examples of the solid tumor xenografts showing either regression (Rh28) or extended tumor growth delay (D456 and D645) are shown in Figure 2. Examples of relative leukemia growth curves for ALL xenografts meeting criteria for intermediate or high activity are shown in Figure 3.

Figure 1.

Quisinostat in vivo objective response activity, left: The colored heat map depicts group response scores. A high level of activity is indicated by a score of 6 or more, intermediate activity by a score of ≥2 but <6, and low activity by a score of <2. Right: representation of tumor sensitivity based on the difference of individual tumor lines from the midpoint response (stable disease). Bars to the right of the median represent lines that are more sensitive, and to the left are tumor models that are less sensitive. Red bars indicate lines with a significant difference in EFS distribution between treatment and control groups, while blue bars indicate lines for which the EFS distributions were not significantly different.

Figure 2.

Quisinostat activity in vivo against individual solid tumor xenografts. Kaplan-Meier curves showing the probability for EFS (left), relative tumor volume (center), and individual tumor volumes (right) graphs are shown for selected xenografts (Rh28, D456, and D645). Controls (gray lines); Treated (black lines), significances of the difference between treated and control groups are included.

Figure 3.

Quisinostat activity in vivo against individual ALL xenografts. Kaplan-Meier curves showing the probability for EFS (left), median leukemia engraftment (center) as detected in peripheral blood (see Materials and Methods), and individual leukemia engraftment (right) graphs are shown for selected xenografts (ALL-2, ALL-3, ALL-8, and ALL19). Controls (gray lines); Treated (black lines), significances of the difference between treated and control groups are included.

Pharmacodynamic studies

To determine whether at the doses and schedule used quisinostat achieved target inhibition, we examined tumor tissue from 3 rhabdomyosarcoma models with differing responses and ALL-8 that had a complete response to treatment (Supplemental Figure 1). Three tumors were assayed at each time point to determine whether there was a drug induced increase in acetylated histones H3 and H4, and p21CIP1. Apoptosis was monitored by assessing cleaved and total PARP. In the most sensitive model, Rh28, there was a marked increase in acetylated H3 4 and 8 hr after dose 1. H3 and H4 acetylation was elevated compared to controls 24 hr after daily dose 3. p21CIP increase paralleled that of acetylated histones, and cleaved PARP was detected in each Rh28 tumor 4 hr after the fourth dose. In the non-responsive Rh30 model increases in H3 and H4 acetylation were similar to those determined in Rh28 xenograft, although p21 induction was seen only after 8 hr, and cleaved PARP was not detected. For the other non-responsive model, Rh41, there was a robust increase in H3 and H4 acetylation at 4 and 8 hr after dose 1 but was deminshed in the 96 and 100 hr samples. Levels of p21CIP1 increased marginally, if at all. For the ALL-8 model, tumor from spleen was analyzed 4 hr after dosing on days 1, 3, and 7. As shown in Supplemental Figure 1D, there was a cumulative increase in acetylated H4 histone.

DISCUSSION

Interest in HDIs and their potential use for cancer treatment has grown substantially in the recent years. In addition to their single agent effects, HDIs have demonstrated sensitization of cancer cells both to radiation and chemotherapeutic treatments [29,30]. A number of reports have described the preclinical activity of HDIs against pediatric cancers, including reports for romidepsin [28,31,32], vorinostat [33,34], and etinostat [35]. Pediatric phase 1 trials of vorinostat and romidepsin have been conducted [36,37], and frontline studies of vorinostat given with radiation therapy have been developed for children with high grade gliomas and diffuse intrinsic pontine glioma. The emerging themes from the prior preclinical reports are that HDIs are strong inducers of apoptosis when tested in vitro, with potency ranging from low micromolar IC₅₀ values for vorinostat to low nanomolar IC₅₀ values for romidepsin. However, in vivo activity for HDIs against pediatric xenografts has been limited. For example, vorinostat induced no objective responses among 30 evaluable solid tumor and 8 evaluable ALL xenografts [33]. Romidepsin at clinically relevant doses induced regressions in 3 of 39 pediatric solid tumor models, with several other xenografts showing prolonged stable disease [28].

Quisinostat demonstrated potent cytotoxic activity in vitro, with T/C% values approaching 0% for all of the cell lines at the highest concentration tested. The 2.2 nM median IC₅₀ value for quisinostat was far lower than the IC₅₀ value observed for the PPTP in vitro cell line panel for vorinostat (1.4 μM) [33]. One of the most sensitive cell lines was MOLT-4, a T-cell ALL cell line, which is consistent with the in vivo testing finding of greatest sensitivity for T-cell ALL xenografts. However, there were only modest differences in drug sensitivity in vitro among the cell lines tested, which represent a range of cancer types. The p53 mutation status showed no relationship with sensitivity to quisinostat, as approximately equal numbers of mutant p53 cell lines had IC₅₀ values above and below the median for the entire panel. These results are consistent with recent reports showing no clear relationship between p53 mutation status and response of pediatric preclinical models to HDIs [28,31].

The treatment of engrafted mice, at the selected doses was well tolerated and toxicity levels were low (no models were excluded because of toxicity). The ALL panel was treated at a reduced dose with respect to the solid tumor panel, as the dose of 5 mg/kg for 21 days exceeded the MTD for non-engrafted NOD/SCID mice. This dose reduction does not seem to have limited the efficacy of the treatments with respect to that observed for the solid tumors.

Promising results obtained with quisinostat for cell line in vitro testing were only partially confirmed when efficacy was tested in vivo against the pediatric cancer xenograft panel, with only one objective response among 33 solid tumor xenografts and 2 objective responses among 8 ALL xenografts observed. We have reported similar discrepancies between in vitro and in vivo findings for other single agents [33,38,39], with the explanation likely reflecting in part the inability in the in vivo setting to achieve the depth and duration of target inhibition produced in vitro. The fact that quisinostat was specifically selected based on its in vivo pharmacodynamic activity suggested that it would overcome this limitation [21], which we have confirmed to some extent. In the case of vorinostat tested as a single agent, there were no xenografts with delays in time to event of two-fold or greater [33], whereas for quisinostat this level of activity was observed in approximately 20% of the solid tumor xenografts. Additionally, objective responses were observed for quisinostat, but not for vorinostat.

The most consistent solid tumor activity signals observed were for the glioblastoma xenograft panel, with both D456 and D645 showing prolonged time to event. These xenografts could be valuable in identifying markers of sensitivity for this histotype. The results for the glioblastoma xenograft D456 are impressive, with control of tumor growth during the 21 days of treatment as well as for an additional 21 days of follow-up. Thus, further preclinical evaluation of the utility of quisinostat for glioblastoma is indicated. The response in the rhabdomyosarcoma xenograft Rh28 was also impressive, but 4 additional rhabdomyosarcoma xenografts showed only limited responses to quisinostat. These results are reminiscent of those obtained for romidepsin, which also induced regressions in a single xenograft (Rh66) from among 10 rhabdomyosarcoma xenografts tested [28]. Of note, Rh28 showed tumor growth delay, but not regression, in response to treatment with romidepsin.

The T-cell ALL responses are particularly interesting, as leukemia burden seemed to increase (as detected in circulation) for a week following initiation of treatment, but thereafter the leukemia cells quickly disappeared and remission was attained. A similar pattern of response was observed for the B-precursor ALL xenograft ALL-3, which showed a clear treatment effect to quisinostat, but failed to meet criteria for objective response. These patterns of blast presence in blood could be indicators of differentiation induction or modification of the expression of homing/attachment molecules which prompt them to exit the bone marrow into peripheral blood. Alteration in the chemokine CXCR4 has been reported by the HDI vorinostat in ALL [40], and a similar effect could help explain our findings. Efficacy testing in vivo on a larger panel of ALL, particularly of T lineage, is required to address the question about a subtype-specific effect that might justify clinical development of JNJ-26481585 for this ALL subtype. Regarding the anti-leukemia effect of quisinostat against the MLL-rearranged (t11;19) ALL-3 xenograft, connectivity map gene expression analyses predicted that HDIs would reverse the leukemia-specific gene signature associated with MLL-AF4 infant ALL [41]. Further evaluation of quisinostat against ALL xenografts with MLL-rearrangement is warranted to develop a more robust dataset testing the role of HDAC inhibition for this high-risk ALL subtype.

At the doses used (5 mg/kg for solid tumors and 2.5 mg/kg for leukemias) quisinostat increased acetylation of histones H3 and H4 in each solid tumor and leukemia model. However, for the rhabdomyosarcoma models induction of acetylated histones was equally robust in responsive as non-responsive models. Similar induction of p21CIP1 was also seen in responsive Rh28 xenografts as in non-responsive Rh30. The only characteristic that distinguished Rh28 xenografts was induction of cleaved PARP after the fourth daily dose. For ALL-8 there was a significant and cumulative increase in acetylated H4 histone with daily dosing. Thus, similar to other studies [28], target inhibition resulting in increased histone acetylation did not correlate with tumor responses.

Single agent activity for quisinostat exceeded that observed for another HDI [33], but nonetheless was restricted to a minority of models. This suggests that identifying biological predictors of response could be a clinically relevant line of future research. Further testing of ALL subtypes such as T-cell ALL and MLL-rearranged ALL is a priority. Additional opportunities arise from the evidence for synergy that this family of drugs has shown with established anticancer agents [30]. While single agent activity tends to be a positive indicator of potential efficacy in combination therapy, the inclusion of this HDI into drug combinations is appealing based on its improved pharmacodynamic profile and the observed effects in vivo. Particularly interesting novel agents that may prove effective in combination with quisinostat in the pediatric setting include hypomethylating agents [42], proteasome inhibitors [43], and pro-apoptotic drugs such as ABT-263, based on the reported changes induced by quisinostat on pro- and anti-apoptotic molecules, and in particular on MCL1 levels [44].