Tasquinimod Is an Allosteric Modulator of HDAC4 survival signaling within the compromised cancer microenvironment

Abstract

Tasquinimod is an orally active antiangiogenic drug that is currently in phase III clinical trials for the treatment of castration-resistant prostate cancer. However, the target of this drug has remained unclear. In this study, we applied diverse strategies to identify the histone deacetylase HDAC4 as a target for the antiangiogenic activity of tasquinimod. Our comprehensive analysis revealed allosteric binding (Kd 10-30 nmol/L) to the regulatory Zn(2+) binding domain of HDAC4 that locks the protein in a conformation preventing HDAC4/N-CoR/HDAC3 complex formation. This binding inhibited colocalization of N-CoR/HDAC3, thereby inhibiting deacetylation of histones and HDAC4 client transcription factors, such as HIF-1α, which are bound at promoter/enhancers where epigenetic reprogramming is required for cancer cell survival and angiogenic response. Through this mechanism, tasquinimod is effective as a monotherapeutic agent against human prostate, breast, bladder, and colon tumor xenografts, where its efficacy could be further enhanced in combination with a targeted thapsigargin prodrug (G202) that selectively kills tumor endothelial cells. Together, our findings define a mechanism of action of tasquinimod and offer a perspective on how its clinical activity might be leveraged in combination with other drugs that target the tumor microenvironment. Cancer Res; 73(4); 1386-99. ©2012 AACR.

Figure 1

A. Chemical structure of tasquinimod (TasQ) (N-ethyl-N-phenyl-5-chloro-1,2-dihydro-4-hydroxy-1-methyl-2-oxo-3-quinoline-carboxamide B. Comparison of anti-cancer efficacy of tasquinimod (10mg/ kg/ day via drinking water) against CWR22-RH human prostate cancer xenografts (n=10/group) growing in castrated hosts initiated immediately at time of tumor inoculation vs. delayed treatment starting on day 49 post tumor inoculation when tumors were 0.250cc in size. Each symbol point is the mean tumor volume at the indicated time. SEM is <15% of the mean for each time point. p<0.05 for all treatment time points beyond day 50 compared to controls. C. Growth inhibitory response of HUVEC and indicated human prostate cell lines to one week treatment with indicated concentration of TasQ. Results are normalized to number of viable cells in control cultures not exposed to TasQ. Asterisk denotes statistical difference (p<0.05) for TasQ treated cells compared to untreated control cells. D. TasQ dose-response inhibition of 3-D endothelial cell (EC) sprouting in fibrin gels over a 7 day (D7) observation period vs. untreated (control) cells.

Figure 2

A. Tas-Q (1µM) and TSA (200nM) prevent lysine deacetylation in H3-histone at position 9 &19 induced by hypoxia in all of the human prostate cancer lines tested. β-actin was used as a loading control. B. Western blots of indicated HDAC in normal human prostate epithelial (i.e., 957E/hTERT) cells and LNCaP, PC-3, and DU-145 human PCs. Vincullin was used as a loading control. C., D.,& E. TasQ(1µM) does not decrease HDAC isotypes expressed by human PC lines (i.e., LNCaP, LAPC-4, and CWR22-R_v1) or HUVECs . β-actin was used as a loading control. In D, western blot of HDAC5 was detected using SuperSignal® West Femto Maximum Sensitivity Substrate.

Figure 3

A. Schematic organization of HDAC4 protein. Numbers refer to amino acid position. Black boxes are known sites of transcription factor binding in N-terminal adapter domain. NLS refers to nuclear localization signal and NES to nuclear export signal. S-P refers to serines which if phosphorylated allows binding to 14-3-3 protein. B. Computer based docking of TasQ to inactive (non N-CoR binding) conformation of regulatory zinc-binding domain (ZRD) within the catalytic domain (amino acids 648-1051) of human HDAC4. Upper magenta color ball is Zn²⁺ in the ZRD and lower magenta ball is Zn²⁺ in the catalytic domain. C. Interaction of TasQ with various amino-acids in the inactive conformation of ZRD. D. Surface plasmon resonance of human full length HDAC4 binding to TasQ immobilized onto a Biacore chip.

Figure 4

A. HDAC4 protein in HDAC4 shRNA2 knock down LNCaP and HUVEC endothelial cells. Number under lane is relative level of expression per cell normalized to untreated control cells. B. HDAC4 shRNA knock-down inhibits sprouting of HUVECs in 3-D assay. C. Upper panel- Growth of control vs. HDAC4 shRNA2 knock-down vs. 1µM TasQ treated LNCaP growth under normoxic conditions and Lower panel- under hypoxic conditions. D. HDAC4 and HIF-1α protein expression in control vs. HDAC4 shRNA2 knock-down vs. knock down restored LNCaPs. β-actin was used as a loading control. E. Tumorigenicity of HDAC4 shRNA2 knock-down vs. restored LNCaPs when xenografted into male nude mice (n=8/ group).

Figure 5

A. SPR determined d TasQ dose-response inhibition of N-CoR binding to HDAC4 immobilized onto a Biacore chip. B. Zn²⁺ dependent binding of full length N-CoR protein binding to recombinant full length HDAC4 protein is prevented by chelation with EDTA. Flag-NCoR used as loading control. C. Binding of recombinant full length HDAC4 binding to full length N-CoR protein in the presence of Zn²⁺ is inhibited by TasQ. GST-HDAC4 used as loading control. TasQ inhibits HDAC4 binding to N-CoR/HDAC3 complexes in HEK-293 cells in both normoxic (D&E) and hypoxic (E) conditions. Flag-HDAC4 used as loading control. F. TasQ inhibits HDAC4 binding to N-CoR/HDAC3 complexes in LNCaPs. Flag-HDAC3 used as loading control.

Figure 6

A. HDAC4 increases and 1µM TasQ decreases HIF-1α protein in HEK-293T cells. β-actin was used as a loading control. B. TasQ lowers level of HDAC4 in the nuclei of LNCaPs in both normoxia and hypoxia. Upper arrow denotes phosphorylated form. β-actin was used as a loading control for total extracts while H3-histone was used as loading control for nuclear extract. C. TasQ lowers endogenous HDAC4 in the nuclei of LNCaPs in hypoxia. H3-histone used as loading control. D. TasQ increases the proportion of HIF-1α which is lysine-acetylated within the nuclei of LNCaPs under hypoxic. Number under lane is the relative level of expression per cell normalized to untreated control cells. β-actin was used as a loading control. E. TasQ inhibits HIF-1α dependent green fluorescent protein (GFP) expression in PC-3 human prostate cancer cells which have stably integrated a HRE/GFP construct in normoxia (signal from 25,000 cells) vs. 24 hr of hypoxia (sign from 1,250 cells). F. TasQ inhibits hypoxia induced HIF-1 α dependent transcriptional stress response of HUVECs detected using a hypoxia cDNA microarray with the results normalized to β-actin. Abreviations for the genes probed are as follows: CA9, carbonic anhydrase IX; CTSD, cathepsin D ; Glut-1, glucose transporter-1; Glut-3, glucose transporter-3; HIF-1, hypoxia induced factor-1α; HIF-2, hypoxia induced factor-2α; MMP2, matrix metalloproteinase 2; MXi-1, Max interactor-1; NDRG, N-myc downstream regulated-1; NDRG2, N-myc downstream regulated-2; PAI-1, plasminogen activator inhibitor-1.

Figure 7

A. Overview of TasQ’s mechanism of action. Under hypoxia/stressful microenvironment conditions, HDAC4/N-CoR/HDAC3 complex binds via HDAC4 to DNA bound transcription factors allowing HDAC3 to deacetylate histones locally repressing basal transcription within the nucleus of endothelial and cancer cells. Also during such stress, HIF-1α accumulates in the nuclei where PCAF acetylates its N-terminal lysines between AA11-21 and at position 674 in its inhibitory domain (ID). HDAC4/N-CoR/HDAC3 complexes bind via HDAC4 to the acetylated lysine 674 in the ID of HIF-1 co-localizing N-CoR/HDAC3 resulting in deacetylation of N-terminal lysines between AA11-21 of HIF-1facilitating binding of p300HAT and HIF-1β needed for formation of HIF-1 complex which via p300 acetylates histones in cell survival/angiogenesis genes stimulating their transcription. TasQ binding allosterically locks ZRD of HDAC4 in an inactive conformation preventing basal gene repression and survival/ angiogenesis gene transcription needed for the angiogenic switch. B. TasQ’s efficacy against a diverse series of human solid cancer xenografts. Results are expressed as percent inhibition of cancer growth based upon comparison of tumor volume (N=5–10 cancer bearing mice per cancer subtype) in TasQ treated (10mg/kg/day via drinking water) vs. untreated mice over a month observation period. C. Therapeutic response of established (0.8cc) MCF-7 human breast cancers growing in mice given two daily IV injections at 56mg/kg of a tumor endothelial targeted cytolytic prodrug (G202) alone and in combination with 10mg/kg/d oral TasQ (N=8 per grouping). Results are presented as relative tumor size normalized to tumor volume at initiation of treatment. p<0.05 for combination (combo) group vs. either monotherapies after day 49.