Preclinical Development of the Class-I–Selective Histone Deacetylase Inhibitor OKI-179 for the Treatment of Solid Tumors

This article is featured in Highlights of This Issue, p. 395

Epigenetic modifications result in changes in gene expression without alteration of DNA sequences and play a key role in cancer progression and resistance to cytotoxic and hormonal cancer therapies (1–3). The activity of epigenetic regulators, such as histone proteins, is manipulated by enzymatic changes; therefore, activity can be altered with pharmacologic inhibitors of specific enzymes (4). Histone acetylation is one mechanism of epigenetic regulation based on the ability of histones to control chromatin organization and switch DNA structure from uncoiled (available for transcription) to densely packed in nucleosomes (unavailable for transcription) and vice versa (4, 5).

Global dysregulation of histone acetylation correlates with tumorigenesis and cancer progression in many hematologic malignancies and solid tumors (1, 6). Histone acetylation is managed by histone acetyltransferases and histone deacetylases (HDACs) that enzymatically add or remove acetyl groups from lysine residues on amino-terminal tails of histones, respectively (5). Acetylation of histones results in increased transcription making HDACs silencers of gene expression. HDAC expression is elevated in many tumor types, including breast and colorectal cancer where overexpression is associated with inferior disease-free and overall survival (1, 7–10).

Human HDACs can be grouped as zinc-dependent [class I (1, 2, 3 and 8), II (4–7, 9–10), and IV (11)] and NAD-dependent (class III; ref. 11). Class I HDACs localize to the nucleus where they are ubiquitously expressed in human tissues (1). Knock down of class I HDACs leads to cell cycle arrest, decreased cell viability, and increased apoptosis in preclinical models of many different solid and hematologic cancers, making class I HDAC inhibitors promising anticancer agents (1, 12). Pharmacologic HDAC inhibitors are active in preclinical models of multiple tumor types where treatment leads to decreased cell proliferation and increased apoptosis (13–15). HDAC inhibitors also lead to decreased angiogenesis, increased cancer cell differentiation, and enhanced cancer cell immunogenicity, including through upregulation of MHC class I and II and natural killer cell activation (16, 17). Many other nonhistone targets of HDACs have been identified, however, the physiologic ramifications of these modifications are not completely characterized (18).

There are currently four FDA-approved HDAC inhibitors (vorinostat, romidepsin, belinostat, and panobinostat) for the treatment of cutaneous T-cell lymphoma (CTCL) and/or peripheral T-cell lymphoma (PTCL) and multiple myeloma (panobinostat) (1, 19, 20). Although active in a subset of hematologic malignancies, the successful clinical development of HDAC inhibitors in solid tumors is perceived to be limited by potency, isoform selectivity, narrow therapeutic indices, and/or a nonoral route of administration (romidepsin and belinostat). The natural product romidepsin has attractive potency and selectivity, but to date attempts to engender oral bioavailability have failed. Similar to romidepsin, largazole is a natural product with potent class-I–selective HDAC inhibition. Although largazole has promising antitumor activity in preclinical solid tumor cancer models, it is not optimized for oral administration (21–23). Therefore, we initiated a multidisciplinary program with a goal to optimize novel largazole analogs discovered previously (24, 25) with superior drug-like properties and expand clinical applications toward both hematologic and nonhematologic malignancies.

Here, we report the structures of largazole-derivatives OKI-006, OKI-005, and OKI-179. OKI-005 and OKI-179 are prodrugs metabolized in cells and in vivo, respectively, to OKI-006 which selectively inhibits class I HDACs with no significant inhibition of class IIa HDACs (26). We present the pharmacokinetic and pharmacodynamic profiles of these drugs in vitro, in vivo, and in efficacy studies in preclinical models of colorectal cancer and breast cancer. Based on the results presented herein, OKI-179 was selected as a clinical development candidate and is currently being evaluated in a phase I dose-escalation clinical trial (NCT03931681, Graphical Abstract).

OKI-005 was dissolved in DMSO to make a 10-mmol/L stock solution in media for in vitro experiments. OKI-005 was dissolved in corn oil, OKI-179 was dissolved in 0.1 M citric acid, and entinostat was dissolved in 0.5% carboxymethylcellulose sodium (CMC-Na) for in vivo experiments. 4-hydroxytamoxifen was dissolved in DMSO to make a 30-mmol/L stock solution for in vitro experiments and tamoxifen was dissolved in peanut oil for in vivo experiments.

For pharmacokinetic studies, male BALB/c nude mice (6–8 weeks) were inoculated subcutaneously in the right flank with HCT-116 cells (5 × 10⁶) in 0.1-mL PBS. When average tumor size reached 200 mm³, mice were treated with a single dose of OKI-005 10, 30, 100 mg/kg by oral gavage or 5 mg/kg intraperitoneally (i.p.) and plasma samples were obtained at 0.5, 1, 2, 4, and 8 hours postdose. Animals were euthanized, and tumor samples were obtained at the same timepoints and 100 mg of tumor tissue was flash frozen for pharmacokinetic (PK) analysis. Plasma and tumor concentrations of the metabolite OKI-006 were determined by Liquid chromatography–mass spectrometry? in at least three individual mice/tumors and concentration-time curves were generated.

For OKI-005 efficacy studies, when the average tumor size reached 200 mm³, BALB/c nude mice were randomized to treatment groups: vehicle control (corn oil), OKI-005 10, 30, or 100 mg/kg orally (p.o.), OKI-005 5 mg/kg i.p., or entinostat 20 mg/kg p.o. daily for 14 days. Body weight was monitored daily, and the dose was reduced in the entire treatment group by 50% for more than 8% body weight loss and stopped for more than 15%. To determine the effect of OKI-005 on tumor metastasis, luciferase-labeled MDA-MB-231 cells (1 × 10⁶/injection) were engrafted into the mammary fat pad, and BALB/c mice were randomized to vehicle or OKI-005 10 mg/kg i.p. three times per week beginning 5 weeks postengraftment. Bioluminescent imaging was used to quantify pulmonary metastasis in the two groups. Briefly, mice were injected with D-luciferin, anesthetized with isofluorane, and imaged 5 minutes after injection using the In Vivo Imaging System (IVIS) Spectrum (Perkin Elmer). For OKI-179 efficacy studies, when average tumor size reached 200 mm³, mice were randomized to vehicle control (0.1 M citric acid), OKI-179 40 or 80 mg/kg p.o. daily or 120 mg orally every other day. Tumor size was measured twice weekly using calipers, and tumor volume was calculated [V = (length × width²) × 0.52]. Tumor samples were collected for pharmacodynamic analysis from nude mice with MDA-MB-231–implanted tumors treated with OKI-179 40 mg/kg, 60 mg/kg, or 80 mg/kg for 3 to 10 days. A one-way ANOVA was performed to compare tumor volume and tumor weight among groups; comparisons between groups were carried out with Games–Howell test using SPSS 17.0.

Human triple-negative breast cancer (TNBC) cell lines MDA-MB-231, MDA-MB-436, MDA-MB-157, HCC1806, HCC1937, Hs578T, HCC38, BT549, HCC1187, SW527, HCC1143, and HCC1395 were obtained from ATCC. CAL-51, CAL-85–1, CAL-120, CAL-148, and HDQ-P1 were obtained from the Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (DSMZ). MDA-MB-468, BT20, and HCC70 were obtained from the University of Colorado Cancer Center (UCCC; Aurora, Colorado) Tissue Culture Core laboratory. Human colorectal cancer cell lines were obtained from ATCC, DSMZ Cell Line Bank, ECACC, and the Korean Cell Line Bank (KCLB; Seoul, South Korea). The Gene Expression Omnibus (GEO) cell line was a generous gift from Dr. Fortunato Ciardiello (Cattedra di Oncologia Medica, Dipartimento Medico-Chirurgico di Internistica Clinica e Sperimentale “F Magrassi e A Lanzara,” Seconda Università degli Studi di Napoli, Naples, Italy). KM20 and KM12C were a generous gift from Dr. Scott Kopetz from MD Anderson Cancer Center (Houston, TX). The 55 human colorectal cancer cell lines used in this study were CL-11(DSMZ), CL-34(DSMZ), COLO205 (ATCC), COLO 741 (ECACC), DLD1 (ATCC), GP5D (SIGMA), HCT116 (ATCC), HCT15 (ATCC), HT15 (SIGMA), HT55 (SIGMA), LOVO (ATCC), LS180 (ATCC), LS513 (ATCC), MDST8 (SIGMA), Mip101 (ECACC), NCI-H716 (ATCC), NCI-H747 (ATCC), RKO (ATCC), SKCO1 (ATCC), SNU-1235 (KCLB), SNU-977 (KCLB), SNU-C1 (KCLB), SNU1460 (KCLB), SW1116 (ATCC), SW1417 (ATCC), SW48 (ATCC), SW480 (ATCC), T84 (ATCC), and WiDr (ATCC). All cell lines were cultured in RPMI media supplemented with 10% FBS, 1% penicillin-streptomycin, and 1% minimum essential medium (MEM) nonessential amino acids and routinely screened monthly for the presence of Mycoplasma (MycoAlert, Cambrex Bio Science). Cell lines were maintained at 37°C with 5% CO₂. Cell lines were authenticated using short tandem repeat DNA Profiling Globalfiler PCR Amplification Kit (Thermo Fisher Scientific) before use and cells were passaged more than 20 times for all experiments.

Breast cancer cells were cultured in DMEM and colorectal cancer cell lines were cultured in RPMI; both supplemented with 10% FBS, 1% penicillin–streptomycin, and 1% minimum essential medium nonessential amino acids (Supplementary Methods). Cells were grown in an incubator at 37°C, which contained 5% CO2 and lines were routinely authenticated and screened for mycoplasma.

The effect of OKI-005 on cell proliferation was evaluated using the Sulforhodamine B (SRB) assay as previously described (27). In brief, 3,000 to 5,000 cells were plated in 96-well plates and allowed to adhere for 24 hours before being exposed to increasing concentrations of OKI-005 (0–5 μmol/L) for 72 hours. The number of cells remained consistent for each cell line through each experiment and experimental arm. The number of cells plated for different cells lines was varied due to differences in doubling times in order to normalize confluency at the time of dosing. Media was then discarded, cells were fixed, washed, and stained with 0.4% SRB (Thermo Fisher Scientific) for 30 minutes. Cells were washed with 1% acetic acid and treated with 10-mmol/L Tris. Plates were read at an absorbance wavelength of 565 nm using a Biotek Synergy 2 96-well plate reader. Raw absorbance data were used to determine IC₅₀ (half-maximal inhibitory concentration) and proliferation curves were generated using GraphPad Prism.

Apoptosis was assessed using the caspase-3/-7 apoptosis assay as previously described (27). In brief, 1,500 to 5,000 cells were plated in black-walled 96-well plates and allowed to adhere overnight. Cells were then exposed to increasing concentrations of OKI-005 (0–5 μmol/L) for 24 or 48 hours, and caspase-3 and -7 activity was assessed using the luminometric Caspase-Glo-3/7 assay (Promega).

For cell cycle analysis, cells were plated in 6-well plates, incubated overnight, and then exposed to increasing concentrations of OKI-005 (0–200 nmol/L) for 24 hours. Culture media was removed, cells were washed 1X with PBS, and harvested by trypsinization. Cells were centrifuged at 1,500 rpm and resuspended in Krishan stain. Cells were incubated at 4°C for 24 hours and analyzed at the UCCC Flow Cytometry Core Facility using a Gallios flow cytometer (22, 23).

Cells were seeded in 6-well plates and allowed to adhere for 24 hours. Cells were then exposed to increasing concentrations of OKI-005 (0, 50, 200, or 500 nmol/L) for 24 hours. Cells were harvested by trypsinization and then lysed in Cell Lysis Buffer (Cell Signaling Technology). Fifteen to 50 μg of total protein was loaded onto a 4% to 20% BIS-TRIS 1.5-mm gradient gel (NuPage, Novex Invitrogen), electrophoresed, and transferred to nitrocellulose using the invitrogen Xcell II blotting apparatus (Bio-Rad). Membranes were blocked in blocking buffer for 1 hour, washed three times for 10 minutes with TBS-Tween (0.1%) and incubated overnight at 4°C with primary antibodies: acetylated histone H3 (Bethyl Lab), acetylated α-tubulin (Cell Signaling Technology), p21 (Cell Signaling Technology), and α-tubulin (Abcam; Supplementary Table S1). Following primary antibody incubation, membranes were washed in TBS-Tween (0.1%) three times and incubated with secondary anti-rabbit or anti-mouse IgG1 horseradish peroxidase–linked antibody at 1:15,000 (Jackson Immuno Research) for 1 hour at room temperature. Blots were developed using the enhanced chemiluminescence detection system 9 pierce, or the Odyssey Infrared Imaging System (LI-COR Biosciences). Experiments were performed in at least triplicate.

Tissue samples were snap-frozen and stored in liquid nitrogen until lysate preparation for immunoblotting. Fifteen to 50 μg of total protein was loaded onto a 4% to 12% bis-tris gel (NuPage, Novex Invitrogen), electrophoresed, and transferred to nitrocellulose using the iPierce G2 Fast Blotter, and analyzed as above.

Male and female BALB/c mice were treated with a single dose orally of 25, 50, or 100 mg/kg OKI-179. Serum samples were obtained at 0.5, 1, 2, and 4 hours postdose. Additional samples were obtained at 8 hours postdose for animals treated at 50 mg/kg and 12 hours for animals treated with 100 mg/kg. Male and female rats were treated with a single 10, 30, or 100 mg/kg dose of OKI-179 orally, and plasma samples were obtained from 0.5 to 24 hours. Male and female beagle dogs were treated with a single dose of 1, 3, or 10 mg/kg OKI-179 orally, and plasma samples were obtained from 0.5 to 24 hours. Quantification of the metabolite OKI-006 and analysis of PK parameters was performed as described above.

Five- to 6-week-old female athymic nude (nu/nu) mice (Envigo) spacing is tight in this line in groups of five per cage and allowed to acclimate for a week before handling. Sterilized food and water were provided ad libitum, and animals were kept on a 12-hour light/dark cycle. Estrogen pellets were prepared in autoclaved medical beeswax packed with estrogen hormone in a biosafety cabinet and transported in sterile sealed containers. The estrogen pellets were implanted a week before the MCF7 cell line was injected to assist in MCF7 tumor growth. Once the estrogen pellet was implanted in the base of the neck by a 10-gauge needle, the neck lesion was sealed by Vetbond Tissue Adhesive to ensure the pellet stayed in place. MCF-7 cells were harvested during logarithmic growth and resuspended in a 1:1 mixture of DMEM and Matrigel (BD Biosciences). The cells were injected bilaterally into the flank using 5 × 10⁶ cells/injection in a volume of 100 μL. Mice were weighed, and tumor measurements were collected twice a week with digital calipers and recorded with the Study Director software package (Studylog Systems). When tumors reached an average volume of 200 mm³, mice were randomized to treatment with vehicle, OKI-179 60 mg/kg orally daily 5 days on, 2 days off, tamoxifen 25 mg/kg subcutaneously three times a week, or the combination of OKI-179 and tamoxifen. At the end of the study, mice were euthanized, and tumor samples were collected.

Results from the proliferation, apoptosis, and cell cycle experiments were analyzed for statistical significance using GraphPad Prism 8.1 using an unpaired t test. In vivo models were analyzed with one-way ANOVA with Dunnett multiple comparison test. Log-rank (Mantel–Cox) test was used to determine significance for in vivo survival. Differences were determined to be statistically significant with P values ≤ 0.05.

All xenograft studies were performed in accordance with the NIH guidelines for the care and use of laboratory animals in a facility accredited by the American Association for Accreditation of Laboratory Animal Care with approval by the Institutional Animal Care and Use Committee prior to the initiation of experiments.

The data generated in this study are available within the article and its supplementary data files.

Largazole is a potent class-I–selective HDAC inhibitor isolated from a marine cyanobacterium, but it is not optimized for clinical drug development. A series of analogs of the active component of largazole were undertaken, resulting in an optimized molecule, OKI-006 (Fig. 1; refs. 24, 25). The biochemical potency and selectivity of OKI-006 for HDACs using a cell-free biochemical assay has been previously published (26). OKI-006 potently inhibits class I HDACs, including HDAC-1 (IC₅₀ 1.2 nmol/L), HDAC-2 (IC₅₀ 2.4 nmol/L), HDAC-3 (IC₅₀ 2.0 nmol/L), and HDAC-8 (IC₅₀ 47 nmol/L); class IIb HDACs, including HDAC-6 (IC₅₀ 47 nmol/L) and HDAC-10 (IC₅₀ 2.8 nmol/L); and the class IV HDAC-11 (IC₅₀ 2.3 nmol/L). OKI-006 does not significantly inhibit class IIa HDACs (HDAC-4,-5,-7 and -9 IC₅₀'s >1000 nmol/L; ref. 26). Although OKI-006 is an optimized HDAC inhibitor, protection of the free sulfhydryl group was felt to be a superior strategy to improve drug delivery. With that objective in mind, we proceeded to make two prodrugs. OKI-005 was optimized for use in cell culture, whereas OKI-179 was optimized for properties for oral delivery in both preclinical models and in humans.

Based on documented higher expression of class I HDACs correlating with clinical outcomes in both colorectal cancer and TNBC, these two disease types were selected for our initial evaluation of the preclinical activity of OKI-005 (7, 10). We tested OKI-005 in vitro against a broad panel of colorectal cancer and TNBC cell lines and found OKI-005 to have potent antiproliferative activity against both TNBC and colorectal cancer cell lines (Fig. 2A; Supplementary Figs. S1 and S2). The majority of cell lines were sensitive to OKI-005 with IC₅₀'s < 500 nmol/L and the most sensitive cell lines exhibited IC₅₀'s < 100 nmol/L.

We selected two sensitive and two resistant TNBC cell lines for further evaluation including OKI-005–induced apoptosis, changes in cell cycle, and effect on histone acetylation. In the CAL-120 and MDA-MB-231 cell lines sensitive to OKI-005 (IC₅₀ < 100 nmol/L), exposure to OKI-005 resulted in a concentration-dependent increase in apoptosis as measured by caspase-3 and -7 activity with up to seven- to nine fold increase compared to no drug control (Fig. 2B). In the resistant HCC1395 and Hs578T cell lines (IC₅₀ to OKI-005 > 500 nmol/L), the degree of apoptosis was much less. As expected, exposure to OKI-005 resulted in a trend toward an increase in G1 and a decrease in the cell fraction in the S phase in CAL-120 cells and a statistically significant difference in the MDA-MB-231 cell line (Fig. 2B).

Exposure to OKI-005 in vitro resulted in a concentration-dependent increase in acetylated histone-H3 and p21 consistent with class I HDAC inhibition in sensitive and resistant TNBC cell lines (Fig. 2D). Interestingly, a small increase in acetylated α-tubulin was observed at the highest dose levels which may be related to class IIb HDAC inhibition.

We selected the MDA-MB-231 (TNBC) and HCT-116 (colorectal cancer) cell lines for xenograft studies to confirm activity of OKI-005 in vivo due to their sensitivity to OKI-005 in vitro and consistent growth patterns in vivo. First, we treated HCT-116 xenografts with OKI-005 at escalating doses orally or i.p. and observed a dose-dependent and statistically significant inhibition of tumor growth at doses of OKI-005 higher than 30 mg/kg, with tumor growth inhibition at 28.1% (P = 0.025) for 30 mg/kg, 75.6% (P < 0.001) for 100 mg/kg compared with vehicle control (Fig. 3A). Mice treated with OKI-005 at 100 mg/kg experienced weight loss related to treatment requiring dose reduction at approximately 1 week, which may explain the initial tumor shrinkage observed in this group that was followed by tumor growth (Supplementary Fig. S3). Treatment with OKI-005 resulted in a dose-dependent increase in acetyl-histone H3 and acetyl H3K9 in OKI-005 treatment groups at 1 hour postdose (Supplementary Fig. S4A). OKI-S005 dosed at 100 mg/kg exerted a time-dependent increase of acetyl-Histone H3 and acetyl H3K9 with highest levels occurring 4 hours posttreatment and returning to baseline 24 hours postdose (Supplementary Fig. S4B and S4C).

Next, we evaluated the PK profile of OKI-005 at escalating doses orally and a single dose i.p. in BALB/c HCT-116 tumor-bearing mice. Treatment with a single oral or i.p. dose of OKI-005 resulted in a dose-proportional increase in the plasma concentration of its active metabolite OKI-006 with highest exposure at 100 mg/kg orally and 5 mg/kg i.p. (Fig. 3B). The half-life in plasma was short with oral dosing (<2 hours), however, levels were more sustained with i.p. dosing. Interestingly, tumor concentrations of OKI-006 exceeded plasma levels at all timepoints and crossed the threshold for antiproliferative activity observed in vitro consistently with OKI-005 dosed at 100 mg/kg orally (Fig. 3C). This finding was consistent with the early tumor shrinkage observed in the HCT-116 xenograft model before dose reduction for weight loss occurred (Fig. 3A).

Finally, we evaluated the ability of OKI-005 to prevent metastasis in addition to inhibit primary tumor growth using an MDA-MB-231 luciferase-labeled metastasis model. Based on sustained plasma exposure with i.p. dosing, this route of administration was selected and the dose was increased to 10 mg/kg to increase exposure. In this model, treatment with OKI-005 10 mg/kg i.p. three times a week for 30 days resulted in statistically significant inhibition of primary tumor growth and metastasis as measured by the number of observed thoracic metastasis (Fig. 3D–G).

Although OKI-005 showed antitumor activity and favorable exposures in blood and tumor in mouse models, prodrug optimization provided OKI-179, which was used for further preclinical in vivo studies and subsequently human clinical trials. Following a single dose of OKI-179 in BALB/c mice and rats, we observed a dose-proportional increase in blood concentration of the active metabolite OKI-006 with no significant difference between male and female animals (Fig. 4A). The maximal concentration (C_max) exceeded 1 μmol/L following a dose of OKI-179 100 mg/kg orally in both models. Exposure was somewhat higher in mice compared with rats. In beagle dogs, lower doses were evaluated based on predicted differences in drug absorption and metabolism across species. The C_max exceeded 1 μmol/L at doses ≥1 mg/kg in male and female dogs (Fig. 4A).

Given the favorable exposure in vivo with OKI-179 when compared with OKI-005, we evaluated the antitumor activity of OKI-179 in HCT-116 and MDA-MB-231 xenograft mouse models. Doses of OKI-179 of 40 to 80 mg/kg orally daily and 120 mg orally every other day were evaluated. Figure 4B depicts the tumor growth curves where we observed in the HCT-116 model, OKI-179 treatment at all doses and schedules evaluated resulted in statistically significant decreased tumor growth at day 24, which compared favorably to the antitumor activity of the same dose of OKI-005 in this model (Fig. 3A). In the MDA-MB-231 model, OKI-179 treatment resulted in statistically significant tumor growth inhibition at higher doses (Fig. 4B). Treatment with OKI-179 in the MDA-MB-231 xenograft model resulted in an increase in acetylated histone H3 consistent with target engagement that was greatest after 10 days of dosing compared with 3 days of dosing and greater with 60 to 80 mg/kg compared with 40 mg/kg (Fig. 4C). We did not observe consistent changes in acetylated α-tubulin following treatment, although a small increase was observed at some dose levels.

In preclinical models of estrogen receptor (ER)-positive (ER⁺) HER2-negative (HER2⁻) breast cancer, HDAC inhibitors potentiate the activity of tamoxifen likely resulting from epigenetic modulation of ER signaling (28, 29). OKI-179 was administered on an intermittent dosing schedule 5 days on 2 days off based on emerging data from the ongoing first-in-human phase I clinical trial on OKI-179 in patients with advanced solid tumors demonstrating dose-limiting thrombocytopenia with continuous dosing (30). Intermittent dosing schedules are generally required for HDAC inhibitors to allow for recovery of thrombocytopenia. As depicted in Fig. 5A, we observed a trend toward improved tumor growth inhibition in an MCF-7 xenograft model with the combination of OKI-179 and tamoxifen compared with either single agent alone, however, both compounds were active in this model. Treatment with the combination resulted in a statistically significant increase in survival with the combination compared with OKI-179 (P = 0.02) and a trend toward increased survival compared with tamoxifen [P = not significant (NS)] (Fig. 5B).

Histone acetylation is dysregulated in a wide array of solid and hematologic cancers, however, to date the development of selective, potent, and active HDAC inhibitors in solid tumors has remained elusive (1). In this manuscript, we report the creation of prodrug OKI-179 that is metabolized in vivo to the largazole-derivative OKI-006. We demonstrated that OKI-005, an alternative prodrug of OKI-006, is optimized for in vitro experimentation as evidenced by potent antiproliferative activity against TNBC and colorectal cancer cell lines with IC₅₀'s in the nanomolar range, robust induction of apoptosis in sensitive cell lines, and proof of target engagement with concentration-proportional increases in histone acetylation. While OKI-005 was active in preclinical in vivo models, OKI-179 was optimized for in vivo experimentation in animal models and humans resulting in a superior PK profile in vivo. Based on these results, OKI-179 was selected to move forward into clinical development.

OKI-179 represents a second-generation prodrug design that leverages sulfhydryl group derivatization of OKI-006 and salt form optimization to aid in absorption with oral dosing. Interestingly, with oral or i.p. administration of OKI-005, tumor concentrations of the active metabolite OKI-006 exceeded plasma concentrations at most doses. Further work will be needed moving forward to investigate the pattern of metabolism in human tissues including in the gastrointestinal tract, liver, and target tumor tissue. The PK profile of OKI-179 in mouse, rat, and beagle dogs was linear with a dose-proportional increase in exposure with oral dosing. OKI-179 is converted to the active metabolite OKI-006 rapidly in preclinical models, which may be an asset clinically. OKI-006 has a short half-life with plasma concentrations more than 100 nmol/L following a single dose persisting for approximately 4 to 12 hours depending on the dose and model (i.e mouse vs. rat vs. dog). The short half-life should be taken into consideration in determining the ideal dosing strategy in the clinic.

Various dosing strategies were employed in our preclinical evaluation of the efficacy of OKI-005 and OKI-179 in xenograft models. In our initial study in the MDA-MD-231 xenograft model, we evaluated OKI-005 administered either i.p. or p.o. once daily (10–100 mg/kg). Our data demonstrate similar efficacy and intratumor drug concentrations for the i.p. and p.o. formulations. At the higher dose of 100 mg/kg oral daily, animals experienced dose-limiting treatment-related toxicity (evidenced by weight loss) and the dose was reduced. This limits our ability to make conclusions regarding the dose response for OKI-005 in this model. For the second-generation compound, OKI-179, we administered doses of 40 to 80 mg/kg p.o. daily or 120 mg/kg p.o. every other day in the HCT-116 and MDA-MB-231 xenograft models. The higher dose of 120 mg/kg with every-other-day dosing was included based on the weight loss observed in animals treated with OKI-179 100 mg/kg oral daily and our desire to increase C_max while maintaining tolerability. This dosing schedule was better tolerated in the animals. For our final in vivo study, we incorporated a dosing schedule currently being evaluated in the ongoing phase I clinical trial, which was an intermittent dosing schedule with 5 days on and 2 days off. This schedule was selected clinically to allow for recovery from drug-related thrombocytopenia. Ultimately, the recommended phase II dose of OKI-179 for future clinical trials will likely be based on clinical toxicity, however, we made efforts in our preclinical experimental design to evaluate dosing strategies anticipated to move forward in the clinic.

One of the promising features of HDAC inhibitors as anticancer agents is the ability to induce apoptosis and convert growth arrest to apoptosis when used in combination with other agents. The apoptotic potential of HDAC inhibitors likely depends on the intrinsic cellular threshold for apoptosis that may be tumor cell–dependent and the mechanism by which death can be triggered by HDAC inhibition in the cell. Given the multiple potential effects of HDACs in cancer cells, there may be differential effects depending on the cellular milieu. In our study, OKI-005 exposure resulted in a robust four- to nine-fold increase in apoptosis in TNBC cell lines sensitive to the antiproliferative activity of OKI-005 compared with a modest two- to three fold increase in resistant cell lines. The pharmacodynamic effect of OKI-005 on histone acetylation was similar in sensitive and resistant cell lines, pointing to another factor in cellular downstream signaling as mediating a proapoptotic response in some cells but not others. Biomarkers predictive of response to HDAC inhibitors are yet to be fully identified, however enhanced JAK/STAT signaling has been reported as associated with resistance to HDAC inhibitor-induced cell death and resistance to vorinostat in patents with cutaneous T-cell lymphoma (CTCL; ref. 31). Biomarkers predictive of a proapoptotic response to OKI-005 and OKI-179 should be actively investigated, as they may aid in selecting a responding patient population in future trials in advanced solid tumors.

OKI-179 is a promising agent for the treatment of TNBC based on preclinical activity and the ongoing clinical need for therapies that may potentiate the activity of programmed cell death 1 (PD-1) and programmed death-ligand 1(PD-L1) inhibitors in TNBC. Although both atezolizumab and pembrolizumab are FDA approved in combination with chemotherapy in patients with metastatic PD-L1–positive TNBC, the rates of 1-year progression-free survival remain in the 30%-to-40% range, highlighting room for improvement in achieving long-term disease control for these patients with immunotherapy (32, 33). HDAC inhibitors potentiate the activity of immunotherapy agents in many preclinical models of solid tumors with evidence of increased tumor immunogenicity as a potential mechanism (13, 34, 35). Supporting these findings, the addition of OKI-179 to nivolumab in a humanized mouse MDA-MB-231 xenograft, resulted in improved tumor growth inhibition and increased T-cell activation in the tumor microenvironment (36). Similar findings were observed for OKI-179 in combination with anti–PD-1 inhibition in preclinical models of B-cell lymphomas resistant to anti–PD-1 where tumor-derived MHC class I was required and MHC class II expression correlated with response (26). Other emerging roles for HDACs include an expanding landscape of the removal of acyl lysine groups from histones (37, 38). This may expand the potential impact of HDAC inhibitors and should be investigated further, as well as the impact on nonhistone proteins, which remains largely uncharacterized. It is possible that these underexplored effects might contribute to the modulation of immune response and potentiation of immunotherapy.

Although there are many opportunities for clinical investigation of OKI-179 as a potent, class-I–selective HDAC inhibitor in solid tumors, we chose to explore the combination of OKI-179 with tamoxifen preclinically. Resistance to endocrine therapy in hormone receptor–positive, HER2⁻ breast cancer remains a clinical challenge, and HDAC inhibitors are promising agents for reversing endocrine resistance (3, 39, 40). In preclinical models, HDAC inhibitors downregulate ER expression and can reverse tamoxifen-induced ER stabilization leading to increased apoptosis (28). Many clinical trials have evaluated HDAC inhibitors in combination with tamoxifen or aromatase inhibitors with some promising results. The combination of vorinostat and tamoxifen was evaluated in a phase II clinical trial in patients with previously treated ER⁺ HER2⁻ metastatic breast cancer and demonstrated an objective response rate of 19% with a clinical benefit rate of 40%, which is higher than expected clinical activity of tamoxifen alone based on historical controls (41). In a phase III clinical trial, the combination of the subtype-selective HDAC inhibitors chidamide and exemestane improved PFS from 3.8 months with exemestane + placebo to 7.4 months with exemestane + chidamide, however, a substantial portion of patients required dose interruption of chidamide for toxicity (42). The isoform selectivity of OKI-179 may lead to improved clinical tolerability, and this is part of the on going investigation in a phase I clinical trial. There are many potential clinical opportunities for OKI-179 in hematologic and solid tumors, including potential combination with hormonal therapy in ER⁺ HER2⁻ breast cancer.

J.R. Diamond reports grants from OnKure, Inc. during the conduct of the study; other support from OnKure, Inc. outside the submitted work; has a current consulting agreement with OnKure, Inc.; and is chief medical officer. T.M. Pitts reports grants from OnKure, Inc. outside the submitted work. C.G. Nasveschuk reports a patent for US20130203681 issued, licensed, and with royalties paid from OnKure, Inc. A.J. Phillips reports grants from the NCI (CA 110246) during the conduct of the study; has a patent for US9422340 issued and licensed to OnKure, Inc.; and is a board member of OnKure, Inc. J.D. Winkler is an employee of OnKure, Inc. W.P. Schiemann reports grants from the Department of Defense during the conduct of the study. S.G. Eckhardt reports grants and personal fees from OnKure, Inc. during the conduct of the study; personal fees from OnKure, Inc. outside the submitted work; and is now on OnKure, Inc.'s scientific advisory board. X. Liu reports personal fees and other support from OnKure, Inc. during the conduct of the study; grants and other support from Vesicle Therapeutics Inc. outside the submitted work; and has a patent for US 9.422,340 B2 issued, licensed, and with royalties paid from OnKure, Inc. No disclosures were reported by the other authors.

J.R. Diamond: Conceptualization, resources, data curation, formal analysis, supervision, investigation, writing–original draft, writing–review and editing. T.M. Pitts: Conceptualization, resources, data curation, formal analysis, supervision, investigation, methodology, writing–review and editing. D. Ungermannova: Conceptualization, resources, data curation, formal analysis, investigation, writing–review and editing. C.G. Nasveschuk: Data curation, formal analysis, investigation, methodology, writing–review and editing. G. Zhang: Data curation, investigation, methodology, writing–review and editing. A.J. Phillips: Conceptualization, resources, data curation, investigation, writing–review and editing. S.M. Bagby: Data curation, formal analysis, supervision, investigation, methodology, writing–review and editing. J. Pafford: Data curation, formal analysis, writing–review and editing. B.W. Yacob: Data curation, formal analysis, investigation, writing–review and editing. T.P. Newton: Data curation, formal analysis, investigation, writing–review and editing. J.J. Tentler: Conceptualization, resources, data curation, formal analysis, supervision, investigation, writing–review and editing. B. Gittleman: Data curation, formal analysis, investigation, writing–review and editing. S.J. Hartman: Conceptualization, resources, data curation, investigation, writing–review and editing. J.A. DeMattei: Conceptualization, resources, data curation, formal analysis, supervision, investigation, methodology, writing–review and editing. J.D. Winkler: Conceptualization, resources, data curation, formal analysis, investigation, writing–original draft, writing–review and editing. M.K. Wendt: Conceptualization, resources, data curation, formal analysis, investigation, writing–review and editing. W.P. Schiemann: Conceptualization, resources, data curation, formal analysis, supervision, investigation, writing–review and editing. S.G. Eckhardt: Conceptualization, resources, data curation, formal analysis, supervision, investigation, writing–review and editing. X. Liu: Conceptualization, resources, data curation, formal analysis, supervision, investigation, methodology, writing–review and editing. A.D. Piscopio: Conceptualization, resources, formal analysis, supervision, investigation, writing–original draft, writing–review and editing.

This work was supported by the NIH and the NCI through grant numbers 5P30CA046934-25 (UCCC Support Grant), 1K23CA172691-01A1 (to J.R. Diamond), R01GM113141 (to X. Liu), W81XWH-10-1-0989 (to X. Liu, A.J. Phillips, W.P. Schiemann), and CPRIT Scholar Award No. RR160093 (to S.G. Eckhardt). Support was also received from OnKure, Inc. The funding bodies played a role in study design, data collection, analysis, and interpretation, and manuscript preparation.