The Tankyrase Inhibitor OM-153 Demonstrates Antitumor Efficacy and a Therapeutic Window in Mouse Models

The catalytic enzymes tankyrase 1 and 2 (TNKS1/2) alter protein turnover by poly-ADP-ribosylating target proteins, which earmark them for degradation by the ubiquitin–proteasomal system. Prominent targets of the catalytic activity of TNKS1/2 include AXIN proteins, resulting in TNKS1/2 being attractive biotargets for addressing of oncogenic WNT/β-catenin signaling. Although several potent small molecules have been developed to inhibit TNKS1/2, there are currently no TNKS1/2 inhibitors available in clinical practice. The development of tankyrase inhibitors has mainly been disadvantaged by concerns over biotarget-dependent intestinal toxicity and a deficient therapeutic window. Here we show that the novel, potent, and selective 1,2,4-triazole–based TNKS1/2 inhibitor OM-153 reduces WNT/β-catenin signaling and tumor progression in COLO 320DM colon carcinoma xenografts upon oral administration of 0.33–10 mg/kg twice daily. In addition, OM-153 potentiates anti–programmed cell death protein 1 (anti–PD-1) immune checkpoint inhibition and antitumor effect in a B16-F10 mouse melanoma model. A 28-day repeated dose mouse toxicity study documents body weight loss, intestinal damage, and tubular damage in the kidney after oral–twice daily administration of 100 mg/kg. In contrast, mice treated oral–twice daily with 10 mg/kg show an intact intestinal architecture and no atypical histopathologic changes in other organs. In addition, clinical biochemistry and hematologic analyses do not identify changes indicating substantial toxicity. The results demonstrate OM-153–mediated antitumor effects and a therapeutic window in a colon carcinoma mouse model ranging from 0.33 to at least 10 mg/kg, and provide a framework for using OM-153 for further preclinical evaluations. Significance: This study uncovers the effectiveness and therapeutic window for a novel tankyrase inhibitor in mouse tumor models.

Clinical tankyrase inhibitor development has so far been hampered by concerns about target-specific and signaling pathway-specific side effects, in particular intestinal toxicity (5,33,34). Despite this, the tankyrase inhibitors E7449 and STP1002 have entered clinical trials in the cancer arena (35)(36)(37)(38). This supports the potential of TNKS1/2 inhibitors and justifies the continued development of drugs directed toward TNKS1/2 inhibition. Recently, we developed the small-molecule 1,2,4-triazole-based TNKSi OM-153 through iterative designmake-test cycles possessing favorable drug properties. These properties include efficacy, off-target liabilities, absorption/distribution/metabolism/excretion (ADME) parameters, and pharmacokinetic profile in mice (32). Here, we evaluate the biological properties and potential toxicity issues of OM-153 in detail. First, we tested the efficacy of OM-153 in a panel of tumor cell lines showing lownanomolar range biotarget engagement and tumor cell growth inhibition. Next, we carried out a standardized dose-escalating experiment using the COLO 320DM colon carcinoma xenograft model, showing significant multidose tumor growth inhibition, and also in an isogenic B16-F10 mouse melanoma model. Subsequently, the oral single-dose maximum tolerated dose (MTD) for OM-153 was established. The MTD was followed by a two-dose (10 mg/kg and 100 mg/kg), oral and twice daily 28-day repeated dose mouse toxicity evaluation, including multiorgan histopathology, biochemistry, and hematology in mice. With the applied methodology used in mice, we show no significant adverse toxicity in mice dosed oral-twice daily with 10 mg/kg, and anti-colon carcinoma efficacy when dosed oral-twice daily at 0.33-10 mg/kg. Hence, in contrast to the earlier studies (33,34), our data propose the existence of a therapeutic window for OM-153 ranging from 0.33 to at least 10 mg/kg in a mouse colon carcinoma model.
The experiment using COLO 320DM (CCL-220, RRID: CVCL_0219, ATCC) xenografts was carried out as previously described (40) at Reaction Biology. On day 3 after subcutanous (s.c.) tumor challenge in CB17-SCID mice (CRL: 561, RRID: IMSR_CRL: 561, Charles River), tumor-bearing animals (mean tumor volume: 12 mm 3 ) were randomized into six groups (all n = 10). The day after, the animals were treated oral-twice daily with 10, 3.3, 1, 0.33, or 0.1 mg/kg OM-153 or vehicle until day 37. Eight mice, evenly distributed between the treatment groups, were euthanized for ethical reasons before experiment termination, due to skin ulcerations in the tumor area (6 mice) or body weight loss (>20%, two mice). Next, the mice were sacrificed, approximately 4 hours after first dose of the day. Protein and RNA extracts were prepared from four moderately sized tumors within each treatment group. Mass spectrometry analyses of OM-153 in plasma and tumors were performed according to the standard protocols of Pharmacelsus.
The experiment using B16-F10 tumors was carried out as previously described (39) at Reaction Biology. On day six after tumor challenge (s.c.) in C57BL/6N mice (CRL: 493, RRID: IMSR_CRL: 493, Charles River), tumor-bearing animals (mean tumor volume: 31 mm 3 ) were randomized into six groups (all n = 15). On the same day, the animals were treated oral-twice daily with vehicle, anti-programmed cell death 1 (anti-PD-1; 10 mg/kg i.p. on day 6, 9, and 12, BE0033-2, RRID: AB_1107747, Bio X Cell), 10 mg/kg OM-153, or combined treatment with anti-PD-1 and 10, 1, and 0.1 mg/kg OM-153 until day 20. Fifteen mice, distributed between the treatment groups, were euthanized for ethical reasons before experiment termination, due to skin ulcerations in the tumor area. Next, the mice were sacrificed, approximately 8 hours after last dosing. Protein and RNA extracts were prepared from 8 tumors within each treatment group.
The dose escalation experiment was performed in single male CD-1 mice (CRL: 022, RRID: IMSR_CRL: 022, Charles River) at BSL BIOSERVICE/Eurofins using their standard protocol. The animals were treated oral-twice daily with escalating doses of 500, 1,000, 1,500, and 2,000 mg/kg OM-153, and observed for 48 hours for clinical signs (7 days for the verification animal, 2,000 mg/kg). Body weight was recorded daily.
The 28-day oral repeated dose toxicity study was performed at Reaction Biology (10 mg/kg) and at Oslo University Hospital (100 mg/kg) using male CD-1 (CRL: 022, RRID: IMSR_CRL: 022, Charles River) and C57BL/6J (JAX: 000664, RRID: IMSR_JAX: 000664, Jackson) mice, respectively. The mice were treated oral-twice daily with 10 (n = 7) or 100 (n = 4) mg/kg OM-153 or vehicle control (both experiments, n = 4) until day 28, or study termination (for 100 mg/kg). Body weight and food consumption were measured thrice weekly. For the experiment using 10 mg/kg oral-twice daily dosing, blood was collected at the end of the experiment and hematologic and clinical biochemistry analyses were performed at IDEXX according to their standard protocols. Duodenum, jejunum, ileum, colon-rectum, lymph nodes, pancreas, heart, kidney, liver, spleen, lung, prostate, and testis were collected, formalin-fixed (10%), paraffin-embedded, stained with hematoxylin and eosin (H&E), and imaged as previously described (39). Staining of ileums using immunofluorescence was performed as previously described (29,39). Donkey anti-rabbit IgG cyanine Cy3 was used as secondary antibody [(550 nm), 1:500, 711165152, RRID: AB_2307443, Jackson ImmunoResearch] for 60 minutes at 37°C. DAPI nuclear dye [(340 nm), 1 μg/μL, D9542, Sigma Aldrich] was added to the final washing solution. RNAscope multiplex fluorescent reagent kit v2 detection (323100, Advanced Cell Diagnostics) was used to detect expression of mouse leucinerich repeat-containing G-coupled receptor 5 (Lgr) mRNA (312171, Advanced Cell Diagnostics) according to the manufacturer's instructions. Imaging was performed using confocal microscopy as described previously (39). All animal experiments were approved by local animal experiment authorities (ethics committee of the Chinese Association for Laboratory Animal Sciences, German Centre for the Protection of Laboratory Animals, and Norwegian Food Safety Authority), and were carried out in compliance with FELASA guidelines and recommendations.

Quantification and Statistical Analysis
Statistical analyses were performed as previously described using t tests and Mann-Whitney rank sum tests (7).

Data Availability Statement
The data generated in this study are available within the article and its Supplementary Data files. Materials generated in this study can be made available upon request to the corresponding author.
Next, in comparison with the selected TNKSi panel, the potential of OM-153 as an antiproliferative agent was tested in a cell growth assay in the adenomatous polyposis coli (APC)-mutated, WNT/β-catenin signaling-dependent, and TNKSi-sensitive colon carcinoma cell line COLO 320DM (7,32,33,40).
To evaluate the specificity of the cell growth inhibition, the APC-wild-type, WNT/β-catenin signaling-independent and TNKSi-insensitive colon carcinoma cell line RKO was used as a control (7,29,32). Treatment with OM-153 decreased cell growth in COLO 320DM cells with a GI 50 value of 10 nmol/L and a GI 25 value of 2.5 nmol/L (concentrations resulting in 50% and 25% growth inhibition, respectively), while cell growth in RKO cells was insubstantially affected by the treatment (GI 50 and GI 25 values >10,000 nmol/L; Fig. 1C; Supplementary Fig. S3).
In conclusion, in addition to our recent report (32), the results demonstrate that OM-153 is a highly potent and specific inhibitor of TNKS1/2, WNT/β-catenin signaling and cell growth in the APC-mutated colon carcinoma cell line COLO 320DM. GI 50 and GI 25 values were calculated relative to controls (100%, 0.01% DMSO) and experiment time 0 values (t 0 , set to 0%) after 5 days of cultivation. Data from one representative experiment of more than three repeated assays, each with six replicates, are shown.

OM-153 Inhibits WNT/β-Catenin, YAP, and MYC Signaling and Shows an Antiproliferative Effect in Human Cancer Cell Lines
Recently, we showed that TNKS1/2 inhibition could attenuate cell growth in a wide range of cancer types in vitro (7). To evaluate cancer cell line growth inhibition by OM-153, the NCI-60 tumor cell line panel was screened. Of the 60 tested cancer cell lines, 16 cell lines showed >25% relative growth inhibition upon treatment with 10 nmol/L OM-153 ( Fig. 2A; Supplementary Fig. S4). These cell lines originated from lung, brain, ovary, and kidney, and included the previously identified TNKSi-sensitive cell lines OVCAR-4 (ovarian can- Recently, we showed that TNKS1/2 inhibition cell-type dependently can block WNT/β-catenin and/or YAP signaling, and consequently MYC protooncogene bHLH transcription factor (MYC) signaling, resulting in attenuated cancer cell growth (7) ing accumulation of β-catenin degradasomes (7,17) in COLO 320DM cells (Fig. 3).
Next, the effect of OM-153 treatment on YAP signaling was examined. Consistent with earlier research (7), an immunoblot analysis showed cytoplasmic stabilization of the TNKS1/2 target protein AMOTL2 in all cell lines upon OM-153 treatment, while nuclear YAP accumulation was only seen in UO-31, OVCAR-4, and ABC-1 cells ( Fig. 2D; Supplementary Fig. S5B). Moreover, the real-time qRT-PCR analysis showed reduced transcription of the YAP signaling target genes AMOTL, cellular communication network factor 1 and 2 (CCN and CCN) in all TNKSi-sensitive cell lines (Fig. 2D). In conclusion, OM-153 functions as a potent cell type-dependent inhibitor of WNT/β-catenin, YAP and MYC signaling and can reduce cell growth in a subset of human cancer cell lines in cell culture.

OM-153 Inhibits WNT/β-Catenin Signaling and Shows Antitumor Effect in a Human Colon Carcinoma Xenograft Model
A mouse pharmacokinetic analysis was designed to evaluate drug exposure for the mouse efficacy and toxicity studies using oral administration of 10 or 100 mg/kg OM-153, dosed twice daily at the start of the experiment and after
The study documented that the combined treatment effect was dependent on loss of β-catenin in the tumor cells and induction of a CD8 + T cell-mediated adaptive antitumor immune response (39).
To examine the effect of combining OM-153 with anti-PD-1 treatment, B16-F10 tumors were established in immunocompetent C57BL/6N mice, followed by oral-twice daily treatment using various doses of OM-153, combined with Tumors treated oral-twice daily with 10 mg/kg OM-153 exhibited significant stabilization of AXIN1 protein, but showed insignificant downregulation of β-catenin in an immunoblot analysis ( Fig. 5B; Supplementary Fig. S7D). The real-time qRT-PCR analysis showed significantly decreased transcription of the target gene Axin, indicating attenuated WNT/β-catenin signaling (Fig. 5C).

Treatment with 10 mg/kg OM-153 Twice Daily Is Tolerated in Mice
TNKS1/2 inhibition has been associated with the induction of intestinal toxicity linked to biotarget and WNT/β-catenin signaling pathway-specific effects (33,34   observed in any animals after oral-twice daily dosing with 500, 1,000, 1,500, or 2,000 mg/kg of OM-153 ( Supplementary Fig. S8A).
To evaluate the long-term effects of OM-153 treatment, a 28-day oral repeated dose toxicity study was carried out using oral-twice daily dosing of 100 and 10 mg/kg. In mice dosed with 100 mg/kg, reduced activity and body weight loss in 3 of 4 mice were documented, and the mice were euthanized for ethical reasons on day 7 ( Fig. 6A; Supplementary Fig. S8B). In contrast, in mice dosed with 10 mg/kg, no clinical signs, body weight loss, or reduced food consumption were observed ( Fig. 6B; Supplementary Fig. S8C and S8D). Next, duodenum, jejunum and ileum, colon-rectum, lymph nodes, pancreas, heart, kidney, liver, spleen, lung, prostate and testis were collected, stained with H&E and the histopathology was evaluated. In the small intestines of 2 out of three 100 mg/kg treated mice, loss of villi height, width, crypt, and as well as surface epithelium, accompanied by inflammation, was documented ( Fig. 6C; Supplementary   Fig. S8E). In the same mice, signs of acute tubular damage in the kidney, presumably caused by intestinal damage, were observed ( Supplementary Fig. S8E).
No abnormal histopathologic changes in the other organs were detected (Supplementary Fig. S8E). On the contrary, the overall intestinal architecture was intact in the 10 mg/kg group (Fig. 6D). Persistent proliferation in the crypt compartments was indicated by visualization of the proliferation and intestinal stem cell (ISC)-like marker antigen identified by mAb Ki 67 (MKI67; Fig. 6E). Keratin 20 (KRT20) staining specified presence of terminally differentiated epithelial cells in the villus termini (Fig. 6E). Similar to previous observations (33,41), expression of the ISC marker and WNT/β-catenin signaling target gene Lgr was reduced (Fig. 6F). No atypical histopathologic changes in other organs were observed ( Supplementary Fig. S8E). Blood was collected at the end of the experiment using 10 mg/kg dosing to perform clinical biochemistry and hematologic analyses. The clinical biochemistry analysis did not suggest any signs of liver damage as indicated by insignificant  alteration of aspartate transaminase (AST), alanine transaminase (ALT), and glutamate dehydrogenase (GLDH) levels (Fig. 7A). However, significant decreases in magnesium, potassium, and triglycerides could indicate a poorer gastrointestinal absorption or altered kidney function (Fig. 7A). The hematologic analysis suggested no signs of inflammation, although a tendency toward reduced leukocytes and reticulocytes, as well as a moderate but significant reduction in erythrocytes and hemoglobin was noted, which may indicate moderately altered hematopoiesis (Fig. 7B). These alterations are possibly caused by changes in WNT/β-catenin and YAP signaling activities, as both pathways are known regulators of hematopoiesis (42,43).
In conclusion, toxicity in mice at 100 mg/kg OM-153 oral-twice daily dosing is, at least in partially, caused by intestinal damage. In contrast, a 10 mg/kg OM-153 oral-twice daily dosing was overall tolerated in mice, and the intestinal architecture was intact despite a diminished Lgr expression in ISCs at the crypt the base.

Discussion
Here we show that OM-153 is a highly potent and specific inhibitor of TNKS1/2, WNT/β-catenin, YAP and MYC signaling capable of reducing cell growth in a subset of human cancer cell lines. Oral-twice daily dosing of 0.33-10 mg/kg OM-153 in COLO 320DM mouse colon carcinoma xenografts attenuated WNT/β-catenin signaling and resulted in a 74%-85% tumor growth inhibition. Moreover, in the isogenic and immunocompetent B16-F10 mouse melanoma model, oral-twice daily dosing of 0.1-10 mg/kg OM-153 potentiated efficacy of anti-PD-1 treatment and resulted in a 51%-65% tumor growth inhibition.
Deregulated WNT/β-catenin signaling contributes to aberrant cell growth and carcinogenesis, but is in parallel also critical for stem cell renewal, proliferation, and differentiation, during both embryogenesis and tissue homeostasis (20,44). In the last decade, two studies using the early-generation TNKS1/2 inhibitors G007-LK and G-631, observed TNKSi-induced anti-colon carcinoma efficacies (33,34). These reports also describe severe side-effects at therapeutic doses including intestinal toxicity limiting the therapeutic window (33,34,45).
In the 28-day repeated dose mouse toxicity study, when dosing mice oral-twice daily with 100 mg/kg OM-153, we indeed documented rapid loss of body weight, intestinal epithelial degeneration, and inflammation, as well as tubular damage in the kidney. However, we were not able to detect substantial side-effects and toxicity in mice treated with oral-twice daily dosing using 10 mg/kg, and importantly, their intestinal architecture was intact. In the ileum of these mice, MKI67 staining documented continual proliferation in the crypts and KRT20 staining showed differentiated epithelial cells. Similar to our previous report, showing TNKSi-induced elimination of differentiated cells traced from WNT/β-catenin signaling-dependent LGR5 + ISCs (41), Lgr expression was lost and the result may propose the presence of a TNKSi-resistant ISC population. Hence, one may hypothesize that intestinal proliferation from dispensable LGR5 + ISCs (46) can be maintained by activated proliferation of an alternative source of stem cells. Quiescent, WNT/β-catenin signaling-independent, LGR5 − /BMI1 proto-oncogene, polycomb ring finger (BMI1) + and crypt +4-positioned ISCs have been described in that role earlier (41,47,48). Comprehensive follow-up studies of possible TNKSi-resistant ISC populations are evidently required.
We note that the primary pharmacologic on-target effect for OM-153 is seen in the low nanomolar range, as indicated by the TNKS1/2, WNT/β-catenin signaling reporter IC 50 -values, and COLO 320DM GI 50 value, while the micromolar range C max values for oral-twice daily dosing with 10 and 100 mg/kg OM-153 are more than 2 logs higher (5.3 and 39.2 μmol/L, respectively). The results propose that the documented toxicity is likely not coupled to on-target inhibition of TNKS1/2 and WNT/β-catenin signaling, but rather caused by concentration-dependent interactions with unknown off-targets, or optionally related to the physicochemical properties of OM-153 (45). Further detailed mechanistic investigation of plausible off-target toxicity is considered necessary. The possible compatibility of the tankyrase biotarget, with advancing tankyrase-specific inhibitors toward clinical trials in the cancer arena, has recently been demonstrated by the initiation of clinical trials with the TNKSi STP1002 (37).
The therapeutic index is normally measured as the ratio of the highest exposure to the drug that results in no toxicity (i.e., toxic dose in 50% of subjects, TD 50 ) to the exposure that produces the desired effect (i.e., efficacious dose in 50% of subjects, ED 50 ; ref. 45). The therapeutic window for a drug is commonly known as the dose range that can treat a disease effectively without having toxic effects. When using monotherapy treatment, we document no adverse toxicity in mice treated with oral-twice daily dosing using 10 mg/kg OM-153, and a significant antitumor efficacy in the COLO 320DM colon carcinoma xenograft model when dosing mice oral-twice daily with 0.33-10 mg/kg OM-153. Together, these data indicate a therapeutic window ranging from 0.33 mg/kg to at least 10 mg/kg.
In conclusion, in our experimental setting, we show that OM-153 exhibits antitumor efficacy and a viable therapeutic window in mouse models, which rationalize further preclinical evaluations to translate TNKS1/2 inhibition to a human therapeutic setting.