Antileukemic Efficacy in Vitro of Talazoparib and APE1 Inhibitor III Combined with Decitabine in Myeloid Malignancies

Malignant hematopoietic cells of myelodysplastic syndromes (MDS)/chronic myelomonocytic leukemias (CMML) and acute myeloid leukemias (AML) may be vulnerable to inhibition of poly(ADP ribose) polymerase 1/2 (PARP1/2) and apurinic/apyrimidinic endonuclease 1 (APE1). PARP1/2 and APE1 are critical enzymes involved in single-strand break repair and base excision repair, respectively. Here, we investigated the cytotoxic efficacy of talazoparib and APE1 inhibitor III, inhibitors of PARP1/2 and APE1, in primary CD34+ MDS/CMML cell samples (n = 8; 4 MDS and 4 CMML) and in primary CD34+ or CD34− AML cell samples (n = 18) in comparison to healthy CD34+ donor cell samples (n = 8). Strikingly, talazoparib and APE1 inhibitor III demonstrated critical antileukemic efficacy in selected MDS/CMML and AML cell samples. Low doses of talazoparib and APE1 inhibitor III further increased the cytotoxic efficacy of decitabine in MDS/CMML and AML cells. Moreover, low doses of APE1 inhibitor III increased the cytotoxic efficacy of talazoparib in MDS/CMML and AML cells. In summary, talazoparib and APE1 inhibitor III demonstrated substantial antileukemic efficacy as single agents, in combination with decitabine, and combined with each other. Hence, our findings support further investigation of these agents in sophisticated clinical trials.

IC 50   The cell proliferation rate of MDS/CMML and AML cells might correlate with the cytotoxic efficacy of talazoparib and APE1 inhibitor III, respectively. Therefore, growth curves of untreated MDS/CMML and AML cells were correlated with the corresponding surviving fractions of talazoparib and APE1 inhibitor III treated MDS/CMML and AML cells ( Figure 1C,D). However, no consistent correlation between cell proliferation and cytotoxic efficacy of talazoparib and APE1 inhibitor III was evident in MDS/CMML and AML cells. These findings suggest that the antileukemic efficacy of talazoparib and APE1 inhibitor III is not strictly dependent on the in vitro proliferation rate of leukemic blasts.

Cytotoxic Efficacy of Decitabine ± Talazoparib, Decitabine ± APE1 Inhibitor III, and Talazoparib ± APE1 Inhibitor III in MDS/CMML and AML Cells
The cytotoxic efficacies of (I) decitabine ± talazoparib, (II) decitabine ± APE1 inhibitor III, and (III) talazoparib ± APE1 inhibitor III were analyzed in CD34+ MDS/CMML cells and in CD34+ or The cell proliferation rate of MDS/CMML and AML cells might correlate with the cytotoxic efficacy of talazoparib and APE1 inhibitor III, respectively. Therefore, growth curves of untreated MDS/CMML and AML cells were correlated with the corresponding surviving fractions of talazoparib and APE1 inhibitor III treated MDS/CMML and AML cells ( Figure 1C,D). However, no consistent correlation between cell proliferation and cytotoxic efficacy of talazoparib and APE1 inhibitor III was evident in MDS/CMML and AML cells. These findings suggest that the antileukemic efficacy of talazoparib and APE1 inhibitor III is not strictly dependent on the in vitro proliferation rate of leukemic blasts.

Analysis of PARP1 and APE1 mRNA Expression in MDS/CMML and AML Cells in Correlation to Cytotoxic Efficacy of Talazoparib and APE1 Inhibitor III
PARP1 and APE1 mRNA expression might correlate with the cytotoxic efficacy of talazoparib and APE1 inhibitor III in MDS/CMML and AML cells and explain why some of the patients did not respond to the drugs. Therefore, PARP1 and APE1 mRNA expression levels were analyzed in MDS/CMML and AML cell samples. However, the PARP1 and APE1 mRNA expression levels did not differ significantly between 'responders' and 'non-responders' (Figure 3A,B).

Analysis of PARP1 and APE1 mRNA Expression in MDS/CMML and AML Cells in Correlation to
Cytotoxic Efficacy of Talazoparib and APE1 Inhibitor III PARP1 and APE1 mRNA expression might correlate with the cytotoxic efficacy of talazoparib and APE1 inhibitor III in MDS/CMML and AML cells and explain why some of the patients did not respond to the drugs. Therefore, PARP1 and APE1 mRNA expression levels were analyzed in MDS/CMML and AML cell samples. However, the PARP1 and APE1 mRNA expression levels did not differ significantly between 'responders' and 'non-responders' (Figure 3A,B).

Analysis of γH2AX Foci in MDS/CMML and AML Cells in Correlation to Cytotoxic Efficacy of Talazoparib and APE1 Inhibitor III
Since both drugs act on the DNA damage response, γH2AX foci (as a readout for DNA damage) were analyzed for their potential correlation with the cytotoxic efficacy of talazoparib and APE1 inhibitor III in MDS/CMML and AML cells. Therefore, γH2AX foci levels were quantified in MDS/CMML and AML cell samples. However, the number of γH2AX foci did not significantly differ between 'responders' and 'non-responders' (Figure 3C,D).

Analysis of γH2AX Foci in MDS/CMML and AML Cells in Correlation to Cytotoxic Efficacy of Talazoparib and APE1 Inhibitor III
Since both drugs act on the DNA damage response, γH2AX foci (as a readout for DNA damage) were analyzed for their potential correlation with the cytotoxic efficacy of talazoparib and APE1 inhibitor III in MDS/CMML and AML cells. Therefore, γH2AX foci levels were quantified in MDS/CMML and AML cell samples. However, the number of γH2AX foci did not significantly differ between 'responders' and 'non-responders' (Figure 3C,D).

Discussion
In the present study, the cytotoxic efficacy of talazoparib and APE1 inhibitor III was investigated using several distinct approaches in CD34+ MDS/CMML cells and in CD34+ or CD34− AML cells in comparison to healthy CD34+ donor cells. In addition, potential predictive markers were analyzed and correlated with the cytotoxic efficacy of talazoparib and APE1 inhibitor III.
Talazoparib demonstrated critical antileukemic efficacy as a single-agent and in combination with decitabine in several MDS/CMML and AML cell samples, which might be mediated by the formation of cytotoxic PARP1-DNA complexes [27], DNMT1-DNA complexes [38], and PARP1-DNMT1-DNA complexes [28], respectively. Notably, our data suggest that decitabine combined with low doses of talazoparib might kill MDS/CMML and AML cells more efficiently than decitabine alone, as demonstrated by an increased antileukemic efficacy for the combination in about 86% of MDS/CMML/AML cell samples. Our observations confirm and extend results of previous experimental studies reporting on the antileukemic efficacy of olaparib [39][40][41] and talazoparib [28] and the synergistic efficacy between talazoparib and DNMTi in AML cell lines and primary AML cells [28]. Currently, the antileukemic efficacy of talazoparib alone and combined with decitabine has been tested in several clinical trials [42][43][44].
Furthermore, APE1 inhibitor III demonstrated antileukemic efficacy in several MDS/CMML and AML cell samples. The 'responder' rate of 25% in MDS/CMML/AML cell samples treated with APE1 inhibitor III was similar to the 'responder' rate of 19% in MDS/CMML/AML cell samples treated with talazoparib. However, APE1 inhibitor III exhibited about 30-fold less potent IC 50 values as compared to talazoparib. Our observations are in accordance with the more advanced development of PARP inhibitors (e.g., talazoparib) as compared to APE1 inhibitors (e.g., APE1 inhibitor III). Despite the inferior potency of APE1 inhibitor III, subtoxic doses increased the antileukemic efficacy of decitabine and talazoparib in about 78% and 68% of MDS/CMML/AML cell samples, respectively. As BER is presumed to be involved in the repair of decitabine-induced DNA lesions [45], we hypothesize that APE1 inhibitor III might interact synergistically with decitabine by interruption of BER. Further, APE1 inhibitor III might interact synergistically with talazoparib by inducing suicidal cross-linking of PARP1 to AP sites [46].
Markers predicting the cytotoxic efficacy of talazoparib and APE1 inhibitor III in MDS/CMML and AML cells are, so far, largely unknown. Therefore, several markers including cell proliferation, PARP1/APE1 mRNA expression, γH2AX foci levels, chromosomal aberrations, and gene mutations were analyzed and correlated with cell survival of MDS/CMML and AML cells after treatment with talazoparib and APE1 inhibitor III, respectively. Cell proliferation, PARP1/APE1 mRNA expression, and γH2AX foci levels demonstrated no consistent correlation with the cytotoxic efficacy of talazoparib and APE1 inhibitor III in MDS/CMML and AML cells. Furthermore, no specific chromosomal aberrations with a predictive impact were detected. However, 100% of the talazoparib 'responders' and 83% of the APE1 inhibitor III 'responders' displayed a normal karyotype. Finally, gene mutation analysis suggested that KMT2A-PTD, NRAS, and U2AF1 mutations might be associated with a sensitivity rate of 50% or more towards talazoparib and APE1 inhibitor III, respectively.
Resistance to PARP inhibitors is a major clinical problem. Several categories of PARP inhibitor resistance mechanisms have been described to date and can be assigned to (I) increased drug efflux (e.g., upregulation of ABC transporters), (II) decreased trapping of PARP1 on DNA (e.g., trapping-diminishing PARP1 mutations), (III) restoration of homologous recombination (e.g., reactivation of BRCA1/2, loss of 53BP1, loss of Shieldin factors), and (IV) stabilization of stalled forks (e.g., loss of PTIP or EZH2) (reviewed in [47]). Further work is needed to clarify the role of these resistance mechanisms in MDS/CMML and AML. Possible mechanisms of resistance towards APE1 inhibitor III might include the expression of translesion polymerases (Pol) capable of bypassing AP sites or the upregulation of DSB repair genes.

Cell Proliferation
Cell proliferation of untreated MDS/CMML and AML cells was determined by trypan blue exclusion assay at days 1, 2, and 3 [51].

Cell Survival Assay
The CellTiter-Glo luminescent cell viability assay (Promega, Southampton, UK) was used for the assessment of cell survival of healthy donor cells, MDS/CMML cells, and AML cells according to the manufacturer's instructions. Cryopreserved CD34+ MDS/CMML cells and CD34+ or CD34− AML cells were thawed and expanded in culture for 3 d followed by 3 d of daily drug exposure. The IC 50 designated the drug concentration capable of killing 50% of the cells and was calculated with GraphPad Prism 5 software (GraphPad Software, La Jolla, US) by linear regression. MDS/CMML and AML patient samples were classified as 'responders' or 'non-responders' using a cutoff value of mean IC 50 of healthy controls minus standard error of mean. Experiments were performed once owing to the limited availability of healthy CD34+ donor cells, CD34+ MDS/CMML cells, and CD34+ or CD34− AML cells. DMSO-treated cells were used as control.

Statistical Analysis
All statistical calculations were done with SAS software, release 9.4 (SAS Institute Inc., Cary, NC, USA). For quantitative variables, mean values and standard errors were calculated. Categorical factors are presented with absolute and relative frequencies. In order to compare more than two groups, Kruskal-Wallis tests were performed. For pairwise comparisons, Wilcoxon two-sample tests were used. Test results with p values < 0.05 were considered as statistically significant.
For analysis of PARP1 and APE1 expressions, relative quantification was used according the ∆∆CT method. Normalization of target genes was performed using the arithmetic mean Ct of the nonregulated housekeeping gene GUSB from 6 healthy donor samples.

Conclusions
Our study performed on primary MDS/CMML and AML patient cells provides important preclinical data for a potential clinical use of talazoparib and APE1 inhibitor III alone and in combination with decitabine. In a next step, the in vivo antileukemic activities of talazoparib and APE1 inhibitor III need to be evaluated in clinical studies. Currently, talazoparib is tested as a single agent in advanced hematologic neoplasias (NCT01399840) [42] and in cohesin-mutated AML and MDS with excess blasts (NCT03974217) [43]. Further, decitabine and talazoparib are tested in patients with untreated and relapsed/refractory AML (NCT02878785) [44]. APE1 inhibitor III may enter similar trials in MDS/CMML/AML in the future.