Azacytidine plus olaparib for relapsed acute myeloid leukaemia, ineligible for intensive chemotherapy, diagnosed with a synchronous malignancy

Abstract Patients with relapsed/refractory acute myeloid leukaemia (AML), ineligible for intensive chemotherapy and allogeneic stem cell transplantation, have a dismal prognosis. For such cases, hypomethylating agents are a viable alternative, but with limited success. Combination chemotherapy using a hypomethylating agent plus another drug would potentially bring forward new alternatives. In the present manuscript, we present the cell and molecular background for a clinical scenario of a 44‐year‐old patient, diagnosed with high‐grade serous ovarian carcinoma, diagnosed, and treated with a synchronous AML. Once the ovarian carcinoma relapsed, maintenance treatment with olaparib was initiated. Concomitantly, the bone marrow aspirate showed 30% myeloid blasts, consistent with a relapse of the underlying haematological disease. Azacytidine 75 mg/m2 treatment was started for seven days. The patient was administered two regimens of azacytidine monotherapy, additional to the olaparib‐based maintenance therapy. After the second treatment, the patient presented with leucocytosis and 94% myeloid blasts on the bone marrow smear. Later, the patient unfortunately died. Following this clinical scenario, we reproduced in vitro the combination chemotherapy of azacytidine plus olaparib, to accurately assess the basic mechanisms of leukaemia progression, and resistance to treatment. Combination chemotherapy with drugs that theoretically target both malignancies might potentially be of use. Still, further research, both pre‐clinical and clinical, is needed to accurately assess such cases.


| INTRODUC TI ON
Acute myeloid leukaemia (AML) is a malignancy of the myeloid hematopoiesis, characterized by the accumulation of accumulating genetic aberrations. [1][2][3][4] Progress in next-generation sequencing (NGS) has successfully risk-classified AML in accordance with specific gene mutations or associations of gene mutations. [5][6][7] Understanding the genetic background of AML yields new therapies that could potentially involve the use of new alternatives. 8,9 In the current manuscript, we present the clinical scenario of a 44-year-old woman, diagnosed with concomitant relapsed ovarian carcinoma and relapsed AML. She received therapy with the poly-ADP ribose polymerase (PARP) inhibitor olaparib for the ovarian cancer. Not eligible for intensive chemotherapy, treatment with azacytidine was subsequently initiated. After two cycles of treatment, the patient succumbed due to infections. Following this therapeutic failure, we aimed to assess the cellular mechanisms of disease progression in vitro. Thus, we investigated the effects of olaparib that causes synthetic apoptosis in cancers with homologous recombination deficiencies (HRD), 10 in combination with daunorubicin (ODC) or azacytidine (AZA), on two AML cell lines. These settings simulated both the first-line chemotherapy for AML and chemotherapyrefractory AML. The first cell line, OCI/AML3, is characterized by the occurrent mutations in both nucleophosmin (NPM1), a gene involved in DNA single-strand break repairs, 11 and DNA methyltransferase 3 alpha (DNMT3A), a gene involved in resistance to chemotherapyinduced DNA damage. 12 Both mutations are common and present a significant impact in the prognosis of AML patients. 5 The second cell line, THP-1, has mutations and deletions in PTEN, MLL-AF9, MLLT3, TP73 and CDKN2A/B. 13,14 The genetic landscape of THP-1 cell line makes it susceptible to the effects of olaparib, particularly through the presence of a partial deletion in PTEN gene. 9 Consequently, we conducted a set of in vitro assays to establish whether OCI/AML3 is affected by ODC in a similar fashion as THP-1 cells, in comparison with the effects generated by the standard cytarabine-daunorubicin regimen (CDR) (therapy given in first line for AML), as well as in comparison with AZA-based chemotherapy (therapy given to patients ineligible for intensive chemotherapy), thus simulating all the possible clinical scenarios in which a PARP inhibitor might be used in the clinic, in order to properly explain the basic mechanisms of disease progression and resistance to chemotherapy, as THP-1 is PTEN mutant and thus susceptible to olaparib treatment.

| CLINI C AL SCENARI O
In the present manuscript, we present the case of a 44-year-old young woman which was diagnosed with high-grade serous carcinoma of ovarian origin, pT3cNxMx FIGO III C, radically operated in 2016 and treated with adjuvant chemotherapy. Following routine medical analysis, bicytopenia was diagnosed, with marked leukocytosis and 95% blasts on the bone marrow smear. Diagnosis of AML was conclusive, both on the myelogram and flow cytometry. In the bone marrow, a percentage of 80% peroxidase-positive blasts was detected. The immunophenotype was CD34+, HLA DR+, CD117+, Molecular biology was negative for the FLT3 mutation, but positive for NPM1. Standard chemotherapy ('3 + 7') regimen was performed. At the end of the induction treatment, a control myelogram was performed, which showed a high percentage of blasts (40%).
For the refractory AML, second-line treatment was administeredthe FLAG-Etoposide regimen. 15 After the second line of treatment, the AML was in complete remission (CR1), with 3%-4% blasts on the control myelogram. Seven months after CR1, the patient presents with relapsed ovarian carcinoma, for which second-line chemotherapy treatment with paclitaxel and carboplatin was started, after which the remission of ovarian carcinoma was obtained. 16,17 The mutational status of BRCA was positive, and subsequent maintenance treatment with olaparib is initiated. [18][19][20] Seven more months after obtaining the second remission (CR2) of the ovarian carcinoma, on maintenance treatment with olaparib, the complete blood count (CBC) showed pancytopenia and the presence of 4% blasts on blood number 13-BM/2020 (grant director Ciprian Tomuleasa); and by an international collaborative grant of the European Economic Space between Romania and Iceland 2020-2022 (Grant No. 19-COP-0031, grant director Ciprian Tomuleasa) unfortunately died. Following this clinical scenario, we reproduced in vitro the combination chemotherapy of azacytidine plus olaparib, to accurately assess the basic mechanisms of leukaemia progression, and resistance to treatment. Combination chemotherapy with drugs that theoretically target both malignancies might potentially be of use. Still, further research, both pre-clinical and clinical, is needed to accurately assess such cases.

K E Y W O R D S
clinical scenario, combination chemotherapy, hypomethylating agents, refractory acute myeloid leukaemia, synthetic lethality smear. The bone marrow aspirate showed 30% myeloid blasts, with relapsed AML. As the patient was not eligible for intensive chemotherapy, due to the altered physical status, the therapeutic options were now chemotherapy with azacytidine (AZA) monotherapy, decitabine monotherapy or low-dose cytarabine. [21][22][23] Taking into consideration the altered physical status of the patients, azacytidine 75 mg/m 2 treatment was started for seven days. Taking into consideration that olaparib is currently under investigation for relapsed/ refractory(R/R) AML in 2 clinical trials (Table 1), as well as considering that olaparib is currently used for the maintenance therapy for ovarian adenocarcinoma, the therapeutics committee decided to keep both olaparib and azacytidine therapy.
The patient was thus given three cycles of azacytidine plus olaparib combination chemotherapy. After the third cycle, the patient presented with leucocytosis and 94% myeloid blasts in the bone marrow. Later, the patient unfortunately died. Following this clinical scenario, we reproduced in vitro the combination of chemotherapy agents, to properly understand the basic mechanisms of leukaemia progression and resistance to chemotherapy.

| Cell Culture
Both cell lines were cultured in vitro to assess drug treatment,

| Cell cycle assessment
For assessing cell cycle arrest, flow cytometry was used after 48h of

| qRT-PCR analysis
qRT-PCR was performed to analyse the gene expression in relation to previously described drug combinations on two AML cell lines. RNA isolation was performed by using TriReagent TM

| Pre-clinical murine testing of azacytidine plus olaparib for AML
Eight-week-old male athymic nude mice purchased from Charles River Laboratories were used in the study. The animals were

| RE SULTS
Cytarabine and daunorubicin are used in first-line chemotherapy for AML, to achieve remission before an allogeneic stem cell transplantation. 30 From the standpoint of the effects of ODC and CDR on OCI/AML3 and THP-1 blast proliferation, ODC proved to be as potent as CDR in decreasing the viability of treated cells that was compared between cell lines ( Figure 1A-OCI/AML3: 54.9% vs 56.1%, P = .999; THP1:38.5% vs 30.1%, P = .982). Statistical analysis for Figure 1C is shown in Supplementary Table S1. Statistical analysis for Figure 1D using ANOVA is shown in Supplementary Table S2, whereas using pairwise t test is shown in Supplementary Table S3.  In order to assess at functional level whether the high rate of response and similarities observed in both therapeutic approaches are in concordance with the gene expression, we performed qRT-PCR and evaluated the expression of genes located downstream of the PARP signalling pathway. [31][32][33][34][35] The results indicate that both ODC and CDR triggered after 48 hour comparable transcript levels of genes related to the DNA repair system. All the genes evaluated were found to display certain levels of down-regulation ( Figure 1D).
Levels lower than 50% of the control for the OCI-AML3 cell line were observed for two genes, ATM and LIG3, for both therapeutic approaches, but a 50% decrease in RAD51 and PARP1 expression was observed only for ODC-treated cells. ( Figure 3A). The mutational profile of NPM1-positive cells is the same as cells isolated from the bone marrow aspirate of our previously described case, with genes associated with DNA damage repair being down-regulated from diagnosis to relapse. Once olaparib plus AZA was introduced, following relapse, the same genes were up-regulated, consisting with patient leukaemia progression and resistance to therapy ( Figure 3B).
In order to investigate the efficiency of 5-azacytidine plus olaparib combination chemotherapy for the treatment of AML, we developed an experimental mouse model of the disease by injecting AML-Luc cells into the cartilage of the mice knee joint. We let the tumours to develop for 20 days to mimic an advanced form of the disease. Considering the highly experimental character of the protocol, F I G U R E 3 RT-PCR assessment for cells treated with azacytidine, olaparib and olaparib plus azacytidine we decided to conduct for the moment a pilot study for proof of principle and include only three mice with the most uniform tumour distribution in the leg as detected through bioluminescent imaging.
The treatment protocol was followed for 5 consecutive days according to Figure 4A and consisted of 5-azacytidine administrated alone or 5-azacytidine given in combination with olaparib ( Figure 4B). The second bioluminescent imaging exposure after the completion of the treatment showed that the mice from the control group had an progressive evolution of the malignant mass, while 5-azacytidine managed to control to some extent the development of the AML-Luc cells, delaying the tumour spread ( Figure 4C). For the cohort with 5-azacytidine plus olaparib combination chemotherapy, we show a decrease in the tumour formation, results confirmed also by automatic ROI measurement of the bioluminescent signal intensity ( Figure 4D). No significant changes in the bodyweight of the mice or other adverse side effects were observed during the experiments, suggesting that the treatment was well tolerated ( Figure 4E).
Considering the aggressive phenotype of AML, where the survival is very poor even the best available therapeutic option, these results can translate in a significant improvement in the clinical management of the patients. Further studies on larger animal cohorts have to be put in place in order to predict the efficiency of the treatment combination for a potential phase I clinical trial. Still, when looking at the R/R AML scenario, for patients unfit or ineligible for intensive chemotherapy, the data show a totally different story. Adding olaparib to AZA has little or no effect for NPM1-positive cells, as shown by in vitro cell proliferation assays in Figure 2A and backed up by cell cycle analysis and RT-PCR of the genes involved in DNA repair ( Figure 2B). This is consistent with the clinical evolution of our patient, who had a FLT3-negative/NPM1positive mutational status.

| D ISCUSS I ON
PARP1 gene expression is also linked to DNA methylation, with important clinical impact in gynaecological malignancies, especially cancers that are BRCA-mutated. 44

ACK N OWLED G EM ENTS
The authors gratefully acknowledge the insightful support of professor Berindan-Neagoe.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of the study are available from the corresponding author upon reasonable request.