In vitro and in vivo evaluation of cytotoxic effect of targeted agents alone or in combination with chemotherapy for the treatment of adenoid cystic carcinoma

Adenoid cystic carcinoma (ACC) is a rare malignancy characterized by high incidence of relapse. When relapsing, ACC has an indolent but relentless behaviour, thus leading to a poor long-term prognosis. The treatment of choice of relapsing ACC remains surgery followed by radiotherapy, whenever feasible. Therapeutic weapons are limited to systemic drugs. The most widely used chemotherapy regimen is the combination of cisplatin and doxorubicin, however with low response rate and not long lasting; there is also a lack of alternatives for second line therapies in case of disease progression. Therefore, a more comprehensive strategy aimed at identifying at preclinical level the most promising drugs or combination is clearly needed. ve combinations one standard chemotherapy plus one targeted were evaluated, in order to nd the best treatment strategy for ACC. that both and olaparib increased the standard chemotherapy


Introduction
Adenoid cystic carcinoma (ACC) is a rare malignancy arising mostly from salivary glands and from other sites, as trachea, bronchi, breast and skin, with an incidence of 4.5 cases per 100,000 individuals (Ellington et al. 2012). It represents 20% of malignant tumors of the major salivary glands and 58% of the minor salivary glands (Coca-Pelaz et al. 2015). ACC is reported to relapse in till 50-60% of the cases, both locally and at distant sites (Xu et al. 2017; Terhaard et al. 2004); mainly metastatic sites are represented by lung (70% of cases), followed by bone (6%), liver (3%) and rarely brain or other sites (Xu et al. 2018).
Generally, when relapsing, ACC has an indolent but relentless behaviour, thus leading to a poor long-term prognosis; 10-year OS ranges from 52% to 65% in several retrospective series (Ciccolallo et al. 2009;van Weert et al. 2013).
Locoregional or distant recurrences represent a crucial part of the patient journey, as negatively impacting on prognosis and with higher risk of patient's quality of life deterioration. The treatment of choice of relapsing ACC remains surgery followed by radiotherapy, whenever feasible. In case of absence of indications for locoregional treatments, with a patient experiencing symptomatic disease or at high risk of complications, therapeutic weapons are limited to systemic drugs.
The most widely used chemotherapy regimen is the combination of cisplatin and doxorubicin, however with response rate (RR) of 25% and not long lasting; there is also a lack of alternatives for second line therapies in case of disease progression (Laurie et al. 2011). Several target agents have been studied in ACC, mostly as single agents and in the absence of molecular selection. Recently, the most promising results have been obtained with multikinase inhibitor lenvatinib and with antiangiogenetic drugs as sorafenib and axitinib; however, these drugs reported a response rate of about 15% with a non-negligible is the NOTCH-1 mutation (Ho et al. 2019); in this regard, tailored targeted approaches have been studied and are currently ongoing in NOTCH-mutated ACC (Ferrarotto et al. 2020). Overall, targeted agents seem to bene t a small quote of recurrent and/or metastatic (RM) ACC, at the price of drug-induced toxicities.
Therefore, a more comprehensive strategy aimed at identifying at preclinical level the most promising drugs or combination is clearly needed.
The primary aim of our study is the evaluation in vitro and in vivo of new therapeutic strategies that can expand therapeutic possibilities available for patients with RM ACC.

Cell culture
The human ACC (hTERT) cell line, derived from a primary untreated and predominantly cribriform ACC of the tongue base immortalized using h-TERT transfection, was kindly given by Prof. Adel El-Naggar (Department of Pathology, The University of Texas, MD Anderson Cancer Center, Houston, TX, U.S.A.). Cells were maintained in a medium containing 82% RPMI-1640 (Sigma-Aldrich , St. Louis, Missouri, USA), 15% fetal bovine serum (FBS, Life Technologies, Carlsbad, California, USA), 1% penicillin/streptomycin (Life Technologies), 2% L-glutamine (Life Technologies), as previously reported (Li et al. 2014). ACC (hTERT) cells doubling time at 37°C and 5% CO 2 was calculated according to ATCC (American Type Culture Collection, Virginia, USA) indication, with the following formula: DT=T ln2/ln(X e /X b ), where T is the incubation time in any units, X b is the cell number at the beginning of the incubation time, X e is the cell number at the end of the incubation time. The same protocol was applied to calculate the doubling time at 32°C. Cells were authenticated by the GenePrint 10 System (Promega Italia, Milan, Italy), according to the protocols suggested by the manufacturer. Cells were used between passage 144 and 155 and periodically tested for mycoplasma.

Cell viability and cell proliferation Assay
Cell viability was assessed by 3-(4,5-Dimethyl-2-thiazol)-2,5-diphenyl-2H-tetrazolium bromide (MTT) dye reduction assay according to the manufacturer protocol (Sigma Aldrich) and performed as described in Abate et al. (2020). Cell proliferation rate was evaluated with TC20 automated cell counter (Bio-Rad, Segrate, Milan, Italy). Brie y, vehicle-and drugs-treated cells were trypsinized and suspended in culture medium followed by trypan blue dilution (1:2). The parameter settings were established according to the manufacturer instructions. Ten microliters of the mixture were loaded into the opening of the TC20 counting slide. The gating was selected as optimal, based on preliminary validation studies performed.

Drug combination experiments
Drug combination experiments were performed to evaluate drug interactions on ACC cell viability, according to the Chou and Talaly method (Chou and Talalay 1984). Cells were treated for 4 days with the following drugs, used alone or in combination: cisplatin (0.018 -13. Collection of Zebra sh eggs and sh maintenance All zebra sh were handled according to national and international guidelines (EU Directive 2010/63/EU), following protocols approved by the local committee (OPBA protocol no. 211B5.24) and authorized by the Ministry of Health (authorization no. 393/2017-PR). Healthy adult wild-type zebra sh (AB strain) were used for egg production. Fishes were maintained under standard laboratory conditions (Wester eld 2000), at 28°C on a constant 14-hour light/10-hour dark cycle. Fish were fed thrice daily with a combination of granular dry food and fresh artemia (Special Diet Services, SDS Diets; LBS Biotech, Horley, UK). Collection of zebra sh eggs and maintenance of developing embryos has been performed as

Statistical analysis
Data analysis was performed using GraphPad Prism software version 5.02 (GraphPad Software, La Jolla, CA). Statistical analysis was carried out using one-way ANOVA analysis with a post hoc test (Bonferroni's test) for multiple comparisons, considering p < 0.05 as threshold for statistical signi cance. IC 50 values for each drug were calculated by non-linear regression of the concentration-response curves. Data are expressed as mean ± SEM of three independent experiments, unless otherwise speci ed. Cytotoxicity experiments were carried out at least three times, each point run in triplicate.

Results
Standard chemotherapy-and targeted therapy-drugs induced cytotoxicity in ACC cell line ACC (hTERT) cell line doubling time was rstly evaluated ( Figure S1). Being the cells doubling time around 50 hours, ACC cells were treated with the drugs for 4 days.
The cytotoxicity of the two standard chemotherapeutic agents (cisplatin and doxorubicin) and ve targeted therapy-drugs (namely, everolimus, palbociclib, olaparib, vorinostat, and lenvatinib) were tested on ACC (hTERT) cell line.
Also, a cell count was performed in order to evaluate the effect on the cell proliferation rate. As reported in Figure 1B-D, cell count con rmed data obtained in the MTT viability test.
We then evaluated the effect of ACC potentially effective target-therapy drugs. Among the drugs tested, the four drugs listed below resulted to be effective in reducing cell viability, considering drug concentrations compatible with clinical practice: vorinostat (0.1-9 µM), olaparib (0.15-18 µM), palbociclib  Table 1. Lenvatinib was as well tested for cytotoxic effects on ACC (hTERT) cells, but its e cacy was very poor, remaining far above the 50% value of cell viability even at high drug concentrations ( Figure S2).

In vivo effects of both standard chemotherapy and target therapy drugs
The cytotoxic effects observed in vitro was then tested in vivo in the experimental model of cell xenograft on zebra sh embryos. As zebra sh embryos were maintained at the temperature of 32°C, ACC cell behavior was studied in terms of viability and doubling time at 32°C and results indicated that this cell line was able to growth and duplicate at this temperature ( Figure S3). The ACC cells were injected in zebra sh embryos and the tumor area was evaluated at different times. In particular, each single drug or solvent were directly added to the sh water and after 3 days of treatment the tumor areas of vehicle-and drug-treated groups at T0 (2 hours after treatment) and T3 (3 days of treatment) were measured. ACC tumor area was acquired at the uorescent microscope and analysed with the DanioScope software ( Figure 3B).

Vorinostat and olaparib enhanced cytotoxicity induced by standard chemotherapy drugs
In order to evaluate the effect on ACC cells of targeted drugs combined with either doxorubicin or cisplatin, the Chou-Talalay method was applied, as described in Methods. The most promising results were obtained with the combination vorinostat plus cisplatin and vorinostat plus doxorubicin. In these cases, the combined treatments in ACC cells induced synergistic cytotoxic effects compared to each single compound. The concentration-response curves of single standard chemotherapy drug and of the combined treatments are reported in Figure 4A Interestingly, olaparib treatment in ACC cells enhanced cytotoxicity induced by standard chemotherapy drugs. Indeed, both olaparib/cisplatin and olaparib/doxorubicin combined treatment increased cytotoxic effect as compared to each single-drug treatment which resulted in an increase in the potency, as reported in concentration-response curves ( Figure 5A-B). The Chou-Thalalay analysis con rmed the additive effect of the combined treatments (combination index = 1, see Figure 5C-D).
All target therapy drugs were tested in combination with standard chemotherapy drugs, but as emerged in Chou-Thalalay analysis (Figure 6), palbociclib and everolimus did not produce any signi cant increase in the single standard chemotherapy treatments. Speci c concentration-response curves are reported in FigureS4.

Discussion
Chemotherapy has been for years the mainstay of treatment of relapsed ACC not eligible to locoregional treatment, but with no proofs of bene t in increasing survival (Licitra et al. 1991(Licitra et al. , 1996 These unsatisfactory results underline the importance of comprehensively dissecting the therapeutic pathways involved in ACC and widely testing in preclinical models the activity of drug(s). The burden of mutations in ACC is low, with a few alterations recognized as therapeutically actionable (Ho et al. 2019). This is mirrored by the paucity of trials aimed at delivering targeted treatments associated with known mutations or with altered molecular pathways. Moreover, the additive or synergistic effect of chemotherapy and targeted agents have not been widely studied at preclinical level.
In the present study, a human ACC cell line immortalized using h-TERT transfection, was used to test in vitro the pharmacological effects of standard chemotherapy and targeted therapy agents used in monotherapy or in combination, to give the preclinical bases for a pharmacological strategy for ACC treatment. We selected the most employed chemotherapy agents in ACC, cisplatin and doxorubicin Our results demonstrated the in vitro cytotoxic activity of the standard chemotherapy drugs. Among targeted agents, vorinostat reported the highest e cacy, with a cell mortality of over 90%, while modest e cacy emerged for olaparib and palbociclib and poor e cacy for everolimus. Finally, lenvatinib did not induce signi cant cytotoxic effects in ACC (hTERT) cell line, and this poor performance could be explained by the lack of vascular components of in vitro models, that are indeed one of the main mechanisms of action of this drug. The in vitro cytotoxic effect of all the drugs tested was con rmed and In the context of ACC tumor, several studies con rmed that a two-drug polychemotherapy gives better response rate compared with monotherapy (Lorini et al. 2021), even if with limited duration and uncertain bene t in overall survival. Therefore, the association of chemotherapy and targeted agents is a logical step to be carried out (Chabner and Roberts 2005) and we tried to demonstrate the activity of these combinations in the in vitro model of ACC cell line. Our results demonstrated that the two standard chemotherapy agents (cisplatin or doxorubicin) plus vorinostat or olaparib were effective. Indeed, vorinostat signi cantly increased both the e cacy and the potency of the standard chemotherapy agents and the combination exerted a synergic effect. Also olaparib improved the performance of the two chemotherapy agents, resulting in an addictive effect. These effects found their rationale in the speci c molecular mechanism of the drugs: both cisplatin and doxorubicin hamper DNA replication and RNA transcription, the rst inducing the formation of intra-and inter-strand cross-link of nuclear DNA (Wang and Lippard 2005), the second interacting with DNA by intercalation and inhibiting macromolecular biosynthesis. This effect nally leads to cell cycle arrest and apoptosis. Vorinostat increases the formation of hyperacetylated histones, which therefore blocks the interaction of histone with DNA and leaves it more accessible by cisplatin or doxorubicin. Therefore, the synergistic effect induced by vorinostat may be due to the action upstream of the chemotherapy agents, which facilitates and greatly enhances their effects and this result in ACC cells is in line with the nding that vorinostat sensitizes ACC Taken together, our results indicate that combined treatment with vorinostat or olaparib with standard chemotherapy agents cisplatin and doxorubicin is signi cantly more effective than monotherapy. These data set the basis for further studies in a dedicated prospective clinical trial.
We acknowledge that a major limitation of this study lies in the experimental model, as we could perform experiments just in a single cell line; however, up to now, the AdCC-hTERT cells are the only cell models available. Also, we did not explore the role of immunotherapy, as in this preclinical cell line model, the absence of tumor microenvironment prevented the possibility to study the interaction with the immune system. In conclusion, we demonstrated the feasibility of preclinical analyses on cell lines for the combination of chemotherapy and targeted agents in ACC, as a model to study combinations and to exploit the synergy between drugs.

Declarations
Funding: This work was supported by private donation and by local grant of the University of Brescia. All other authors declare that they have no con icts of interest in the area of adenoid cystic carcinoma.
Availability of data and material: All data generated or analysed during this study are included in this published article [and its supplementary information les]. Further details are available on request from the corresponding author.
Code availability: not applicable Author's Contribution: T.S. and A.A. contributed to the collection of data, preparation of the gures and Table 1