[0001] This application claims the priority benefit of U.S. Provisional Patent Application Serial No. 63/141,308, filed January 25, 2021, which is hereby incorporated by reference in its entirety.

[0002] This invention was made with government support under U54OD020353 awarded by National Institutes of Health. The government has certain rights in the invention.

FIELD

[0003] The present disclosure relates to methods and compositions for treating adenoid cystic carcinoma.

BACKGROUND

[0004] Adenoid cystic carcinoma (ACC) is a relatively rare neoplasm that metastasizes frequently and widely. ACC is the most common malignant tumor of the minor salivary glands and the second most common of the major salivary glands (Coca-Pelaz et al., “Adenoid Cystic Carcinoma of the Head and Neck— An update,” Oral Oncol. 51 :652-661 (2015)). Despite early dissemination, it is relatively slow growing. In the United States, approximately 20,000 patients are living with ACC in various stages of progression. One thousand two hundred new cases are reported annually; approximately 60% of those affected are women. On average, patients with ACC present in their 40s and therefore may live with their cancer for decades depending on the rate of progression, with consequent emotional and financial costs to family and society. The median survival is 85% at 5 years and 34% at 15 years, with lymphovascular invasion most associated with poor prognosis (Ouyang et al., “Risk Factors and Prognosis for Salivary Gland Adenoid Cystic Carcinoma in Southern China: A 25-Year Retrospective Study,” Medicine (Baltimore) 96:e5964 (2017)).

[0005] Patients with ACC have few therapeutic options. Treatment goals are limited and focused on achieving local or regional control through combinations of surgery, radiotherapy, and chemotherapy. Once disseminated or regionally recurrent, there are no effective therapies (Gatta et al., “Major and Minor Salivary Gland Tumours,” Crit. Rev. Oncol. Hematol.

152: 102959 (2020)). Chemotherapy and targeted therapies have proven poorly effective with arbitrary and transient responses (Tchekmedyian et al., “Phase II Study of Lenvatinib in Patients with Progressive, Recurrent or Metastatic Adenoid Cystic Carcinoma,” J. Clin. Oncol. 37: 1529- 1537 (2019)), while regional, focused therapies such as radiotherapy and surgery are used primarily to reduce symptoms (palliation) to address, e.g., bronchial obstruction and symptomatic bone metastases.

[0006] Recent advances in genetic studies have pointed to further challenges: most ACC tumors contain the fusion myeloblastosis viral oncogene homolog-nuclear factor IB (MYB- NF1B (Persson et al., “Recurrent Fusion of MYB and NFIB Transcription Factor Genes in Carcinomas of the Breast and Head and Neck,” Proc. Natl. Acad. Sci. USA 106: 18740-18744 (2009) and Andersson and Stenman, “The Landscape of Gene Fusions and Somatic Mutations in Salivary Gland Neoplasms - Implications for Diagnosis and Therapy,” Oral Oncol. 57:63-69 (2016))) but also include multiple other cancer-associated gene mutations (Ho et al., “The Mutational Landscape of Adenoid Cystic Carcinoma,” Nat. Genet. 45:791-798 (2013); Stephens et al., “Whole Exome Sequencing of Adenoid Cystic Carcinoma,” J. Clin. Invest. 23:2965-2968 (2013); Ross et al., “Comprehensive Genomic Profiling of Relapsed and Metastatic Adenoid Cystic Carcinomas by Next-Generation Sequencing Reveals Potential New Routes to Targeted Therapies,” Am. J. Surg. Pathol. 38:235-238 (2014); and Mitani et al., “Novel MYBL1 Gene Rearrangements with Recurrent MYBL1-NFIB Fusions in Salivary Adenoid Cystic Carcinomas Lacking t(6; 9) Translocations,” Clin. Cancer Res. 22:725-733 (2016)). MYB is a transcriptional activator with a C-terminal inhibitory domain (Sakura et al., “Delineation of Three Functional Domains of the Transcriptional Activator Encoded by the c-myb Protooncogene,” Proc. Natl. Acad. Sci. USA 86:5758-5762 (1989); Weston and Bishop, “Transcriptional Activation by the v- myb Oncogene and its Cellular Progenitor, c-myb,” Cell 58:85-93 (1989); and Dubendorff et al., “Carboxy-Terminal Elements of c-Myb Negatively Regulate Transcriptional Activation in cis and in trans,” Genes Dev. 6(12B):2524-2535 (1992)). Most ACC tumors show activation of MYB through gene fusion of MYB with the transcription factor NFIB due to a 6; 9 translocation (Persson et al., “Recurrent Fusion of MYB and NFIB Transcription Factor Genes in Carcinomas of the Breast and Head and Neck,” Proc. Natl. Acad. Sci. USA 106: 18740-18744 (2009) and Andersson and Stenman, “The Landscape of Gene Fusions and Somatic Mutations in Salivary Gland Neoplasms - Implications for Diagnosis and Therapy,” Oral Oncol. 57:63-69 (2016)) or, less often, by truncation or copy number gain (Persson et al., “Clinically Significant Copy Number Alterations and Complex Rearrangements of MYB and NFIB in Head and Neck Adenoid Cystic Carcinoma,” Genes Chromosomes Cancer 51 :805-817 (2012). Fusion or truncation leads to loss of MYB’ s C-terminus, which is sufficient to generate a constitutively active MYB protein (Gonda et al., “Activation of c-myb by Carboxy-Terminal Truncation: Relationship to Transformation of Murine Haemopoietic Cells in vitro,” EMBO J. 8: 1777-1783 (1989)); though unlikely (Persson et al., “Recurrent Fusion of MYB and NFIB Transcription Factor Genes in Carcinomas of the Breast and Head and Neck,” Proc. Natl. Acad. Sci. USA 106: 18740-18744 (2009)), a function for the accompanying small (5 amino acids) C-terminal fragment of NFIB has not been ruled out. Most patients have additional mutations in other cancer-related genes such as activating mutations in the NOTCH 1 and ERBB3 receptors and regulators of signal transduction and cell cycle (Ho et al., “The Mutational Landscape of Adenoid Cystic Carcinoma,” Nat. Genet. 45:791-798 (2013); Stephens et al., “Whole Exome Sequencing of Adenoid Cystic Carcinoma,” J. Clin. Invest. 23:2965-2968 (2013); Ross et al., “Comprehensive Genomic Profiling of Relapsed and Metastatic Adenoid Cystic Carcinomas by Next-Generation Sequencing Reveals Potential New Routes to Targeted Therapies,” Am. J. Surg. Pathol. 38:235-238 (2014); and Mitani et al., “Novel MYBL1 Gene Rearrangements with Recurrent MYBL1-NFIB Fusions in Salivary Adenoid Cystic Carcinomas Lacking t(6;9) Translocations,” Clin. Cancer Res. 22:725-733 (2016)). Recent studies, both basic and clinical (Bangi et al., “Functional Exploration of Colorectal Cancer Genomes using Drosophila,” Nat. Commun. 7: 13615 (2016); Levine and Cagan, “Drosophila Lung Cancer Models Identify Trametinib plus Statin as Candidate Therapeutic,” Cell Rep. 14: 1477-1487 (2016); and Levinson and Cagan, “Drosophila Cancer Models Identify Functional Differences Between ret Fusions,” Cell Rep. 2016;16:3052-3061), are consistent with a growing body of work demonstrating that tumor heterogeneity and genetic complexity can lead to drug resistance.

[0007] The present disclosure is directed to overcoming these and other deficiencies in the art.

SUMMARY

[0008] One aspect of the present disclosure relates to a method of treating adenoid cystic carcinoma (ACC). This method involves administering, to a subject in need of treatment for adenoid cystic carcinoma, a therapeutically effective amount of a histone deacetylase inhibitor, a beta blocker, and a janus kinase inhibitor, where said administering is effective to treat the adenoid cystic carcinoma in the subject.

[0009] Another aspect of the present disclosure relates to a pharmaceutical composition comprising a histone deacetylase inhibitor, a beta blocker, a janus kinase inhibitor, and a pharmaceutically acceptable carrier.

[0010] A personalized fly-to-bedside therapeutic discovery platform was recently described (Bangi et al., “Functional Exploration of Colorectal Cancer Genomes using Drosophila,” Nat. Commun. 7: 13615 (2016), which is hereby incorporated by reference in its entirety) (FIG. 1 A). Modeling the disease of a patient with colorectal cancer in a personalized Drosophila transgenic model, a novel two-drug cocktail that proved effective in both Drosophila and in the modeled patient was identified. Building on this work, the Examples of the present disclosure provide a fly-to-bedside platform for ACC, a tumor that has resisted targeted therapies. A personalized Drosophila line that targeted five genes altered in a patient’s tumor was developed. This “personalized avatar” exhibited aspects of transformation. This line was used as a screening tool to identify a three-drug cocktail comprising a histone deacetylase inhibitor, a beta blocker, and a j anus kinase inhibitor (vorinostat-pindolol-tofacitinib), that rescued transgene-mediated lethality in the fly avatar and led to stable disease and a metabolic response in the patient lasting for 12 months.

BRIEF DESCRIPTION OF THE DRAWINGS

[0011] FIGs. 1 A-1G show the development of a personalized Drosophila avatar screening platform. FIG. 1 A is a schematic overview of the personalized Drosophila avatar approach. Genomic analysis of a patient’s tumor identified predicted tumor drivers used to develop a personalized fly avatar. Robotics-based drug screening identified a three-drug cocktail that was vetted for safety by a tumor board and internal review board. FIG. IB is a table showing prioritized oncogenes and tumor suppressors that emerged from the genomic analysis. FA TA, ERCC2, and FAT1/FAT3 were heterozygous. FIG. 1C is an image showing that immunohistochemistry (brown) identified high levels of plasma membrane and nuclear NOTCH1, indicating elevated NOTCH1 protein and activity in patient tumor sections obtained prior to treatment. Similar immunohistochemical assays failed to validate elevated MAP2K2 activity (pERK) or loss of MAX, and neither were included in the final avatar model. FIG. ID is a schematic of a transformation vector used to target 4 of 5 cancer genes to different Drosophila tissues. Inducible Notch overexpression (UAS-Notch) was introduced by standard genetic crosses. FIG. IE is a graph showing that small hairpins targeting xpd. f and kug in CPCT012.2 led to a -50% reduction in expression as assessed with qPCR. This line has the best model of heterozygosity. FIG. IF is a plot quantifying results of directing ptc > CPCT012 expression on the wing’s ptc domain, which led to expansion of the domain including a loss of the sharp boundary. Results are represented as the ratio of the ptc domain area to total wing disc area. FIG. 1G is an image showing an example of ptc > CPCT012-vaQ<33?L\.Q<4 expansion. The ptc domain was visualized with an included UAS-GFP marker (green). Insets highlight expansion; dotted lines indicate added black background to square images. Error bars represent standard error of the mean. [0012] FIGs. 2A-2D show the results of a screen for candidate combinations of FDA- approved drugs. FIG. 2A is a flowchart of a multi-step drug screen. An initial screen of the Focused FDA Library yielded tofacitinib and docetaxel as weak single agent hits. Subsequent screens identified tofacitinib, vorinostat, and pindolol as an effective 3 -drug combination. FIG. 2B is a plot showing data demonstrating initial rescue by docetaxel and tofacitinib as single agents. FIG. 2C is a plot showing data demonstrating CPCT012 rescue to adulthood by gemcitabine plus tofacitinib, an effective two-drug combination. Tofacitinib was used at a dose below that required for significant rescue. FIG. 2D is a plot showing data demonstrating CPCT012 rescue to adulthood by tofacitinib, vorinostat, and pindolol. The 3 -drug combination proved the most effective at rescuing CPCT012 to adulthood. Asterisks (*) in FIGs. 2B-2D indicate p < 0.05 as assessed by Student’s t-test. Error bars represent standard error of the mean. [0013] FIGs. 3A-3C show body PET scans from baseline and after 6 months of treatment. FIG. 3 A is a pair of graphs showing that prior to the start of treatment, tumor volume (left panel) and standardized uptake value (SUV; right panel) of the 2-deoxy-2-[ ¹⁸ F]fluoro-D- glucose (FDG) tracer were increasing over time, indicating progressive disease. Initiation of treatment led to stabilization of total tumor volume and reduction of lung SUV. FIG. 3B is an image showing control scans just prior to treatment, which highlights extensive tumor metastases in the bone and lung. FIG. 3C is an image showing that glucose tracer (FDG) uptake in the lung and bone metastases was substantially reduced after 6 months of therapy in imaged sites; further, no new lesions appeared. These data indicate clinical benefit from the drug treatment, manifested as reduced FDG uptake and absence of progression, h = heart, r = renal tubules, b = bladder.

[0014] FIGs. 4A-4B show patient somatic genomic profiles. The patient tumor samples from 2016 to 2019 exhibited significant genomic differences. Somatic protein-altering molecular variants (SNVs and indels with AF > 0.05) (FIG. 4A) and somatic copy number variant (sCNV) (FIG. 4B) profiles of the four tumor samples are summarized, as assessed with saasCNV (Zhang and Ho, “SAAS-CNV: A Joint Segmentation Approach on Aggregated and Allele Specific Signals for the Identification of Somatic Copy Number Alterations with Next- Generation Sequencing Data,” PLoS Comput. Biol. 11 :el004618 (2015), which is hereby incorporated by reference in its entirety). The 2019 specimens contained de novo variants and more unstable sCNV profiles.

[0015] FIG. 5 shows sCNV profiles of 2019 specimens. GATK4 Somatic CNV algorithm -based profiles of the 2019 specimens are consistent with results from the saasCNV analysis shown in FIG. 4B. [0016] FIG. 6 is a table listing variants identified for the four patient tumor samples (related to FIGs. 1A-1G). Note conserved alterations in ANKAR, ANXA8L2, CNTRL, EP300, and GEMIN5 in the 2019 samples. Only the variant in PCDH9 was common to all samples; this was not included in the avatar models because it was not predicted to alter protein function.

[0017] FIG. 7 is a table showing GATK4 somatic gene variant analysis (related to FIGs. 1 A-1G). sCNV segmentation profiles of 2019 specimens as determined by the GATK4 somatic CNV algorithm.

[0018] FIG. 8 is a table showing COSMIC Cancer Gene Census analysis of germline variants (related to FIGs. 1 A-1G). The list of germline variants in annotated tumor suppressor genes (TSGs) by COSMIC Cancer Gene Census were also examined for the loss of sCNV status to identify any bi-allelic inactivated gene candidates. The sole germline variant with gnomAD allele frequency <= 0.1% in a loss sCNV segment (KMT2D p.Arg228Gly) was found to have lost the mutant allele, not the wild type allele. Therefore, no functionally relevant, bi-allelic inactivated TSGs were identified in the 2019 specimens.

[0019] FIG. 9 is a table showing quality control metrics (related to FIGs. 1 A-1G). The sequencing QC metrics of tumor and matching normal patient specimens used in the study.

[0020] FIG. 10 is a table showing the construction of knockdown vectors (related to FIGs. 1 A-1G). Sequences for the hairpin, spacer and assembled knockdown vectors are listed for CPCT012.1 and CPCT012.2. The assembled vector was then placed into a transformation vector.

DETAILED DESCRIPTION

Definitions

[0021] Unless otherwise indicated, the definitions and embodiments described in this and other sections are intended to be applicable to all embodiments and aspects of the present application herein described for which they are suitable as would be understood by a person skilled in the art.

[0022] Unless defined otherwise, all technical and scientific terms used in this disclosure have the same meanings as commonly understood by one of ordinary skill in the art to which this disclosure belongs.

[0023] Preferences and options for a given aspect, feature, embodiment, or parameter of the invention should, unless the context indicates otherwise, be regarded as having been disclosed in combination with any and all preferences and options for all other aspects, features, embodiments, and parameters of the invention described in the present disclosure. [0024] As used herein, the singular forms “a,” “an,” and “the” and the like include plural referents unless the context clearly dictates otherwise. Thus, for example, reference to “a compound” includes both a single compound and a plurality of different compounds.

[0025] The term “and/or” as used herein means that the listed items are present, or used, individually or in combination. In effect, this term means that “at least one of’ or “one or more” of the listed items is used or present.

[0026] As will be understood by one skilled in the art, for any and all purposes, such as in terms of providing a written description, all ranges disclosed herein also encompass any and all possible subranges and combinations of subranges thereof. Any listed range can be easily recognized as sufficiently describing and enabling the same range being broken down into at least equal halves, thirds, quarters, fifths, tenths, and so on. As a non-limiting example, each range discussed herein can be readily broken down into a lower third, middle third, and upper third, and so on. As will also be understood by one skilled in the art all language such as “up to,” “at least,” and the like include the number recited and refer to ranges which can be subsequently broken down into subranges as discussed above. Finally, as will be understood by one skilled in the art, a range includes each individual member.

[0027] In understanding the scope of the present disclosure, the term “comprising” and its derivatives, as used herein, are intended to be open ended terms that specify the presence of the stated features, elements, components, groups, integers, and/or steps, but do not exclude the presence of other unstated features, elements, components, groups, integers and/or steps. The foregoing also applies to words having similar meanings such as the terms, “including”, “involving”, “having”, and their derivatives. The term “consisting” and its derivatives, as used herein, are intended to be closed terms that specify the presence of the stated features, elements, components, groups, integers, and/or steps, but exclude the presence of other unstated features, elements, components, groups, integers and/or steps. The term “consisting essentially of’, as used herein, is intended to specify the presence of the stated features, elements, components, groups, integers, and/or steps as well as those that do not materially affect the basic and novel characteristic(s) of features, elements, components, groups, integers, and/or steps. In embodiments or claims where the term comprising (or the like) is used as the transition phrase, such embodiments can also be envisioned with replacement of the term “comprising” with the terms “consisting of’ or “consisting essentially of.” The methods, kits, systems, and/or compositions of the present disclosure can comprise, consist essentially of, or consist of, the components disclosed. [0028] In embodiments comprising a “further”, “additional”, or “second” component, the second component as used herein is different from the other components or first component. A “third” component is different from the other, first, and second components, and further enumerated or “additional” components are similarly different.

[0029] Also provided herein are embodiments wherein any embodiment described herein may be combined with any one or more other embodiments, provided the combination is not mutually exclusive.

[0030] The term “avatar” refers to an organism (e.g., an organism of the Drosophila genus or subgenus, such as the species Drosophila melariogaster) whose genomic, proteomic and/or phenomic profile is modified as compared to non-modified organism. The genomic, proteomic and/or phenomic profile of the organism (e.g., Drosophila may be modified such that the activity of one or more orthologs of disease drivers (e.g., tumor drivers) are altered in the organism as compared to the non-modified organism (e.g., Drosophila). The genomic, proteomic, and/or phenomic profile of the organism may be modified such that the activity of one or more orthologs of disease drivers (e.g., tumor drivers) is altered to correspond to the profile of a subject suffering from cancer; sometimes referred to herein as a “personalized avatar.”

[0031] The terms “disease driver” and “disease drivers” refer to any gene product(s) whose altered activity in a subject contributes to progression of disease. The terms “tumor driver” and “tumor drivers” refer to any gene product(s) whose altered activity in a subject contributes to the progression of cancer; for example, increased activity of one or more oncogene products and/or reduced activity of one or more tumor suppressor gene products that contribute to tumor progression.

[0032] The term “genome” refers to an organism’s complete set of DNA, including all of its genes. Each genome contains all of the information needed to build and maintain that organism.

[0033] The terms “therapies” and “therapy” can refer to any protocol(s), method(s), and/or agent(s) that can be used in the treatment of a disease (e.g., cancer). In certain embodiments, the terms “therapies” and “therapy” refer to chemotherapy, immune modulatory drugs, and/or other therapies useful in the treatment of a cancer. In a specific embodiment, a therapy includes adjuvant therapy. For example, using a therapy in conjunction with a drug therapy, biological therapy, hormonal therapy, surgery, and/or supportive therapy.

[0034] The term “transgene” may refer to a cDNA of an ortholog of a disease driver (e.g., tumor driver), a cDNA of an activator of an ortholog of a disease driver, a cDNA of an inhibitor of an ortholog of a disease driver, a small interfering RNA (siRNA) that targets an ortholog of the disease driver (e.g., tumor driver), a short hairpin RNA (shRNA) that targets an ortholog of the disease driver (e.g., tumor driver), or any combination thereof.

Methods and Compositions for Treating Adenoid Cystic Carcinoma

[0035] Adenoid Cystic Carcinoma (ACC) is a rare cancer type that may originate in glands of the head and neck. Tumors commonly invade along nerve tracks in the head and neck, making surgery challenging. Follow-up treatments for recurrence or metastasis including chemotherapy and targeted therapies have shown limited efficacy, emphasizing the need for new therapies.

[0036] The Examples of the present disclosure describe the analysis of tumor/cancer cells from a human subject to characterize the patient’s mutations, particularly those predicted to increase activity of one or more oncogene products and/or reduced activity of one or more tumor suppressor gene products. This information was used to design and construct a Drosophila avatar that recapitulated the patient’s phenome and served as a screening tool which identified a three-drug cocktail that rescued transgene-mediated lethality in the fly avatar and led to stable disease and a metabolic response in the patient lasting for 12 months.

[0037] Accordingly, one aspect of the present disclosure relates to a method of treating adenoid cystic carcinoma (ACC). This method involves administering, to a subject in need of treatment for adenoid cystic carcinoma, a therapeutically effective amount of a histone deacetylase inhibitor, a beta blocker, and a j anus kinase inhibitor, where said administering is effective to treat the adenoid cystic carcinoma in the subject.

[0038] The adenoid cystic carcinoma may be a salivary gland adenoid cystic carcinoma (SACC), a lacrimal gland adenoid cystic carcinoma (LACC), adenoid cystic carcinoma of the ceruminous gland (ACCCG), or an adenoid cystic carcinoma of the female genital tract (see, e.g., Xu et al., “MYB Promotes the Growth and Metastasis of Salivary Adenoid Cystic Carcinoma,” Int. J. Oncol. 54: 1579-1590 (2019); Chen et al., “Adenoid Cystic Carcinoma of the Lacrimal Gland is Frequently Characterized by MYB Rearrangement,” Eye (Lond) 31 :720-725 (2017); Conlin et al., “Ceruminous Gland Adenoid Cystic Carcinoma With Contralateral Metastasis to the Brain,” Arch. Pathol. Lab. Med. 126:87-89 (2002); Ebelhar et al., “Ceruminous Adenoid Cystic Carcinoma of External Auditory Canal,” J. Int. Adv. Otol. 13:292-294 (2017); Rais et al., “Adenoid Cystic Carcinoma of the Uterine Cervix: A Report of 2 Cases,” Case Rep. Pathol. 2017:8401741 (2017); and Xing and Lu, “Distinctive Clinicopathological Features and Disease-Specific Survival of Adenoid Cystic Carcinoma and Adenoid Basal Carcinoma in the Lower Female Genital Tract,” Oncol. Rep. 41 : 1769-1778 (2019), which are hereby incorporated by reference in their entirety). In some embodiments, the adenoid cystic carcinoma is a salivary gland adenoid cystic carcinoma.

[0039] The terms “treat” and “treating” in the context of the administration of a therapeutically effective amount of a combination of agents (e.g., a histone deacetylase inhibitor, a beta blocker, and a j anus kinase inhibitor) refers to a treatment/therapy from which a subject in need of treatment for a disease (e.g., adenoid cystic carcinoma) receives a beneficial effect, such as the reduction, decrease, attenuation, diminishment, stabilization, remission, suppression, inhibition or arrest of the development or progression of the disease e.g., adenoid cystic carcinoma), or a symptom thereof. In some embodiments, the treatment/therapy that a subject receives does not cure the disease e.g., adenoid cystic carcinoma), but prevents the progression or worsening of the disease. In certain embodiments, the treatment/therapy that a subject receives does not prevent the onset/development of disease e.g., adenoid cystic carcinoma), but may prevent the onset of disease e.g., adenoid cystic carcinoma) symptoms.

[0040] The terms “subject” or “patient” are used interchangeably. As used herein, the terms “subject” and “subjects” refer to an animal. For example, the subject may be a mammal. Suitable mammals include non-human mammals (e.g., a camel, donkey, zebra, cow, horse, horse, cat, dog, rat, and mouse, etc.), non-human primates (e.g., a monkey, chimpanzee, etc.), and a human. In some embodiments of the methods according to the present disclosure, the subject is a non-human mammal. In certain embodiments of the methods according to the present disclosure, the subject is a pet (e.g., dog or cat) or farm animal (e.g., a horse, pig or cow). In other specific embodiments of the methods according to the present disclosure, the subject is a human.

[0041] In some embodiments of the methods disclosed herein, the subject treated in accordance with the methods described herein has been diagnosed with adenoid cystic carcinoma. Techniques for diagnosing adenoid cystic carcinoma are known to one of skill in the art and include, without limitation, biopsy (e.g., fine needle biopsy), magnetic resonance imaging (MRI), computation tomography (CT or CAT scan), positron emission tomography (PET) or PET-CT scan, etc.

[0042] In some embodiments of the methods disclosed herein, the method further involves selecting a subject in need of treatment for adenoid cystic carcinoma prior to said administering.

[0043] As described herein supra, adenoid cystic carcinoma is a rare tumor of secretory glands, which is characterized by slow growth kinetics and perineural invasion. Many adenoid systemic carcinoma tumors have slow growth kinetics and do not benefit from systemic chemotherapy (Dillon et al., “Adenoid Cystic Carcinoma: A Review of Recent Advances, Molecular Targets, and Clinical Trials,” Head Neck 38:620-627 (2016), which is hereby incorporated by reference in its entirety). Nevertheless, several chemotherapy studies have been performed over the years. The results show consistently low response rates to cytotoxic chemotherapy for metastatic disease. Thus, there is no accepted standard systemic chemotherapy for patients with adenoid systemic carcinoma tumors. A consensus is to reserve chemotherapy for palliation of patients with symptomatic metastases or rapidly progressing disease who are not candidates for other treatment modalities or clinical trials (Dillon et al., “Adenoid Cystic Carcinoma: A Review of Recent Advances, Molecular Targets, and Clinical Trials,” Head Neck 38:620-627 (2016), which is hereby incorporated by reference in its entirety).

[0044] In some embodiments, the adenoid cystic carcinoma is resistant to treatment with one or more chemotherapeutic agents. For example, the adenoid cystic carcinoma may be resistant to treatment with one or more chemotherapeutic agents selected from the group consisting of Imatinib Mesylate (Gleevec™), Cisplatin (Platinol®, Platinol®-AQ), Gefitinib (Iressa®), Bortezomib (Velcade®), Laptinib (TYKERB®), Cetuximab (Erbitux®), Sunitinib (Sutent®), Gemcitabine (Gemzar®), Carboplatin (Paraplatin®), Paclitaxel (Taxol®, Onxal™), and combinations thereof. In some embodiments, the adenoid cystic carcinoma is resistant to treatment with one or more chemotherapeutic agents selected from the group consisting of Gemcitabine (Gemzar®), Carboplatin (Paraplatin®), Paclitaxel (Taxol®, Onxal™), and combinations thereof.

[0045] About 40% of patients with adenoid cystic carcinoma develop metastatic disease (Dodd and Slevin,” Salivary Gland Adenoid Cystic Carcinoma: A Review of Chemotherapy and Molecular Therapies,” Oral Oncol. 42:759-769 (2006), which is hereby incorporated by reference in its entirety). In some embodiments, the subject has metastatic adenoid cystic carcinoma. The most common sites of metastases are the lungs followed by bone, liver, skin, breast, and rarely the brain (Spiro, R., “Distant Metastasis in Adenoid Cystic Carcinoma of Salivary Origin,” Am. J. Surg. 174:495-498 (1997), which is hereby incorporated by reference in its entirety). In some embodiments, the metastasis is selected from the group consisting of a lung metastasis, a bone metastasis, a liver metastasis, a skin metastasis, a breast metastasis, a brain metastasis, and combinations thereof.

[0046] The adenoid cystic carcinoma may be characterized by a mutation in one or more genes selected from the group consisting of FAT1, FAT2, FAT3, FAT4, ERCC2, and combinations thereof. The human FAT gene family consists of the FAT1, FAT2, FATS, and FAT4 genes (Katoh, M., “Function and Cancer Genomics of 7’47' Family Genes,” Int. J. Oncol. 41 : 1913-1918 (2012), which is hereby incorporated by reference in its entirety). FAT1 is most homologous to FAT3. Human FAT family genes as well as Drosophila fat family genes encode large proteins with extracellular Cadherin repeats, EGF-like domains, and Laminin G-like domain(s). FAT1, FAT3, and FAT4 are atypical cadherins and key regulators of cancer-relevant cellular processes including planar cell polarity and MST/HIPPO signaling, which regulates organ size. They are associated as tumor suppressors in a variety of tumors including ovarian, medullary thyroid, gastric, cervical, colorectal, bladder, and squamous cell carcinomas, as well as ACC (Ho et al., “The Mutational Landscape of Adenoid Cystic Carcinoma,” Nat.

Genet. 45:791-798 (2013); Gao et al., “Genetic Landscape of Esophageal Squamous Cell Carcinoma,” Nat. Genet. 46: 1097-1102 (2014); Chen et al., “A Novel Gene Signature Combination Improves the Prediction of Overall Survival in Urinary Bladder Cancer,” J. Cancer 10:5744-5753 (2019); Huang et al., “Comprehensive Genomic Variation Profiling of Cervical Intraepithelial Neoplasia and Cervical Cancer Identifies Potential Targets for Cervical Cancer Early Warning,” J. Med. Genet. 56: 186-194 (2019); Skuja et al., “Deletions in Metastatic Colorectal Cancer with Chromothripsis,” Exp. Oncol. 41 :323-327 (2019); Malgundkar et al., “FAT4 Silencing Promotes Epithelial-to-Mesenchymal Transition and Invasion via Regulation of YAP and P-Catenin Activity in Ovarian Cancer,” BMC Cancer 20:374 (2020); Melis et al., “Mutations in Long-Lived Epithelial Stem Cells and their Clonal Progeny in Pre-Malignant Lesions and in Oral Squamous Cell Carcinoma,” Carcinogenesis 41 : 1553 (2020); Qu et al., “Genomic and Transcriptomic Characterization of Sporadic Medullary Thyroid Carcinoma,” Thyroid 30: 1025-1036 (2020); and Wang et al., “Integrated Characterisation of Cancer Genes Identifies Key Molecular Biomarkers in Stomach Adenocarcinoma,” J. Clin. Pathol. 73:579-586 (2020), which are hereby incorporated by reference in their entirety).

[0047] The human ERCC2 gene encodes XPD, a protein associated with regulation of TFIIH-mediated transcription and DNA damage. XPD has effects on progression of a broad palette of tumor types when mutated or altered — increased or decreased — in expression, and variants in ERCC2 have been associated with altered drug response, especially platinum-based therapies (Du et al., “Associations of Polymorphisms in DNA Repair Genes and MDR1 Gene with Chemotherapy Response and Survival of Non-Small Cell Lung Cancer,” PLoS one 9:e99843 (2014); Fu et al., “Association between the Asp312Asn, Lys751Gln, and Argl56Arg Polymorphisms in XPD and the Risk of Prostate Cancer,” Technol. Cancer Res. Treat. 16:692- 704 (2017); Tan et al., “Genetic Polymorphisms and Platinum-Based Chemotherapy Treatment Outcomes in Patients with Non-Small Cell Lung Cancer: A Genetic Epidemiology Study Based Meta- Analysis,” Sci. Rep. 7:5593 (2017); Pajuelo-Lozano et al., “XPA, XPC, and XPD Modulate Sensitivity in Gastric Cisplatin Resistance Cancer Cells,” Front. Pharmacol. 9: 1197 (2018); and Liu et al., “Association of XPD Asp312Asn Polymorphism and Response to Oxaliplatin-Based First-Line Chemotherapy and Survival in Patients with Metastatic Colorectal Cancer,” Adv. Clin. Exp. Med. 28: 1459-1468 (2019), which are hereby incorporated by reference in their entirety).

[0048] The mutation may be a missense mutation, a nonsense mutation, a frameshift mutation, a splice site mutation, an indel mutation, a translocation mutation (inter-chromosomal translocation or intra-chromosomal translocation).

[0049] In some embodiments, the mutation is a missense mutation. As used herein, the term “missense mutation” refers to a change in the DNA sequence that changes a codon in the mRNA that is normally translated as one amino acid into a codon that is translated as a different amino acid. Some but not all missense mutations result in a non-functional gene-product. Some missense mutations may also result in a gain of function. In any embodiment of the methods according to the present disclosure, the adenoid cystic carcinoma comprises a missense mutation in FA TE ERCC2, and/or FAT 1/3.

[0050] In some embodiments, the mutation is a nonsense mutation. As used herein, the term “nonsense mutation” refers to the substitution of a single base pair that leads to the appearance of a stop codon where previously there was a codon specifying an amino acid. The presence of this premature stop codon results in the production of a shortened, and likely nonfunctional, protein.

[0051] In some embodiments, the mutation is a frameshift mutation. As used herein, the term “frameshift mutation” refers to a type of mutation involving the insertion or deletion of a nucleotide in which the number of deleted base pairs is not divisible by three. If a mutation disrupts the reading frame, then the entire DNA sequence following the mutation will be read incorrectly.

[0052] In some embodiments, the mutation is an indel. As used herein, the term “indel” refers to the insertion and/or deletion of nucleotides into genomic DNA and include events less than 1 kb in length.

[0053] In some embodiments, the mutation is a chromosomal translocation mutation. As used herein, the term “translocation mutation” refers to a type of abnormality in which a chromosome breaks and a portion of it reattaches to a different chromosome. The translocation may results in the transfer of a segment from one chromosome to a non-homologous chromosome (inter-chromosomal translocation) or to a new site on the same chromosome (intra- chromosomal translocation). In some embodiments of the methods according to the present disclosure, the adenoid cystic carcinoma comprises MYB-NFIB inter-chromosomal translocation mutation.

[0054] In some embodiments of the methods according to the present disclosure, the adenoid cystic carcinoma is characterized by one or more mutations in the MYB/MYC signaling pathway and/or the NOTCH signaling pathway. Suitable such mutations are well known in the art (see, e.g., Ho et al., “The Mutational Landscape of Adenoid Cystic Carcinoma,” Nat.

Genet. 45:791-798 (2013) and Rettig et al., “Whole-Genome Sequencing of Salivary Gland Adenoid Cystic Carcinoma,” Cancer Prev. Res. (Phila) 9:265-274 (2016), which are hereby incorporated by reference in their entirety).

[0055] In some embodiments of the methods according to the present disclosure, the adenoid cystic carcinoma is characterized by one or more mutations in the MYB/MYC signaling pathway in one or more genes selected from the group consisting of MYB, NFIB, MYBL1, MYCN, MYCBP2, MGA, and MCM4. In accordance with such embodiments, the one or more mutations in the MYB/MYC signaling pathway is MYB-NFIB inter-chromosomal translocation mutation, e.g., a t(6;9) (q22-23;p23-24) translocation (Persson et al., “Recurrent Fusion of MYB and NFIB Transcription Factor Genes in Carcinomas of the Breast and Head and Neck,” Proc. Natl. Acad. Set. USA. 106: 18740-18744 (2009) and Andersson and Stenman, “The Landscape of Gene Fusions and Somatic Mutations in Salivary Gland Neoplasms - Implications for Diagnosis and Therapy,” Oral Oncol. 57:63-69 (2016)), which are hereby incorporated by reference in their entirety). Such translocations predominately result in MYB-NF1B gene fusions, which are present in nearly half of adenoid cystic carcinoma tumors (McIntyre et al., “MYB-NF1B Fusions Identified in Archival Adenoid Cystic Carcinoma Tissue Employing NanoString Analysis: An Exploratory Study,” Diagn. Pathol. 14:78 (2019), which is hereby incorporated by reference in its entirety). Additional exemplary MYB/MYC signaling pathway mutations include, without limitation, splice site and coding mutations involving exon 10 of MYB and truncating mutations in the CTF/NFI transcription modulation domains (e.g., Y249*, P390fs) (Persson et al., “Recurrent Fusion of MYB and NFIB Transcription Factor Genes in Carcinomas of the Breast and Head and Neck,” Proc. Natl. Acad. Sci. USA 106:18740- 18744 (2009), which is hereby incorporated by reference in its entirety)

[0056] In some embodiments of the methods according to the present disclosure, the adenoid cystic carcinoma is characterized by one or more mutations in the NOTCH signaling pathway in one or more genes selected from the group consisting of NOTCH1, FOXP2, DTX4, FBXW7, CNTN6, MAML3, and combinations thereof. In accordance with such embodiments, the one or more mutations in the NOTCH signaling pathway is a NOTCH1 gain of function mutation. Suitable NOTCH1 gain of function mutations include, without limitation, in-frame mutations in exons 25 to 28 that disrupt the negative regulatory region and lead to ligandindependent Notchl activation and stop-codon or nonsense mutations in exon 34 that result in deletion of the C-terminal degron domain (e.g., Pro-Glu-Ser-Thr-rich domain). Exemplary such mutations include, without limitation, S2467fs*, L1600Q, L1600Q/S2467fs, and combinations thereof (Ferrarotto et al., “Activating NOTCH1 Mutations Define a Distinct Subgroup of Patients With Adenoid Cystic Carcinoma Who Have Poor Prognosis, Propensity to Bone and Liver Metastasis, and Potential Responsiveness to Notchl Inhibitors,” Biology of Neoplasia 35:352- 360 (2016), which is hereby incorporated by reference in its entirety).

[0057] In some embodiments of the methods according to the present disclosure, the subject has a mutation in FAT4, FAT 1/3, ERCC2, MYB, and NOTCH 1.

[0058] The histone deacetylase inhibitor, the beta blocker, and the j anus kinase inhibitor may be administered to the subject by any route known to one of skill in the art. For example, the histone deacetylase inhibitor, the beta blocker, and the j anus kinase inhibitor may be administered locally or systemically to a subject. The histone deacetylase inhibitor, the beta blocker, and/or the j anus kinase inhibitor may be administered to the subject parenterally, orally, topically, transdermally, intranasally, intraperitoneally, intratumorally, intradermally, or intracerebrally. In some embodiments, the histone deacetylase inhibitor, the beta blocker, and/or the j anus kinase inhibitor is administered subcuteanously, intravenously, or intramuscularly. If the histone deacetylase inhibitor, the beta blocker, and/or the j anus kinase inhibitor is already FDA approved or approval from another regulatory authority, the histone deacetylase inhibitor, the beta blocker, and/or the j anus kinase inhibitor may be administered by a route previously approved by the FDA or another regulatory authority.

[0059] The histone deacetylase inhibitor, the beta blocker, and the j anus kinase inhibitor may be administered concurrently or sequentially. For example, one drug may be administered prior to (e.g., 5 minutes, 15 minutes, 30 minutes, 45 minutes, 1 hour, 2 hours, 3 hours, 4 hours, 6 hours, 12 hours, 18 hours, 24 hours, 2 days, 5 days, 7 days, 14 days, or 21 days prior to) the administration of another drug. The histone deacetylase inhibitor, the beta blocker, and the j anus kinase inhibitor may be administered via the same route of administration or different routes of administration to the adenoid cystic carcinoma patient.

[0060] The amount of the histone deacetylase inhibitor, the beta blocker, and the j anus kinase inhibitor used to treat the subject may be extrapolated by one of skill in the art from the avatar. Alternatively or in addition, the amount of the histone deacetylase inhibitor, the beta blocker, and the j anus kinase inhibitor to administer to the subject may be determined based on cell culture studies using patient adenoid cystic carcinoma cells exposed to one or more of the histone deacetylase inhibitor, the beta blocker, and the j anus kinase inhibitor and/or animal model studies (e.g., patient xenograft studies) using the histone deacetylase inhibitor, the beta blocker, and the j anus kinase inhibitor. If the histone deacetylase inhibitor, the beta blocker, and/or the j anus kinase inhibitor has been FDA approved or approved by another regulatory authority, the dose and/or frequency of administration previously approved can be used to determine the dosing regimen for the drug or combination of drugs. In some embodiments, a low dose of the histone deacetylase inhibitor, the beta blocker, and the janus kinase inhibitor is initially administered to the subject and the dose is increased over time. In addition or alternatively, the frequency of the histone deacetylase inhibitor, the beta blocker, and the janus kinase inhibitor administered to the patient may be altered (e.g., increased or decreased). In certain embodiments, the dose and/or frequency of administration of the histone deacetylase inhibitor, the beta blocker, and the janus kinase inhibitor to the subject is adjusted over time depending on the condition of the subject (e.g., the growth of a tumor, whether the adenoid cystic carcinoma has metastasized, whether the patient is experiencing any side effects, such as adverse events, etc.).

[0061] In some embodiments, the dose and/or frequency of administration of the histone deacetylase inhibitor, the beta blocker, and/or the janus kinase inhibitor to the subject is adjusted based on the expression or activity of a disease driver in a sample from the subject after administration of the histone deacetylase inhibitor, the beta blocker, and/or the janus kinase inhibitor to the subject. For example, the dosing frequency of the histone deacetylase inhibitor, the beta blocker, and/or the janus kinase inhibitor may be increased if the expression or activity of a disease driver (e.g., oncogene) is increased in a sample from the patient after administration of the drug or combination of drugs to the subject as compared to the expression or activity of the disease driver (e.g., oncogene) prior to administration of the drug or combination of drugs to the subject. The dosing frequency of the histone deacetylase inhibitor, the beta blocker, and/or the janus kinase inhibitor may be increased if the expression or activity of a disease driver (e.g., tumor suppressor) is decreased in the sample from the subject after administration of the histone deacetylase inhibitor, the beta blocker, and/or the janus kinase inhibitor to the subject as compared to the expression or activity of the disease driver (e.g., tumor suppressor) prior to administration of the histone deacetylase inhibitor, the beta blocker, and the janus kinase inhibitor to the subject.

[0062] The dosing frequency of the histone deacetylase inhibitor, the beta blocker, and/or the janus kinase inhibitor may be decreased or maintained if the expression or activity of a disease driver (e.g., oncogene) is decreased in the sample from the subject after administration of the histone deacetylase inhibitor, the beta blocker, and the j anus kinase inhibitor to the subject as compared to the activity of the disease driver (e.g., oncogene) prior to administration of the histone deacetylase inhibitor, the beta blocker, and the j anus kinase inhibitor to the subject. The dosing frequency of the histone deacetylase inhibitor, the beta blocker, and/or the janus kinase inhibitor may be decreased or maintained if the expression or activity of a disease driver (e.g., tumor suppressor) is increased in the sample from the subject after administration of the histone deacetylase inhibitor, the beta blocker, and the janus kinase inhibitor to the subject as compared to the activity of the disease driver (e.g., tumor suppressor) prior to administration of the histone deacetylase inhibitor, the beta blocker, and the janus kinase inhibitor to the subject.

[0063] In certain embodiments, the efficacy of the histone deacetylase inhibitor, the beta blocker, and the janus kinase inhibitor for the treatment of a cancer patient is assessed by measuring one or more biomarkers of the adenoid cystic carcinoma. Biomarkers for different adenoid cystic carcinoma are known in the art and can be used to assess the efficacy of the histone deacetylase inhibitor, the beta blocker, and the janus kinase inhibitor to treat adenoid cystic carcinoma in a subject. Techniques known to one of skill in the art can be used to measure the biomarkers at the RNA and/or protein level.

[0064] In some embodiments of the methods according to the present disclosure, said administering (e.g., the histone deacetylase inhibitor, a beta blocker, and a janus kinase inhibitor) results in one, two, three or more of the following effects: complete response, partial response, increase in overall survival, increase in disease free survival, increase in objective response rate, increase in time to progression, increase in progression-free survival, increase in time-to- treatment failure, and improvement or elimination of one or more symptoms of adenoid cystic carcinoma.

[0065] In some embodiments, said administering (e.g., the histone deacetylase inhibitor, a beta blocker, and a janus kinase inhibitor) is effective to prolong overall survival and/or progression-free survival in the subject. In some embodiments, a method of treating adenoid cystic carcinoma as described herein results in an increase in overall survival. In other embodiments, a method of treating adenoid cystic carcinoma as described herein results in an increase in progression-free survival. In other specific embodiments, a method of treating adenoid cystic carcinoma as described herein results in an increase in overall survival and an increase in progression-free survival.

[0066] The term “complete response” refers to an absence of clinically detectable disease with normalization of any previously abnormal imaging or serum studies. [0067] The term “partial response” refers to at least about a 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, or 90% decrease in all measurable tumor burden (/.< ., the number of malignant cells present in the subject, or the measured bulk of tumor masses or the quantity of abnormal monoclonal protein) in the absence of new lesions.

[0068] The term “overall survival” is defined as the time from randomization until death from any cause, and is measured in the intent-to-treat population. Overall survival should be evaluated in randomized controlled studies. Demonstration of a statistically significant improvement in overall survival can be considered to be clinically significant if the toxicity profile is acceptable, and has often supported new drug approval.

[0069] Several endpoints are based on tumor assessments. These endpoints include disease free survival (DFS), objective response rate (ORR), time to progression (TTP), progression-free survival (PFS), and time-to-treatment failure (TTF). The collection and analysis of data on these time-dependent endpoints are based on indirect assessments, calculations, and estimates (e.g., tumor measurements).

[0070] Generally, “disease free survival” or “DFS” is defined as the time from randomization until recurrence of tumor or death from any cause. Although overall survival is a conventional endpoint for most adjuvant settings, DFS can be an important endpoint in situations where survival may be prolonged, making a survival endpoint impractical. DFS can be a surrogate for clinical benefit or it can provide direct evidence of clinical benefit. This determination is based on the magnitude of the effect, its risk-benefit relationship, and the disease setting. The definition of DFS can be complicated, particularly when deaths are noted without prior tumor progression documentation. These events can be scored either as disease recurrences or as censored events. Although all methods for statistical analysis of deaths have some limitations, considering all deaths (deaths from all causes) as recurrences can minimize bias. DFS can be overestimated using this definition, especially in patients who die after a long period without observation. Bias can be introduced if the frequency of long-term follow-up visits is dissimilar between the study arms or if dropouts are not random because of toxicity.

[0071] As used herein, “objective response rate” or “ORR” is defined as the proportion of patients with tumor size reduction of a predefined amount and for a minimum time period. Response duration usually is measured from the time of initial response until documented tumor progression. Generally, the FDA has defined ORR as the sum of partial responses plus complete responses. When defined in this manner, ORR is a direct measure of drug antitumor activity, which can be evaluated in a single-arm study. If available, standardized criteria should be used to ascertain response. A variety of response criteria have been considered appropriate (e.g., RECIST criteria) (Therasse et al., “New Guidelines to Evaluate the Response to Treatment in Solid Tumors. European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada,” J. Natl. Cancer Inst. 92:205- 216 (2000), which is hereby incorporated by reference in its entirety). The significance of ORR is assessed by its magnitude and duration, and the percentage of complete responses (no detectable evidence of tumor).

[0072] As used herein, “time to progression” or “TTP” and “progression-free survival” or “PFS” have served as primary endpoints for drug approval. TTP is defined as the time from randomization until objective tumor progression; TTP does not include deaths. PFS is defined as the time from randomization until objective tumor progression or death. Compared with TTP, PFS is the preferred regulatory endpoint. PFS includes deaths and thus can be a better correlate to overall survival. PFS assumes patient deaths are randomly related to tumor progression. However, in situations where the majority of deaths are unrelated to cancer, TTP can be an acceptable endpoint.

[0073] As an endpoint to support drug approval, PFS can reflect tumor growth and be assessed before the determination of a survival benefit. Its determination is not confounded by subsequent therapy. For a given sample size, the magnitude of effect on PFS can be larger than the effect on overall survival. However, the formal validation of PFS as a surrogate for survival for the many different malignancies that exist can be difficult. Data are sometimes insufficient to allow a robust evaluation of the correlation between effects on survival and PFS. Cancer trials are often small, and proven survival benefits of existing drugs are generally modest. The role of PFS as an endpoint to support licensing approval varies in different cancer settings. Whether an improvement in PFS represents a direct clinical benefit or a surrogate for clinical benefit depends on the magnitude of the effect and the risk-benefit of the new treatment compared to available therapies.

[0074] As used herein, “time-to-treatment failure” or “TTF” is defined as a composite endpoint measuring time from randomization to discontinuation of treatment for any reason, including disease progression, treatment toxicity, and death. TTF is not recommended as a regulatory endpoint for drug approval. TTF does not adequately distinguish efficacy from these additional variables. A regulatory endpoint should clearly distinguish the efficacy of the drug from toxicity, patient, or physician withdrawal, or patient intolerance.

[0075] In some embodiments, treating a subject comprises an improvement in and/or the elimination of one or more symptoms of adenoid cystic carcinoma in the subject. The one or more symptoms of adenoid cystic carcinoma may include, but are not limited to, a lump on the roof of the mouth, under the tongue, or in the bottom of the mouth; an abnormal area on the lining of the mouth; numbness of the upper jaw, palate, face, or tongue; difficulty swallowing or chewing; hoarseness; dull pain; a bump or nodule in front of the ear or underneath the jaw; paralysis of a facial nerve; fatigue; pain; weakness; loss of appetite; swollen lymph nodes; persistent bad breath; and unintended weight loss.

[0076] In some embodiments, said administering is effective to induce regression of a primary tumor and/or a metastatic tumor in the subject. Techniques for evaluating adenoid cystic carcinoma tumor regression and/or metastatic disease are known to one of skill in the art and include, without limitation, biopsy (e.g., fine needle biopsy), magnetic resonance imaging (MRI), computation tomography (CT or CAT scan), positron emission tomography (PET) or PET-CT scan, etc.

[0077] In some embodiments, the method further involves administering a chemotherapeutic agent to the selected subject. The chemotherapeutic agent may be selected from the group consisting of alkylating agents, anthracyclines, cytoskeletal disruptors, inhibitors of topoisomerase I, inhibitors of topoisomerase II, kinase inhibitors, nucleotide analogs and precursor analogs, peptide antibiotics, platinum-based agents, retinoids, and vinca alkaloids. Exemplary alkylating agents include, but are not limited to, bifunctional alkylators including but not limited to Cyclophosphamide, Meehl orethamine, Chlorambucil and Melphalan; and monofunctional alkylators including but not limited to Nitrosoureas and Temozolomide. Exemplary anthracyclines include, but are not limited to, Daunorubicin, Doxorubicin, Epirubicin, Idarubicin, Mitoxantrone, Varubicin. Exemplary cytoskeletal disruptors include, but are not limited to, taxanes, including but not limited to Paclitaxel and Docetaxol; as well as Epothilones. Exemplary inhibitors of topoisomerase I , include but are not limited to Irenotecan and Topotecan. Exemplary inhibitors of topoisomerase II include, but not limited to, Etoposide, Teniposide and Tafluposide. Exemplary kinase inhibitors include, but are not limited to, Bortezomib, Erlotinib, Gefitinib, Imatinib, Vemurafenib, and Vismodegib. Exemplary nucleotide analogs and precursor analogs include, but are not limited to, Azacitidine, Azathioprine, Capecitabine, Cytarabine, Doxifluridine, Fluorouracil, Gemcitabine, Hydroxyurea, Mercaptopurine, Methotrexate, and Thioguanine (aka Tioguanine). Exemplary peptide antibiotics include, but are not limited to, Bleomycin and Actinomycin. Exemplary platinumbased agents include, but are not limited to, Carboplatin, Cisplatin, and Oxaliplatin. Exemplary retinoids include, but not limited to, Tretinoin, Alitretinin, and Bexarotene. Exemplary vinca alkaloids and derivatives include, but are not limited to Vinblastine, Vincristine, Vindesine, and Vinorelbine. [0078] In some embodiments, the chemotherapeutic agent is selected from the group consisting of Gemcitabine (Gemzar®), Carboplatin (Paraplatin®), and Paclitaxel (Taxol®, Onxal™).

[0079] Another aspect of the present disclosure relates to a pharmaceutical composition comprising: a histone deacetylase inhibitor, a beta blocker, a j anus kinase inhibitor, and a pharmaceutically acceptable carrier.

[0080] As used herein, the term “histone deacetylase” or “HD AC” refers to any one of a family of enzymes that remove acetyl groups from the 8-amino groups of lysine residues at the N-terminus of a histone. Unless otherwise indicated by con-text, the term “histone” is meant to refer to any histone protein, including Hl, H2A, H2B, H3, H4, and H5, from any species. Preferred histone deacetylases include class I and class II enzymes. Preferably the histone deacetylase is a human HDAC, including, but not limited to, HDAC-1, HDAC-2, HDAC-3, HDAC-4, HDAC-5, HDAC-6, HDAC-7, and HDAC-8.

[0081] The term “histone deacetylase inhibitor” refers to a compound that is capable of inhibiting the deacetylation of histones in vivo, in vitro, or both. As such, histone deacetylase inhibitors inhibit the activity of at least one histone deacetylase. As a result of inhibiting the deacetylation of at least one histone, an increase in acetylated histone occurs and accumulation of acetylated histone is a suitable biological marker for assessing the activity of HDAC inhibitors. Therefore, procedures that can assay for the accumulation of acetylated histones can be used to determine the HDAC inhibitory activity of compounds of interest.

[0082] Suitable histone deacetylase inhibitors are well known in the art and include, without limitation, (i) short-chain fatty acids for example butyrate, 4-phenylbutyrate, or valproic acid; (ii) hydroxamic acids, for example, suberoylanilide hydroxamic acid (SAHA), biaryl hydroxamate A-161906, bicyclic aryl-N-hydroxycarboxamides, CG-1521, PXD-101, sulfonamide hydroxamic acid, LAQ-824, oxamflatin, scriptaid, m-carboxy cinnamic acid bishydroxamic acid, trapoxin-hydroxamic acid analogue, trichostatins like trichostatin A (TSA), TSA analogues such as 5-(4-dimethylaminobenzoyl)-aminovaleric acid hydroxamide (4-Me2N- BAVAH), m-carboxycinnamic acid bis-hydroxamideoxamflatin (CBHA), ABHA, Scriptaid, pyroxamide, and propenamides; (iii) epoxyketone-containing cyclic tetrapeptides, e.g., trapoxins, apidicin, depsipeptide, HC -toxin, chlamydocin, diheteropeptin, WF-3161, Cyl-1 and Cyl-2; (iv) benzamides or non-epoxyketone-containing cyclic tetrapeptides for example FR901228; apicidin, cyclic-hydroxamic-acid-containing peptides (CHAPs), benzamides, MS-275 (MS-27- 275), and CI-994; (v) depudecin; (vi) PXD101; and (vii) organosulfur compounds. Additional examples of HDAC inhibitors include TSA, TPXA and B, oxamflatin, FR901228 (FK228), trapoxin B, CHAP1, aroylpyrrol ylhydroxy-amides (APHAs), apicidin, and depudecin.

[0083] In some embodiment of the methods and compositions according to the present disclosure, the histone deacetylase inhibitor is selected from the group consisting of Vorinostat (Zolinza®, suberoylanilide hydroxamic acid), Romidepsin (Istodax®, NSC 630176, FR901228, FK228, depsipeptide), Trichostatin A (TSA), Belinostat (Beleodaq®, PXD-101), Entiostat (MS- 275, SNDX-275), Mocetinostat (MGCD0103), valproic acid, and sodium phenylbutyrate.

[0084] Beta-1 receptors are G-protein-coupled receptors (specifically Gs alpha subunit) whose action is exerted through the cyclic AMP (cAMP) and cAMP-dependent protein kinase action with resultant calcium ion concentration increases. Increased intracellular calcium increase inotropy in the heart through calcium-induced exchange facilitated by the sarcoplasmic reticulum.

[0085] Beta-2 receptors are found in smooth muscle (especially bronchial, vascular, gastrointestinal, and uterine), skeletal muscle, the myocardium, and the liver. Stimulation of these receptors causes smooth muscle relaxation, which may result in peripheral vasodilation with subsequent hypotension and reflex tachycardia.

[0086] As used herein, the term “beta blocker” refers to a compound that blocks the action of substances, such as adrenaline, on nerve cells and causes blood vessels to relax and dilate (widen). This allows blood to flow more easily and lowers blood pressure and the heart rate. Beta-blockers are traditionally used to treat high blood pressure, chest pain (angina), abnormal heart rate (arrhythmia), congestive heart failure, and several other conditions. Beta blockers may be classified as non-selective (blocking both beta-1 and beta-2 receptors) and selective (blocking only beta-1 receptors).

[0087] Suitable beta blockers are well known in the art and include, without limitation, (i) non-selective beta blockers such as Propranolol (Inderal®, InnoPran XL®), Carvedilol (Coreg®), nadolol (Corgard®), timolol maleate (Blocadren®), penbutolol sulfate (Levatol®), sotalol hydrochloride (Betapace®), Labetalol (Normodyne®, Trandate®), and Pindolol (Visken®) and (ii) selective beta blockers such as Atenolol (Tenormin®), nebivolol (Bystolic®), metoprolol (v), Bisoprolol (Zebeta®), Esmolol (Brevibloc®), betaxolol hydrochloride (Kerlone), and Acebutolol (Sectral®).

[0088] In some embodiments of the methods and compositions according to the present disclosure, the beta blocker is selected from the group consisting of Pindolol (Visken®), Acebutolol (Sectral®), Atenolol (Tenormin®), Bisoprolol (Zebeta®), Carvedilol (Coreg®), Esmolol (Brevibloc®), Labetalol (Normodyne®, Trandate®), Metoprolol (Lopressor®, Toprol XL®), Nadolol (Corgard®), Nebivolol (Bystolic®), and Propranolol (Inderal®, InnoPran XL®). [0089] Janus kinases are kinds of tyrosine kinases that are bound to the cytoplasmic regions of type I and II cytokine receptors. Multimerization of receptors occurs when ligands bind to their receptors. Subunits of some receptor are expressed as homodimers, e.g. erythropoietin and growth hormone; while other receptor subunits are expressed as heteromultimers, such as interferons (IFN) and Interleukins (IL) (Rawlings et al., “The JAK/STAT Signaling Pathway,” J. Cell Sci. 117: 1281-1283 (2004), which is hereby incorporated by reference in its entirety). Activation of the receptors that are associated with JAKs is critical to initiate the JAK transphosphorylation and subsequent recruitment of one or more STATs to be phosphorylated (O’Shea and Murray,” Cytokine Signaling Modules in Inflammatory Responses,” Immunity 28:477-487 (2008), which is hereby incorporated by reference in its entirety). Eventually, dimerized STATs enter to the nucleus and regulate transcription of myriad target genes.

[0090] The human JAK family contains four JAKS: JAK1, JAK2, JAK3 and TYK2 (Stark et al., “How Cells Respond to Interferons,” Annu. Rev. Biochem. 67:227-264 (1998), which is hereby incorporated by reference in its entirety). Each JAK member comprises several distinct domains which are described as follows: N-terminal FERM domain containing three subdomains Fl, F2 &F3, which is responsible for protein-protein interactions, such as adaptor and scaffolding interactions with membrane associated proteins; the SH2(Src homology 2) domain, which is a motif containing approximately 100 residues that binds to phosphotyrosine residues and which plays a role in the activation and dimerization of STATs; the central pseudokinase domain; and a conserved PTK domain is located at the C-terminus, which contains approximately 250-300 residues and an ATP -binding site juxtaposing a catalytic region and which is responsible for phosphorylation of specific tyrosine residues positioned on special downstream substrates.

[0091] As used herein, the term “janus kinase inhibitor” refers to a compound (e.g., a small molecule compound) that interacts with and thereby reduces the signaling activity of a Janus kinase (JAK), preferably by reducing its enzymatic activity. Suitable janus kinase inhibitors are well known in the art and include, without limitation, those targeting (i) JAK1/JAK2 such as Ruxolitinib (INC424), Baricitinib (INCB028050), Momelitinib, INCB039110, AZD1480, LY3009104 (formerly INCB028050), CYT387, GLPG-0634, and SAR302503 (TG101348); (ii) JAK1/JAK3 such as Peficitinib; (iii) JAK1 such as GLPG0634 and GSK2586184; (iv) JAK2 such as CEP-33779, AC-430, R723, BMS911543, Lestaurtinib (CEP-701), BMS-911543, and Pacritinib (SB 1518); (v) JAK3 such as VX-509, R348, and R- 348; and (vi) JAK3>JAK1»JAK2 such as Tofactinib (CP690,550) (see, e.g., Kontzias et al., “Jakinibs: A New Class of Kinase Inhibitors in Cancer and Autoimmune Disease,” Curr. Opin. Pharmacol. 12:464-470 (2012); Deuse et al., “Novel immunosuppression: R348, a JAK3- and Syk-Inhibitor Attenuates Acute Cardiac Allograft Rejection,” Transplantation 85:885-892 (2008); DeVries et al., “GSK2586184, A JAK1 Selective Inhibitor, in two Patients with Ulcerative Colitis,” BMJ Case Rep. 2017:bcr2017221078 (2017); and Wan et al., “Discovery of a Highly Selective JAK2 Inhibitor, BMS-911543, for the Treatment of Myeloproliferative Neoplasms,” ACS Med. Chem. Lett. 6:850-855 (2015), which are hereby incorporated by reference in their entirety).

[0092] In some embodiments of the methods and compositions according to the present disclosure, the janus kinase inhibitor is selected from the group consisting of Tofactinib (CP690,550), CYT387, Baricitinib (INCB028050), Ruxolitinib (INCB018424), TG101348 (SAR302503), Lestaurtinib (CEP-701), AZD1480, R348, VX-509, GLPG0634, GSK2586184, AC-430, Pacritinib (SB1518), and BMS-911543.

[0093] In some embodiments of the methods and compositions according to the present disclosure, the histone deacetylase inhibitor is Vorinostat (Zolinza®, suberoylanilide hydroxamic acid), the beta blocker is Pindolol (Visken®), and the janus kinase inhibitor is Tofactinib (CP690,550).

EXAMPLES

[0094] The examples below are intended to exemplify the practice of embodiments of the disclosure but are by no means intended to limit the scope thereof.

Materials and Methods

Enrollment

[0095] This work was regulated by a biorepository protocol that managed specimen acquisition, processing, and inventory and a separate protocol that governed analysis of genome data, model building validation, and drug screening. Patient treatment was regulated by a third protocol and a personalized treatment consent written for the therapy recommended by the multidisciplinary tumor board. All protocols and consent forms were approved by the Mount Sinai Institutional Review Board. Whole Exome Sequencing

[0096] Genome assays were performed on a fresh frozen tumor specimen and whole blood collected at the time of consent to serve as a patient matched normal (i.e. germline) control. Protocols for sample processing, genomic assays, and analysis were as previously described (Uzilov et al., “Development and Clinical Application of an Integrative Genomic Approach to Personalized Cancer Therapy,” Genome Med. 8(1):62 (2016) and Bangi et al., “A Personalized Platform Identifies Trametinib plus Zoledronate for a Patient with KRAS-Mutant Metastatic Colorectal Cancer,” Set. Adv. 5(5):eaav6528 (2019), which are hereby incorporated by reference in their entirety).

[0097] Paired-end (2x100 nt) whole exome sequencing (WES) was carried using Illumina HiSeq 2500 or 4000 instruments. Hybridization capture was carried out using SureSelect Human All Exon V5 (Agilent) for the 2016 specimen and Twist Human Core Exome for all three 2019 specimens. Libraries for tumor and normal samples were multiplexed in a 2: 1 or 3 : 1 ratio of tumor to normal. The same normal (blood) sample was used as the matching normal for all four tumor specimens. Full details of the resulting sequencing QC result are given in FIG. 9.

[0098] Alignment of de-multiplexed FASTQ files and calling of molecular variants (somatic or germline SNVs and small insertions/deletions) were carried out using a Sema4 inhouse pipeline (Tigris version 2.2.0). This pipeline implemented Broad Institute’s best practices for running the Genome Analysis Toolkit (GATK) version 4.0.4.0 (McKenna et al., “The Genome Analysis Toolkit: A MapReduce Framework for Analyzing Next-Generation DNA Sequencing Data,” Genome Res. 20(9): 1297-303 (2010) and DePristo et al., “A Framework for Variation Discovery and Genotyping using Next-Generation DNA Sequencing Data,” Nature Genet. 43(5):491-498 (2011), which are hereby incorporated by reference in their entirety). The pipeline was written in Workflow Description Language (WDL) and executed using the Cromwell workflow engine using the Amazon AWS Batch backend. Sequencing reads were aligned to the hgl9 human reference genome (USC Genome Browser Downloads) using BWA- MEM version 0.7.17 (Li, H., “Aligning Sequence Reads, Clone Sequences and Assembly Contigs with BWA-MEM,” arXiv.1303.3997 (2013), which is hereby incorporated by reference in its entirety), duplicate reads were marked out using Picard MarkDuplicates, and base quality recalibration was carried out using GATK4’s BaseRecalibrator, resulting in BAM files used for all further analysis. Somatic mutations were called using GATK4’s Mutect2 in tumor/normal analysis mode with default settings, then filtered using GATK4’s FilterMutectCalls and FilterByOrientationBias. The resulting VCFs were loaded into a custom MySQL database schema using in-house scripts and annotated using RVS (Hakenberg et al., “Integrating 400 Million Variants from 80,000 Human Samples with Extensive Annotations: Towards a Knowledge Base to Analyze Disease Cohorts,” BMC Bioinformatics 17:24 (2016), which is hereby incorporated by reference in its entirety) and SnpEff 4.0b (Cingolani et al., “A Program for Annotating and Predicting the Effects of Single Nucleotide Polymorphisms, SnpEff: SNPs in the Genome of Drosophila melanogaster Strain, wl 118; iso-2; iso-3,” Fly (Austin) 6:80-92 (2012), which is hereby incorporated by reference in its entirety), using the Ensembl version 75/GRCh37 resource bundle. The germline protein-altering variant KMT2D (p.Arg228Gly) was manually reviewed in IGV (Thorvaldsdottir et al., “Integrative Genomics Viewer (IGV): High- Performance Genomics Data Visualization and Exploration,” Brief Bionform. 14: 178-192 (2013), which is hereby incorporated by reference in its entirety) to inspect supporting alignment quality. Of note, the WES sequencing hybridization capture region was different in the 2016 specimen, and GATK4 Somatic CNV could only be applied to the 2019 specimens (FIG. 5).

Somatic Copy Number Variants (sCNV) Calling and Analysis

[0099] All 2016 and 2019 specimens were analyzed for sCNV using saasCNV (Zhang and Ho, “SAAS-CNV: A Joint Segmentation Approach on Aggregated and Allele Specific Signals for the Identification of Somatic Copy Number Alterations with Next-Generation Sequencing Data,” PLoS Comput. Biol. 11 :el004618 (2015), which is hereby incorporated by reference in its entirety). In addition, since internal normal (blood) samples were available from other individuals sequenced via the same hybridization capture kit (Twist Human Core Exome) similar to the three 2019 specimens, a panel of control samples was created; therefore, GATK4 somatic CNV was used as an alternative sCNV calling tool to confirm the sCNV in all three 2019 specimens. NOTCH1 amplification was functionally confirmed by identifying nuclear- localized protein using immunohistochemical assays on tumor sections using previously described staining and scoring protocols (Donovan et al., “A Systems Pathology Model for Predicting Overall Survival in Patients with Refractory, Advanced Non-Small-Cell Lung Cancer Treated with Gefitinib,” Eur. J. Cancer 45: 1518-1526 (2009) and Bangi et al., “A Personalized Platform Identifies Trametinib plus Zoledronate for a Patient with KRAS-Mutant Metastatic Colorectal Cancer,” Sci. Adv. 5(5):eaav6528 (2019), which are hereby incorporated by reference in their entirety)

Model Building and Validation

[0100] In order to generate the patient model, a previously reported multigenic Drosophila transformation vector (Ni et al., “A Genome-Scale shRNA Resource for Transgenic RNAi in Drosophila,” Nat. Methods 8:405-407 (2011) and Bangi et al., “A Personalized Platform Identifies Trametinib plus Zoledronate for a Patient with KRAS-Mutant Metastatic Colorectal Cancer,” Sci. Adv. 5(5):eaav6528 (2019), which are hereby incorporated by reference in their entirety) that contains three UAS cassettes was used. Each UAS cassette in the vector contains a unique multiple cloning site (MCS) flanked by a UAS promoter and SV40 transcription terminator sequence. The coding sequence for the truncated MYB protein observed in the patient was generated by PCR from a cDNA clone of the human MYB gene using the following forward and reverse primer sequences respectively: atgcGGCCGGCCcaaaATGGCCCGAAGACCCCG (SEQ ID NO: 1) and atgcTTAATTAATTACTGCAAGGGGCTCGCCA (SEQ ID NO:2). These primers also included restriction sites for enzymes Fsel and PacI to the 5' and 3' ends of the product respectively, which were used to clone the amplified product into MCS1 of the multi genic vector (FIG. ID).

[0101] For gene knockdown, a synthetic short hairpin cluster design where individual short hairpins were separated by spacer sequences found 5' to well-expressed endogenous microRNAs in the Drosophila genome was used. Criteria for short hairpin selection and the protocol for short cluster design have both been previously reported (Vert et al., “An Accurate and Interpretable Model for siRNA Efficacy Prediction,” BMC Bioinformatics 7:520 (2006); Ni et al., “A Genome-Scale shRNA Resource for Transgenic RNAi in Drosophila.'' Nat. Methods 8:405-407 (2011); and Bangi et al., “A Personalized Platform Identifies Trametinib plus Zoledronate for a Patient with KRAS-Mutant Metastatic Colorectal Cancer,” Sci. Adv. 5(5):eaav6528 (2019), which are hereby incorporated by reference in their entirety). In order to increase the likelihood of success, two synthetic clusters that target the same three genes using different short hairpin sequences were designed (012.1 and 012.2). The hairpin, spacer and final cluster sequences can be found in FIG. 10. Cluster synthesis was outsourced to GENEWIZ. Sequence confirmed synthetic clusters 012.1 and 012.2 were separately cloned into a UAS cassette of the multigenic vector that already contained the truncated human MYB coding sequence using Xbal and Notl enzymes using restriction sites that were appended to the 5' and 3' ends of the clusters respectively. The UAS cassette that the clusters cloned into was specifically designed for short hairpin expression (Vert et al., “An Accurate and Interpretable Model for siRNA Efficacy Prediction,” BMC Bioinformatics 7:520 (2006); Ni et al., “A Genome-Scale shRNA Resource for Transgenic RNAi in Drosophila ' Nat. Methods 8:405-407 (2011); and Bangi et al., “A Personalized Platform Identifies Trametinib plus Zoledronate for a Patient with KRAS-Mutant Metastatic Colorectal Cancer,” Sci. Adv. 5(5):eaav6528 (2019), which are hereby incorporated by reference in their entirety). The resulting two multigenic vectors both contained the same MYBAC transgene and a different hairpin cluster targeting the same three genes using different hairpin sequences (FIG. 10). [0102] After the final vectors were sequence confirmed, transgenic flies were generated by PhiC31 -mediated targeted integration (Bischof et al., “An Optimized Transgenesis System for Drosophila using Germ-Line-Specific phiC31 Integrases,” Proc. Natl. Acad. Sci. USA 104:3312-3317 (2007), which is hereby incorporated by reference in its entirety) into the second Drosophila chromosome using the landing site attp40. Tranegensis was outsourced to BestGene. After transgenic lines containing the multigenic vectors were generated, a previously established UAS-Notch line on chromosome three (Matsuno et al., “Involvement of a Proline- Rich Motif and RING-H2 Finger of Deltex in the Regulation of Notch Signaling,” Development 129: 1049-1059 (2002), which is hereby incorporated by reference in its entirety) was introduced into the multigenic vector background to generate the final patient models: w; UAS-multigenic [UAS-MYBAC, UAS-3sh(Xpd,ft,kug)] attp40; UAS-N/S-T, Cy, Hu, Tb.

[0103] The efficacy of gene knock-down induced by the hairpin clusters were evaluated by qPCR analysis. Experimental animals were generated by crossing patient models containing hairpin clusters 012.1 and 012.2 to a ubiquitously expressed gal4 line that also contains a. gal 80' (tub-gal4, tiib-gal8(D) to transiently induce transgene expression for three days during larval development. Whole larvae with genotypes 1) tub-gal4, tub-gal8(f ^s >UAS-012.1; UAS-N 2) tub-gal4, tub- gal8(f ^s >UAS-012.2; UAS-N, and 3) tub-gal4, tub-gal8(f ^s /+ as controls were collected for RNA extraction (three biological replicates per genotype; six larvae per replicate) and stored in 300 pl RNALater (Life Technologies). For RNA extraction, Qiagen’s RNeasy Plus Kit with RNase-free DNase Set for on-column DNA digestion was used following the manufacturer’s instructions. RNA concentration was measured using Qubit. For qPCR analysis, 1 pg of RNA per replicate was used to generate complementary DNA (cDNA) using the High-Capacity RNA-to-cDNA kit (Life Technologies). qPCR assays were performed using the PerfeCTa SYBR Green FastMix for IQ (VWR Scientific). To identify the best housekeeping control, a panel of 5 candidate genes (rpl32, hsp83, sdha, rpl!3a and cyp33) were assayed. Of these, rpl!3a produced the most robust and consistent result across replicates and genotypes and was selected as the housekeeping control. qPCR data were analyzed using the AAC(t) method as previously described (Sopko et al., “Combining Genetic Perturbations and Proteomics to Examine Kinase-Phosphatase Networks in Drosophila Embryos,” Dev. Cell 31 : 114-127 (2014), which is hereby incorporated by reference in its entirety).

Quantifying ptc>CPCT012

[0104] Wing discs from wandering w~ (control) and ptc>CPCT012 L3 larvae grown at 27°C were fixed with 4% paraformaldehyde, mounted with Vectashield plus DAPI and imaged on a Leica DM5500 Q microscope. The same exposure (220 ms, 545 ms) and gain (2.8, 6.0) was used within each channel for all images. The ptc region was quantified using FIJI (ImageJ, v2.1.0) and represented as a ratio of the ptc region to total wing disc area to account for varying sizes of each individual wing disc. An outline was drawn around (i) the ptc region (GFP) or (ii) the entire wing disc (DAPI) manually using the freehand selection tool to exclude extraneous signal from the peripodial membrane, trachea, etc. Areas within the outlines were measured by establishing a threshold (adjust-+ threshold), then quantifying the GFP-labeled area (analyze- + analyze particles'). Area Ratios were calculated, graphed, and analyzed by Student’s t-test (Microsoft Excel).

Drug Screening

[0105] A focused FDA Library was custom-made in house using drugs individually purchased as powder from Selleck Chemicals, LC Laboratories, TOcris Bioscience or MedChemExpress depending on availability. Stock solutions were made by dissolving drugs in water or 100% dimethyl sulfoxide (DMSO) at the highest possible concentration based on solubility information provided by the manufacturers. For each drug, the highest dose with no discernible toxicity (Maximum Tolerable Dose or MTD) on wildtype animals was selected for screening. Full FDA Library used for this patient was purchased from Selleck Chemicals in a 96- well format (100 pl of 10 mM solution). Both libraries were aliquoted into 384-well plates for screening. Drug screens were conducted at a single dose for each drug for both libraries along with DMSO and no DMSO controls (8 replicates per condition for Focused FDA screens and 4 replicates per condition for the Full FDA screens)

[0106] Drug-food was prepared using the PerkinElmer automated liquid handling workstation by mixing 0.7 pl of drug from the screening plate with 700 pl of semi-defined Drosophila medium (recipe available from the Bloomington Drosophila Stock Center) in 12 mm by 75 mm round-bottom test tubes (Sarstedt). As a result, each drug was diluted 1 : 1000 in the food, also bringing the DMSO concentration to 0.1% .

[0107] Drug food for combination screens were prepared by first adding 0.7 pl of 18 mM tofacitinib into each tube followed by 0.7 pl of each drug in the library and finally 700 pl Drosophila medium and mixing by repeated pipetting using the PerkinElmer automated liquid handling workstation, resulting in a 1 : 1000 dilution of each drug in the food and a DMSO concentration of 0.2%. Drug food for hit retests were prepared by hand, using a new batch of powder drug.

[0108] Experimental animals for drug screening were generated from the following cross: w/Y; UAS-012.2; UAS-N Xw; ptc-gal4, tub-gal80 ^ts and directly aliquoted into the drugfood tubes as embryos after the food was solidified. Crosses were set up en masse in cages which produced embryos for four to five consecutive days laid on apple-juice plates supplemented with fresh yeast paste. Egg lays were performed at 22°C for 24 hours to minimize transgene expression during embryogenesis and prevent embryonic lethality or irreversible developmental defects that could not be rescued by drug feeding during larval development. [0109] Embryos were collected from fresh apple juice plates every day and embryo suspensions were generated in an “embryo buffer” designed to minimize embryo clumping and setting while aliquoting (15% glycerol, 1% Bovine Serum Albumin, 0,1% Tween-20 in water). 7.5-10 pl embryo suspension was added to each drug-food tube using a single-channel, variable volume multi-dispense electronic pipette.

[0110] After the embryos were aliquoted, drug tubes were transferred to 29°C to induce transgene expression. Tubes were scored for survival to adult stage 12 days later by counting both total number of experimental pupae (EP) and those that were empty, which reflected the number of experimental adults (EA). Drugs that showed significantly higher survival to the adult stage compared to controls based on multiple t tests (PRISM software) were considered hits.

Example 1 - Clinical History

[oni] The patient presented with an extensive left-sided maxillary sinus adenoid cystic carcinoma in February, 2013. At the time of presentation the patient was a 54 year old Caucasian male with no other significant medical problems. The tumor was found to invade the skin of the face as well as the base of the skull on imaging. The patient underwent an extensive surgical resection (removal) of gross disease and reconstruction of the face and orbit. The tumor was staged as T3N0M0 - indicating primarily localized disease - with perineural invasion and pathologically positive post-surgical margins.

[0112] The patient was treated with adjuvant proton beam radiation and weekly carboplatin and paclitaxel for 7 weeks for local regional control, completing therapy in June of 2013. He underwent periodic surveillance imaging; enlarging pulmonary metastases were identified in March, 2015. He was followed expectantly for symptoms and growth with periodic positron emission tomography and X-ray computed tomography (PET/CT) scans.

[0113] In January of 2016, the patient was consented for participation in the “Personalized Cancer Therapy for Patients with Metastatic Medullary Thyroid or Metastatic Colon Cancer (NCT02363647)” protocol under a rare cancer cohort. A sample was obtained by excisional biopsy of a lung metastasis. This sample confirmed the diagnosis of ACC and was used for genetic analysis and confirmatory studies of genetic findings. Symptomatic bone metastases were identified in February of 2017 by positron emission tomography (PET) imaging and magnetic resonance imaging. The patient received radiotherapy in March, May, and October, 2017 to thoracic bone metastases and a scapular metastasis, respectively. Imaging by PET/CT demonstrated increasing size and number of bone and pulmonary metastases with increasing 2-deoxy-2-[ ¹⁸ F]fluoro-D-glucose (FDG) uptake during this period of time and prior to starting therapy.

[0114] The patient was consented for treatment on the protocol for a personalized treatment plan after reviewing results of drug screening on his tumor-matched fly avatar line. Assessment of his tumor by PET and CT imaging immediately prior to treatment demonstrated continued progression with growth of metastases, new metastases and rising standard uptake value (SUV) indicating increased metabolic activity by the tumor.

Example 2 - Genomic Analysis and Variant Selection

[0115] The overall study approach is summarized in FIG. 1 A. The first step towards building a personalized Drosophila model for the patient was a comprehensive analysis of the tumor genomic landscape. DNA and RNA were extracted from a freshly frozen specimen from a lung metastasis obtained from a 2016 specimen prior to treatment in the present study, DNA from a blood sample was also extracted as matched control. For genomic analysis of small molecular variants and copy number variants (CNVs), whole-exome sequencing (WES) was performed using tumor and matched normal blood DNA as well as RNA sequencing (RNA-seq). [0116] The ACC genomic landscape is typically diverse and includes many low frequency drivers (Ho et al., “The Mutational Landscape of Adenoid Cystic Carcinoma,” Nat. Genet. 45:791-798 (2013) and Rettig et al., “Whole-Genome Sequencing of Salivary Gland Adenoid Cystic Carcinoma,” Cancer Prev. Res. (Phila) 9:265-274 (2016), which are hereby incorporated by reference in their entirety). As a result, identifying driver alterations and building representative models are challenging. Commonly altered pathways in ACC tumors include over activation of the MYB/MYC, NOTCH, and FGF/IGF/PI3K pathways, as well as alterations in DNA damage repair and chromatin remodeling pathway components (Ho et al., “The Mutational Landscape of Adenoid Cystic Carcinoma,” Nat. Genet. 45:791-798 (2013) and Rettig et al., “Whole-Genome Sequencing of Salivary Gland Adenoid Cystic Carcinoma,” Cancer Prev. Res. (Phila) 9:265-274 (2016), which are hereby incorporated by reference in their entirety). Whole exome sequencing of the patient’s tumor DNA from the 2016 specimen - obtained prior to treatment - identified 11 nonsynonymous somatic mutations (SNVs/indels) with allelic fraction (AF) > 0.05 (FIG. 6), none of which were in genes previously associated with ACC. Most were novel, functionally uncharacterized variants in genes that were not previously associated with cancer.

[0117] In the absence of experimental data, functional prediction algorithms were used to determine the likelihood that each variant was deleterious (i.e., had a negative impact on protein function) (Kircher et al., “A General Framework for Estimating the Relative Pathogenicity of Human Genetic Variants,” Nat. Genet. 46:310-315 (2014) and Liu et al., “dbNSFP v3.0: A One- Stop Database of Functional Predictions and Annotations for Human Nonsynonym ous and Splice-Site SNVs,” Hum. Mutat. 37:235-241 (2016), which are hereby incorporated by reference in their entirety). Most whole exome sequencing variants found in the patient’s tumor were predicted to be benign by two different functional prediction tools and were eliminated. Finally, variants that were not detected by RNA-seq, suggesting that they were either false positives or they were not expressed in the tumor, were discarded. At the end of this analysis, it was concluded that none of the somatic variants were appropriate for model building.

[0118] In addition to somatic mutations, 935 rare germline variants were identified in the patient’s non-tumor (z.e., blood) DNA. Given this large number, analysis was focused on variants in genes previously associated with cancer as well as those encoding components of cancer relevant pathways and cellular processes. Of these, it was found that heterozygous missense mutations in four genes - FA TE FA TE FAT3, and ERCC2 - were predicted to be deleterious and also detected by tumor RNA sequence data, indicating that the mutant alleles were expressed in the tumor. These were selected for the fly model (FIG. IB).

[0119] NOTCH1 is a component of the Notch signaling pathway frequently activated in adenoid cystic carcinoma (Ho et al., “The Mutational Landscape of Adenoid Cystic Carcinoma,” Nat. Genet. 45:791-798 (2013), which is hereby incorporated by reference in its entirety). Although copy number analysis was inconclusive, elevated NOTCH1 protein in the tumor specimen was established by immunohistochemistry (FIG. 1C). Two other potential variants, MAP2K2 (predicted gain) and MAX (predicted loss), were rejected by similar immunohistochemical criteria (FIG. 1C). In addition, RNA-seq data revealed a t(6;9) (q22- 23;p23-24) MYB-NFIB fusion event in the patient’s tumor, a commonly observed cancer driver in patients with ACC (Ho et al., “The Mutational Landscape of Adenoid Cystic Carcinoma,” Nat. Genet. 45:791-798 (2013), which is hereby incorporated by reference in its entirety). In the patient’s tumor, the fusion event resulted in an out-of-frame transcript encoding a truncated, constitutively active MYB protein that lacked the C-terminal cytoplasmic domain required to regulate its activity (West et al., “MYB Expression and Translocation in Adenoid Cystic Carcinomas and other Salivary Gland Tumors with Clinicopathologic Correlation,” Am. J. Surg. Pathol. 35:92-99 (2011), which is hereby incorporated by reference in its entirety). NOTCH1 and truncated MYB - which together define a common subtype of adenoid cystic carcinoma (Ho et al., “The Mutational Landscape of Adenoid Cystic Carcinoma,” Nat. Genet. 45:791-798 (2013), which is hereby incorporated by reference in its entirety) - were also selected for the fly model, bringing the final number of modeled cancer drivers to six patient variants modeled by five targeted fly genes (FIG. IB).

Example 3 - Model Building and Validation

[0120] To create a patient-specific model that represents the six alterations identified in the genomic analysis (FIG. IB), a multigenic vector platform tailored for this purpose was utilized (Bangi et al., “A Personalized Platform Identifies Trametinib plus Zoledronate for a Patient with KRAS-Mutant Metastatic Colorectal Cancer,” Set. Adv. 5(5):eaav6528 (2019), which are hereby incorporated by reference in their entirety). This vector carries three different multiple cloning sites, each flanked by promoter and transcription terminator sequences: two are designed for protein expression to model oncogenes, and one is reserved for short hairpin- mediated knockdown of Drosophila tumor suppressor orthologs (FIG. ID). Transgenes were cloned downstream of a GAL4-inducible UAS promoter, a well-established ectopic expression system in Drosophila that allows both spatial and temporal control of transgene expression (Brand and Perrimon, “Targeted Gene Expression as a Means of Altering Cell Fates and Generating Dominant Phenotypes,” Development 118:401-415 (1997), which is hereby incorporated by reference in its entirety). Transgene expression was targeted by crossing the patient-specific transgenic line to transgenic fly lines that express GAL4 in specific tissues, such as ptc-GAL4. The result was a “ /c > CPCT012.2” transgenic fly line that expressed the transgenes - targeting five fly orthologs of six patient variants - in discrete regions across the developing fly.

[0121] Previous work found that expressing full-length c-MYB in Drosophila can disrupt developing tissues including aspects of the cell cycle (Lipsick et al., “Functional Evolution of the Myb Oncogene Family,” Blood Cells Mol. Dis. 27:456-458 (2001); Fitzpatrick et al., “Drosophila myb Exerts Opposing Effects on S Phase, Promoting Proliferation and Suppressing Endoreduplication,” Development 129:4497-4507 (2002); and Davidson et al., “Functional Evolution of the Vertebrate Myb Gene Family: B-Myb, but Neither A-Myb nor c-Myb, Complements Drosophila Myb in Hemocytes,” Genetics 169:215-229 (2005), which are hereby incorporated by reference in their entirety). To model the MYB-NFIB fusion, a truncated MYB construct that represents the product of the translocation event QMYBAC) was generated. NOTCH1 copy gain was modeled by overexpressing a wild-type Drosophila Notch cDNA. Heterozygous missense variants in FAT4, FA T1 3, and ERCC2 were modeled by targeting their Drosophila orthologs using short hairpins designed to achieve moderate knockdown (to model the heterozygous nature of each variant). Individual hairpins targeting each gene were selected using previously reported protocols (Vert et al., “An Accurate and Interpretable Model for siRNA Efficacy Prediction,” BMC Bioinformatics 7:520 (2006) and Ni et al., “A Genome-Scale shRNA Resource for Transgenic RNAi in Drosophila ' Nat. Methods 8:405-407 (2011), which are hereby incorporated by reference in their entirety) and stitched together as a synthetic multihairpin cluster using the microRNA-inspired design disclosed herein (Bangi et al., “A Personalized Platform Identifies Trametinib plus Zoledronate for a Patient with KRAS-Mutant Metastatic Colorectal Cancer,” Sci. Adv. 5(5):eaav6528 (2019), which is hereby incorporated by reference in its entirety). As hairpin selection relies on algorithms to predict efficacy (Vert et al., “An Accurate and Interpretable Model for siRNA Efficacy Prediction,” BMC Bioinformatics 7:520 (2006) and Ni et al., “A Genome-Scale shRNA Resource for Transgenic RNAi in Drosophila,” Nat. Methods 8:405-407 (2011), which are hereby incorporated by reference in their entirety), two hairpin clusters targeting the same three Drosophila genes with different hairpins were generated — 12.1 and 12.2 — to increase the likelihood of success.

[0122] To build the patient-specific multigenic vector, the MYBAC coding sequence and the hairpin cluster were cloned into their respective multiple cloning sites, each downstream of their own inducible UAS promoter (FIG. ID). Two different versions of the patient model were generated: CPCT012.1 and CPCT012.2. Both versions carried the same MYBAC transgene but a different hairpin cluster designed to reduce expression of Drosophila orthologs of FAT4, FAT1/3, and ERCC2. Two transgenic lines were established using a site-specific chromosomal integration method mediated by standard $C31-based integration (Bischof et al., “An Optimized Transgenesis System for Drosophila using Germ-Line-Specific phiC31 Integrases,” Proc. Natl. Acad. Sci. USA 104:3312-3317 (2007), which is hereby incorporated by reference in its entirety). Once transgenic lines were established, an existing Notch transgenic construct that expresses the full-length Drosophila Notch protein under UAS control (Matsuno et al., “Involvement of a Proline-Rich Motif and RING-H2 Finger of Deltex in the Regulation of Notch Signaling,” Development 129: 1049-1059 (2002), which is hereby incorporated by reference in its entirety) was introduced into each line by standard genetic crosses.

[0123] Once the two final patient models were established, each multigenic construct ubiquitously expressed (tub > CPCT012) in developing larvae to determine whether the hairpin clusters generated were effective. Quantitative polymerase chain reaction (qPCR) analysis indicated that the short hairpin cluster in CPCT012.2 was effective in moderately reducing expression of all three genes, consistent with the goal for modeling heterozygous variants; CPCT012.1 did not show significant knockdown of the Drosophila FAT4 ortholog// (FIG. IE). Therefore, the CPT012.2 patient-specific transgenic line was selected for further characterization.

[0124] To further validate the CPCT012.2 line, ptc-GAL4 was used to direct transgene expression within several tissues including a discrete stripe of expression at the developing wing epithelium’s anterior/posteri or boundary (FIG. IF and FIG. 1G). In previous work, expressing oncogenes with the ptc-GAL4 driver led to expansion of the ptc domain in the developing wing, reflecting overproliferation (Vidal et al., “Csk-Deficient Boundary Cells are Eliminated from Normal Drosophila Epithelia by Exclusion, Migration, and Apoptosis,” Dev. Cell 10:33-44 (2006); Vidal et al., “Differing Src Signaling Levels have Distinct Outcomes in Drosophila,” Cancer Res. 67:10278-10285 (2007); Levinson and Cagan, “Drosophila Cancer Models Identify Functional Differences Between ret Fusions,” Cell Rep. 16:3052-3061 (2016); and Sonoshita et al., “A Whole-Animal Platform to Advance a Clinical Kinase Inhibitor into New Disease Space,” Nat. Chem. Biol. 14:291-298 (2018), which are hereby incorporated by reference in their entirety). Consistent with promoting at least one aspect of transformation, expressing the transgenes at the larval wing boundary (ptc > CPCT012.2') led to expansion of the ptc domain (FIG. IF and FIG. 1G). It was concluded that this model could prove useful as an accessible whole animal platform to screen for candidate therapeutics.

Example 4 - Drug Screening

[0125] Rescue from lethality is a useful primary read-out for whole animal drug screens, providing rapid, quantitative data (Rudrapana et al., “A Jnk-Rho- Actin Remodeling Positive Feedback Network Directs Src-Driven Invasion,” Oncogene 33:2801-2806 (2014); Levine and Cagan, “Drosophila Lung Cancer Models Identify Trametinib plus Statin as Candidate Therapeutic,” Cell Rep. 14: 1477-1487 (2016); and Bangi et al., “A Personalized Platform Identifies Trametinib plus Zoledronate for a Patient with KRAS-Mutant Metastatic Colorectal Cancer,” Sci. Adv. 5(5):eaav6528 (2019), which are hereby incorporated by reference in their entirety). ptc-gal4 has been successfully used in previous genetic and drug screens of Drosophila cancer models (Dar et al., “Chemical Genetic Discovery of Targets and Anti-Targets for Cancer Polypharmacology,” Nature 486:80-84 (2012) and Sonoshita et al., “A Whole- Animal Platform to Advance a Clinical Kinase Inhibitor into New Disease Space,” Nat. Chem. Biol. 14:291-298 (2018), which are hereby incorporated by reference in their entirety). Therefore, ptc > CPCT012.2 flies were calibrated for lethality, using temperature to alter GAL4 activity to a level of near-complete animal lethality.

[0126] It was previously found that most drugs are not effective as single agents against genetically complex cancer models (Bangi et al., “A Personalized Platform Identifies Trametinib plus Zoledronate for a Patient with KRAS-Mutant Metastatic Colorectal Cancer,” Set. Adv. 5(5):eaav6528 (2019), which is hereby incorporated by reference in its entirety). Therefore, an iterative screening process was used to progressively identify drug combinations (FIG. 2A). A custom-built ‘Focused FDA-Approved Library’ of 122 drugs/drug combinations enriched for cancer relevant activities was first screened. Two hits with weak efficacy were identified and confirmed: the chemotherapy drug docetaxel is an anti -microtubule agent; the JAK inhibitor tofacitinib (Changelian et al., “Prevention of Organ Allograft Rejection by a Specific Janus Kinase 3 Inhibitor,” Science 302:875-878 (2003) and Vyas et al., “Tofacitinib: the First Janus Kinase (JAK) Inhibitor for the Treatment of Rheumatoid Arthritis,” Ann. Pharmacother 47: 1524-1531 (2013), which are hereby incorporated by reference in their entirety) is FDA approved for rheumatoid arthritis, ulcerative colitis, and psoriatic arthritis (FIG. 2A and FIG. 2B). The patient was previously treated with the anti -microtubule agent paclitaxel as part of a combination therapy with carboplatin, and therefore docetaxel was not pursued further.

[0127] Tofacitinib is a promising candidate for combination screens. Its target pathway, JAK/STAT signaling, can be activated by NOTCH in cancer cells (Jin et al., “Non-Canonical Notch Signaling Activates IL-6/JAK/STAT Signaling in Breast Tumor Cells and is Controlled by p53 and IKKa/IKKp,” Oncogene 32:4892-4902 (2013), which is hereby incorporated by reference in its entirety), which was elevated in the patient’s tumor. Re-screening the Focused Library in the presence of low dose tofacitinib identified the DNA synthesis inhibitor and FDA- approved chemotherapy agent gemcitabine as an effective partner for tofacitinib (FIGs. 2A-2C). A commercially available “FDA- Approved Drug Library” of 1280 drugs approved for all indications was also screened, again in combination with low dose tofacitinib. These combination screens identified several tofacitinib drug combinations with low efficacy. Next, all confirmed hits were tested in double and triple combinations (FIG. 2A). From these screens, a three-drug cocktail emerged (FIG. 2D): tofacitinib, vorinostat (histone deacetylase inhibitor, anti-cancer agent), and pindolol (a non-selective beta blocker used to treat high blood pressure). [0128] Each drug identified the screens described herein has been reported to have targets or anti -tumor effects that could be relevant to the patient tumor’s genomic profile. Tofacitinib is an inhibitor of JAK/STAT signaling (Vyas et al., “Tofacitinib: the First Janus Kinase (JAK) Inhibitor for the Treatment of Rheumatoid Arthritis,” Ann. Pharmacother 47: 1524-1531 (2013), which is hereby incorporated by reference in their entirety), a cancerrelevant pathway that can be activated downstream of Notch signaling in tumor cells (Jin et al., “Non-Canonical Notch Signaling Activates IL-6/JAK/STAT Signaling in Breast Tumor Cells and is Controlled by p53 and IKKa/IKKp,” Oncogene 32:4892-4902 (2013), which is hereby incorporated by reference in its entirety). Gemcitabine is a nucleoside analog that interferes with DNA synthesis (Plunkett et al., “Gemcitabine: Metabolism, Mechanisms of Action, and SelfPotentiation,” Semin. Oncol. 22:3-10 (1995), which is hereby incorporated by reference in its entirety) and is approved for treatment of multiple cancer types. The histone deacetylase (HD AC) inhibitor vorinostat may be particularly relevant for treatment of ACC, as deregulation of chromatin remodeling is observed in about 35% of sequenced tumors (Duvic and Vu, “Vorinostat: A New Oral Histone Deacetylase Inhibitor Approved for Cutaneous T-Cell Lymphoma,” Expert Opin. Investig. Drugs 16: 1111-1120 (2007); Marks and Breslow, “Dimethyl Sulfoxide to Vorinostat: Development of this Histone Deacetylase Inhibitor as an Anticancer Drug,” Nat. Biotechnol. 25:84-90 (2007); and Ho et al., “The Mutational Landscape of Adenoid Cystic Carcinoma,” Nat. Genet. 45:791-798 (2013), which are hereby incorporated by reference in their entirety). A mechanism of action clearly relevant to cancer has not been reported for pindolol. However, there is some evidence suggesting that inhibition of P- adrenergic signaling by beta blockers can have anti-tumor effects, including inhibition of tumor cell proliferation, migration, invasion, angiogenesis, and metastases (Creed et al., “P2- Adrenoceptor Signaling Regulates Invadopodia Formation to Enhance Tumor Cell Invasion,” Breast Cancer Res. 17: 145 (2015); Partecke et al., “Chronic Stress Increases Experimental Pancreatic Cancer Growth, Reduces Survival and can be Antagonised by Beta-Adrenergic Receptor Blockade,” Pancreatology 16:423-433 (2016); and Wrobel et al., “Propranolol Induces a Favourable Shift of Anti-Tumor Immunity in A Murine Spontaneous Model of Melanoma,” Oncotarget 7:77825-77837 (2016), which are hereby incorporated by reference in their entirety). A positive correlation between beta blocker use and cancer-specific survival has been documented (Na et al., “The Effects of Beta-Blocker Use on Cancer Prognosis: A Meta-Analysis Based on 319,006 Patients,” Onco Targets Ther. 11 :4913-4944 (2018), which is hereby incorporated by reference in its entirety), although no causal relationship between the two has been reported.

[0129] These findings were reviewed by a multidisciplinary tumor board that included pharmacists and oncologists with expertise in clinical trial design and dosing. The tumor board raised some concerns regarding myelosuppression, the major dose limiting toxicity associated with gemcitabine. Given the concerns with gemcitabine and both the clinical relevance of the signaling nodes targeted by the triple drug combination and their particular importance for this patient’s tumor genome landscape, the tumor board unanimously selected the tofacitinib/ vorinostat/pindolol triple combination as the first line recommendation for the patient.

Example 5 - Patient Treatment

[0130] The patient initiated treatment orally with 400 mg of vorinostat daily (2800 mg/week), 10 mg of tofacitinib daily, and 10 mg of pindolol daily beginning on 4/19/18. At four weeks, grade 1 thrombocytopenia, creatinine elevation, and folliculitis were observed likely due to vorinostat; minor fatigue was reported likely due to pindolol. All the drugs were held and then restarted 10 days later with 400 mg of vorinostat reduced to five days per week (2000 mg/week). Recurrent rash, thrombocytopenia, and creatinine elevation developed after restarting vorinostat, which resolved after halting and then restarting vorinostat with a further dose reduction to 300 mg four days per week (1200 mg/week) after which creatine and platelet counts remained within normal limits.

[0131] In response to treatment, the patient exhibited documented stable disease (FIG. 3) with no new bone lesions and mild regression of lung lesions. Quantitative imaging of a FDG tracer with PET/CT scans was used to measure relative glucose uptake by the tumor.

Cumulative FDG data over time indicated a significant reduction (49%) in SUV in pulmonary and bone metastases (FIGs. 3 A-3B), indicating a significant metabolic response.

[0132] The patient continued on therapy until 4/19/2019 (12 months), when he exhibited documented progression by progression by response evaluation criteria in solid tumors (RECIST), version 1.1, criteria. He had subsequent palliative surgery, chemotherapy, and radiation for further rapid progression. The patient passed away on 3/17/2020.

Example 6 - Post-Treatment Analysis

[0133] To better understand the nature of the emergent resistance at 12 months, two additional biopsy samples were obtained from the patient shortly after he stopped receiving the drug cocktail treatment in April 2019. Similar to the original sample, WES was performed on both biopsies: two specimens were obtained from the same formalin-fixed paraffin-embedded (FFPE) tumor block in June 2019 biopsy; a bone biopsy specimen was obtained in October 2019. The somatic non-synonymous small molecular variants with AF >0.05 and CNVs from the posttreatment 2019 specimens were compared to the original pre-treatment 2016 specimen.

Significant genomic differences were observed (FIGs. 4A-4B). [0134] A small number of somatic variants were common to all 2019 specimens but not the 2016 specimen (FIG. 4 A and FIG. 6). These included variants in genes encoding ANKAR, ANXA8L2, CNTRL, EP300, GEMIN5, which are therefore candidates to play a role in the resistance that emerged during treatment. Somatic copy number variants (sCNVs), identified by saasCNV (Zhang and Hao, “SAAS-CNV: A Joint Segmentation Approach on Aggregated and Allele Specific Signals for the Identification of Somatic Copy Number Alterations with Next- Generation Sequencing Data,” PLoS Comput. Biol. 11 :el004618 (2015), which is hereby incorporated by reference in its entirety), found that all three 2019 specimens had a significantly larger number of sCNVs than the original 2016 sample, which had relatively few sCNVs (FIG. 4B). Loss of a large segment on chromosome 14 in the original specimen was the sole sCNV retained in 2019 specimens; in contrast, e.g., an aneuploid gain of chromosome 19 from the original specimen was lost in all 2019 specimens, suggesting the aneuploid gain was contained within a subclonal alteration that was selected against as the patient's disease advanced. GATK4 sCNV was used to confirm the sCNV calling in the 2019 specimens (FIG. 5, FIG. 6, and FIG. 7); see also methods); the sCNV profiles between the two callers largely agreed with each other. [0135] Finally, to further identify driver genes responsible for emergent resistance, annotated tumor suppressor genes by Catalogue of Somatic Mutations in Cancer (COSMIC) Cancer Gene Census (Sondka et al., “The COSMIC Cancer Gene Census: Describing Genetic Dysfunction Across All Human Cancers,” Nat. Rev. Cancer 18:696-705 (2018), which is hereby incorporated by reference in its entirety) were checked against the biallelic inactivated genes (defined by the intersection between loss sCNVs and germline protein-altering variants with gnomAD allele frequency < 0.05%). No additional examples of functionally significant variants were found unique to all three 2019 specimens vs. the 2016 specimen (FIG. 6). Overall, candidates for emergent drug resistance include multiple somatic mutated loci plus significant changes in the CNV landscape.

Discussion of Examples 1-6

[0136] Adenoid cystic carcinoma (ACC) has proven to be a challenging disease, with limited overall response to traditional and targeted therapies. An experimental approach based on efforts to model its genomic complexity in the context of a personalized approach was therefore undertaken. The Examples presented herein report the results of a study treating a patient with progressive disease that failed to respond to standard-of-care treatments: a novel three-drug cocktail provided stable disease for 12 months, followed by treatment resistance and extensive genomic alterations observed in biopsy samples. [0137] Building on an approach reported for a patient with advanced colorectal cancer adenocarcinoma (Bangi et al., “A Personalized Platform Identifies Trametinib plus Zoledronate for a Patient with KRAS-Mutant Metastatic Colorectal Cancer,” Sci. Adv. 5(5):eaav6528 (2019), which is hereby incorporated by reference in its entirety), a decision tree approach was used to prioritize six genes that were identified as key for affecting tumor progression and, potentially, drug response. Additional tools such immunohistochemistry (FIGs. 1 A-1F) were used to help validate gene choices, but especially given the presence of CNVs and the timeline required for patient treatment, this analysis was necessarily incomplete.

[0138] ACC is commonly associated with truncation of MYB and elevated NOTCH1 expression/activity, and this patient's tumor presented with both. ACC tumors are not associated with large numbers of somatic mutations (Ho et al., “The Mutational Landscape of Adenoid Cystic Carcinoma,” Nat. Genet. 45:791-798 (2013), which is hereby incorporated by reference in its entirety) and, again, this patient’s tumor reflected this as few other somatic changes, including CNVs, were observed. Germline mutations are often not accounted for in similar analyses. Nevertheless, germline analysis identified four genes with known functions likely to contribute to tumor progression and potentially drug response: FAT1 and FAT3, FA T4, and ERCC2, modeled by targeting kug, ft, and xpd, respectively.

[0139] FAT1, FAT3, and FAT4 are atypical cadherins and key regulators of cancerrelevant cellular processes including planar cell polarity and MST/HIPPO signaling, which regulates organ size. They are associated as tumor suppressors in a variety of tumors including ovarian, medullary thyroid, gastric, cervical, colorectal, bladder, and squamous cell carcinomas, as well as ACC (Ho et al., “The Mutational Landscape of Adenoid Cystic Carcinoma,” Nat. Genet. 45:791-798 (2013); Gao et al., “Genetic Landscape of Esophageal Squamous Cell Carcinoma,” Nat. Genet. 46: 1097-1102 (2014); Chen et al., “ A Novel Gene Signature Combination Improves the Prediction of Overall Survival in Urinary Bladder Cancer,” J. Cancer 10:5744-5753 (2019); Huang et al., “Comprehensive Genomic Variation Profiling of Cervical Intraepithelial Neoplasia and Cervical Cancer Identifies Potential Targets for Cervical Cancer Early Warning,” J. Med. Genet. 56: 186-194 (2019); Skuja et al., “Deletions in Metastatic Colorectal Cancer with Chromothripsis,” Exp. Oncol. 41 :323-327 (2019); Malgundkar et al., “FAT4 Silencing Promotes Epithelial-to-Mesenchymal Transition and Invasion via Regulation of YAP and P-Catenin Activity in Ovarian Cancer,” BMC Cancer 20:374 (2020); Melis et al., “Mutations in Long-Lived Epithelial Stem Cells and their Clonal Progeny in Pre-Malignant Lesions and in Oral Squamous Cell Carcinoma,” Carcinogenesis 41 : 1553 (2020); Qu et al., “Genomic and Transcriptomic Characterization of Sporadic Medullary Thyroid Carcinoma,” Thyroid 30: 1025-1036 (2020); and Wang et al., “Integrated Characterisation of Cancer Genes Identifies Key Molecular Biomarkers in Stomach Adenocarcinoma,” J. Clin. Pathol. 73:579-586 (2020), which are hereby incorporated by reference in their entirety). Reducing these atypical cadherins by approximately 50% expression was therefore considered potentially impactful to the drug screening platform. ERCC2 encodes XPD, a protein associated with regulation of TFIIH-mediated transcription and DNA damage. XPD has effects on progression of a broad palette of tumor types when mutated or altered — increased or decreased — in expression, and variants in ERCC2 have been associated with altered drug response, especially platinum-based therapies (Du et al., “Associations of Polymorphisms in DNA Repair Genes and MDR1 Gene with Chemotherapy Response and Survival of Non-Small Cell Lung Cancer,” PLoS one 9:e99843 (2014); Fu et al., “Association between the Asp312Asn, Lys751Gln, and Argl56Arg Polymorphisms in XPD and the Risk of Prostate Cancer,” Technol. Cancer Res. Treat. 16:692- 704 (2017); Tan et al., “Genetic Polymorphisms and Platinum-Based Chemotherapy Treatment Outcomes in Patients with Non-Small Cell Lung Cancer: A Genetic Epidemiology Study Based Meta- Analysis,” Sci. Rep. 7:5593 (2017); Pajuelo-Lozano et al., “XPA, XPC, and XPD Modulate Sensitivity in Gastric Cisplatin Resistance Cancer Cells,” Front. Pharmacol. 9: 1197 (2018); and Liu et al., “Association of XPD Asp312Asn Polymorphism and Response to Oxaliplatin-Based First-Line Chemotherapy and Survival in Patients with Metastatic Colorectal Cancer,” Adv. Clin. Exp. Med. 28: 1459-1468 (2019), which are hereby incorporated by reference in their entirety). This may have contributed to the patient’s failure to respond to earlier carboplatin-based therapy.

[0140] A key advantage of this approach is the ability to identify novel drug combinations selected solely on the basis of efficacy in a whole animal model. The three-drug cocktail — tofacitinib, pindolol, and vorinostat — is especially interesting to consider from a mechanism standpoint. Vorinostat has been reported to provide some benefit to a subset of patients with ACC (Goncalves et al., “A Phase 2 Study of Vorinostat in Locally Advanced, Recurrent, or Metastatic Adenoid Cystic Carcinoma,” Oncotarget 8:32918-32929 (2017), which is hereby incorporated by reference in its entirety), though it did not display activity as a single agent in the screens carried out in the study. The primary hit was tofacitinib, a LAK inhibitor used primarily for rheumatoid arthritis. As part of the JAK/STAT pathway, JAK is a signaling kinase with broad effects on development and disease including cancer (Thomas et al., “The Role of JAK/STAT Signalling in the Pathogenesis, Prognosis and Treatment of Solid Tumours,” Br. J. Cancer 113:365-371 (2015), which is hereby incorporated by reference in its entirety). JAK is positively regulated by NOTCH signaling activity in multiple contexts (Monsalve et al., “Notch- 1 Up-Regulation and Signaling Following Macrophage Activation Modulates Gene Expression Patterns Known to Affect Antigen-Presenting Capacity and Cytotoxic Activity,” J. Immunol. 176:5362-5373 (2006); Jin et al., “Non-Canonical Notch Signaling Activates IL-

6/JAK/STAT Signaling in Breast Tumor Cells and is Controlled by p53 and IKKa/IKKp,” Oncogene 32:4892-4902 (2013); and Ho et al., “The Notch-Mediated Hyperplasia Circuitry in Drosophila Reveals a Src-JNK Signaling Axis,” eLife 4:e05996 (2015), which are hereby incorporated by reference in their entirety), suggesting a mechanism for tofacitinib activity. [0141] Finally, genomic analysis of later patient biopsies highlights the challenge that selection-based resistance poses, even for a three-drug therapeutic cocktail. The analysis presented herein confirmed retention of the MYB-NFIB fusion; however, the CNV landscape was significantly altered including a number of new CNVs in the post-treatment samples. This provides potential insight into the mechanisms by which ACC has proven recalcitrant to most drug treatments: selection for progressive clones can lead to significant genomic changes that can in turn subvert therapeutic activity. One potential response would have been to follow with treatment of a second-line drug cocktail, tofacitinib-gemcitabine. However, follow-up radiationbased therapy and subsequent rapid patient decline made use of gemcitabine contraindicated. [0142] In conclusion, the Examples presented herein provide a personalized approach to treating ACC. Genomic analysis followed by construction and screening of a “personalized fly avatar” led to a unique three-drug cocktail that promoted stable disease and reduced tumor metabolic activity for 12 months. This drug cocktail is unique, suggesting at least some specificity for this patient. Assessing the broad utility of this whole animal platform approach — and this specific drug cocktail — for patients with ACC will require a larger study. This personalized approach is adaptable to a broad palette of tumor types and may prove especially useful for rare cancers that do not have a standard-of-care second-line therapy or a clear treatment guidance protocol.

[0143] Although preferred embodiments have been depicted and described in detail herein, it will be apparent to those skilled in the relevant art that various modifications, additions, substitutions, and the like can be made without departing from the spirit of the invention and these are therefore considered to be within the scope of the invention as defined in the claims which follow.