Genome-matched treatments and patient outcomes in the Maine Cancer Genomics Initiative (MCGI)

Genomic tumor testing (GTT) is an emerging technology aimed at identifying variants in tumors that can be targeted with genomically matched drugs. Due to limited resources, rural patients receiving care in community oncology settings may be less likely to benefit from GTT. We analyzed GTT results and observational clinical outcomes data from patients enrolled in the Maine Cancer Genomics Initiative (MCGI), which provided access to GTTs; clinician educational resources; and genomic tumor boards in community practices in a predominantly rural state. 1603 adult cancer patients completed enrollment; 1258 had at least one potentially actionable variant identified. 206 (16.4%) patients received a total of 240 genome matched treatments, of those treatments, 64% were FDA-approved in the tumor type, 27% FDA-approved in a different tumor type and 9% were given on a clinical trial. Using Inverse Probability of Treatment Weighting to adjust for baseline characteristics, a Cox proportional hazards model demonstrated that patients who received genome matched treatment were 31% less likely to die within 1 year compared to those who did not receive genome matched treatment (HR: 0.69; 95% CI: 0.52-0.90; p-value: 0.006). Overall, GTT through this initiative resulted in levels of genome matched treatment that were similar to other initiatives, however, clinical trials represented a smaller share of treatments than previously reported, and "off-label" treatments represented a greater share. Although this was an observational study, we found evidence for a potential 1-year survival benefit for patients who received genome matched treatments. These findings suggest that when disseminated and implemented with a supportive infrastructure, GTT may benefit cancer patients in rural community oncology settings, with further work remaining on providing genome-matched clinical trials.

High-throughput deconvolution of 3D organoid dynamics at cellular resolution for cancer pharmacology with Cellos

Three-dimensional (3D) organoid cultures are flexible systems to interrogate cellular growth, morphology, multicellular spatial architecture, and cellular interactions in response to treatment. However, computational methods for analysis of 3D organoids with sufficiently high-throughput and cellular resolution are needed. Here we report Cellos, an accurate, high-throughput pipeline for 3D organoid segmentation using classical algorithms and nuclear segmentation using a trained Stardist-3D convolutional neural network. To evaluate Cellos, we analyze ~100,000 organoids with ~2.35 million cells from multiple treatment experiments. Cellos segments dye-stained or fluorescently-labeled nuclei and accurately distinguishes distinct labeled cell populations within organoids. Cellos can recapitulate traditional luminescence-based drug response of cells with complex drug sensitivities, while also quantifying changes in organoid and nuclear morphologies caused by treatment as well as cell-cell spatial relationships that reflect ecological affinity. Cellos provides powerful tools to perform high-throughput analysis for pharmacological testing and biological investigation of organoids based on 3D imaging.

The Maine Cancer Genomics Initiative: Implementing a Community Cancer Genomics Program Across an Entire Rural State

PURPOSE

The Maine Cancer Genomics Initiative (MCGI) aimed to overcome patient- and provider-level barriers to using genomic tumor testing (GTT) in rural practices by providing genomic tumor boards (GTBs), clinician education, and access to comprehensive large-panel next-generation sequencing to all patients with cancer in Maine. This paper describes the successful implementation of the initiative and three key services made operative between 2016 and 2020.

METHODS

A community-inclusive, hub-and-spoke approach was taken to implement the three program components: (1) a centralized GTB program; (2) a modular online education program, designed using an iterative approach with broad clinical stakeholders; and (3) GTT free of charge to clinicians and patients. Implementation timelines, participation metrics, and survey data were used to describe the rollout.

RESULTS

The MCGI was launched over an 18-month period at all 19 oncology practices in the State. Seventy-nine physicians (66 medical oncologists, 5 gynecologic oncologists, 1 neuro-oncologist, and 7 pediatric oncologists) enrolled on the study, representing 100% of all practicing oncologists in Maine. Between July 2017 and September 2020, 1610 patients were enrolled. A total of 515 cases were discussed by 47 (73%) clinicians in 196 GTBs. Clinicians who participated in the GTBs enrolled significantly more patients on the study, stayed in Maine, and reported less time spent in clinical patient care.

CONCLUSION

The MCGI was able to engage geographically and culturally disparate cancer care practices in a precision oncology program using a hub-and-spoke model. By facilitating access to GTT, structured education, and GTBs, we narrowed the gap in the implementation of precision oncology in one of the most rural states in the country.

Mouse models for immuno-oncology

Realizing the clinical promise of cancer immunotherapy is hindered by gaps in our knowledge of in vivo mechanisms underlying treatment response as well as treatment limiting toxicity. Preclinical in vivo model systems and technologies are required to address these knowledge gaps and to surmount the challenges faced in the clinical application of immunotherapy. Mice are commonly used for basic and translational research to support development and testing of immune interventions, including for cancer. Here, we discuss the advantages and the limitations of current models as well as future developments.

Cellos: High-throughput deconvolution of 3D organoid dynamics at cellular resolution for cancer pharmacology

Three-dimensional (3D) culture models, such as organoids, are flexible systems to interrogate cellular growth and morphology, multicellular spatial architecture, and cell interactions in response to drug treatment. However, new computational methods to segment and analyze 3D models at cellular resolution with sufficiently high throughput are needed to realize these possibilities. Here we report Cellos (Cell and Organoid Segmentation), an accurate, high throughput image analysis pipeline for 3D organoid and nuclear segmentation analysis. Cellos segments organoids in 3D using classical algorithms and segments nuclei using a Stardist-3D convolutional neural network which we trained on a manually annotated dataset of 3,862 cells from 36 organoids confocally imaged at 5 μm z-resolution. To evaluate the capabilities of Cellos we then analyzed 74,450 organoids with 1.65 million cells, from multiple experiments on triple negative breast cancer organoids containing clonal mixtures with complex cisplatin sensitivities. Cellos was able to accurately distinguish ratios of distinct fluorescently labelled cell populations in organoids, with ≤3% deviation from the seeding ratios in each well and was effective for both fluorescently labelled nuclei and independent DAPI stained datasets. Cellos was able to recapitulate traditional luminescence-based drug response quantifications by analyzing 3D images, including parallel analysis of multiple cancer clones in the same well. Moreover, Cellos was able to identify organoid and nuclear morphology feature changes associated with treatment. Finally, Cellos enables 3D analysis of cell spatial relationships, which we used to detect ecological affinity between cancer cells beyond what arises from local cell division or organoid composition. Cellos provides powerful tools to perform high throughput analysis for pharmacological testing and biological investigation of organoids based on 3D imaging.

Lipid-laden lung mesenchymal cells foster breast cancer metastasis via metabolic reprogramming of tumor cells and natural killer cells

While the distant organ environment is known to support metastasis of primary tumors, its metabolic roles in this process remain underdetermined. Here, in breast cancer models, we found lung-resident mesenchymal cells (MCs) accumulating neutral lipids at the pre-metastatic stage. This was partially mediated by interleukin-1β (IL-1β)-induced hypoxia-inducible lipid droplet-associated (HILPDA) that subsequently represses adipose triglyceride lipase (ATGL) activity in lung MCs. MC-specific ablation of the ATGL or HILPDA genes in mice reinforced and reduced lung metastasis of breast cancer respectively, suggesting a metastasis-promoting effect of lipid-laden MCs. Mechanistically, lipid-laden MCs transported their lipids to tumor cells and natural killer (NK) cells via exosome-like vesicles, leading to heightened tumor cell survival and proliferation and NK cell dysfunction. Blockage of IL-1β, which was effective singly, improved the efficacy of adoptive NK cell immunotherapy in mitigating lung metastasis. Collectively, lung MCs metabolically regulate tumor cells and anti-tumor immunity to facilitate breast cancer lung metastasis.

Adaptive Sentinel Testing in Workplace for COVID-19 Pandemic

Testing and isolation of infectious employees is one of the critical strategies to make the workplace safe during the pandemic for many organizations. Adaptive testing frequency reduces cost while keeping the pandemic under control at the workplace. However, most models aimed at estimating test frequencies were structured for municipalities or large organizations such as university campuses of highly mobile individuals. By contrast, the workplace exhibits distinct characteristics: employee positivity rate may be different from the local community because of rigorous protective measures at workplace, or self-selection of co-workers with common behavioral tendencies for adherence to pandemic mitigation guidelines. Moreover, dual exposure to COVID-19 occurs at work and home that complicates transmission modeling, as does transmission tracing at the workplace. Hence, we developed bi-modal SEIR (Susceptible, Exposed, Infectious, and Removed) model and R-shiny tool that accounts for these differentiating factors to adaptively estimate the testing frequency for workplace. Our tool uses easily measurable parameters: community incidence rate, risks of acquiring infection from community and workplace, workforce size, and sensitivity of testing. Our model is best suited for moderate-sized organizations with low internal transmission rates, no-outward facing employees whose position demands frequent in-person interactions with the public, and low to medium population positivity rates. Simulations revealed that employee behavior in adherence to protective measures at work and in their community, and the onsite workforce size have large effects on testing frequency. Reducing workplace transmission rate through workplace mitigation protocols and higher sensitivity of the test deployed, although to a lesser extent. Furthermore, our simulations showed that sentinel testing leads to only marginal increase in the number of infections even for high community incidence rates, suggesting that this may be a cost-effective approach in future pandemics. We used our model to accurately guide testing regimen for three campuses of the Jackson Laboratory.

A Genomically and Clinically Annotated Patient-Derived Xenograft Resource for Preclinical Research in Non-Small Cell Lung Cancer

Patient-derived xenograft (PDX) models are an effective preclinical in vivo platform for testing the efficacy of novel drugs and drug combinations for cancer therapeutics. Here we describe a repository of 79 genomically and clinically annotated lung cancer PDXs available from The Jackson Laboratory that have been extensively characterized for histopathologic features, mutational profiles, gene expression, and copy-number aberrations. Most of the PDXs are models of non-small cell lung cancer (NSCLC), including 37 lung adenocarcinoma (LUAD) and 33 lung squamous cell carcinoma (LUSC) models. Other lung cancer models in the repository include four small cell carcinomas, two large cell neuroendocrine carcinomas, two adenosquamous carcinomas, and one pleomorphic carcinoma. Models with both de novo and acquired resistance to targeted therapies with tyrosine kinase inhibitors are available in the collection. The genomic profiles of the LUAD and LUSC PDX models are consistent with those observed in patient tumors from The Cancer Genome Atlas and previously characterized gene expression-based molecular subtypes. Clinically relevant mutations identified in the original patient tumors were confirmed in engrafted PDX tumors. Treatment studies performed in a subset of the models recapitulated the responses expected on the basis of the observed genomic profiles. These models therefore serve as a valuable preclinical platform for translational cancer research.

SIGNIFICANCE

Patient-derived xenografts of lung cancer retain key features observed in the originating patient tumors and show expected responses to treatment with standard-of-care agents, providing experimentally tractable and reproducible models for preclinical investigations.

Animals, quality and the pursuit of relevance

In 2021, the National Institutes of Health Advisory Committee to the Director (ACD) announced recommendations to improve the reproducibility of biomedical research using animals. In response, The Jackson Laboratory faculty and institutional leaders identified key strategies to further address this important issue. Taking inspiration from the evolution of clinical trials over recent decades in response to similar challenges, we identified opportunities for improvement, including establishment of common standards, use of genetically diverse populations, requirement for robust study design with appropriate statistical methods, and improvement in public databases to facilitate meta-analyses. In this Perspective, we share our response to ACD recommendations, with a specific focus on mouse models, with the aim of promoting continued active dialogue among researchers, using any animal system, worldwide. Such discussion will help to inform the biomedical community about these recommendations and further support their much-needed implementation.

Genomic and epigenomic BRCA alterations predict adaptive resistance and response to platinum-based therapy in patients with triple-negative breast and ovarian carcinomas

Triple-negative breast cancer (TNBC) and ovarian carcinomas (OvCas) with BRCA1 promoter methylation ( BRCA1 meth) respond more poorly to alkylating agents compared to those bearing mutations in BRCA1 and BRCA2 ( BRCA mut). This is a conundrum given the biologically equivalent homologous recombination deficiency (HRD) induced by these genetic and epigenetic BRCA perturbations. We dissected this problem through detailed genomic analyses of TNBC and OvCa cohorts and experimentation with patient-derived xenografts and genetically engineered cell lines. We found that despite identical downstream genomic mutational signatures associated with BRCA1 meth and BRCA mut states, BRCA1 meth uniformly associates with poor outcomes. Exposure of BRCA1 meth TNBCs to platinum chemotherapy, either as clinical treatment of a patient or as experimental in vivo exposure of preclinical patient derived xenografts, resulted in allelic loss of BRCA1 methylation and increased BRCA1 expression and platinum resistance. These data suggest that, unlike BRCA mut cancers, where BRCA loss is a genetically “fixed” deficiency state, BRCA1 meth cancers are highly adaptive to genotoxin exposure and, through reversal of promoter methylation, recover BRCA1 expression and become resistant to therapy. We further found a specific augmented immune transcriptional signal associated with enhanced response to platinum chemotherapy but only in patients with BRCA-proficient cancers. We showed how integrating both this cancer immune signature and the presence of BRCA mutations results in more accurate predictions of patient response when compared to either HRD status or BRCA status alone. This underscores the importance of defining BRCA heterogeneity in optimizing the predictive precision of assigning response probabilities in TNBC and OvCa.

Abstract P5-01-05: Transcriptional signature of metastatic triple negative breast cancer in humanized mice

Triple-negative breast cancer (TNBC) is a particularly aggressive form of breast cancer with high risk of recurrence and approximately 22% rate of five-year survival when the disease becomes metastatic. Thus, understanding of mechanisms supporting metastatic colonization of distant organs is of critical importance for the development of new therapies and possibly improved outcomes. Syngeneic mouse models suggest the role of innate immune cells, particularly neutrophils, in support of metastatic dissemination of TNBC. However, it is not possible to study human cancer in immunocompetent mice. Furthermore, organoids or other 3D tissue models do not allow investigations of distant organs colonization with metastatic TNBC tumors. Here, we used humanized mice and patient-derived xenograft (PDX) from treatment naïve primary TNBC tumors to investigate the mechanisms that promote metastasis. NSG mice with transgenic expression of human hematopoietic cytokines SCF/GM-CSF/IL-3 were engrafted with human CD34+ hematopoietic progenitor cells (HPCs) to generate humanized (h)NSG-SGM3 mice. All PDX tumors grew after orthotopic implantation at week 8-12. The presence of distant metastasis was determined by macroscopic evaluation of distant organs and further confirmed by E-cadherin and cytokeratin 19 expression using polychromatic immunofluorescence on frozen tissue section. Among ten PDX tumors tested, four did not develop metastasis, four developed only lung metastasis and two developed multi-organ metastases (lung and liver). We find that different TNBC PDX tumors have different metastatic potential. Their metastatic potential is linked with differences in cellular composition and transcriptional signatures at the level of the primary tumor. Interferon Response signature is enriched in primary TNBC PDXs with metastatic potential, while non-metastatic primary tumors display a TNF signature as well as allograft rejection signature. Furthermore, liver metastases were enriched in myeloid transcripts. Thus, our model enables mechanistic and pre-clinical studies of human TNBC metastasis.

Citation Format: Chun I Yu, Te-Chia Wu, Francesca Menghi, Joshy George, Kyung In Kim, Florentina Marches, Edison T Liu, Jacques Banchereau, Karolina Palucka. Transcriptional signature of metastatic triple negative breast cancer in humanized mice [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr P5-01-05.

Phase II prospective open label study of neoadjuvant pertuzumab, trastuzumab, and nab-paclitaxel in patients with HER-2 positive advanced breast cancer

583

Background: HER2 overexpression occurs in 20-25% of breast cancers (BC) and is associated with poor prognosis. The addition of trastuzumab (trast) to chemotherapy significantly improves disease-free (DFS) and overall survival (OS) in the adjuvant setting. Pertuzumab (pert) inhibits ligand-activated signaling and in combination with trast has synergistic inhibition of BC cells overexpressing HER2. In the neoadjuvant therapy (NT) setting, the combination of trast, pert, and docetaxel can improve the pCR rate. PCR may predict for improved DFS and OS. De-escalation with weekly paclitaxel combined with trast and pert appeared to be safe and efficacious but requires steroid premedication, whereas nab-paclitaxel (nab) does not require steroid premedication. To decrease treatment-associated toxicity in patients with HER2+ BC, we utilized a non-anthracycline regimen with pert, trast, and nab as NT. The objectives of this study were to evaluate the safety and efficacy of pert added to trast and nab in HER2+ locally advanced BC (LABC) to determine the pCR, as well as DFS and OS. Methods: A total of 45 patients with biopsy-confirmed HER2+ LABC or inflammatory BC were enrolled from 2013-2017, and were treated with 6 cycles of neoadjuvant pert (840 mg loading dose, then 420 mg IV day 1 every 21 days), weekly trast (4 mg/kg loading dose, then 2 mg/kg), and weekly nab (100 mg/m2 IV). Patient characteristics, including age, race, menopausal status, grade, stage, and prior surgery and radiation were recorded. Median treatment cycles determined, and events (AE) were identified for each arm. PCR rate, DFS and OS were calculated. Results: Median age was 56 (31-78) years. 1/45 (2%) was stage I, 30/45 (67%) were stage II, 14/45 (31%) were stage III. pCR rate was 29/45 (64.4%). The initial primary tumor size was similar in pCR and non-pCR patients (mean 4.1 cm vs. 3.2 cm, respectively). Median follow-up was 36.1 months (95% CI [27.1, 41.8]). Median treatment cycles completed was 6 (1-6). A total of 4/45 (9%) patients had >1 cycle delayed, and 32/45 (71%) patients had >1 cycle modified. For the patients achieving pCR, the DFS (95% CI) at 3 years was 85.9% (66.7%, 94.4%) and for those without pCR, it was 87.5% (58.6%, 96.7%). OS was not reached (95% CI [NR, NR]). Grade 3 AEs (> 2 patients) included 7/45 (16%) of patients with hypertension; 4/45 (9%) with anemia; and 2/45 (4%) with diarrhea, ALT, fatigue, or rash. Conclusions: This anthracycline-free regimen which included nab achieved great pCR rate of 64.4% in HER2+ BC patients with fewer treatment-related toxicities. The pCR rate is comparable with docetaxel, carboplatin, trast, and pert (TCHP) therapy in NT setting, but without the treatment-associated toxicities. This suggests we may be able to safely avoid anthracyclines and carboplatin for NT in HER2+ BC patients. The improved pCR did not translate into DFS benefit. Clinical trial information: NCT01730833.

FANCM regulates repair pathway choice at stalled replication forks

Repair pathway "choice" at stalled mammalian replication forks is an important determinant of genome stability; however, the underlying mechanisms are poorly understood. FANCM encodes a multi-domain scaffolding and motor protein that interacts with several distinct repair protein complexes at stalled forks. Here, we use defined mutations engineered within endogenous Fancm in mouse embryonic stem cells to study how Fancm regulates stalled fork repair. We find that distinct FANCM repair functions are enacted by molecularly separable scaffolding domains. These findings define FANCM as a key mediator of repair pathway choice at stalled replication forks and reveal its molecular mechanism. Notably, mutations that inactivate FANCM ATPase function disable all its repair functions and "trap" FANCM at stalled forks. We find that Brca1 hypomorphic mutants are synthetic lethal with Fancm null or Fancm ATPase-defective mutants. The ATPase function of FANCM may therefore represent a promising "druggable" target for therapy of BRCA1-linked cancer.

Employees' Views and Ethical, Legal, and Social Implications Assessment of Voluntary Workplace Genomic Testing

Employers have begun to offer voluntary workplace genomic testing (wGT) as part of employee wellness benefit programs, but few empirical studies have examined the ethical, legal, and social implications (ELSI) of wGT. To better understand employee perspectives on wGT, employees were surveyed at a large biomedical research institution. Survey respondents were presented with three hypothetical scenarios for accessing health-related genomic testing: via (1) their doctor; (2) their workplace; and 3) a commercial direct-to-consumer (DTC) genetic testing company. Overall, 594 employees (28%) responded to the survey. Respondents indicated a preference for genomic testing in the workplace setting (70%; 95% CI 66-74%), followed by doctor's office (54%; 95% CI 50-58%), and DTC testing (20%; 95% CI 17-24%). Prior to participating in wGT, respondents wanted to know about confidentiality of test results (79%), existence of relevant laws and policies (70%), and privacy protection (64%). Across scenarios, 92% of respondents preferred to view the test results with a genetic counselor. These preliminary results suggest that many employees are interested and even prefer genetic testing in the workplace and would prefer testing with support from genetic health professionals. Confirmation in more diverse employer settings will be needed to generalize such findings.

Chromatin topology reorganization and transcription repression by PML-RARα in acute promyeloid leukemia

BACKGROUND

Acute promyeloid leukemia (APL) is characterized by the oncogenic fusion protein PML-RARα, a major etiological agent in APL. However, the molecular mechanisms underlying the role of PML-RARα in leukemogenesis remain largely unknown.

RESULTS

Using an inducible system, we comprehensively analyze the 3D genome organization in myeloid cells and its reorganization after PML-RARα induction and perform additional analyses in patient-derived APL cells with native PML-RARα. We discover that PML-RARα mediates extensive chromatin interactions genome-wide. Globally, it redefines the chromatin topology of the myeloid genome toward a more condensed configuration in APL cells; locally, it intrudes RNAPII-associated interaction domains, interrupts myeloid-specific transcription factors binding at enhancers and super-enhancers, and leads to transcriptional repression of genes critical for myeloid differentiation and maturation.

CONCLUSIONS

Our results not only provide novel topological insights for the roles of PML-RARα in transforming myeloid cells into leukemia cells, but further uncover a topological framework of a molecular mechanism for oncogenic fusion proteins in cancers.

Abstract P1-03-05: Patient-derived xenografts in humanized mice classify metastatic potential of primary triple negative breast cancer

Triple-negative breast cancer (TNBC) is a particularly aggressive form of breast cancer with high risk of recurrence and approximately 22% rate of five-year survival when the disease becomes metastatic. Thus, understanding of mechanisms supporting metastatic colonization of distant organs is of critical importance for the development of new therapies and possibly improved outcomes. Syngeneic mouse models suggest the role of innate immune cells, particularly neutrophils, in support of metastatic dissemination of TNBC. However, it is not possible to study human cancer in immunocompetent mice. Furthermore, organoids or other 3D tissue models do not allow investigations of distant organs colonization with metastatic TNBC tumors. Here, we used humanized mice and patient-derived xenograft (PDX) from treatment naïve primary TNBC tumors to investigate the mechanisms that promote metastasis. NSG mice with transgenic expression of human hematopoietic cytokines SCF/GM-CSF/IL-3 were engrafted with human CD34+ hematopoietic progenitor cells (HPCs) to generate humanized (h)NSG-SGM3 mice. All eleven (11) analyzed to date PDX tumors grew after orthotopic implantation at week 8-12. The presence of distant metastasis was determined by macroscopic evaluation of distant organs and further confirmed by E-cadherin and cytokeratin 19 expression using polychromatic immunofluorescence on frozen tissue section. Among 11 PDX tumors tested, five did not develop metastasis, four developed only lung metastasis and two developed multi-organ metastasis (lung and liver). Transcriptional profiling with RNAseq revealed significant differences in the immune landscape of primary and metastatic tumors. In particular, liver metastases were enriched in myeloid and plasma cell transcripts. Further analysis is ongoing to uncover specific pathways involved. Thus, our model enables mechanistic and pre-clinical studies of human TNBC metastasis.

Citation Format: Chun I Yu, Te-Chia Wu, Kyung In Kim, Francesca Menghi, Vanessa Oliveira, Florentina Marches, Edison T Liu, Jacques Banchereau, Karolina Palucka. Patient-derived xenografts in humanized mice classify metastatic potential of primary triple negative breast cancer [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P1-03-05.

Prospective study for comprehensive genomic profiling (GP) in cancer of unknown primary (CUP): Feasibility, molecular landscape, and clinical utility in current era

834

Background: CUP presents a unique niche and challenge for application of GP. Absence of a primary limits consensus regarding site-directed and availability of tissue specific targeted therapy. We evaluated the real time feasibility and clinical utility of GP in CUP. Methods: Treatment eligible CUP pts were prospectively enrolled. A novel next-gen targeted sequencing (NGS) assay (ActionSeq/FusionSeq) was used to find genomic alterations (GAs) (somatic mutations in 212 and fusions in 53 cancer-associated genes). The primary objective was to determine prevalence of GAs and to assess the clinical impact via change in pre-test planned therapy (either referral to biomarker pertinent clinical trial (CT) or off label use of FDA approved drugs). With 54 pts we achieved 80% power (α 0.05) to a treatment change in 10% (5% to 15%) pts. Results: Between 9/2016 and 8/2019, 150 pts were consented. Tissue for GP was available in 59 (39%) pts (for 91 pts, samples had exhausted or insufficient tissue). Test was successfully performed on 54 (92%) pts. Cohort characteristics include: median age: 58 yr, male 43%, ECOG PS ≤1 96%, median IHC 8 (range 2-26), median survival 33 m (95% CI 18-47). Median reporting time was 23 days. Four (7%) pts had no identifiable GAs. A fusion ( PTRPK) was seen in 1 (2%) pt. Among 50 pts, total number of GAs were 487; 123 (26%) were “clinically relevant” (median 2.5/pt, range 1-11) while 364 (76%) were variants of unknown significance. Of the 123 GAs, 94 were mutations and 29 were amplifications. The 5 most common mutations were TP53, KRAS, PIK3CA, ARID1A, and NRAS and amplifications were CCND1, FGFR3, ERBB2, EGFR, and MYC. Planned therapy change post ActionSeq occurred in 13 pts (22%, 95% CI 13-34) (2 received an off-label drug; 9 were CT eligible [2 enrolled, 5 had PS decline, 2 pending]; 2 were lost to follow-up). Responses were seen in 2 of 4 pts who received GP based treatment. Conclusions: Comprehensive GP should be offered early to CUP pts. GP can help identify novel therapy and clinical trial options. Given the high rate of insufficient tissue cases, integrating a tissue sensitive algorithm involving IHC and GP in therapeutic management of CUP is merited.

Abstract P3-06-09: BRCA mutations and not type 1 tandem duplicator phenotypes are associated with pathological complete response in patients with triple negative breast cancer undergoing neoadjuvant carboplatin/nab-paclitaxel

Background. We recently described six distinct genomic configurations characterized by large numbers of distributed somatic tandem duplications (TDs) known as Tandem Duplicator Phenotypes (TDPs). Different TDPs feature TDs of different span sizes, and are enriched in TNBC, ovarian, and uterine cancers. Type 1 TDPs (i.e. groups 1, 1/2mix and 1/3mix) feature short span TDs (˜11Kb in size), invariably show abrogation of BRCA1 (via mutation or methylation) and of TP53, and affect ˜40% of TNBCs. We had observed, in limited in vitro and preclinical PDX models, that TDP status correlates with platinum sensitivity (1). Here, we assess TDP status across a cohort of 42 TNBC patients (pts) undergoing neoadjuvant carboplatin and NAB-paclitaxel to test the hypothesis that type 1 TDP status may be predictive of optimal response to platinum-based therapy.

Methods. 42 pts with TNBC were enrolled in a phase II study of neoadjuvant carboplatin/nab-paclitaxel at the City of Hope National Medical Center (NCT01525966).Pathological complete response (pCR) was achieved in 50% of pts (21/42). WGS was performed using standard Illumina protocols. Structural variants were called using Crest, Delly and BreakDancer, and high confidence breakpoints were selected when called by at least two tools and by requiring split-read support. TDP status was ascertained as recently described (2). BRCA1 methylation was determined by methylation-specific PCR.

Results. 45% of the tumors classified as TDP (19/42). Consistent with our previous observation, the vast majority were type 1 TDPs with short span TDs (n=17) and were strongly associated with BRCA1 mutation or methylation (16/17, P= 1.4E-8). However, there was no correlation between TDP status and pCR (OR=1.1, NS). In a more detailed analysis, we found that BRCA1 mutation correlated with pCR rate (6/7 pCR, P=0.01), whereas promoter methylation did not (4/11 pCR, NS). Moreover, both pts with mutant BRCA2 achieved pCR. Thus, as a group, pts with BRCA1/2 mutations (but not BRCA1 methylation) were more likely to achieve pCR than those with wild type BRCA1/2 (OR=11.9, P=1.7E-2). Results were unchanged when using RCB 0 and 1 vs. RCB 2 and 3 as the response criteria.

Conclusions. This study confirmed that reduction of BRCA1 activity via either mutation or methylation robustly associates with type 1 TDPs in TNBC. However, TDP status did not predict good response, suggesting the separation of BRCA effects on genomic instability and platinum sensitivity. This indicates that genomic signature assessments, such as TDP and HRD, may not be sufficient in predicting pCR in TNBC. Importantly, we found that BRCA1/2 mutated TNBC pts were more likely to experience pCR (8/9) compared with pts with either BRCA1 methylation (4/11) or wild type BRCA1/2 (8/21). The exact genetic underpinnings of response in non-BRCA pts are currently under investigation.

References.

1) Menghi et al, The Tandem Duplicator Phenotype is a Prevalent Genome-Wide Cancer Configuration Driven by Distinct Gene Mutations, Cancer Cell (2018).

2) Menghi et al, The tandem duplicator phenotype as a distinct genomic configuration in cancer, Proc Natl Acad Sci (2016).

Citation Format: Menghi F, Yuan Y, Somlo G, Liu ET. BRCA mutations and not type 1 tandem duplicator phenotypes are associated with pathological complete response in patients with triple negative breast cancer undergoing neoadjuvant carboplatin/nab-paclitaxel [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P3-06-09.

Mouse Models for Cancer Immunotherapy Research

Immunotherapy has revolutionized cancer therapy, largely attributed to the success of immune-checkpoint blockade. However, there are subsets of patients across multiple cancers who have not shown robust responses to these agents. A major impediment to progress in the field is the availability of faithful mouse models that recapitulate the complexity of human malignancy and immune contexture within the tumor microenvironment. These models are urgently needed across all malignancies to interrogate and predict antitumor immune responses and therapeutic efficacy in clinical trials. Herein, we seek to review pros and cons of different cancer mouse models, and how they can be used as platforms to predict efficacy and resistance to cancer immunotherapies.Significance: Although immunotherapy has shown substantial benefit in the treatment of a variety of malignancies, a key hurdle toward the advancement of these therapies is the availability of immunocompetent preclinical mouse models that recapitulate human disease. Here, we review the evolution of preclinical mouse models and their utility as coclinical platforms for mechanistic interrogation of cancer immunotherapies. Cancer Discov; 8(11); 1358-65. ©2018 AACR.

Development, Function, and Clinical Significance of Plasmacytoid Dendritic Cells in Chronic Myeloid Leukemia

Plasmacytoid dendritic cells (pDC) are the main producers of a key T-cell-stimulatory cytokine, IFNα, and critical regulators of antiviral immunity. Chronic myeloid leukemia (CML) is caused by BCR-ABL, which is an oncogenic tyrosine kinase that can be effectively inhibited with ABL-selective tyrosine kinase inhibitors (TKI). BCR-ABL-induced suppression of the transcription factor interferon regulatory factor 8 was previously proposed to block pDC development and compromise immune surveillance in CML. Here, we demonstrate that pDCs in newly diagnosed CML (CML-pDC) develop quantitatively normal and are frequently positive for the costimulatory antigen CD86. They originate from low-level BCR-ABL-expressing precursors. CML-pDCs also retain their competence to maturate and to secrete IFN. RNA sequencing reveals a strong inflammatory gene expression signature in CML-pDCs. Patients with high CML-pDC counts at diagnosis achieve inferior rates of deep molecular remission (MR) under nilotinib, unless nilotinib therapy is combined with IFN, which strongly suppresses circulating pDC counts. Although most pDCs are BCR-ABL-negative in MR, a substantial proportion of BCR-ABL ⁺ CML-pDCs persists under TKI treatment. This could be of relevance, because CML-pDCs elicit CD8⁺ T cells, which protect wild-type mice from CML. Together, pDCs are identified as novel functional DC population in CML, regulating antileukemic immunity and treatment outcome in CML.Significance: CML-pDC originates from low-level BCR-ABL expressing stem cells into a functional immunogenic DC-population regulating antileukemic immunity and treatment outcome in CML. Cancer Res; 78(21); 6223-34. ©2018 AACR.

The Tandem Duplicator Phenotype Is a Prevalent Genome-Wide Cancer Configuration Driven by Distinct Gene Mutations

The tandem duplicator phenotype (TDP) is a genome-wide instability configuration primarily observed in breast, ovarian, and endometrial carcinomas. Here, we stratify TDP tumors by classifying their tandem duplications (TDs) into three span intervals, with modal values of 11 kb, 231 kb, and 1.7 Mb, respectively. TDPs with ∼11 kb TDs feature loss of TP53 and BRCA1. TDPs with ∼231 kb and ∼1.7 Mb TDs associate with CCNE1 pathway activation and CDK12 disruptions, respectively. We demonstrate that p53 and BRCA1 conjoint abrogation drives TDP induction by generating short-span TDP mammary tumors in genetically modified mice lacking them. Lastly, we show how TDs in TDP tumors disrupt heterogeneous combinations of tumor suppressors and chromatin topologically associating domains while duplicating oncogenes and super-enhancers.

Abstract 5676: Patient-derived tumor xenografts in humanized NSG-SGM3 mice: An improved immuno-oncology platform

The JAX® Onco-Hu® platform utilizes humanized mice engrafted with tumors to enable in vivo investigation of the interactions between the human immune system and human cancer. We have recently shown that humanized NOD-scid IL2Rγnull (NSG™) mice bearing patient-derived xenografts (PDX) allow efficacy studies of checkpoint inhibitors. A major avenue of our investigation is to generate murine humanized models containing a more complete human hematopoietic system and robust innate immune cell population. Next-generation NSG strains include triple transgenic NSG mice (NSG-SGM3) expressing myelosupportive human cytokines KITLG, CSF2, and IL-3. When engrafted with CD34+ human hematopoietic progenitor cells (HPCs) from CD3-depleted umbilical cord blood, NSG-SGM3 mice produce higher myeloid and Treg populations in the circulation as compared to NSG mice over 18 weeks post engraftment. We implanted an array of PDX tumors into humanized NSG-SGM3 mice at 2-3 months post engraftment. Tumors were dissociated and single-cell infiltrates were analyzed by multicolor flow cytometry with a focus on examining overall immune cell infiltration and the levels of hCD33+ myeloid cells. In the PS4050 melanoma PDX model, we found that hCD45+ cell infiltration was significantly increased in hu-NSG-SGM3 mice as compared to hu-NSG mice engrafted with the same HPC donor (3.7% vs. 1% of viable cells). The majority of tumor-infiltrating cells in hu-NSG-SGM3 mice expressed hCD33 (55% of hCD45+) and the percentage was significantly higher than that in hu-NSG mice (13%). hCD3+T cell infiltration level was similar between these two strains (~20% of hCD45+). PS4050-bearing hu-NSG-SGM3 mice treated with the anti-PD-1 antibody pembrolizumab (Keytruda) showed a significant reduction in tumor growth and the PD-1 levels in tumor-infiltrating T cells were greatly reduced by flow cytometry analysis. The overall hCD45+ cell infiltration and the frequencies of hCD4+T, hCD8+T, and hCD33+myeloid cells in tumors remained similar after treatment. Lastly, we observed that the effect of Keytruda on tumor growth reduction in hu-NSG-SGM3 mice is PD-L1-dependent using the human lung carcinoma cell line NCI-H460 depleted of PD-L1 expression by CRISPR. Keytruda treatment significantly reduced mock-transfected NCI-H460 cell growth. By comparison, PD-L1 KO NCI-H460 cells grew more slowly than the mock cells and lost the response to Keytruda. Together, these results indicate that PDX tumor-implanted hu-NSG-SGM3 mice serve as an important platform for understanding human immune system and tumor microenvironment interactions and for preclinical immuno-oncology efficacy studies.

Citation Format: Li-Chin Yao, Mingshan Cheng, Ken-Edwin Aryee, Pooja Kumar, Nicole Walsh, Dale Greiner, Leonard Shultz, Edison T. Liu, Michael Brehm, James G. Keck. Patient-derived tumor xenografts in humanized NSG-SGM3 mice: An improved immuno-oncology platform [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 5676.

Picky comprehensively detects high-resolution structural variants in nanopore long reads

Acquired genomic structural variants (SVs) are major hallmarks of cancer genomes, but they are challenging to reconstruct from short-read sequencing data. Here we exploited the long reads of the nanopore platform using our customized pipeline, Picky ( https://github.com/TheJacksonLaboratory/Picky ), to reveal SVs of diverse architecture in a breast cancer model. We identified the full spectrum of SVs with superior specificity and sensitivity relative to short-read analyses, and uncovered repetitive DNA as the major source of variation. Examination of genome-wide breakpoints at nucleotide resolution uncovered micro-insertions as the common structural features associated with SVs. Breakpoint density across the genome is associated with the propensity for interchromosomal connectivity and was found to be enriched in promoters and transcribed regions of the genome. Furthermore, we observed an over-representation of reciprocal translocations from chromosomal double-crossovers through phased SVs. We demonstrate that Picky analysis is an effective tool for comprehensive detection of SVs in cancer genomes from long-read data.

Abstract P6-15-07: Not presented

This abstract was not presented at the symposium.

Abstract GS1-05: Tandem duplicator phenotypes define 50% of triple negative breast cancers

Background. We recently discovered a unique chromotype, the Tandem Duplicator Phenotype (TDP), characterized by hundreds of somatic tandem duplications (TDs) scattered throughout the genome of a large percentage of triple negative breast cancers (TNBCs). Importantly, we observed that the TDP associates with a better response to cisplatin therapy in vitro and in vivo, suggesting that it is a tractable and quantitative biomarker of response to platinum-based therapy. Here, we expand on our initial findings by analyzing Whole-Genome (WG) sequences of over 2,700 tumors.

Methods. TD coordinates from WG sequences relative to 2717 tumors were assembled from over 30 independent studies representing several cancer types, including 254 TNBCs. WG sequencing of mouse breast tumors was carried out using standard Illumina protocols. The number, distribution and span-size of somatic TDs from a training set of 992 tumors were used to develop a TDP classifier that identifies highly recurrent but clearly distinct TDP profiles. The TDP classifier was then applied to the remaining tumor sequences. WG mutation and copy number datasets were investigated to identify the genetic drivers associated with each TDP profile, and the genomic consequences of different TDPs were evaluated through identification of genomic hotspots for gene duplication and transection.

Results. We describe six different TDPs featuring distinct TD span size distributions, with peaks at 10Kb (group 1), 300Kb (group 2) and 3Mb (group 3), or different combinations of these (mix12, mix13 and mix23). More than half of all TNBC display a TDP. Of these, 55% classify as group 1, 14% as group 2 and 15% as group mix12. Whereas all TDP groups show a higher TP53 mutation rate compared to non-TDP tumors, each TDP profile is characterized by specific additional gene perturbations, with loss of BRCA1 occurring in groups 1, mix12 and mix13; CCNE1 amplification in group 2; and CDK12 mutations in group mix23. We show that different TDPs are subject to the perturbation of specific oncogenic networks resulting from the duplication of oncogenes by larger TDs (>300Kb) or the disruption of tumor suppressors via double transections by shorter TDs (10Kb). Indeed, tumor suppressor genes such as PTEN, RB1 and MLL3 are frequently disrupted by TDs in TNBC TDP group 1 tumors, whereas TNBC TDP group 2 tumors commonly feature duplication of oncogenes such as MYC and MALAT1. Finally, through WG analyses of 18 mouse models (GEMMs) of breast cancer, we provide the first mechanistic evidence of the driving role of conjoint loss of TP53 and BRCA1, but not of BRCA2, in inducing the TDP group 1 profile.

Conclusions. Our study shows a definitive genetic induction of one specific form of TDP (group 1) characterized by 10kb TD span. Different TDP profiles are characterized by alternative somatic genetic origins but always couple with disruptive TP53 mutations. The consequences of the massive TD formation in TDP TNBCs suggest a systems strategy to tumor induction involving heterogeneous combinations of oncogenes and tumor suppressors. That these TDP forms, accounting for ˜50% of TNBC, are associated with significant sensitivity to cisplatin suggest that this chromotype may identify TNBC patients who would benefit from upfront platinum-based chemotherapy.

Citation Format: Liu ET, Menghi F, Barthel F, Yadav V, Tang M, Ji B, Carter G, Jonkers J, Verhaak R. Tandem duplicator phenotypes define 50% of triple negative breast cancers [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr GS1-05.

Humanized mice in studying efficacy and mechanisms of PD-1-targeted cancer immunotherapy

Establishment of an in vivo small animal model of human tumor and human immune system interaction would enable preclinical investigations into the mechanisms underlying cancer immunotherapy. To this end, nonobese diabetic (NOD).Cg-Prkdc^scidIL2rg^tm1Wjl/Sz (null; NSG) mice were transplanted with human (h)CD34⁺ hematopoietic progenitor and stem cells, which leads to the development of human hematopoietic and immune systems [humanized NSG (HuNSG)]. HuNSG mice received human leukocyte antigen partially matched tumor implants from patient-derived xenografts [PDX; non–small cell lung cancer (NSCLC), sarcoma, bladder cancer, and triple-negative breast cancer (TNBC)] or from a TNBC cell line-derived xenograft (CDX). Tumor growth curves were similar in HuNSG compared with nonhuman immune-engrafted NSG mice. Treatment with pembrolizumab, which targets programmed cell death protein 1, produced significant growth inhibition in both CDX and PDX tumors in HuNSG but not in NSG mice. Finally, inhibition of tumor growth was dependent on hCD8⁺ T cells, as demonstrated by antibody-mediated depletion. Thus, tumor-bearing HuNSG mice may represent an important, new model for preclinical immunotherapy research.—Wang, M., Yao, L.-C., Cheng, M., Cai, D., Martinek, J., Pan, C.-X., Shi, W., Ma, A.-H., De Vere White, R. W., Airhart, S., Liu, E. T., Banchereau, J., Brehm, M. A., Greiner, D. L., Shultz, L. D., Palucka, K., Keck, J. G. Humanized mice in studying efficacy and mechanisms of PD-1-targeted cancer immunotherapy.

Mechanism of tandem duplication formation in BRCA1-mutant cells

Small, approximately 10-kilobase microhomology-mediated tandem duplications are abundant in the genomes of BRCA1-linked but not BRCA2-linked breast cancer. Here we define the mechanism underlying this rearrangement signature. We show that, in primary mammalian cells, BRCA1, but not BRCA2, suppresses the formation of tandem duplications at a site-specific chromosomal replication fork barrier imposed by the binding of Tus proteins to an array of Ter sites. BRCA1 has no equivalent role at chromosomal double-stranded DNA breaks, indicating that tandem duplications form specifically at stalled forks. Tandem duplications in BRCA1 mutant cells arise by a replication restart-bypass mechanism terminated by end joining or by microhomology-mediated template switching, the latter forming complex tandem duplication breakpoints. Solitary DNA ends form directly at Tus-Ter, implicating misrepair of these lesions in tandem duplication formation. Furthermore, BRCA1 inactivation is strongly associated with ~10 kilobase tandem duplications in ovarian cancer. This tandem duplicator phenotype may be a general signature of BRCA1-deficient cancer.

Molecular features of influenza A (H1N1)pdm09 prevalent in Mexico during winter seasons 2012-2014

Since the emergence of the pandemic H1N1pdm09 virus in Mexico and California, biannual increases in the number of cases have been detected in Mexico. As observed in previous seasons, pandemic A/H1N1 09 virus was detected in severe cases during the 2011-2012 winter season and finally, during the 2013-2014 winter season it became the most prevalent influenza virus. Molecular and phylogenetic analyses of the whole viral genome are necessary to determine the antigenic and pathogenic characteristics of influenza viruses that cause severe outcomes of the disease. In this paper, we analyzed the evolution, antigenic and genetic drift of Mexican isolates from 2009, at the beginning of the pandemic, to 2014. We found a clear variation of the virus in Mexico from the 2011-2014 season due to different markers and in accordance with previous reports. In this study, we identified 13 novel substitutions with important biological effects, including virulence, T cell epitope presented by MHC and host specificity shift and some others substitutions might have more than one biological function. The systematic monitoring of mutations on whole genome of influenza A pH1N1 (2009) virus circulating at INER in Mexico City might provide valuable information to predict the emergence of new pathogenic influenza virus.

Abstract IA13: The Patient-Derived Xenograft Program at The Jackson Laboratory

The Patient Derived Xenograft (PDX) Program at The Jackson Laboratory have as its focus, the development of quality surrogates of primary human tumors for the research community, the construction and optimization of murine hosts for human tumor xenografts, and the conduct of research into cancer biology using these PDX models. Begun in 2010, the PDX Program is based on the foundational murine host, the NOD-scid IL2rgnull or NSG (NOD.Cg-Prkdcscid IL2rgtm1wjl/SzJIL2rgtm1wjl) mouse, which lacks T, B, and NK cells and harbors deficiencies in innate immune response due to defects in multiple cytokine receptors including the IL2. 4, 7, 9, 15, and 21 receptors1,2. NSG mice have superior engraftment profiles as compared to other immunodeficient strains. These tumor xenografts in NSG hosts maintain the features of primary of engrafted human tumors including, tumor heterogeneity, three-dimensional architecture, and tumor biology including response to therapeutics. The engraftment rates vary according to tumor types with prostate cancers being among the lowest (3%), and colorectal cancers (66%), melanomas, and glioblastomas being among the best, according to whether the tumors are from metastatic sites, show a higher tumor grade, or have larger amounts of starting tumor material.

We currently have 446 PDX models in stock with ~100 in the pipeline, and when coupled with 218 models from our Korean cooperative institution (Ewha University), our total collection is ~664 covering 15 tumor types with none beyond passage number 4 (P4). Most have genomic data available and the Korean PDXs have matching normal tissue and DNA sequence data. Importantly, the human immune system can be engrafted in NSG mice and in the presence of a human tumor, can be used to study tumor-immune cell interactions including response to immune therapeutics. In addition, our colleagues and we have developed genetic modifications of the NSG to provide unique engraftment options: MHC class I and II knockouts to reduce xenogeneic graft-vs.-host disease, expressing HLA-A2 and other HLA class I and II transgenes to develop HLA-restricted human cytotoxic T cells, and IL-6 and prolactin to support engraftment of tumors needing these factors. NSG mice expressing human IL3/GM-CSF/SCF supports the immune reconstitution of myeloid progenitors providing a more complete immune system and has been useful in testing cancer immune therapeutics. With 26 staff dedicated to PDX and NSG experimentation, we provide strains of NSG mice with or without human immune cell reconstitution, and conduct large-scale pharmacological studies in cohorts of PDX models.

1Hidalgo M, Amant F, Biankin AV, Budinská E, Byrne AT, Caldas C, Clarke RB, de Jong S, Jonkers J, Mælandsmo GM, Roman-Roman S, Seoane J, Trusolino L, Villanueva A. Patient-derived xenograft models: an emerging platform for translational cancer research. Cancer Discov. 2014 Sep;4(9):998-1013.

2Shultz LD, Goodwin N, Ishikawa F, Hosur V, Lyons BL, Greiner DL. Human cancer growth and therapy in immunodeficient mouse models. Cold Spring Harbor Protoc. 2014 Jul 1;2014(7):694-708.

Citation Format: Edison T. Liu, James Keck, Susie Airhart, Lenny Shultz, Charles Lee, Carol Bult. The Patient-Derived Xenograft Program at The Jackson Laboratory. [abstract]. In: Proceedings of the AACR Special Conference: Patient-Derived Cancer Models: Present and Future Applications from Basic Science to the Clinic; Feb 11-14, 2016; New Orleans, LA. Philadelphia (PA): AACR; Clin Cancer Res 2016;22(16_Suppl):Abstract nr IA13.

Abstract IA29: Mice host selection for patient-derived xenograft (PDX) model development and other critical factors for success

The Jackson Laboratory PDX program engages 26 staff members, 5 PIs, and runs thousands of PDX models per year in pre-clinical experiments both for internal scientific experiments and external projects. Over the last 6 years of operation of the PDX program, we have gained a significant amount of experience in all aspects of the process. This experience includes the genomic analysis of the tumors and the establishment of an extensive database annotating many of the 446 PDX models in our US inventory. Herein, we describe some of the characteristics of the system that enhances successful experimentation in this platform.

Several factors significantly improve the engraftment rate of tumors1: 1) the degree of immunodeficiency the host mouse - the NOD.Cg-Prkdcscid IL2rgtm1wjl/SzJ (aka, NSG), the NOD.CB17-Prkdcscid/J (aka,NOD-scid), or the NOD.Cg-Rag1tm1momIL2rgtm1wjl/SzJ (aka, NRG) engrafting better than the beige-scid and athymic nude mice; 2) the greater the amount of tissue engrafted, 3) the late or metastatic nature of the tumor, 4) the shorter time from surgery to implantation, 5) the absence of enzymatic dissociation, and 6) orthotopic implantation. Sequencing and expression analysis show the maintenance of the core genomic configuration between PDX and the primary tumors though some genetic differences are noted of indeterminate significance. Human stromal cells, however, tend to be replaced by murine stroma after the first passage. Tumor genetic heterogeneity is maintained through the fourth passage in NSG PDX models as confirmed by deep sequencing of bulk tumor and of individual progenitor clones. Most importantly, however, the tumor responses to systemic therapies appear to reflect the patient response2. Therefore, the use of PDX models to test novel agents may speed pre-clinical drug development.

The most important use of PDX systems may be in immuno-oncology given that “humanized” PDX models where implantation of a primary human tumor is implanted in a mouse with a reconstituted human cellular immune system provides a powerful preclinical experimental platform to test new immune modulators in human cancers. We, and others, have shown that humanized NSG mice bearing human tumor xenografts exhibit dramatic responses to immune checkpoint inhibitors that are immune cell and drug dependent3. Immune reconstitution is more complex because of the requirement of human cytokines that are not substituted by their murine counterparts. Engineered mice expressing human cytokines, e.g., IL3, CSF2 (GM-CSF), and KITLG (stem cell factor) in the NSG background (aka, NSG-SGM3 mice), and those expressing CSF1 (M-CSF), IL3, CSF2, and THPO in the C;129S4-Rag2tm1.1Flv IL2rgtm1.1Flv/J background (aka, MITRG mice) after engraftment with human hematopoietic stem cells have been shown to support myeloid cells4 including macrophages absent in standard NSG5. It is anticipated that these “next generation” humanized mice will provide a more nuanced picture of the tumor-immune system interaction6.

It is important to understand the challenges and limitations of the PDX platform that can be mitigated to a degree by quality control and study design. There are simple caveats. ~5 % of engrafted solid tumors give rise to EBV positive lymphomas and not the primary tumor. Moreover, ~5-10% of PDX tumors are overgrown by a transformed murine cell especially in late passaged. Thus, stringent histological quality control is necessary which includes the assessment of human cytokeratin, which provides an assessment of murine cancer incursion of solid human cancers. Another concern is that drug dosing for PDX experiments is often very different from that used in human studies. The NSG mice are more sensitive to certain DNA damaging agents and to radiation than NRG or Rag1null mice. Therefore, the structuring of combination studies using genotoxic agents is complicated and should be interpreted with appropriate care.

In terms of study design, we have found that each PDX model from an individual patient will give rise to individual tumors in an NSG cohort with significant growth and response variations. Thus, for each treatment arm we have calculated that between 6-8 animals is the minimal number required to attain statistical power of 95-99% to identify efficacy between arms. Moreover, in any treatment arm, it is necessary to assess the response of each individual PDX bearing mouse since a few individual tumors in a cohort may be resistant to a drug whereas the average of the arm shows an overall response. Recently, Gao, et al7 presented an alternative way to conduct PDX studies for drug development where only one mouse per PDX model was used per drug. The overall data (not whether a drug specifically was efficacious in a specific disease type) provided important strategic information in development. They raised an important point and that is that PDX preclinical studies should be structured differently from classical clinical trials to make best use of the platform. Not only the trial design, but even how to call a response should be reexamined. The partial responses seen in PDX experiments that can be precisely quantified but that do not qualify using RECIST criteria provide potentially important information about drug efficacy.

Taken together, the PDX platform using severely immunodeficient mice is a powerful tool that can significantly accelerate the development of new therapeutics by dramatically facilitating the advancement of innovative therapies with a high likelihood for success8.

References:

1Shultz LD, Goodwin N, Ishikawa F, Hosur V, Lyons BL, Greiner DL. Human cancer growth and therapy in immunodeficient mouse models. Cold Spring Harbor Protoc. 2014 Jul 1;2014(7):694-708.

2Garralda E, Paz K, López-Casas PP, Jones S, Katz A, Kann LM, López- Rios F, Sarno F, Al-Shahrour F, Vasquez D, Bruckheimer E, Angiuoli SV, Calles A, Diaz LA, Velculescu VE, Valencia A, Sidransky D, Hidalgo M. Integrated next-generation sequencing and avatar mouse models for personalized cancer treatment. Clin Cancer Res. 2014 May 1;20(9):2476-84 Epub 2014 Mar 14.

3Wang M, Keck JG, Cheng M, Cai D, Shultz L, Palucka K, Banchereau J, Bult C, Huntress R. Patient-derived tumor xenografts in humanized NSG mice: a model to study immune responses in cancer therapy. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr LB-050.

4Billerbeck E, Barry WT, Mu K, Dorner M, Rice CM, Ploss A. Development of human CD4+FoxP3+ regulatory T cells in human stem cell factor-, granulocyte-macrophage colony-stimulating factor-, and interleukin-3- expressing NOD-SCID IL2Rγ(null) humanized mice. Blood. 2011 Mar 17;117(11):3076-86.

5Rongvaux A, Willinger T, Martinek J, Strowig T, Gearty SV, Teichmann LL, Saito Y, Marches F, Halene S, Palucka AK, Manz MG, Flavell RA. Development and function of human innate immune cells in a humanized mouse model. Nat Biotechnol. 2014 Apr;32(4):364-72.

6Shultz LD, Brehm MA, Garcia-Martinez JV, Greiner DL. Humanized mice for immune system investigation: progress, promise and challenges. Nat Rev Immunol. 2012 Nov;12(11):786-98.

7Gao H, Korn JM, Ferretti S, Monahan JE, Wang Y, Singh M, Zhang C, Schnell C, Yang G, Zhang Y, Balbin OA, Barbe S, Cai H, Casey F, Chatterjee S, Chiang DY, Chuai S, Cogan SM, Collins SD, Dammassa E, Ebel N, Embry M, Green J, Kauffmann A, Kowal C, Leary RJ, Lehar J, Liang Y, Loo A, Lorenzana E, Robert McDonald E 3rd, McLaughlin ME, Merkin J, Meyer R, Naylor TL, Patawaran M, Reddy A, Röelli C, Ruddy DA, Salangsang F, Santacroce F, Singh AP, Tang Y, Tinetto W, Tobler S, Velazquez R, Venkatesan K, Von Arx F, Wang HQ, Wang Z, Wiesmann M, Wyss D, Xu F, Bitter H, Atadja P, Lees E, Hofmann F, Li E, Keen N, Cozens R, Jensen MR, Pryer NK, Williams JA, Sellers WR. High- throughput screening using patient-derived tumor xenografts to predict clinical trial drug response. Nat Med. 2015 Nov;21(11):1318-25.

8Gandara DR, Mack PC, Bult C, Li T, Lara PN Jr, Riess JW, Astrow SH, Gandour-Edwards R, Cooke DT, Yoneda KY, Moore EH, Pan CX, Burich RA, David EA, Keck JG, Airhart S, Goodwin N, de Vere White RW, Liu ET. Bridging tumor genomics to patient outcomes through an integrated patient-derived xenograft platform. Clin Lung Cancer. 2015 May;16(3):165- 72.

Citation Format: Edison T. Liu, Carol Bult, Jeff Chuang, Pooja Kumar, Mingshan Cheng, R. Krishna Murthy Karuturi, Vivek Philip, James Keck, Karolina Palucka, Larry Shultz. Mice host selection for patient-derived xenograft (PDX) model development and other critical factors for success. [abstract]. In: Proceedings of the AACR Special Conference: Patient-Derived Cancer Models: Present and Future Applications from Basic Science to the Clinic; Feb 11-14, 2016; New Orleans, LA. Philadelphia (PA): AACR; Clin Cancer Res 2016;22(16_Suppl):Abstract nr IA29.

CTCF-Mediated Human 3D Genome Architecture Reveals Chromatin Topology for Transcription

Spatial genome organization and its effect on transcription remains a fundamental question. We applied an advanced chromatin interaction analysis by paired-end tag sequencing (ChIA-PET) strategy to comprehensively map higher-order chromosome folding and specific chromatin interactions mediated by CCCTC-binding factor (CTCF) and RNA polymerase II (RNAPII) with haplotype specificity and nucleotide resolution in different human cell lineages. We find that CTCF/cohesin-mediated interaction anchors serve as structural foci for spatial organization of constitutive genes concordant with CTCF-motif orientation, whereas RNAPII interacts within these structures by selectively drawing cell-type-specific genes toward CTCF foci for coordinated transcription. Furthermore, we show that haplotype variants and allelic interactions have differential effects on chromosome configuration, influencing gene expression, and may provide mechanistic insights into functions associated with disease susceptibility. 3D genome simulation suggests a model of chromatin folding around chromosomal axes, where CTCF is involved in defining the interface between condensed and open compartments for structural regulation. Our 3D genome strategy thus provides unique insights in the topological mechanism of human variations and diseases.

Xenopus tropicalis Genome Re-Scaffolding and Re-Annotation Reach the Resolution Required for In Vivo ChIA-PET Analysis

Genome-wide functional analyses require high-resolution genome assembly and annotation. We applied ChIA-PET to analyze gene regulatory networks, including 3D chromosome interactions, underlying thyroid hormone (TH) signaling in the frog Xenopus tropicalis. As the available versions of Xenopus tropicalis assembly and annotation lacked the resolution required for ChIA-PET we improve the genome assembly version 4.1 and annotations using data derived from the paired end tag (PET) sequencing technologies and approaches (e.g., DNA-PET [gPET], RNA-PET etc.). The large insert (~10Kb, ~17Kb) paired end DNA-PET with high throughput NGS sequencing not only significantly improved genome assembly quality, but also strongly reduced genome “fragmentation”, reducing total scaffold numbers by ~60%. Next, RNA-PET technology, designed and developed for the detection of full-length transcripts and fusion mRNA in whole transcriptome studies (ENCODE consortia), was applied to capture the 5' and 3' ends of transcripts. These amendments in assembly and annotation were essential prerequisites for the ChIA-PET analysis of TH transcription regulation. Their application revealed complex regulatory configurations of target genes and the structures of the regulatory networks underlying physiological responses. Our work allowed us to improve the quality of Xenopus tropicalis genomic resources, reaching the standard required for ChIA-PET analysis of transcriptional networks. We consider that the workflow proposed offers useful conceptual and methodological guidance and can readily be applied to other non-conventional models that have low-resolution genome data.

Abstract P6-06-02: Molecular characterization of a patient-derived xenograft (PDX) resource for triple negative breast cancer

The Jackson Laboratory (JAX) has developed a resource of human tumors implanted into immune deficient mice (patient derived xenografts; PDX) as a platform for testing standard of care and novel therapeutic options for Triple Negative Breast Cancer. PDX models provide an advantage over cell culture based models for testing therapeutic interventions because they retain properties such as tumor cell heterogeneity that are critical to the biological properties of a patient’s tumor and response to treatment.

Tumor material acquired from biopsy or surgical resection was implanted subcutaneously into the flank of immune deficient NOD-scid IL2r gamma-chain null (NSG) mice. The PDX resource currently contains 21 established breast cancer PDX models (12 TNBC) with 24 additional models currently in development. Two of the established TNBC PDX models have BRCA1 mutations. The median age of the patients from whom tumor material was obtained for all breast models is 53 (45-89).

Tumors that successfully engrafted were characterized for somatic mutations using the new JAX Clinical Cancer Panel, Copy Number Variants using the Affymetrix human 6.0 SNP array, and gene expression using both Affymetrix U133 plus v2 and RNA-Seq. Normalized gene expression was analyzed for characteristic patterns in a pan-cancer approach across all PDX models and further compared with the previously identified TNBC molecular subtypes (Lehmann et al. 2011. JCI 121:2750-2767). The combination of principal components analysis and classification via expression pattern resulted in putative matches of models to most of the known molecularly defined subtypes of TNBC tumors.

Tumor bearing mice for the TNBC PDX models have been treated with docetaxel, cisplatin, cyclophosphamide and doxorubicin. Preliminary studies of tumor response to these treatment regimes revealed systematic differences that can be correlated with features of the genomic analysis, including expression subtype characterization.

The JAX collection of TNBC cancer PDX models is a well-annotated, publically available resource of models with deep genomic characterization and standard of care therapy response data for use in the development of advanced therapeutic options. Genomically defined subgroups within the collection suggest strategies to refine patient selection and treatment algorithms. Information about the models along with summarized genomic data is publicly available at the Mouse Tumor Biology database PDX web portal (http://tumor.informatics.jax.org).

Citation Format: Joel H Graber, James G Keck, Susan D Airhart, Carol J Bult, Edison T Liu. Molecular characterization of a patient-derived xenograft (PDX) resource for triple negative breast cancer [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P6-06-02.

Bridging tumor genomics to patient outcomes through an integrated patient-derived xenograft platform

New approaches to optimization of cancer drug development in the laboratory and the clinic will be required to fully achieve the goal of individualized, precision cancer therapy. Improved preclinical models that more closely reflect the now recognized genomic complexity of human cancers are needed. Here we describe a collaborative research project that integrates core resources of The Jackson Laboratory Basic Science Cancer Center with genomics and clinical research facilities at the UC Davis Comprehensive Cancer Center to establish a clinically and genomically annotated patient-derived xenograft (PDX) platform designed to enhance new drug development and strategies for targeted therapies. Advanced stage non-small-cell lung cancer (NSCLC) was selected for initial studies because of emergence of a number of "druggable" molecular targets, and recent recognition of substantial inter- and intrapatient tumor heterogeneity. Additionally, clonal evolution after targeted therapy interventions make this tumor type ideal for investigation of this platform. Using the immunodeficient NOD scid gamma mouse, > 200 NSCLC tumor biopsies have been xenotransplanted. During the annotation process, patient tumors and subsequent PDXs are compared at multiple levels, including histomorphology, clinically applicable molecular biomarkers, global gene expression patterns, gene copy number variations, and DNA/chromosomal alterations. NSCLC PDXs are grouped into panels of interest according to oncogene subtype and/or histologic subtype. Multiregimen drug testing, paired with next-generation sequencing before and after therapy and timed tumor pharmacodynamics enables determination of efficacy, signaling pathway alterations, and mechanisms of sensitivity-resistance in individual models. This approach should facilitate derivation of new therapeutic strategies and the transition to individualized therapy.

Patient-derived xenograft platform to guide precision medicine in bladder cancer

315

Background: The prognosis for bladder cancer has not changed in 30 years. No targeted agents have been approved even though reproducible genetic abnormalities have been identified. The goal of this project was to develop and characterize a patient-derived xenograft (PDX) platform to determine the efficacy of molecularly guided targeted and chemotherapy therapy, drug re-purpose, study resistance mechanisms, and design novel therapy to overcome resistance. Methods: PDXs were developed from direct implantation of uncultured patient bladder cancer specimens into immunodeficient NSG mice. Deep sequencing in combination with computational biology was performed to characterize PDXs and identify druggable genetic aberrations that guided efficacy screening and mechanistic studies. Results: Nineteen PDXs have been established with annotated clinical information. PDXs retained morphology and 92-97% genetic aberrations of parental patient cancers. Deep sequencing revealed multiple druggable genetic aberrations, including the fibroblast growth factor receptor 3 (FGFR3) and other tyrosine kinase receptor pathways. Compared to the progression-free survival (PFS) of 9.5 days in the control arm, matched therapy with an FGFR3 inhibitor BGJ398 prolonged PFS to 18.5 days (p=2.61 X 10-6) in PDXs overexpressing FGFR3. Serial biopsies during treatment revealed reactivation of the downstream pathways coincided with development of resistance while targeting these downstream effectors reversed resistance (12 vs. 22 days, p=0.001). Efficacy studies also revealed that PDXs had differential response to chemotherapeutic drugs that could potentially guide selection of chemotherapeutic drugs for first- and second-line therapies. To determine the clinical applicability of non-myoinvasive bladder cancer, we further developed an orthotopic PDX model that mimiced disease progression to invasive and metastatic bladder cancer. Conclusions: The PDX platform allows screening for multiple targeted therapy, chemotherapy or combinations simultaneously for the most efficacious drugs or combination, and serial biopsies during treatment to study drug resistance, a task not possibly replicable at the clinical setting.

Patient-specific driver gene prediction and risk assessment through integrated network analysis of cancer omics profiles

Extensive and multi-dimensional data sets generated from recent cancer omics profiling projects have presented new challenges and opportunities for unraveling the complexity of cancer genome landscapes. In particular, distinguishing the unique complement of genes that drive tumorigenesis in each patient from a sea of passenger mutations is necessary for translating the full benefit of cancer genome sequencing into the clinic. We address this need by presenting a data integration framework (OncoIMPACT) to nominate patient-specific driver genes based on their phenotypic impact. Extensive in silico and in vitro validation helped establish OncoIMPACT's robustness, improved precision over competing approaches and verifiable patient and cell line specific predictions (2/2 and 6/7 true positives and negatives, respectively). In particular, we computationally predicted and experimentally validated the gene TRIM24 as a putative novel amplified driver in a melanoma patient. Applying OncoIMPACT to more than 1000 tumor samples, we generated patient-specific driver gene lists in five different cancer types to identify modes of synergistic action. We also provide the first demonstration that computationally derived driver mutation signatures can be overall superior to single gene and gene expression based signatures in enabling patient stratification and prognostication. Source code and executables for OncoIMPACT are freely available from http://sourceforge.net/projects/oncoimpact.

Functional chromatin features are associated with structural mutations in cancer

BACKGROUND

Structural mutations (SMs) play a major role in cancer development. In some cancers, such as breast and ovarian, DNA double-strand breaks (DSBs) occur more frequently in transcribed regions, while in other cancer types such as prostate, there is a consistent depletion of breakpoints in transcribed regions. Despite such regularity, little is understood about the mechanisms driving these effects. A few works have suggested that protein binding may be relevant, e.g. in studies of androgen receptor binding and active chromatin in specific cell types. We hypothesized that this behavior might be general, i.e. that correlation between protein-DNA binding (and open chromatin) and breakpoint locations is common across divergent cancers.

RESULTS

We investigated this hypothesis by comprehensively analyzing the relationship among 457 ENCODE protein binding ChIP-seq experiments, 125 DnaseI and 24 FAIRE experiments, and 14,600 SMs from 8 diverse cancer datasets covering 147 samples. In most cancers, including breast and ovarian, we found enrichment of protein binding and open chromatin in the vicinity of SM breakpoints at distances up to 200 kb. Furthermore, for all cancer types we observed an enhanced enrichment in regions distant from genes when compared to regions proximal to genes, suggesting that the SM-induction mechanism is independent from the bias of DSBs to occur near transcribed regions. We also observed a stronger effect for sites with more than one protein bound.

CONCLUSIONS

Protein binding and open chromatin state are associated with nearby SM breakpoints in many cancer datasets. These observations suggest a consistent mechanism underlying SM locations across different cancers.

Systems consequences of amplicon formation in human breast cancer

Chromosomal structural variations play an important role in determining the transcriptional landscape of human breast cancers. To assess the nature of these structural variations, we analyzed eight breast tumor samples with a focus on regions of gene amplification using mate-pair sequencing of long-insert genomic DNA with matched transcriptome profiling. We found that tandem duplications appear to be early events in tumor evolution, especially in the genesis of amplicons. In a detailed reconstruction of events on chromosome 17, we found large unpaired inversions and deletions connect a tandemly duplicated ERBB2 with neighboring 17q21.3 amplicons while simultaneously deleting the intervening BRCA1 tumor suppressor locus. This series of events appeared to be unusually common when examined in larger genomic data sets of breast cancers albeit using approaches with lesser resolution. Using siRNAs in breast cancer cell lines, we showed that the 17q21.3 amplicon harbored a significant number of weak oncogenes that appeared consistently coamplified in primary tumors. Down-regulation of BRCA1 expression augmented the cell proliferation in ERBB2-transfected human normal mammary epithelial cells. Coamplification of other functionally tested oncogenic elements in other breast tumors examined, such as RIPK2 and MYC on chromosome 8, also parallel these findings. Our analyses suggest that structural variations efficiently orchestrate the gain and loss of cancer gene cassettes that engage many oncogenic pathways simultaneously and that such oncogenic cassettes are favored during the evolution of a cancer.

Detection of chromosomal breakpoints in patients with developmental delay and speech disorders

Delineating candidate genes at the chromosomal breakpoint regions in the apparently balanced chromosome rearrangements (ABCR) has been shown to be more effective with the emergence of next-generation sequencing (NGS) technologies. We employed a large-insert (7-11 kb) paired-end tag sequencing technology (DNA-PET) to systematically analyze genome of four patients harbouring cytogenetically defined ABCR with neurodevelopmental symptoms, including developmental delay (DD) and speech disorders. We characterized structural variants (SVs) specific to each individual, including those matching the chromosomal breakpoints. Refinement of these regions by Sanger sequencing resulted in the identification of five disrupted genes in three individuals: guanine nucleotide binding protein, q polypeptide (GNAQ), RNA-binding protein, fox-1 homolog (RBFOX3), unc-5 homolog D (C.elegans) (UNC5D), transmembrane protein 47 (TMEM47), and X-linked inhibitor of apoptosis (XIAP). Among them, XIAP is the causative gene for the immunodeficiency phenotype seen in the patient. The remaining genes displayed specific expression in the fetal brain and have known biologically relevant functions in brain development, suggesting putative candidate genes for neurodevelopmental phenotypes. This study demonstrates the application of NGS technologies in mapping individual gene disruptions in ABCR as a resource for deciphering candidate genes in human neurodevelopmental disorders (NDDs).