Integrative Analyses of Tumor and Peripheral Biomarkers in the Treatment of Advanced Renal Cell Carcinoma

The phase III JAVELIN Renal 101 trial demonstrated prolonged progression-free survival (PFS) in patients (N = 886) with advanced renal cell carcinoma treated with first-line avelumab + axitinib (A+Ax) versus sunitinib. We report novel findings from integrated analyses of longitudinal blood samples and baseline tumor tissue. PFS was associated with elevated lymphocyte levels in the sunitinib arm and an abundance of innate immune subsets in the A+Ax arm. Treatment with A+Ax led to greater T-cell repertoire modulation and less change in T-cell numbers versus sunitinib. In the A+Ax arm, patients with tumors harboring mutations in ≥2 of 10 previously identified PFS-associated genes (double mutants) had distinct circulating and tumor-infiltrating immunologic profiles versus those with wild-type or single-mutant tumors, suggesting a role for non-T-cell-mediated and non-natural killer cell-mediated mechanisms in double-mutant tumors. We provide evidence for different immunomodulatory mechanisms based on treatment (A+Ax vs. sunitinib) and tumor molecular subtypes.

SIGNIFICANCE

Our findings provide novel insights into the different immunomodulatory mechanisms governing responses in patients treated with avelumab (PD-L1 inhibitor) + axitinib or sunitinib (both VEGF inhibitors), highlighting the contribution of tumor biology to the complexity of the roles and interactions of infiltrating immune cells in response to these treatment regimens. This article is featured in Selected Articles from This Issue, p. 384.

HLA-A*03 and response to immune checkpoint blockade in cancer: an epidemiological biomarker study

BACKGROUND

Predictive biomarkers could allow more precise use of immune checkpoint inhibitors (ICIs) in treating advanced cancers. Given the central role of HLA molecules in immunity, variation at the HLA loci could differentially affect the response to ICIs. The aim of this epidemiological study was to determine the effect of HLA-A*03 as a biomarker for predicting response to immunotherapy.

METHODS

In this epidemiological study, we investigated the clinical outcomes (overall survival, progression free survival, and objective response rate) after treatment for advanced cancer in eight cohorts of patients: three observational cohorts of patients with various types of advanced tumours (the Memorial Sloan Kettering Integrated Mutation Profiling of Actionable Cancer Targets [MSK-IMPACT] cohort, the Dana-Farber Cancer Institute [DFCI] Profile cohort, and The Cancer Genome Atlas) and five clinical trials of patients with advanced bladder cancer (JAVELIN Solid Tumour) or renal cell carcinoma (CheckMate-009, CheckMate-010, CheckMate-025, and JAVELIN Renal 101). In total, these cohorts included 3335 patients treated with various ICI agents (anti-PD-1, anti-PD-L1, and anti-CTLA-4 inhibitors) and 10 917 patients treated with non-ICI cancer-directed therapeutic approaches. We initially modelled the association of HLA amino-acid variation with overall survival in the MSK-IMPACT discovery cohort, followed by a detailed analysis of the association between HLA-A*03 and clinical outcomes in MSK-IMPACT, with replication in the additional cohorts (two further observational cohorts and five clinical trials).

FINDINGS

HLA-A*03 was associated in an additive manner with reduced overall survival after ICI treatment in the MSK-IMPACT cohort (HR 1·48 per HLA-A*03 allele [95% CI 1·20-1·82], p=0·00022), the validation DFCI Profile cohort (HR 1·22 per HLA-A*03 allele, 1·05-1·42; p=0·0097), and in the JAVELIN Solid Tumour clinical trial for bladder cancer (HR 1·36 per HLA-A*03 allele, 1·01-1·85; p=0·047). The HLA-A*03 effect was observed across ICI agents and tumour types, but not in patients treated with alternative therapies. Patients with HLA-A*03 had shorter progression-free survival in the pooled patient population from the three CheckMate clinical trials of nivolumab for renal cell carcinoma (HR 1·31, 1·01-1·71; p=0·044), but not in those receiving control (everolimus) therapies. Objective responses were observed in none of eight HLA-A*03 homozygotes in the ICI group (compared with 59 [26·6%] of 222 HLA-A*03 non-carriers and 13 (17·1%) of 76 HLA-A*03 heterozygotes). HLA-A*03 was associated with shorter progression-free survival in patients receiving ICI in the JAVELIN Renal 101 randomised clinical trial for renal cell carcinoma (avelumab plus axitinib; HR 1·59 per HLA-A*03 allele, 1·16-2·16; p=0·0036), but not in those receiving control (sunitinib) therapy. Objective responses were recorded in one (12·5%) of eight HLA-A*03 homozygotes in the ICI group (compared with 162 [63·8%] of 254 HLA-A*03 non-carriers and 40 [55·6%] of 72 HLA-A*03 heterozygotes). HLA-A*03 was associated with impaired outcome in meta-analysis of all 3335 patients treated with ICI at genome-wide significance (p=2·01 × 10-8) with no evidence of heterogeneity in effect (I2 0%, 95% CI 0-0·76) INTERPRETATION: HLA-A*03 is a predictive biomarker of poor response to ICI. Further evaluation of HLA-A*03 is warranted in randomised trials. HLA-A*03 carriage could be considered in decisions to initiate ICI in patients with cancer.

FUNDING

National Institutes of Health, Merck KGaA, and Pfizer.

Avelumab maintenance in advanced urothelial carcinoma: biomarker analysis of the phase 3 JAVELIN Bladder 100 trial

In a recent phase 3 randomized trial of 700 patients with advanced urothelial cancer (JAVELIN Bladder 100; NCT02603432 ), avelumab/best supportive care (BSC) significantly prolonged overall survival relative to BSC alone as maintenance therapy after first-line chemotherapy. Exploratory biomarker analyses were performed to identify biological pathways that might affect survival benefit. Tumor molecular profiling by immunohistochemistry, whole-exome sequencing and whole-transcriptome sequencing revealed that avelumab survival benefit was positively associated with PD-L1 expression by tumor cells, tumor mutational burden, APOBEC mutation signatures, expression of genes underlying innate and adaptive immune activity and the number of alleles encoding high-affinity variants of activating Fcγ receptors. Pathways connected to tissue growth and angiogenesis might have been associated with reduced survival benefit. Individual biomarkers did not comprehensively identify patients who could benefit from therapy; however, multi-parameter models incorporating genomic alteration, immune responses and tumor growth showed promising predictive utility. These results characterize the complex biologic pathways underlying survival benefit from immune checkpoint inhibition in advanced urothelial cancer and suggest that multiple biomarkers might be needed to identify patients who would benefit from treatment.

Expanding control of the tumor cell cycle with a CDK2/4/6 inhibitor

The CDK4/6 inhibitor, palbociclib (PAL), significantly improves progression-free survival in HR⁺/HER2^- breast cancer when combined with anti-hormonals. We sought to discover PAL resistance mechanisms in preclinical models and through analysis of clinical transcriptome specimens, which coalesced on induction of MYC oncogene and Cyclin E/CDK2 activity. We propose that targeting the G₁ kinases CDK2, CDK4, and CDK6 with a small-molecule overcomes resistance to CDK4/6 inhibition. We describe the pharmacodynamics and efficacy of PF-06873600 (PF3600), a pyridopyrimidine with potent inhibition of CDK2/4/6 activity and efficacy in multiple in vivo tumor models. Together with the clinical analysis, MYC activity predicts (PF3600) efficacy across multiple cell lineages. Finally, we find that CDK2/4/6 inhibition does not compromise tumor-specific immune checkpoint blockade responses in syngeneic models. We anticipate that (PF3600), currently in phase 1 clinical trials, offers a therapeutic option to cancer patients in whom CDK4/6 inhibition is insufficient to alter disease progression.

Molecular medicine tumor board: whole-genome sequencing to inform on personalized medicine for a man with advanced prostate cancer

PURPOSE

Molecular profiling of cancer is increasingly common as part of routine care in oncology, and germline and somatic profiling may provide insights and actionable targets for men with metastatic prostate cancer. However, all reported cases are of deidentified individuals without full medical and genomic data available in the public domain.

PATIENT AND METHODS

We present a case of whole-genome tumor and germline sequencing in a patient with advanced prostate cancer, who has agreed to make his genomic and clinical data publicly available.

RESULTS

We describe an 84-year-old Caucasian male with a Gleason 10 oligometastastic hormone-sensitive prostate cancer. Whole-genome sequencing provided insights into his tumor's underlying mutational processes and the development of an SPOP mutation. It also revealed an androgen-receptor dependency of his cancer which was reflected in his durable response to radiation and hormonal therapy. Potentially actionable genomic lesions in the tumor were identified through a personalized medicine approach for potential future therapy, but at the moment, he remains in remission, illustrating the hormonal sensitivity of his SPOP-driven prostate cancer. We also placed this patient in the context of a large prostate-cancer cohort from the PCAWG (Pan-cancer Analysis of Whole Genomes) group. In this comparison, the patient's cancer appears typical in terms of the number and type of somatic mutations, but it has a somewhat larger contribution from the mutational process associated with aging.

CONCLUSION

We combined the expertise of medical oncology and genomics approaches to develop a molecular tumor board to integrate the care and study of this patient, who continues to have an outstanding response to his combined modality treatment. This identifiable case potentially helps overcome barriers to clinical and genomic data sharing.

Efficacy and correlative analyses of avelumab plus axitinib versus sunitinib in sarcomatoid renal cell carcinoma: post hoc analysis of a randomized clinical trial

BACKGROUND

Among patients with advanced renal cell carcinoma (RCC), those with sarcomatoid histology (sRCC) have the poorest prognosis. This analysis assessed the efficacy of avelumab plus axitinib versus sunitinib in patients with treatment-naive advanced sRCC.

METHODS

The randomized, open-label, multicenter, phase III JAVELIN Renal 101 trial (NCT02684006) enrolled patients with treatment-naive advanced RCC. Patients were randomized 1 : 1 to receive either avelumab plus axitinib or sunitinib following standard doses and schedules. Assessments in this post hoc analysis of patients with sRCC included efficacy (including progression-free survival) and biomarker analyses.

RESULTS

A total of 108 patients had sarcomatoid histology and were included in this post hoc analysis; 47 patients in the avelumab plus axitinib arm and 61 in the sunitinib arm. Patients in the avelumab plus axitinib arm had improved progression-free survival [stratified hazard ratio, 0.57 (95% confidence interval, 0.325-1.003)] and a higher objective response rate (46.8% versus 21.3%; complete response in 4.3% versus 0%) versus those in the sunitinib arm. Correlative gene expression analyses of patients with sRCC showed enrichment of gene pathway scores for cancer-associated fibroblasts and regulatory T cells, CD274 and CD8A expression, and tumors with The Cancer Genome Atlas m3 classification.

CONCLUSIONS

In this subgroup analysis of JAVELIN Renal 101, patients with sRCC in the avelumab plus axitinib arm had improved efficacy outcomes versus those in the sunitinib arm. Correlative analyses provide insight into this subtype of RCC and suggest that avelumab plus axitinib may increase the chance of overcoming the aggressive features of sRCC.

Integrating peripheral biomarker analyses from JAVELIN Renal 101: Avelumab + axitinib (A + Ax) versus sunitinib (S) in advanced renal cell carcinoma (aRCC)

4547

Background: In the phase 3 JAVELIN Renal 101 trial (NCT02684006), treatment-naive patients (pts) with aRCC demonstrated prolonged progression-free survival (PFS) and a higher objective response rate with A + Ax vs S. We report the association of blood-based biomarkers with differential responses to treatment. Methods: Biomarkers in pretreatment (pre-tx) and on-treatment (on-tx) blood samples from 886 enrolled pts were correlated with clinical outcomes and molecular profiling data from corresponding tumor samples. Analyses include blood counts of unique populations, T-cell receptor sequencing, circulating cytokines, and serum proteomics by mass spectrometry MALDI-TOF. Results: At baseline, higher pre-tx monocyte counts were associated with shorter PFS in the A + Ax arm (Table). In the S arm, higher pre-tx levels of multiple T-cell–related metrics, including the percent of productively rearranged peripheral T cells, were associated with longer PFS but had no association in the A + Ax arm (Table). Higher pre-tx neutrophil counts were associated with shorter PFS in both arms, but neutrophil-to-lymphocyte ratio (NLR) was only associated with PFS for the S arm (Table). On-therapy biomarkers showed differential post-tx changes in T-cell numbers and clones at C2D1. Tx-specific differences were also seen in non–T-cell populations such as monocytes and neutrophils at multiple time points through C3D1. Serum levels of pre- and on-tx VEGF, CRP, and several interleukins showed differential associations with PFS (eg, higher pre-tx VEGF was associated with shorter PFS in only the S arm) (Table). Specific genomic alterations in tumor tissues were associated with differences in several pre- and on-tx angiokines & cytokines. Conclusions: Response to treatment with first-line A + Ax or S was associated with immune fitness and tx-specific immunomodulation. We identified peripheral biomarkers in pts with aRCC associated with the presence of impactful genomic alterations and differential clinical outcomes. Clinical trial information: NCT02684006. [Table: see text]

Standard machine learning approaches outperform deep representation learning on phenotype prediction from transcriptomics data

BACKGROUND

The ability to confidently predict health outcomes from gene expression would catalyze a revolution in molecular diagnostics. Yet, the goal of developing actionable, robust, and reproducible predictive signatures of phenotypes such as clinical outcome has not been attained in almost any disease area. Here, we report a comprehensive analysis spanning prediction tasks from ulcerative colitis, atopic dermatitis, diabetes, to many cancer subtypes for a total of 24 binary and multiclass prediction problems and 26 survival analysis tasks. We systematically investigate the influence of gene subsets, normalization methods and prediction algorithms. Crucially, we also explore the novel use of deep representation learning methods on large transcriptomics compendia, such as GTEx and TCGA, to boost the performance of state-of-the-art methods. The resources and findings in this work should serve as both an up-to-date reference on attainable performance, and as a benchmarking resource for further research.

RESULTS

Approaches that combine large numbers of genes outperformed single gene methods consistently and with a significant margin, but neither unsupervised nor semi-supervised representation learning techniques yielded consistent improvements in out-of-sample performance across datasets. Our findings suggest that using l2-regularized regression methods applied to centered log-ratio transformed transcript abundances provide the best predictive analyses overall.

CONCLUSIONS

Transcriptomics-based phenotype prediction benefits from proper normalization techniques and state-of-the-art regularized regression approaches. In our view, breakthrough performance is likely contingent on factors which are independent of normalization and general modeling techniques; these factors might include reduction of systematic errors in sequencing data, incorporation of other data types such as single-cell sequencing and proteomics, and improved use of prior knowledge.

Molecular characteristics of renal cell carcinoma (RCC) risk groups from JAVELIN Renal 101

744

Background: The phase 3 JAVELIN Renal 101 trial in patients (pts) with advanced RCC demonstrated a progression-free survival (PFS) benefit and higher objective response rate (ORR) with avelumab + axitinib (A+Ax) vs sunitinib (S) (Motzer NEJM 2019). PFS and ORR favored A+Ax in all MSKCC risk groups, but median PFS varied. Here, we report results from analyses of baseline tumor samples to define molecular characteristics underlying risk group classifications. Methods: Nephrectomy or tumor samples from pts enrolled in the study were characterized by immunohistochemistry (CD8 and PD-L1), whole exome sequencing, or gene expression profiling (n = 705–850). Gene expression signatures (GES), pathway activation status, and mutational profiles were examined in relation to MSKCC risk groups. Results: Of the 886 total pts enrolled, 23%, 66%, and 11% had favorable (F), intermediate (I), or poor (P) MSKCC risk factors at baseline. In the F, I, and P groups, the ORR was 66%/38%, 50%/24%, and 31%/9%; median PFS was NR/16.7 mo, 13.3 mo/7.9 mo, and 5.6 mo/2.8 mo for the A+Ax/S arms, respectively. Neither the presence of PD-L1+ immune cells nor CD8+ cells differentiated the subgroups; however, the presence of PD-L1+ tumor cells was highest in the P group (p=0.0159). When compared to the I and P groups, the F group was enriched for NOTCH2 mutations (p=0.0002), displayed high FLT1 expression (p=0.007), and showed a trend favoring angiogenesis GES (JAVELIN Renal 101 and IMmotion150). The I group displayed few distinguishing characteristics (low neutrophil GES [p=0.02] and elevated homeobox gene expression [p=0.00052]). Relative to the F group, the P group showed higher cell cycle gene expression (p=0.0057) and PTEN mutation frequency, wild-type NOTCH2 genotype, elevated IMmotion150 Myeloid inflamed GES (p=0.0024), and macrophage-specific GES (p=0.0327), as well as GES for TH2 (p=0.0002), hypoxia (p=0.0199), glycolysis (p=0.0066), and the lowest expression of a dendritic cell GES (p=0.0209). Conclusions: In advanced RCC patients, biological differences within each MSKCC risk category may impact response to treatment and may help explain why these groups perform differently. Clinical trial information: NCT02684006.

Causal modeling of TCGA, NSCLC, and HNSCC data to identify network drivers of tumor immune subtypes

68

Background: Immune checkpoint inhibitors have achieved unprecedented success in several cancer types, yet only a subset of patients derives clinical benefit. Better understanding of tumor-immune interactions is imperative to improve clinical outcomes. Bayesian causal machine learning was applied on real world data to elucidate the molecular drivers of immune subtypes of tumors. Methods: Using a Reverse Engineering Forward Simulation (REFS) platform, ensembles of causal models were built on genomics, transcriptomics, and clinical data from 681 non-small cell lung carcinoma (NSCLC), 328 lung adeno (LUAD) and 353 lung squamous cell (LUSC), and 413 head and neck squamous cell carcinoma (HNSCC) patients from TCGA. The outcomes modeled were six tumor immune subtypes (Thorsson et al., 2018): wound-healing, IFNγ dominant, inflammatory, lymphocyte-depleted, immunologically quiet, and TGFβ dominant. Causal drivers of immune subtypes were identified from average causal effect (ACE) of the variables, as computed from the counterfactual simulations. ACE was defined as median of posterior distribution of odds ratio (1 vs 0 for discrete; 95th vs. 5th %ile for continuous variables). Results: The models showed impressive k-fold cross validation predictive performance (AUC ~ 0.8-0.9) for the most prevalent immune subtypes in TCGA: wound-healing and IFNg dominant in both LUAD and HNSCC, as well as inflammatory in LUAD. The potential causal drivers of the tumor immune subtypes and their ACE are listed in Table. The findings suggest that macrophage activation and polarization, which is driven in part by metabolic reprogramming, is a primary driver of tumor immune subtypes. Conclusions: Bayesian causal modeling revealed literature-supported hypotheses regarding predictors of response (CXCL13) and resistance (STK11 mutation, mTOR pathway) to PD-(L)1 blockade therapy. The additional target pathways such as AKT1/mTOR may be actionable for altering immunogenicity of tumors. [Table: see text]

Abstract 3783: Effective MEK inhibitor combinations to target tumor heterogeneity in EGFR mutant transformed SCLC

Background: Epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs) are effective for the treatment of EGFR-mutant non-small lung cancer (NSCLC). In almost all cases, however, the response duration is limited and resistance ultimately emerges. Transformation of NSCLC to small cell lung cancer (SCLC) occurs as one of the resistance mechanisms to EGFR-TKIs. To date, the mechanism associated with this transformation is largely unknown, necessitating the need to establish an effective treatment strategy for patients with transformed SCLC (tSCLC).

Methods: We isolated cancer cells that grow as floating aggregates (F) and adherent monolayers (Ad) from the pleural effusion of three patients with tSCLC to establish patient derived cell lines (DFCI112F, DFI112ad, DFCI190F, DFCI190ad, DFCI283F, DFCI283ad). In parallel, patient derived xenograft (PDX) models from these patients were successfully established by subcutaneous implantation of pleural effusions in NSG mice. Transcriptomic and genomic characterization of the cell lines were performed using whole-exome sequencing (WES) and RNA sequencing. These cell lines were further treated with a panel of agents, to screen for effective combination therapies.

Results: Biological characterization of the created cell lines revealed that the floating aggregates have neuroendocrine (NE) features, while the adherent cell population presents non-NE/mesenchymal characteristics. The histology of the PDX tumors was confirmed as SCLC. According to WES, all tSCLC cell lines retained the original activating EGFR mutations. RB1 and TP53 were universally inactivated. Co-culturing the floating-NE cells with the adherent non-NE cells promoted the proliferation of the floating-NE cells in vitro. NE cells (DFCI112F) and non-NE cells (DFCI112ad) were injected into nude mice either as single or admixed populations. Tumor growth was observed only in admixed cell populations. In an effort to target both NE and non-NE populations, we performed a drug screen and demonstrated that MEK inhibitors are selectively efficacious against non-NE populations. We further evaluated the MEK inhibitor combination therapies with agents targeting NE populations (cisplatin, ABT-263 and JQ1), which effectively eliminated both NE and non-NE components in vitro. Our further study demonstrated that ABT-263 in combination with MEK inhibitors presented a synergistic effect, which could be a promising therapeutic combination for tSCLC patients.

Conclusion: We demonstrate that the cancer cell lines established from tSCLC patients consist of heterogeneous population of cells with NE and non-NE/mesenchymal characteristics. The crosstalk between these two cell populations is observed and our findings underscore the importance of targeting tumor heterogeneity as a promising treatment strategy for patients with tSCLC.

Citation Format: Atsuko Ogino, Xinmeng Jasmine Mu, Mika Lin, Antonio Calles, Stephen Wang, Man Xu, Lynette M. Scholl, Geoffrey R. Oxnard, Paul Kirschmeier, Pasi A. Jänne. Effective MEK inhibitor combinations to target tumor heterogeneity in EGFR mutant transformed SCLC [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 3783.

Biomarker analyses from JAVELIN Renal 101: Avelumab + axitinib (A+Ax) versus sunitinib (S) in advanced renal cell carcinoma (aRCC)

101

Background: The phase 3 JAVELIN Renal 101 trial in previously untreated patients (pts) with aRCC demonstrated a progression-free survival (PFS) benefit and higher objective response rate with A+Ax vs S (Motzer, ESMO 2018; LBA6_PR). Here, we report outcomes from biomarker analyses of baseline tumor samples. Methods: We correlated efficacy with the results of molecular analyses of tissue samples from all 886 pts enrolled in JAVELIN Renal 101. Nephrectomy or tumor samples were characterized by immunohistochemistry (CD8 and PD-L1), whole-exome sequencing (WES), and RNAseq. WES and RNAseq were used to examine somatic mutations and analyze relevant gene expression signatures (GES) in relation to clinical outcomes. GES analyses included published and de novo signatures: effector T cell (Teff), angiogenesis (angio),T cell-inflamed (Tinf), and a novel immune-related signature incorporating pathway indicators for T- and NK-cell activation and IFNγ signaling, among others. Results: PD-L1 expression (≥1% immune cells) was associated with the longest PFS in the A+Ax arm and the shortest in the S arm (HR, 0.63; 95% CI, 0.49, 0.81). Significant treatment arm–specific differences in PFS were observed relative to wildtype when mutations in genes such as CD1631L, PTEN, or DNMT1 were present. Tumor mutational burden did not distinguish pts with respect to PFS. High-angio GES was associated with significantly improved PFS in the S arm but did not differentiate PFS in the A+Ax arm. In the low-angio subset, A+Ax improved PFS vs S. Pts with high Teff and Tinf in the A+Ax arm had longer PFS vs the S arm. In the A+Ax arm, PFS was enhanced in patients with immune GES–positive tumors vs those in the negative group (HR, 0.63; 95% CI, 0.46, 0.86; 2-sided p = 0.004), as well as in an independent dataset (JAVELIN Renal 100; Choueiri, Lancet Oncol, 2018) (HR, 0.46; 95% CI, 0.20, 1.05; 2-sided p = 0.064). Updated efficacy, including overall survival, will be presented. Conclusions: These findings define molecular features that differentiate therapy-specific outcomes in first-line aRCC and may inform personalized therapy strategies for pts with aRCC. Funding: Pfizer and Merck KGaA. Clinical trial information: NCT02684006.

Abstract PR03: Genetic mechanisms of immune evasion in colorectal cancer

Background: Immune checkpoint blockade has shown activity in microsatellite-instability high (MSI-H) colorectal cancers (CRCs). However, despite a very high mutational and neoantigen load among virtually all MSI-H tumors, the response rate is around 40-50%. In addition, for the majority of CRCs, which are microsatellite stable (MSS), immune checkpoint inhibitors have so far proven ineffective. Thus, to better understand the genetic drivers of immune evasion in CRC, we integrated next generation sequencing data from over 1200 tumors with transcriptional and immunohistochemical measures of immune infiltration.

Methods: We molecularly characterized 1,211 colorectal cancers, including 592 tumors from The Cancer Genome Atlas with whole exome sequencing (WES) and whole transcriptome (RNAseq) data and 619 cancers from two prospective cohort studies with WES and immunohistochemical (IHC) annotations. To identify driver gene alterations and their selection pressures specific to MSI-high tumors, we developed a statistical method to identify significantly mutated microsatellite tracts and we further developed a method to identify copy neutral loss of heterozygosity (CN-LOH) events. We used an established immune-gene transcriptional signature as well as IHC stains against specific subsets of immune-infiltrating cells to identify genetic events associated with immune evasion.

Results: We demonstrated that WNT-signaling and immune-related genes are significantly mutated in colorectal cancer. In MSI-H CRCs, we found biallelic antigen-presentation machinery (APM) mutations in the HLA Class I genes, B2M and TAP2, in addition to recurrent mutations in NLRC5 and RFX5, which downregulated HLA Class I expression. In all CRCs, we showed that WNT-signaling activity and APC-biallelic mutations are inversely associated with both transcriptional and IHC measures of T-cell infiltration. Specifically, nuclear CTNNB1 expression was inversely correlated with overall tumor-infiltrating lymphocytes (p = 0.027), CD8+ subset of T-cells (p = 0.0019) and CD45RO+ subset of T-cells (0.0080). Meanwhile, colorectal tumors with biallelic disruptive mutations in APC had a significantly decreased activated T-cell transcriptional signature (p = 4e-12) relative to samples with no disruptive mutations in APC. We further showed that in MSS tumors, AXIN2 (a key WNT-signaling effector) super-enhancer hypomethylation, independent of the APC mutation status, was further associated with decreased T-cell activity.

Conclusions: In this largest CRC genomic analysis to date, we identify genetic events that are associated with immune evasion in this disease. Specifically, we find evidence of immuno-editing in MSI-H tumors through disruptive mutations in APM. In the MSS and MSI-H subtype of CRCs, we use transcriptional and immunohistochemical orthogonal analyses to demonstrate exclusion of an effective immune infiltrate in CRC through an active WNT-signaling pathway. Our results shed light on the underlying molecular mechanisms of immune exclusion in CRC and have direct implications for novel combination immunotherapy trials for patients with this disease.

This abstract is also being presented as Poster B23.

Citation Format: Marios Giannakis, Catherine Grasso, Daniel Wells, Tsuyoshi Hamada, Xinmeng Jasmine Mu, Michael Quist, Jonathan Nowak, Reiko Nishihara, Charles M. Connolly, Sachet Shukla, William M. Grady, David Wheeler, Catherine J. Wu, Jesse Zaretsky, Levi Garraway, Thomas Hudson, Charles Fuchs, Antoni Ribas, Riki Peters, Shuji Ogino. Genetic mechanisms of immune evasion in colorectal cancer [abstract]. In: Proceedings of the AACR Special Conference on Tumor Immunology and Immunotherapy; 2017 Oct 1-4; Boston, MA. Philadelphia (PA): AACR; Cancer Immunol Res 2018;6(9 Suppl):Abstract nr PR03.

Abstract 1224: Deep targeted tumor sequencing of colorectal cancer cases to study associations of molecular subtypes with clinical, genetic, and lifestyle risk factors

Colorectal cancer (CRC), a common malignancy, is a biologically heterogeneous disease. Next-generation sequencing (NGS) has enabled CRC characterization by identifying somatically mutated genes which now allow us to better define colorectal tumor subtypes (e.g. by mutated pathways). However, the relationship of such CRC subtypes to patient survival and genetic and lifestyle risk factors has not been comprehensively studied.

To identify somatic mutations in CRC cases, we designed a targeted AmpliSeq panel of CRC related genes and genomic regions informed by whole exome sequencing data from ~1,200 CRC cases. The sequencing was conducted on Illumina HiSeq 2500 with a mean coverage of 740x and 240x for DNA extracted from FFPE tumor tissues and matched normal samples, respectively. Strelka, MuTect, VarDict, and Varscan2 were used to identify somatic single nucleotide variants and indels. Sanger sequencing was performed to validate a subset of variants. To date, we have sequenced ~2,400 CRC tumors and matched control tissues from four studies participating in the Colon Cancer Family Registry (CCFR) and Genetics and Epidemiology of Colorectal Cancer Consortium (GECCO). In most tumors, we identified non-silent mutations in genes belonging to the WNT (77%), p53 (44%), IGF2/PI3K (22%), RTK-RAS (47%), and TGF-beta (26%) signaling pathways. Among the 15% of tumors that could be classified as hypermutated, based on the number of mutations, 39% exhibited non-silent mutations in MLH1, MLH3, MSH2, MSH6, and PMS2 and 41% exhibited non-silent mutations in POLE and POLD1.

In a subset of studies with available survival data, we used Cox regression to assess the association of hypermutation status and the presence of non-silencing mutations in key signaling pathways with overall (OS) and disease-specific (DSS) survival. OS and DSS were significantly more favorable in cases with hypermutated vs. non-hypermutated CRC (HR=0.77, 95% CI: 0.60-0.98, p=0.04 and HR=0.35, 95% CI: 0.22-0.57, p=2x10-5, respectively); these associations were most pronounced for POLE/POLD1 mutated hypermutated CRC (HR=0.69, 95% CI: 0.46-1.02, p=0.06, HR=0.21, 95% CI: 0.08-0.56, p=2x10-3, respectively). There was no significant association of mutations in WNT, p53, IGF2/PI3K, RTK-RAS, or TGF-beta pathways with survival (p>0.05).

The comprehensive molecular characterization of this large panel of CRC cases will support further studies of molecular subtypes of CRC with clinical, lifestyle, and environmental factors. A better understanding of molecular mechanisms of CRC will be valuable in developing strategies for prevention, diagnosis, and treatment of this life-threatening disease.

Citation Format: Syed Zaidi, Amanda Phipps, Tabitha Harrison, Catherine Grasso, Robert Steinfelder, Quang Trinh, Charles Connolly, Barbara Banbury, Adilya Rafikova, Philipp Hofer, Stefanie Brezina, Marios Giannakis, Xinmeng Jasmine Mu, Michael Quist, Charles Fuchs, Levi Garraway, Li Hsu, Lincoln Stein, Andrea Gsur, Shuji Ogino, Steven Gallinger, Polly Newcomb, Peter Campbell, Wei Sun, Thomas Hudson, Ulrike Peters. Deep targeted tumor sequencing of colorectal cancer cases to study associations of molecular subtypes with clinical, genetic, and lifestyle risk factors [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 1224.

TIME (Tumor Immunity in the MicroEnvironment) classification based on tumor CD274 (PD-L1) expression status and tumor-infiltrating lymphocytes in colorectal carcinomas

Inhibitors targeting the PDCD1 (programmed cell death 1, PD-1) immune checkpoint pathway have revolutionized cancer treatment strategies. The TIME (Tumor Immunity in the MicroEnvironment) classification based on tumor CD274 (PDCD1 ligand 1, PD-L1) expression and tumor-infiltrating lymphocytes (TIL) has been proposed to predict response to immunotherapy. It remains to be determined clinical, pathological, and molecular features of TIME subtypes of colorectal cancer. Using 812 colon and rectal carcinoma cases from the Nurses' Health Study and Health Professionals Follow-up Study, we examined the association of tumor characteristics and survival outcomes with four TIME subtypes (TIME 1, CD274^low/TIL^absent; TIME 2, CD274^high/TIL^present; TIME 3, CD274^low/TIL^present; and TIME 4, CD274^high/TIL^absent). In survival analyses, Cox proportional hazards models were adjusted for potential confounders, including microsatellite instability (MSI) status, CpG island methylator phenotype (CIMP) status, LINE-1 methylation level, and KRAS, BRAF, and PIK3CA mutation status. TIME subtypes 1, 2, 3 and 4 had 218 (27%), 117 (14%), 103 (13%), and 374 (46%) colorectal cancer cases, respectively. Compared with TIL-absent subtypes (TIME 1 and 4), TIL-present subtypes (TIME 2 and 3) were associated with high-level MSI, high-degree CIMP, BRAF mutation, and higher amounts of neoantigens (p < 0.001). TIME subtypes were not significantly associated with colorectal cancer-specific or overall survival. In conclusion, TIL-present TIME subtypes of colorectal cancer are associated with high levels of MSI and neoantigen load, supporting better responsiveness to cancer immunotherapy. Further studies examining tumor molecular alterations and additional factors in the tumor microenvironment may inform development of immunoprevention and immunotherapy strategies.

Genetic Mechanisms of Immune Evasion in Colorectal Cancer

To understand the genetic drivers of immune recognition and evasion in colorectal cancer, we analyzed 1,211 colorectal cancer primary tumor samples, including 179 classified as microsatellite instability-high (MSI-high). This set includes The Cancer Genome Atlas colorectal cancer cohort of 592 samples, completed and analyzed here. MSI-high, a hypermutated, immunogenic subtype of colorectal cancer, had a high rate of significantly mutated genes in important immune-modulating pathways and in the antigen presentation machinery, including biallelic losses of B2M and HLA genes due to copy-number alterations and copy-neutral loss of heterozygosity. WNT/β-catenin signaling genes were significantly mutated in all colorectal cancer subtypes, and activated WNT/β-catenin signaling was correlated with the absence of T-cell infiltration. This large-scale genomic analysis of colorectal cancer demonstrates that MSI-high cases frequently undergo an immunoediting process that provides them with genetic events allowing immune escape despite high mutational load and frequent lymphocytic infiltration and, furthermore, that colorectal cancer tumors have genetic and methylation events associated with activated WNT signaling and T-cell exclusion.Significance: This multi-omic analysis of 1,211 colorectal cancer primary tumors reveals that it should be possible to better monitor resistance in the 15% of cases that respond to immune blockade therapy and also to use WNT signaling inhibitors to reverse immune exclusion in the 85% of cases that currently do not. Cancer Discov; 8(6); 730-49. ©2018 AACR.This article is highlighted in the In This Issue feature, p. 663.

Effect of tissue frozen on quantitative optical properties using optical coherence tomography

The purpose is to demonstrate the optical charactering concerning nasopharyngeal tissue of pig by fresh sections and frozen correlating sections with optical coherence tomography (OCT). After being imaged on a fresh specimen, samples are then stored in low temperature refrigerators (-80°C) for one year for the second OCT measurement. The OCT structure of the epithelium, lamina propria, and the basement membrane are still resolvable; the median scattering coefficients and anisotropy factors fitting from OCT images based on the multiple scattering effects for epithelium are 27.6  mm^-1 [interquartile range (IQR) 23.6 to 29.3  mm^-1] versus 22.5  mm^-1 (IQR 20.5 to 24.4  mm^-1), 0.86 (IQR 0.81 to 0.9) versus 0.88 (IQR 0.87 to 0.9) for fresh and frozen tissue, respectively; and 10.2  mm^-1 (IQR 8.1 to 13.6  mm^-1) versus 9.6  mm^-1 (IQR 8.1 to 13.8  mm^-1), 0.96 (IQR 0.93 to 0.98) versus 0.92 (IQR 0.9 to 0.98) for lamina propria, respectively. The results show that the frozen storage method can be used for OCT research.

Abstract 1564: A computational framework for removing mouse contamination in tumors sequenced from patient-derived xenografts

Introduction: Using patient derived xenograft (PDX) models has become an effective way for investigating response to standard or new therapeutics in cancer. Human cancer cells injected in mice are allowed to establish tumors and subjected to desired treatments. The PDX tumors are later harvested and characterized, often by massive parallel sequencing. However, a major challenge with analyzing these data is the presence of stromal mouse genomic material, frequently resulting in artifacts in downstream variant detection. We present a computational method to eliminate mouse contamination in PDX.

Method: We used the Burrows-Wheeler Aligner to map reads obtained from sequencing the PDX samples to a combined human and mouse reference genome . We remove reads that are mapped to the mouse reference. The remaining reads are then processed through standard mutation calling pipelines for somatic mutation detection. To test the efficacy of our method, we created in silico mixtures of human and mouse whole-exome sequencing reads from a melanoma patient’s tumor and an immortalized mouse cell line captured with human exome baits. We then carried out a sensitivity analysis to examine how changing the mean target coverage of sequencing, or mouse contamination levels affects our results. For each of the computational experiments, we evaluated somatic mutations detected from the synthetic samples in comparison to the original human sample.

Results: We calculated the sensitivity and specificity of detecting somatic mutations to determine our algorithm’s performance. In all instances, we found greater than 99% for both sensitivity and specificity.

Conclusions: Our results demonstrate that our method works accurately towards removing mouse reads in PDX samples. This task could also be applied to separating sequence reads from other species.

Citation Format: Ali Amin-Mansour, Judit Jané-Valbuena, Xinmeng Jasmine Mu, Levi Garraway. A computational framework for removing mouse contamination in tumors sequenced from patient-derived xenografts [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 1564. doi:10.1158/1538-7445.AM2017-1564

Clinical actionability of germline testing in patients with limited colorectal polyps

e13027

Background: As classic adenoma-carcinoma sequence is the main process underlying most colorectal cancer (CRC), early detection and removal of colorectal adenomas is crucial in preventing CRC. Although adenomatous polyps are usually sporadic, several inherited CRC syndromes such as Familial adenomatous polyposis, MUTYH-associated polyposis and Lynch syndrome can present initially with colon polyps. The identification of germline defects in patients with colon polyps is thus critical for proper cancer risk counseling and CRC prevention. Current guidelines recommend germline testing for patients with more than 20 polyps or those with more than 10 polyps and a family history of CRC. However, the diagnostic yield of germline testing on otherwise healthy individuals with 10 or fewer colon polyps has not been well studied. Here, we performed a pilot study to evaluate the clinical actionability of germline genetic testing on these patients. Methods: A total of 13 cancer-free adults, who presented with colon polyps (n < 10) and who otherwise were not selected based on age of onset or family history underwent germline Exome Sequencing. Variants in 13 well-established CRC risk genes ( APC, CHEK2, MYH, MLH1, MSH2, MSH6, PMS2, NTHL1, BMPR1A, SMAD4, PTEN, STK11, TP53) were evaluated for pathogenicity. Results: A total of 13 patients (12 male, 1 female) were evaluated. The median age of presentation was 69 (range 49 to 88). The median number of adenomatous colon polyps was 1 (range 1 to 8). Two (15.4%, 95% CI = 1.9 - 45.4, Binomial Exact) patients had at least one disruptive mutation in the examined genes. One of these patients had a truncating mutation in APC (p.Arg216*) and presented with two tubulovillous adenomas at age 49. The second patient had 8 adenomas in distal colon and rectum at age 64, and harbored a known pathogenic mutation in MSH6(p.Arg1035*). Conclusions: This pilot study provides evidence that a relatively high percentage of patients presenting with a few colon polyps may have inherited defects in highly actionable genes. If validated in larger cohorts with appropriate population controls, these findings may influence the clinical care of such patients and their families and suggest germline molecular testing in those patients.

Enrichment of germline DNA-repair gene mutations in patients with colorectal cancer

1500

Background: Twin studies showed that 30% of all colorectal cancer (CRC) patients have an inherited genetic susceptibility. Several CRC predisposition genes have been described to date. However, mutations in these genes explain the risk in only 5-10% of CRC cases. In this study, we hypothesized that some of the CRC heritability may be explained by excess disruptive germline mutations in DNA repair genes (DRGs). Methods: Exome sequencing data of 716 in the discovery cohort (CanSeq and NHS/HPFS studies) and 1609 CRC patients in the validation cohort (TCGA and NSCCG studies) were used to evaluate germline variants in a pre-selected group of 42 DRGs and 12 known CRC risk genes. Frequencies of disruptive mutations in our cohorts were examined relative to 27173 non-Finnish European cancer-free adults from the ExAC cohort to evaluate for enrichment. Results: Of 716 patients in the discovery cohort, 27 (3.8%) patients harbored germline mutations in APC (n = 11), MSH6 (n = 2), MUTYH (n = 11), CHEK2 (n = 1) and TP53 (n = 2). Interestingly, germline mutations in ATM and PALB2 were significantly enriched in our CRC discovery cohort (OR = 2.7; P = 0.044; and OR = 4.8; P = 0.026, respectively). Evaluation of germline data from another 1609 CRC patients (validation cohort) also showed significantly higher rates of ATM mutations (5; 0.7%; OR = 2.1; P = 0.044), and a trend for enrichment of PALB2 mutations (3; 0.4%; OR = 2.8; P = 0.056). Secondary analysis of actionable germline mutations in a highly penetrant cancer risk gene set ( ATM, BRCA1, BRCA2, BRIP1 and PALB2) suggest a broader enrichment trend in CRC patients for these genes (Discovery: OR = 1.7; P = 0.06; Validation: OR = 2; P = 1.96e-04). Conclusions: Our analysis of germline variants in 2325 CRC patients showed the first robust evidence for germline ATM mutations to confer a higher risk of developing CRC. We also presented evidence to support PALB2 as a potential novel CRC risk gene. Overall, our study shows that mutations in some DRGs may explain some of the missing CRC heritability. It also indicates that a significant percentage of CRC patients, who carry mutations in highly actionable genes where cancer screening recommendations for patients and families do exist, are not captured with current testing recommendations.

Clinical, Molecular, and Immune Analysis of Dabrafenib-Trametinib Combination Treatment for BRAF Inhibitor-Refractory Metastatic Melanoma: A Phase 2 Clinical Trial

IMPORTANCE

Combined treatment with dabrafenib and trametinib (CombiDT) achieves clinical responses in only about 15% of patients with BRAF inhibitor (BRAFi)-refractory metastatic melanoma in contrast to the higher response rate observed in BRAFi-naïve patients. Identifying correlates of response and mechanisms of resistance in this population will facilitate clinical management and rational therapeutic development.

OBJECTIVE

To determine correlates of benefit from CombiDT therapy in patients with BRAFi-refractory metastatic melanoma.

DESIGN, SETTING, AND PARTICIPANTS

Single-center, single-arm, open-label phase 2 trial of CombiDT treatment in patients with BRAF V600 metastatic melanoma resistant to BRAFi monotherapy conducted between September 2012 and October 2014 at the University of Texas MD Anderson Cancer Center. Key eligibility criteria for participants included BRAF V600 metastatic melanoma, prior BRAFi monotherapy, measurable disease (RECIST 1.1), and tumor accessible for biopsy.

INTERVENTIONS

Patients were treated with dabrafenib (150 mg, twice daily) and trametinib (2 mg/d) continuously until disease progression or intolerance. All participants underwent a mandatory baseline biopsy, and optional biopsy specimens were obtained on treatment and at disease progression. Whole-exome sequencing, reverse transcription polymerase chain reaction analysis for BRAF splicing, RNA sequencing, and immunohistochemical analysis were performed on tumor samples, and blood was analyzed for levels of circulating BRAF V600.

MAIN OUTCOMES AND MEASURES

The primary end point was overall response rate (ORR). Progression-free survival (PFS) and overall survival (OS) were secondary clinical end points.

RESULTS

A total of 28 patients were screened, and 23 enrolled. Among evaluable patients, the confirmed ORR was 10%; disease control rate (DCR) was 45%, and median PFS was 13 weeks. Clinical benefit was associated with duration of prior BRAFi therapy greater than 6 months (DCR, 73% vs 11% for ≤6 months; P = .02) and decrease in circulating BRAF V600 at day 8 of cycle 1 (DCR, 75% vs 18% for no decrease; P = .02) but not with pretreatment mitogen-activated protein kinase (MAPK) pathway mutations or activation. Biopsy specimens obtained during treatment demonstrated that CombiDT therapy failed to achieve significant MAPK pathway inhibition or immune infiltration in most patients.

CONCLUSIONS AND RELEVANCE

The baseline presence of MAPK pathway alterations was not associated with benefit from CombiDT in patients with BRAFi-refractory metastatic melanoma. Failure to inhibit the MAPK pathway provides a likely explanation for the limited clinical benefit of CombiDT in this setting. Circulating BRAF V600 is a promising early biomarker of clinical response.

TRIAL REGISTRATION

clinicaltrials.gov Identifier: NCT01619774.

Genomic Correlates of Immune-Cell Infiltrates in Colorectal Carcinoma

Large-scale genomic characterization of tumors from prospective cohort studies may yield new insights into cancer pathogenesis. We performed whole-exome sequencing of 619 incident colorectal cancers (CRCs) and integrated the results with tumor immunity, pathology, and survival data. We identified recurrently mutated genes in CRC, such as BCL9L, RBM10, CTCF, and KLF5, that were not previously appreciated in this disease. Furthermore, we investigated the genomic correlates of immune-cell infiltration and found that higher neoantigen load was positively associated with overall lymphocytic infiltration, tumor-infiltrating lymphocytes (TILs), memory T cells, and CRC-specific survival. The association with TILs was evident even within microsatellite-stable tumors. We also found positive selection of mutations in HLA genes and other components of the antigen-processing machinery in TIL-rich tumors. These results may inform immunotherapeutic approaches in CRC. More generally, this study demonstrates a framework for future integrative molecular epidemiology research in colorectal and other malignancies.

FunSeq2: a framework for prioritizing noncoding regulatory variants in cancer

Identification of noncoding drivers from thousands of somatic alterations in a typical tumor is a difficult and unsolved problem. We report a computational framework, FunSeq2, to annotate and prioritize these mutations. The framework combines an adjustable data context integrating large-scale genomics and cancer resources with a streamlined variant-prioritization pipeline. The pipeline has a weighted scoring system combining: inter- and intra-species conservation; loss- and gain-of-function events for transcription-factor binding; enhancer-gene linkages and network centrality; and per-element recurrence across samples. We further highlight putative drivers with information specific to a particular sample, such as differential expression. FunSeq2 is available from funseq2.gersteinlab.org.

An integrated map of structural variation in 2,504 human genomes

Structural variants are implicated in numerous diseases and make up the majority of varying nucleotides among human genomes. Here we describe an integrated set of eight structural variant classes comprising both balanced and unbalanced variants, which we constructed using short-read DNA sequencing data and statistically phased onto haplotype blocks in 26 human populations. Analysing this set, we identify numerous gene-intersecting structural variants exhibiting population stratification and describe naturally occurring homozygous gene knockouts that suggest the dispensability of a variety of human genes. We demonstrate that structural variants are enriched on haplotypes identified by genome-wide association studies and exhibit enrichment for expression quantitative trait loci. Additionally, we uncover appreciable levels of structural variant complexity at different scales, including genic loci subject to clusters of repeated rearrangement and complex structural variants with multiple breakpoints likely to have formed through individual mutational events. Our catalogue will enhance future studies into structural variant demography, functional impact and disease association.

Analysis of deletion breakpoints from 1,092 humans reveals details of mutation mechanisms

Investigating genomic structural variants at basepair resolution is crucial for understanding their formation mechanisms. We identify and analyze 8,943 deletion breakpoints in 1,092 samples from the 1000 Genomes Project. We find breakpoints have more nearby SNPs and indels than the genomic average, likely a consequence of relaxed selection. By investigating the correlation of breakpoints with DNA methylation, Hi-C interactions, and histone marks and the substitution patterns of nucleotides near them, we find that breakpoints with the signature of non-allelic homologous recombination (NAHR) are associated with open chromatin. We hypothesize that some NAHR deletions occur without DNA replication and cell division, in embryonic and germline cells. In contrast, breakpoints associated with non-homologous (NH) mechanisms often have sequence micro-insertions, templated from later replicating genomic sites, spaced at two characteristic distances from the breakpoint. These micro-insertions are consistent with template-switching events and suggest a particular spatiotemporal configuration for DNA during the events.

RNF43 is frequently mutated in colorectal and endometrial cancers

We report somatic mutations of RNF43 in over 18% of colorectal adenocarcinomas and endometrial carcinomas. RNF43 encodes an E3 ubiquitin ligase that negatively regulates Wnt signaling. Truncating mutations of RNF43 are more prevalent in microsatellite-unstable tumors and show mutual exclusivity with inactivating APC mutations in colorectal adenocarcinomas. These results indicate that RNF43 is one of the most commonly mutated genes in colorectal and endometrial cancers.

Abstract 2371: Heterogeneity of resistance mechanisms in lung adenocarcinoma patients with acquired resistance to EGFR inhibitors

Drug resistance is a major obstacle in overcoming cancer. Understanding the origin, extent and heterogeneity of resistance mechanisms within a patient is crucial to guide treatment and developing drugs/combinations to circumvent resistance. To this end, we studied autopsy samples collected from four EGFR-mutant lung adenocarcinoma patients who developed resistance to EGFR inhibitors. We used whole-exome sequencing to characterize genomic profiles of multiple metastasis tumors within each patient. To address the challenge of non-uniform variant detection power at sequenced genetic loci across multiple samples, we developed a computational framework that tackles variant detection from pooled tumor followed by a genotyping module. After the identification of somatic alterations, we determined the extent of heterogeneity of resistance mechanisms between multiple tumor samples. Our results revealed that multiple metastasis samples in a single patient can potentially use one common resistance mechanism. For example, one patient harbored a resistant EGFR T790M mutation in all 11 metastasis samples with a correlated allele frequency to the original L858R mutation. However, in some other cases, we found heterogeneous resistance mechanisms coexisting in multiple metastasis samples with various allele frequencies. For instance, one patient harbored both the resistance EGFR T790M mutation and an activating PIK3CA E542K mutation that appear with a range of allelic frequencies in multiple metastatic samples. Furthermore, branched evolutionary patterns of tumors were constructed using a phylogenetic approach, which inform resistance in the context of tumor evolution. Our findings are consistent with the model that cells harboring a variety of resistance mechanisms coexist in each metastasis tumor, and those subclones with higher selective advantage outgrow the others.

Citation Format: Xinmeng Jasmine Mu, Eli Van Allen, Nikhil Wagle, Curtis Chong, Mohit Butaney, Deborah Farlow, Gad Getz, Pasi A. Jänne, Levi A. Garraway. Heterogeneity of resistance mechanisms in lung adenocarcinoma patients with acquired resistance to EGFR inhibitors. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 2371. doi:10.1158/1538-7445.AM2014-2371

Integrative annotation of variants from 1092 humans: application to cancer genomics

Interpreting variants, especially noncoding ones, in the increasing number of personal genomes is challenging. We used patterns of polymorphisms in functionally annotated regions in 1092 humans to identify deleterious variants; then we experimentally validated candidates. We analyzed both coding and noncoding regions, with the former corroborating the latter. We found regions particularly sensitive to mutations ("ultrasensitive") and variants that are disruptive because of mechanistic effects on transcription-factor binding (that is, "motif-breakers"). We also found variants in regions with higher network centrality tend to be deleterious. Insertions and deletions followed a similar pattern to single-nucleotide variants, with some notable exceptions (e.g., certain deletions and enhancers). On the basis of these patterns, we developed a computational tool (FunSeq), whose application to ~90 cancer genomes reveals nearly a hundred candidate noncoding drivers.

The origin, evolution, and functional impact of short insertion-deletion variants identified in 179 human genomes

Short insertions and deletions (indels) are the second most abundant form of human genetic variation, but our understanding of their origins and functional effects lags behind that of other types of variants. Using population-scale sequencing, we have identified a high-quality set of 1.6 million indels from 179 individuals representing three diverse human populations. We show that rates of indel mutagenesis are highly heterogeneous, with 43%–48% of indels occurring in 4.03% of the genome, whereas in the remaining 96% their prevalence is 16 times lower than SNPs. Polymerase slippage can explain upwards of three-fourths of all indels, with the remainder being mostly simple deletions in complex sequence. However, insertions do occur and are significantly associated with pseudo-palindromic sequence features compatible with the fork stalling and template switching (FoSTeS) mechanism more commonly associated with large structural variations. We introduce a quantitative model of polymerase slippage, which enables us to identify indel-hypermutagenic protein-coding genes, some of which are associated with recurrent mutations leading to disease. Accounting for mutational rate heterogeneity due to sequence context, we find that indels across functional sequence are generally subject to stronger purifying selection than SNPs. We find that indel length modulates selection strength, and that indels affecting multiple functionally constrained nucleotides undergo stronger purifying selection. We further find that indels are enriched in associations with gene expression and find evidence for a contribution of nonsense-mediated decay. Finally, we show that indels can be integrated in existing genome-wide association studies (GWAS); although we do not find direct evidence that potentially causal protein-coding indels are enriched with associations to known disease-associated SNPs, our findings suggest that the causal variant underlying some of these associations may be indels.

The GENCODE pseudogene resource

Background

Pseudogenes have long been considered as nonfunctional genomic sequences. However, recent evidence suggests that many of them might have some form of biological activity, and the possibility of functionality has increased interest in their accurate annotation and integration with functional genomics data.

Results

As part of the GENCODE annotation of the human genome, we present the first genome-wide pseudogene assignment for protein-coding genes, based on both large-scale manual annotation and in silico pipelines. A key aspect of this coupled approach is that it allows us to identify pseudogenes in an unbiased fashion as well as untangle complex events through manual evaluation. We integrate the pseudogene annotations with the extensive ENCODE functional genomics information. In particular, we determine the expression level, transcription-factor and RNA polymerase II binding, and chromatin marks associated with each pseudogene. Based on their distribution, we develop simple statistical models for each type of activity, which we validate with large-scale RT-PCR-Seq experiments. Finally, we compare our pseudogenes with conservation and variation data from primate alignments and the 1000 Genomes project, producing lists of pseudogenes potentially under selection.

Conclusions

At one extreme, some pseudogenes possess conventional characteristics of functionality; these may represent genes that have recently died. On the other hand, we find interesting patterns of partial activity, which may suggest that dead genes are being resurrected as functioning non-coding RNAs. The activity data of each pseudogene are stored in an associated resource, psiDR, which will be useful for the initial identification of potentially functional pseudogenes.

A systematic survey of loss-of-function variants in human protein-coding genes

Genome-sequencing studies indicate that all humans carry many genetic variants predicted to cause loss of function (LoF) of protein-coding genes, suggesting unexpected redundancy in the human genome. Here we apply stringent filters to 2951 putative LoF variants obtained from 185 human genomes to determine their true prevalence and properties. We estimate that human genomes typically contain ~100 genuine LoF variants with ~20 genes completely inactivated. We identify rare and likely deleterious LoF alleles, including 26 known and 21 predicted severe disease-causing variants, as well as common LoF variants in nonessential genes. We describe functional and evolutionary differences between LoF-tolerant and recessive disease genes and a method for using these differences to prioritize candidate genes found in clinical sequencing studies.

Genomics and privacy: implications of the new reality of closed data for the field

Open source and open data have been driving forces in bioinformatics in the past. However, privacy concerns may soon change the landscape, limiting future access to important data sets, including personal genomics data. Here we survey this situation in some detail, describing, in particular, how the large scale of the data from personal genomic sequencing makes it especially hard to share data, exacerbating the privacy problem. We also go over various aspects of genomic privacy: first, there is basic identifiability of subjects having their genome sequenced. However, even for individuals who have consented to be identified, there is the prospect of very detailed future characterization of their genotype, which, unanticipated at the time of their consent, may be more personal and invasive than the release of their medical records. We go over various computational strategies for dealing with the issue of genomic privacy. One can “slice” and reformat datasets to allow them to be partially shared while securing the most private variants. This is particularly applicable to functional genomics information, which can be largely processed without variant information. For handling the most private data there are a number of legal and technological approaches—for example, modifying the informed consent procedure to acknowledge that privacy cannot be guaranteed, and/or employing a secure cloud computing environment. Cloud computing in particular may allow access to the data in a more controlled fashion than the current practice of downloading and computing on large datasets. Furthermore, it may be particularly advantageous for small labs, given that the burden of many privacy issues falls disproportionately on them in comparison to large corporations and genome centers. Finally, we discuss how education of future genetics researchers will be important, with curriculums emphasizing privacy and data security. However, teaching personal genomics with identifiable subjects in the university setting will, in turn, create additional privacy issues and social conundrums.

The real cost of sequencing: higher than you think!

Advances in sequencing technology have led to a sharp decrease in the cost of 'data generation'. But is this sufficient to ensure cost-effective and efficient 'knowledge generation'?

Analysis of genomic variation in non-coding elements using population-scale sequencing data from the 1000 Genomes Project

In the human genome, it has been estimated that considerably more sequence is under natural selection in non-coding regions [such as transcription-factor binding sites (TF-binding sites) and non-coding RNAs (ncRNAs)] compared to protein-coding ones. However, less attention has been paid to them. To study selective pressure on non-coding elements, we use next-generation sequencing data from the recently completed pilot phase of the 1000 Genomes Project, which, compared to traditional methods, allows for the characterization of a full spectrum of genomic variations, including single-nucleotide polymorphisms (SNPs), short insertions and deletions (indels) and structural variations (SVs). We develop a framework for combining these variation data with non-coding elements, calculating various population-based metrics to compare classes and subclasses of elements, and developing element-aware aggregation procedures to probe the internal structure of an element. Overall, we find that TF-binding sites and ncRNAs are less selectively constrained for SNPs than coding sequences (CDSs), but more constrained than a neutral reference. We also determine that the relative amounts of constraint for the three types of variations are, in general, correlated, but there are some differences: counter-intuitively, TF-binding sites and ncRNAs are more selectively constrained for indels than for SNPs, compared to CDSs. After inspecting the overall properties of a class of elements, we analyze selective pressure on subclasses within an element class, and show that the extent of selection is associated with the genomic properties of each subclass. We find, for instance, that ncRNAs with higher expression levels tend to be under stronger purifying selection, and the actual regions of TF-binding motifs are under stronger selective pressure than the corresponding peak regions. Further, we develop element-aware aggregation plots to analyze selective pressure across the linear structure of an element, with the confidence intervals evaluated using both simple bootstrapping and block bootstrapping techniques. We find, for example, that both micro-RNAs (particularly the seed regions) and their binding targets are under stronger selective pressure for SNPs than their immediate genomic surroundings. In addition, we demonstrate that substitutions in TF-binding motifs inversely correlate with site conservation, and SNPs unfavorable for motifs are under more selective constraints than favorable SNPs. Finally, to further investigate intra-element differences, we show that SVs have the tendency to use distinctive modes and mechanisms when they interact with genomic elements, such as enveloping whole gene(s) rather than disrupting them partially, as well as duplicating TF motifs in tandem.

PEMer: a computational framework with simulation-based error models for inferring genomic structural variants from massive paired-end sequencing data

Personal-genomics endeavors, such as the 1000 Genomes project, are generating maps of genomic structural variants by analyzing ends of massively sequenced genome fragments. To process these we developed Paired-End Mapper (PEMer; http://sv.gersteinlab.org/pemer). This comprises an analysis pipeline, compatible with several next-generation sequencing platforms; simulation-based error models, yielding confidence-values for each structural variant; and a back-end database. The simulations demonstrated high structural variant reconstruction efficiency for PEMer's coverage-adjusted multi-cutoff scoring-strategy and showed its relative insensitivity to base-calling errors.

Interleukin-29 functions cooperatively with interferon to induce antiviral gene expression and inhibit hepatitis C virus replication

The interferon (IFN)-related cytokine interleukin (IL)-29 (also known as IFN-lambda1) inhibits virus replication by inducing a cellular antiviral response similar to that activated by IFN-alpha/beta. However, because it binds to a unique receptor, this cytokine may function cooperatively with IFN-alpha/beta or IFN-gamma during natural infections to inhibit virus replication, and might also be useful therapeutically in combination with other cytokines to treat chronic viral infections such as hepatitis C (HCV). We therefore investigated the ability of IL-29 and IFN-alpha or IFN-gamma to cooperatively inhibit virus replication and induce antiviral gene expression. Compared with the individual cytokines alone, the combination of IL-29 with IFN-alpha or IFN-gamma was more effective at blocking vesicular stomatitis virus and HCV replication, and this cooperative antiviral activity correlated with the magnitude of induced antiviral gene expression. Although the combined effects of IL-29 and IFN-alpha were primarily additive, the IL-29/IFN-gamma combination synergistically induced multiple genes and had the greatest antiviral activity. Two different mechanisms contributed to the enhanced gene expression induced by the cytokine combinations: increased activation of ISRE promoter elements and simultaneous activation of both ISRE and GAS elements within the same promoter. These findings provide new insight into the coregulation of a critical innate immune response by functionally distinct cytokine families.

[Effect of expression of exogenous PDGF-A chain on growth and transformation of CHO cells]

CHO cells were transfected with plasmid pSV2-PDGF-A (containing human PDGF-A cDNA) by calcium phosphate method. Twenty transfected cell lines were obtained after G418 selection. The selected 2 cell lines At1 and Aot7), with prominent changes in morphology and growth behaviour, showed transcription of PDGF-A chain mRNA much higher than CHO cells, strong fluorescent PDGF-specific reaction, appearing that PDGF-like proteins were synthesized in cytoplasm of these cells. At1 and Aot7 cells not only had increased growth rate, but also formed large colonies in soft agar and grew into fibrosarcomas in nude mice. These results suggested that the expression of exogenous PDGF-A gene might cause the uncontrolled growth and malignant transformation of CHO cells.

[Expression of exogenous platelet-derived growth factor B chain gene in CHO cells]

CHO cells were transfected with plasmid pSM-1 (containing human c-sis cDNA) singly or co-transfected with pSV 2 neo DNA by calcium phosphate method. After low serum or G418 selection several cell lines with expression of platelet-derived growth factor (PDGF) were obtained. One among them, FB5, was of the highest PDGF expression and showed the following biological characteristics when compared with CHO cells: (1) a prominent change in morphology from spindle to round in shape: (2) increase of growth rate; (3) growth in low serum (2%) medium as a semisuspension culture; (4) growth on soft agar to larger colonies; (5) synthesis of PDGF in cytoplasm identified by immunofluorescent method; (6) the conditioned medium stimulated DNA synthesis of NRK cells; (7) RNA dot hybridization showing high transcription of PDGF mRNA; (8) southern blot showing integration of human c-sis gene was still stable after 7 months. These results indicated that intergration of exogenous c-sis gene and its high expression might cause CHO cells to high growth rate and even transformation. The establishment of this stable transformed cell line, FB5 is thought to be a good model for further study on the function of PDGF in cell growth control and cell transformation.