Background

The analysis of recurrent genetic mutations in cancer driver genes represents a standard approach for patients diagnosed with lung cancer. Yet, only about 15% of patients harbor activating mutations in oncogenic drivers that can be exploited therapeutically. Despite substantial progress in immune check point inhibition, the survival data for a majority of mutation-negative patients remains unsatisfactory.

Methods

Clonal evolution: Multi-region copy number data from the TRACERx study was used. The copy number values were adjusted by ploidy and log-transformed. Gains with a copy number state > log2(4/2) were classified as high-level gains. Clonal gains were ubiquitously identified across every tumour region sequenced from a given tumour, while subclonal gains were identified as present in at least one tumour region, but not all. To identify significant frequent gains and losses across the cohort, the 95% quantile of the corresponding null distribution was used as a threshold. A threshold of 20 Mb was used to distinguish between focal and broad gains. In order to check which of the genes located in the gained area around MCL- 1 show an increase in expression, the RNAseq data from the TRACERx cohort were analysed. The mRNA levels were measured as TPM (Transcripts Per Kilobase Million) values.

Results

Here we show that an improved understanding of genetic liabilities beyond oncogenic driver alterations might improve patient outcomes. We provide evidence that non-oncogene addiction represents a hallmark of lung cancer and the identification of such liabilities might provide an unprecedented treatment approach for lung cancer patients. The TRACERx data acquisition and analysis platform has provided a variety of novel insights into lung cancer genetics, immune evasion and into the frequency and composition of chromosomal aberrations. We find that chromosomal gains range amongst the most common genomic events in lung cancer patients and their identification might represent a viable option for drug targeting.

Conclusions

Overall, targeting genetic liabilities that serve as non-oncogene addiction in lung cancer and that lie beyond oncogenic driver alterations likely hold therapeutic promise.

Background

In recent years, immunotherapy has been greatly developed, and the regulatory role of epigenetics has been confirmed. However, the role of 5- methylcytosine (m5C) in tumor microenvironment and immunotherapy response is still unclear.

Methods

Based on 11 m5C regulators, we evaluated m5C modification patterns of 572 lung adenocarcinoma patients. And m5C score was constructed by principal component analysis algorithms to quantify the m5C modification pattern of individual lung adenocarcinoma patients.

Results

Two m5C methylation modification patterns were revealed according to 11 m5C regulators. The two patterns had remarkably distinct tumor microenvironment (TME) immune cell infiltration characterization. Then 226 differentially expressed genes related to m5c phenotype were screened. Patients were divided into three different gene cluster subtypes based on these genes, which had different TME immune cell infiltration and prognosis characteristics. M5C score was constructed to quantify the m5C modification pattern of individual lung adenocarcinoma patients. It can be seen that a high m5C score group had a better prognosis. The role of m5C score in predicting prognosis was also verified in GSE31210.

Conclusions

Our study revealed that m5C modification played a significant role in TME regulation of lung adenocarcinoma. The study of m5C regulation mode may have some implications for tumor immunotherapy in the future.

Background

Tumor heterogeneity has been long considered to shape tumor fitness and the metastatic potential of early stage tumor cells as well as the therapeutic response in advanced stages.

Methods

All primary tumors were implanted in immunocompromised NOD-SCID mice within 6 hours from surgery and passaged at least once before freezing. Both the primary tumor and the patient-derived xenograft were characterized with targeted exome sequencing against a panel of 70 genes which encompasses standard oncogenes and tumor suppressor genes involved in lung cancer, DNA damage/DNA repair proteins including those involved in homologous recombination, and others. Moreover, preoperative patients’ blood was analyzed for circulating tumor cells (CTCs) using the Parsortix platform and gene expression analysis.

Results

We surgically implanted more than 50 primary specimens from patients with early stage operable NSCLC. Over 50% of those grafted successfully. Our results indicate that the ability of primary tumors to successfully graft correlates significantly with tumor stage and grade, primary tumor size, and the detection of CTCs. Moreover, we observed that squamous cell carcinomas are more likely to graft than adenocarcinomas. Other than mutations associated tumor type (squamous vs. adeno), we did not observe any combination of mutations favors or disfavors grafting ability. The comparison between the genetic profile of the primary tumor and the PDX revealed differences in some cases. Detailed results will be presented.

Conclusions

Primary early stage tumor samples from NSCLC patients graft efficiently in immunocompromised mice while the grafting efficiency is affected by the tumoral clinic-pathological characteristics.

Background

Background:

Pleomorphic carcinoma (PC) is a rare subtype of non-small-cell lung cancer (NSCLC). PC is defined as highly heterogeneous and poorly differentiated adenocarcinoma, squamous cell carcinoma, or large cell carcinoma containing at least 10% of sarcomatoid components of giant and/or spindle cells. PC has a more aggressive clinical course and a worse outcome than other NSCLCs. In comparison to pure LUAD and LUSC, it is not sufficiently characterized on a genomic and molecular level. Understanding the molecular and genetic basis of PCs is necessary to find new treatment options. Here, we performed whole exome sequencing (WES) to investigate the genetic heterogeneity of this rare tumor type.

Methods

Methods:

FFPE (formalin-fixed, paraffin-embedded) sections were used to dissect tumor tissue from six primary PCs (separate for each component) and from three matched metastases and were subjected to WES separately.

Results

Results:

The fourteen analyzed tumor areas including eleven areas from six primary tumors and three matched metastases harbored a median of 334 (range 159 - 478) non-synonymous somatic mutations and a median of tumor mutation burden (TMB) of mutations/Mb 10,1 (range 4,8-14,5). KRAS and TP53 were the most frequently altered genes (4/6 patients), followed by FAT1 and KMT2D (3/6 patients). RBM10 was mutated in 2/6 patients, followed by ALK, FBXW7, PIK3CA, ATM and MET altered in 1/6 patient. Furthermore, 3/6 patients showed concurrent mutations of KRAS and TP53. Clonality analysis showed that most of cancer gene mutations found were clonal and truncal in both components of PCs. KRAS gene amplification was identified in 1/6 patients with no simultaneous KRAS mutation. Mutational signature analysis showed that most of the mutations were dominated by ageing and smoking-related signatures number 1, 4 and 5.

Conclusions

Conclusions:

The two different tumor components of PC are clonally related suggesting underlying transdifferentiation. The presence of the high frequency of the KRAS gene alterations including mutations and amplifications points to their primary role in the evolution of this tumor type. Studies are ongoing to explore the role epigenomic changes in the morphologically striking heterogeneity of PC.

Background

Circulating microRNAs (miRNAs) modulate immune response and the crosstalk between tumor cells and the immune system. We herein investigated the association of miR-26a, let-7c, miR-30d, miR-195, miR-202, miR-98, involved in the polarization of tumor associated macrophages (TAMs), with the outcome of non-small cell lung cancer (NSCLC) patients (N=125) treated with first-line platinum-based chemotherapy.

Methods

MiRNA expression levels were analyzed by RT-qPCR in plasma obtained prior to the initiation of chemotherapy. U6 snRNA was used as reference gene for normalization and fold change of each miRNA expression relative to the expression in healthy controls was calculated by the 2^-ΔΔCt method.

Results

High miR-202 expression was associated with disease progression as best response to treatment (p=0.030) and independently predicted for both shorter progression free survival (PFS, p=0.021) and overall survival (OS, p=0.024), in the whole group of patients. In the non-squamous (non-SqCC) subgroup, high miR-202 was also revealed as an independent predictor for shorter OS (p=0.008). High miR-26a was associated with a shorter OS in the SqCC subgroup (p=0.033).

Conclusions

These results suggest that the expression of circulating miRNAs involved in the regulation of TAMs, is associated with outcomes of patients with NSCLC treated with platinum-based chemotherapy. Further studies are required to clarify whether their role is prognostic or predictive role and to establish their clinical validity.

Background

The significance of the immune homeostasis modulator thymic stromal lymphopoietin (TSLP) in lung cancer has still not been explored in detail. The aim of this study was to assess the prognostic and predictive potential of TSLP in EGFR-TKI treated non-small cell lung cancer.

Methods

The interactive publicly available database the Human Protein Atlas was used to analyze the Cancer Genome Atlas (TCGA) transcriptome data and assess the expression of TSLP in lung cancer. 101 advanced EGFR mutated lung adenocarcinoma patients were treated with TKIs until progression or unacceptable toxicity. Primary resistance was defined as the absence of any response within the first 3 months of therapy. Progression-free survival (PFS) was defined as the time from the start of therapy to date of clinical progression, and overall survival (OS) as the time from diagnosis to death from any cause. TSLP gene expression was determined by real-time qPCR. Receiver operating characteristics analysis/area under the curve with 95% confidence interval was applied for the investigation of the discriminatory potential of TSLP expression (p < 0.05).

Results

In silico analysis showed that high TSLP expression is a favorable prognostic factor in unselected lung adenocarcinoma patients with a 5-year survival rate for high expression of 51%, and 36% for patients with low TSLP-expressing tumors (p=0.0043, median follow up time: 1.79 years). However, in our study group of EGFR-TKI treated advanced lung adenocarcinoma patients, primary resistance was detected in 14 patients and correlated with significantly higher pre-treatment levels of TSLP compared to patients with disease stabilization and complete or partial response (1.04 ± 0.03 vs. 0.96 ± 0.04, p=0.0001). Low TSLP expression predicted a favorable response to TKIs with a ROC cut-off value of 1.006 (AUC 0.713, 95% CI 0.569-0.831, p=0.0015).

Conclusions

Increased levels of TSLP might lead to a switch to Th2-mediated immunity inducing a poor response to TKIs in advanced lung adenocarcinoma. The combined effect of other Th2-related cytokines (IL-33, IL-25) on the response to TKIs is currently under investigation at our Institute.

Background

By using high-throughput technologies many potentially prognostic genes can rapidly be identified. However, new prognostic markers must be validated in independent patient cohorts, add prognostic information to already existing tools, and be assessed by a technique that is applicable in clinical laboratories. By using a multicohort discovery and validation strategy, we identified genes with expression linked to overall survival (OS) in adenocarcinoma (AC) and subsequently validated candidate prognostic markers with immunohistochemistry (IHC).

Methods

Six publicly available gene expression datasets from microarray-based studies were used. In four discovery datasets the samples were, for each probe, dichotomized into two groups based on the median gene expression values. The log-rank test was used to test for differences in OS between the two groups and probes with p-value <0.05 were collected for each dataset. Only probes with p-value <0.05 in all four datasets could proceed to the validation step, where the probes were tested as above in two validation datasets. Probes with p-value <0.05 in both datasets were considered as candidate markers. Out of these markers, three were selected for IHC validation. IHC stainings were performed on tissue microarrays consisting of 124 surgically treated AC cases. For each marker, a cut-off for classifying samples into positive or negative protein expression was defined. Also, a combined score system of 0-3 points, with one point per positive marker, was constructed. Expression was correlated to OS using log-rank test with Kaplan-Meier plots. Multivariable Cox proportional hazards regression models were used to adjust for stage, adjuvant treatment, smoking, age and gender.

Results

We identified 19 candidate markers, and three were chosen for IHC validation; Ki67, MCM4 and TYMS. Protein expression of TYMS, as well as the combined score, was significantly associated to OS (also in multivariable analysis), while a trend towards worse prognosis was seen in patients positive for Ki67 and MCM4 protein expression.

Conclusions

In summary, our gene expression-based strategy and subsequent validation with IHC successfully identified three markers with prognostic potential in AC.

Background

KRAS is one of the most common oncoproteins in human cancer but therapeutic strategies tailored to KRAS-mutant cancers has remained an unmet medical need. One approach to treat KRAS-driven tumors is to exploit collateral vulnerabilities or synthetic lethal partners that are essential in the context of oncogenic KRAS. PLK1 has been identified as a codependency in KRAS-mutant cancer cells as its inhibition is synthetic lethal with KRAS mutations. However, PLK1 inhibitors alone have only demonstrated limited efficacy in clinical studies, suggesting that additional targets are needed to improve PLK1-targeted therapy in KRAS-mutant cancers.

Methods

We performed chemical synthetic lethality screens to identify complementary targets that enhance the efficacy of PLK1-targeted agents in KRAS-mutant cancer cells. Newly identified drug pairs were investigated in various tumor models of KRAS-mutant cancers.

Results

Among others, we identified AZD4547(FGFRs inhibitor) synergizes with PLK1 inhibitors (BI2536/BI6727, a selective PLK1 inhibitor currently investigated in phase II/III clinical trials) to drive KRAS-mutant cancer cell death. We provided evidence showing that genetic and pharmacological inhibition of FGFR1 synergistically enhances the effect of PLK1 inhibitors in KRAS-mutant lung and pancreatic cancer cells. Mechanistically, the combinatorial effect is ascribed the surge of metabolic ROS and subsequently activation of JNK/p38-dependent apoptosis .

Conclusions

Our results suggest a synergistic combination strategy for the treatment of KRAS-mutant cancers.

Background

Kirsten rat sarcoma viral oncogene (KRAS) mutations occur frequently in NSCLC (non-small cell lung cancer), but clinically applicable targeted therapy is limited until now. This study set out to evaluate a novel strategy to treat KRAS-driven lung cancer.

Methods

The in vitro effects of the combined use of anlotinib and trametinib were evaluated by cell viability assay, wound healing scratch assay, clonogenic survival assay and flow cytometry. A xenograft model was used to evaluate the inhibitory effect of the drug combination on KRAS-driven NSCLC growth in vivo. The transcriptome-wide differential gene expression analysis and protein microarray were applied to discover underlying molecular mechanisms.

Results

Anlotinib was found to synergize with trametinib in vitro and in vivo to further exert anti-tumor effects on KRAS-mutant lung cancer. Furthermore, we reported that the expression of IGFBP2 decreased significantly in the dual-drug group, which was a key protein that mediated the synergistic effect by regulating the cell cycle and EMT process. Recombinant IGFBP2 protein could restore the inhibitory effects of dual drugs on KRAS-mutant tumor proliferation, invasion and migration.

Conclusions

Overall, the findings of this research provide insights for clinical application of anlotinib and trametinib in treating KRAS-driven NSCLC.

Background

Aberrant expression of Anaplastic lymphoma kinase (ALK) has been recognized as potent drivers of oncogenesis in non-small cell lung cancer (NSCLC). Detecting ALK gene rearrangements in patients with newly diagnosed NSCLC is critical, as the presence of this oncogene influences treatment opportunities. Given the recent approval of multiple drugs targeting ALK-rearranged NSCLC, optimal first-line therapy is an area of active debate. The recently published CROWN trial showed that lorlatinib retains potency against all known single ALK resistance mutations, including ALK G1202R. Here, we present a comprehensive analysis of ALK alterations in NSCLC.

Methods

Patient-specific ALK alterations were analyzed using the open-source AACR Project GENIE Cohort v8.1. Using cbioportal as a query database, we analyzed 11,107 samples from 10,082 patients of lung adenocarcinoma for the prevalence of ALK fusions, mutations, and copy number alterations in NSCLC.

Results

584 (5% of queried samples) ALK alterations were detected, including 354 missense mutations (60.6%), 265 fusions (45.4%), 51 truncating mutations (8.7%), and 1 in-frame mutation (0.17%). Following the exclusion of alterations with mutations and copy number alterations of unknown significance, 284 (2%) of samples still had significant ALK alterations. Among them, there were 259 fusions and 25 missenses mutations identified. Of the 25 missense mutations, we identified only 7/284 (2.46%) mutations with G1202R protein change, 7/284 (2.46%) mutations with L1196M, and 4/284 with I1171N which were all identified as oncogenic. The most common ALK gene upstream partners were EML4, KIF5B, and SQSTM1 (81.5%, 1.5%, and 1.1%) of identified ALK fusions, respectively. ALK fusions were significantly co-altered with HSP90AB1 mutations (p=2.51E-04), IDH2 mutations (p=0.027), and NRG1 mutations (p=0.024).

Conclusions

Most ALK variants are described as VUS, limiting the impact of precision oncology. ALK fusions occur in 2.6%% of the lung adenocarcinomas, with EML4 being the most common upstream partner. Meanwhile, G1202R mutations occur only among 0.07% of the ALK alterations. Heat shock protein and Neuregulin-1 pathway may present additional opportunities for combination targeted therapies in the future for ALK-positive NSCLC.

Background

EGFR mutations (EGFRm) represent 10-15% of advanced non-small cell lung cancer (NSCLC) in European patients (pts).

Tissue molecular profiling is the gold-standard, but liquid biopsy (LB) offers a non-invasive alternative. Digital droplet PCR (ddPCR) is a fast, high-sensitive and low-cost LB to detect specific molecular alterations.

We aimed to describe ddPCR clinical utility for EGFRm detection in advanced NSCLC.

Methods

Prospective blood sample collection in advanced NSCLC pts harboring EGFRm either at baseline and/or at progression (PD) between Jan/16 and Sep/20 at Gustave Roussy. LB was performed by ddPCR (Stilla®): sensitizing (exon19 deletion; exon21 [L858R]) and T790M resistance EGFRm. We defined high tumor burden as >2 metastatic sites. We analyzed EGFRm detection by ddPCR at these timepoints.

Results

A total of 252 samples were collected (N=140 pts): 27 at baseline, 144 at PD and 81 under response. Median of samples/pts was 1 [1-8]. Median age was 66 [36-92]; 29% were female, 62% non-smokers and 97% had adenocarcinoma histology.

At baseline, sensitizing EGFRm were detected in 59% (16/27) of samples: 12 ex19del and 4 ex21. Median number of sites was 3 [1-5]. In pts with intracranial and/or thoracic isolated lesions, EGFRm were detected in 33% (2/6); in those with high tumor burden detection was 78% (7/9).

At PD, sensitizing EGFRm were detected in 57% (82/144) of samples: 64 ex19del and 18 ex21. Median number of PD sites was 3 [1-8]; lung (70%) and bone (64%) were the most common. In pts with intracranial/thoracic isolated lesions, EGFRm were detected in 31% (11/35); in those with high tumor burden detection was 67% (54/81).

At PD to 1^st-2^nd generation tyrosine-kinase inhibitors, sensitizing EGFRm were detected in 49% (40/81) of samples. The T790M mutation was found in 22% (18/81) overall, all of them with positive ddPCR sensitizing EGFRm (45%, 18/40); detection rate was 9% (2/23) for intracranial/thoracic vs. 32% (13/41) for high tumor burden cases.

Conclusions

ddPCR is a sensitive liquid biopsy for sensitizing and resistance EGFRm detection. ddPCR positivity was more likely observed in systemic PD cases with high tumor burden. It can provide a rapid EGFRm result to guide treatment in NSCLC, however metastatic profile should be taken into account.