Abstract
While patients with advanced-stage non-small cell lung cancers (NSCLCs) harboring MET exon 14 skipping mutations (METex14) often benefit from MET tyrosine kinase inhibitor (TKI) treatment, their long-term survival is limited by drug resistance. The molecular basis for this resistance remains incompletely understood. Through targeted sequencing analysis of cell-free circulating tumor DNA obtained from 289 patients with advanced-stage METex14 NSCLC, we find prominent co-occurring RAS pathway gene alterations (e.g. in KRAS/NRAS, NF1). Clinical resistance to MET TKI treatment was associated with the presence of these co-occurring alterations. In a new preclinical model expressing a METex14 mutation, KRAS overexpression promoted MET TKI resistance, which was overcome by co-treatment with crizotinib and the MEK inhibitor trametinib. Our study provides a genetic landscape of co-occurring alterations in advanced-stage METex14-mutated NSCLC and suggests a potential combination therapy strategy to enhance clinical outcomes.
Statement of Significance This report describes targeted sequencing of the largest reported cohort of advanced-stage NSCLC with a MET exon 14 skipping mutation. The data uncover certain RAS pathway genetic alterations as potential mediators of MEK TKI resistance, which could be overcome by MEK and MET co-inhibition in NSCLC.
Introduction
Somatic mutations near splice sites in MET lead to splicing-mediated loss of exon 14, and a stabilized and possibly hyperactive MET protein. These METex14 mutations are an emerging therapeutic target in NSCLC and play a driver role in 2-4% of lung adenocarcinomas (1,2). MET tyrosine kinase inhibitor (TKI) treatment was associated with improved overall survival in a retrospective study of patients with METex14 NSCLC. (3). Response rates to treatment with the multi-kinase inhibitor crizotinib, which has activity against MET, were as high as 68% in a phase I clinical trial (3,4), with a progression free survival of 7.3 months in one case series (5).
Both intrinsic and acquired resistance limit the long-term survival of patients with METex14 NSCLC. Second-site MET mutations occur at acquired resistance in these patients after an initially favorable response to MET TKI treatment. These mutations include the second-site MET Y1230C/H/S, D1228A/H/N, and G1163R mutations at residues critical for type I MET TKI binding (e.g. crizotinib), against which type II MET TKIs may retain activity (6-8). Activation of downstream and parallel signaling pathways can bypass inhibition of the primary oncogenic driver in other NSCLC genetic subtypes (e.g. EGFR-mutant, ALK rearrangement positive), comprising a second main class of resistance mechanisms (9). While the signaling pathways downstream of activated MET include the RAS-RAF-MAPK and PI3K pathways (10), their role in targeted therapy resistance is not well-understood, and potential bypass track-based resistance mechanisms have yet to be reported in METex14 NSCLC.
Recent studies show that advanced-stage NSCLCs often harbor multiple oncogenic alterations that may impact therapeutic response (11-13). A more detailed understanding of the mutational landscape which co-exists with oncogenic MET alterations in NSCLC, including amongst these parallel and bypass signaling pathways, may facilitate an improved understanding of MET TKI resistance determinants and identify rational polytherapy strategies to improve clinical outcomes. Analysis of cell-free circulating tumor DNA (cfDNA) provides one avenue to describe the genomic landscape within a tumor and offers the potential to capture genomic changes reflecting heterogeneity across distinct metastatic tumor sites (11,14).
Here, we describe the spectrum of co-occurring genomic alterations observed within cfDNA of patients with METex14-harboring, advanced-stage NSCLC and identify prominent alteration of RAS pathway genes, including KRAS, in association with MET TKI resistance.
Results
o-occurring genomic alterations are common in advanced-stage METex14 mutated lung cancer
We studied a cohort of 289 patients with advanced-stage NSCLC with a METex14 mutation identified upon cfDNA analysis of 68 cancer-relevant genes using a clinically validated assay, to evaluate the frequency with which METex14 mutations co-occur with other cancer-associated mutations (Supplemental Table S1) (11,15). Synonymous mutations and those with predicted neutral or unknown functional impact were excluded as previously described (11), as were mutations previously associated with clonal hematopoiesis (16). 86.5% of samples contained co-occurring genomic alterations, with a mean of 2.74 alterations per sample (range 0-22) in addition to the METex14 mutation. The most commonly altered genes, co-occurring with the METex14 mutation in at least 10% of patients, were TP53 (49.5% of patients), EGFR (16.3%), NF1 (15.6%), BRAF (10.7%), and CDK4 (10.4%) (Figure 1A). Additional MET gene alterations were also common in this patient population, in which 9.3% of patients had co-occurring MET copy number gain and 12.5% had a co-occurring second-site MET mutation (Figure 1A). Including all identified MET second-site mutations, regardless of predicted functional impact, 17 of 34 mutations were located in the tyrosine kinase domain. Of these, fourteen (G1163R, L1195V/F, F1200I, D1228H/N/Y, and Y1230H/S) were located at residues previously associated with MET TKI resistance (6-8,17-19). Nine of these occurred in patients with known prior MET TKI exposure and five in patients with unknown prior treatment history. The remaining three mutations (H1094Y, R1336W, and I1084L) occurred in patients without available prior treatment history. While the MET H1094Y mutation is known to lead to MET activation (20), the other two mutations remain uncharacterized.
The co-occurrence of other established oncogenic driver alterations (KRAS, EGFR, ALK, ROS-1, BRAF) was uncommon except for the presence of activating KRAS mutations in 5.2% of patients (G12C/D/S/V 3.5%, G13C 1%, Q22K 0.3%, and Q61H 0.3%), canonical EGFR-activating mutations in 3.5% (del19 3.1%, L858R 0.7%, T790M 2%), and an ALK fusion in 0.7%. In addition, a HER2 exon 20 insertion was detected in one patient. In a patient with known clinical outcomes data, with both EGFR del19/EGFR T790M mutations and a METex14 mutation, there was partial response (RECIST 1.1) to second-line treatment with the EGFR TKI osimertinib, which was ongoing at 11 months treatment.
The spectrum of co-occurring gene alterations differs between NSCLCs with a MET exon 14 skipping mutation or an activating EGFR mutation
To understand how the genomic landscape in advanced-stage METex14 NSCLC compares to NSCLCs with a different targetable oncogenic driver mutation, we used identical cfDNA profiling to compare a cohort of patients with known METex14 NSCLC to a separate, stage-matched cohort of 1653 samples from 1489 patients with advanced-stage EGFR-mutant (del19, L858R) NSCLC (Supplemental Table S1) with identical filters to remove mutations with predicted unknown or absent function impact and mutations previously associated with clonal hematopoiesis. This comparison demonstrated differential frequency of co-occurring genomic alterations in 17 genes (Figure 1B). In the METex14 cohort, NF1, CDK4, STK11, ALK, KRAS, ATM, CDKN2A, NRAS, TSC1, and ESR 1 were more commonly identified. In the EGFR-mutated cohort, co-occurring alterations in AR, ERBB2, CCNE1, PIK3CA, BRAF, CTNNB1, and MYC were more commonly identified, consistent with our prior findings identifying these as common co-occurring alterations in EGFR-mutant NSCLC (11). Among the 10 genes with a greater frequency of genomic alteration in patients with a METex14 mutation, three (NF1, KRAS, and NRAS) were involved in the RAS-RAF-MAPK signaling pathway.
Both MET second-site mutations and gene alterations promoting bypass pathway activation are acquired following MET TKI exposure
We identified twelve patients with cfDNA obtained following treatment with crizotinib, with an available matched comparator sample obtained prior to known crizotinib exposure (Table 1, Supplemental Table S2). A potential mechanism of resistance to MET TKI therapy was identified in all but four patients, including both on-target resistance via MET second-site mutations and activating changes in potential parallel signaling pathways. Details regarding one patient (patient 5) in this dataset have previously been published by other groups (6,8).
MET second-site mutations (Y1230H/S, D1228H/N, F1200I, L1195V) were seen in four of twelve patients (Table 1, Supplemental Table S2). With the exception of the MET F1200I mutation, these second-site mutations have been reported in patients with acquired MET TKI resistance (6-8,17-19). Like the MET L1195V mutation, in preclinical studies the MET F1200I mutation is predicted to alter the kinase domain conformation such that it interferes with the binding of type II MET TKIs within the DFG-out binding pocket, and to a lesser extent to interfere with type I MET TKI effect through loss of an autoinhibitory MET conformation (Supplemental Figure S1) (6).
Potential parallel and downstream pathway alterations were found in 7 patients (Table 1). These included an activating KRAS mutation and KRAS copy number gain (Figures 2A and 2B). Also within the RAS pathway, an NF1 frameshift mutation (NF1 p.Gln2636fs) was found in one patient. Loss of function mutations in NF1 lead to RAS/RAF/MAPK pathway hyperactivation and are a potential therapeutic target for MEK inhibition in NSCLC (NCT03232892) (9). Copy number gain of the receptor tyrosine kinases EGFR and KIT that act upstream of the RAS pathway were present in four patients. Both KIT and EGFR encode receptor tyrosine kinases capable of providing alternative oncogenic signaling through, in part, RAS/RAF/MAPK activation, and are implicated in resistance to crizotinib in ALK-rearranged NSCLC (21), but have not been described in the METex14-mutated setting.
Emergence of RAS pathway genetic activation during MET TKI resistance
We identified a patient within the described cfDNA cohort on whom serial plasma cfDNA analysis was performed during treatment with a MET TKI (Figure 2A, patient one in Table 1). The patient was an 84 year-old male with a 30 pack-year history of tobacco use and presented with metastatic pulmonary adenocarcinoma involving the lungs, lymph nodes, axial skeleton, and brain. While there was insufficient tissue upon diagnostic tumor biopsy for molecular testing, evaluation of plasma cfDNA using the Guardant360® clinical assay identified a canonical METex14 skipping mutation (c.3028+3A>G; variant allele frequency 7.6%). The patient received stereotactic radiosurgery to brain lesions, followed by the MET TKI crizotinib at 250 mg twice daily. After four weeks treatment repeat imaging demonstrated a radiographic partial response (RECIST 1.1) with a decrease in METex14 cfDNA variant allele frequency (VAF) to an undetectable level. Crizotinib was subsequently dose-reduced and then held for treatment-associated fluid retention, which was followed by radiographic disease progression and a return of the METex14 variant to levels detectable in plasma within four weeks (Figure 2A, Supplemental Table S3). Progression continued despite return to full dose crizotinib treatment.
Targeted sequencing of 479 cancer-associated genes using the UCSF500 next generation sequencing (NGS) assay (22) was performed on a tumor biopsy of a progressing metastatic soft tissue rib lesion and identified a high level wild-type KRAS copy number gain (~22x fold), in addition to other genetic variants of unknown significance (Supplemental Table S4) and the known METex14 skipping mutation. The UCSF500 assay utilizes control sequencing of peripheral blood mononuclear cells, which allows for more precise identification of somatic variants and reduces risk of false positive results related to detection of genomic alterations associated with clonal hematopoiesis. Serial cfDNA analysis also showed wild-type KRAS copy number gain in a sample obtained 8 weeks after this biopsy, followed by emergence of the MET Y1230S and F1200I mutations on subsequent samples. (Figure 2A, Supplemental Table S3). Pathology review of the progressing rib lesion biopsy sample confirmed persistent adenocarcinoma histology, without evidence of histologic transformations (Supplemental Figure S2). The FDA-approved MEK inhibitor trametinib (2 mg daily) was added to crizotinib (250 mg twice a day) in an attempt to target the activity of the downstream high level KRAS copy number gain that was hypothesized to drive MEK TKI resistance in this patient. A molecular response was observed after two weeks of combined trametinib and crizotinib treatment, with decrease in the cfDNA VAF of all identified MET mutations and return to undetectable KRAS copy number gain (Figure 2A, Supplemental Table S3). However, the patient continued to decline clinically with fatigue, fluid retention, and diarrhea consistent with toxicity of the combination therapy regimen despite dose-reduction of trametinib to 2 mg every other day, and expired before radiographic response could be assessed.
KRAS amplification and KRAS G12D mutation upon MET TKI resistance
A second patient was identified with a canonical MET exon 14 skipping mutation and acquired KRAS gene alterations in the setting of crizotinib resistance (Figure 2B, patient two in Table 1). This 67 year-old male without a history of tobacco use was diagnosed with pulmonary adenocarcinoma metastatic to the adrenal gland. Targeted sequencing of a dominant right upper lobe lung lesion demonstrated a METex14 mutation (2888-19_2895del27), with wild-type KRAS (Figure 2B, Supplemental Table S5). Plasma cfDNA testing using the FoundationACT® clinical assay demonstrated the known METex14 mutation and an activating KRAS G12D mutation (Supplemental Table S5). He was treated with crizotinib in the first-line setting with a mixed tumor response after two months. While the biopsied right upper lobe pulmonary lesion responded to treatment (RECIST 1.1), a second smaller right upper lobe pulmonary lesion and an adrenal lesion enlarged, consistent with primary MET TKI resistance (Figure 2B). Confirmatory droplet digital PCR (ddPCR) testing verified absence of the KRAS G12D mutation within the pre-treatment biopsy sample to a sensitivity of <0.02%, while confirming the presence of the known METex14 mutation (present in 11% of alleles).
Following stereotactic body radiotherapy (SBRT) to the progressing lung and adrenal lesions, the patient continued to receive crizotinib with radiographic response until 7 months of treatment, when a progressing abdominal lesion adjacent to the prior adrenal lesion and progressing lung nodules were detected. A biopsy of the progressing abdominal lesion identified the known METex14 mutation (77.9% VAF) and KRAS G12D mutations (47.4% VAF), and also found KRAS copy number gain (~11-fold), along with other gene alterations of less certain significance (Supplemental Table S5). Next-generation DNA sequencing of the patient’s peripheral blood mononuclear cells (PBMCs) confirmed absence of detectable KRAS G12D mutation within the hematopoietic lineage. He was subsequently treated with platinum-based chemotherapy with radiographic response.
KRAS overexpression induces resistance to crizotinib that is overcome by co-treatment with trametinib
The clinical data suggested that hyperactivation of certain RAS pathway genes (e.g. KRAS) may induce resistance to MEK TKI treatment in METex14 NSCLC. We engineered a new Ba/F3 cell-based system to assess the impact of RAS pathway hyperactivation on sensitivity to MET TKI treatment. The IL-3 dependent Ba/F3 cell line, while not of epithelial origin, is an established system to assess oncogenic capacity and putative drug resistance mechanisms (23,24). Stable expression of human METex14 within the Ba/F3 cell line in the presence of the MET ligand human HGF induced IL-3 independent growth, to a greater extent than stable expression of wild type MET (Figure 3A). Increased downstream Erk phosphorylation was observed with METex14 versus wild type MET (Supplemental Figure S3). IL-3 independent growth was not seen in the absence of HGF treatment (Figure 3B), consistent with the known functional consequences of METex14 skipping, which leads to reduced MET degradation but not necessarily to ligand-independent MET activation (25).
METex14 expressing cells were sensitive to treatment with crizotinib, as measured by reduced cell growth on cell viability assays (Figure 3C), and decreased Erk phosphorylation was observed upon crizotinib treatment with associated increase in cleaved PARP indicative of apoptosis (Supplemental Figure S3). To model the clinical cases, we then tested whether the forced overexpression of wild type KRAS (to mimic copy number gain or mutational activation) in METex14-expressing cells had an effect on response to crizotinib. We generated KRAS-overexpressing cells in the METex14 background and performed viability experiments, which showed that the cells became resistant to single-agent MET TKI treatment, with an increase in IC50 to crizotinib (IC50 of 0.16 μM in control cells versus 1.68 μM in KRAS overexpressing cells, p-value < 0.001) (Figures 3C and 3E). We then treated these cells that harbor both METex14 and KRAS overexpression with a combination of crizotinib and trametinib in order to block both MET signaling and downstream MAPK pathway signaling. METex14-expressing cells overexpressing KRAS were re-sensitized to MET inhibition when co-treated with the MEK inhibitor trametinib (crizotinib IC50 of 1.68 μM in the absence of trametinib versus 0.25 μM with trametinib co-treatment, p-value < 0.001) (Figure 3D and 3E). The selected trametinib dose reduced but did not eliminate cell growth in the absence of crizotinib (Supplemental Figure S4). Immunoblot analysis demonstrated Erk phosphorylation despite crizotinib treatment in samples with KRAS overexpression, which was abrogated by the addition of trametinib. The combination treatment was associated with increased cleaved PARP indicative of tumor cell apoptosis (Supplemental Figure S3). Overall, these results show that METex14 expression confers oncogene addiction in the presence of HGF. These cells are sensitive to MET inhibition and are rescued by RAS pathway hyperactivation via KRAS overexpression. Combination treatment with both MET and MEK inhibitors overcomes MET TKI resistance induced by KRAS overexpression in this system.
Discussion
The challenge of therapeutic resistance is emerging as a key problem in the management of patients harboring METex14 mutations as they increasingly receive treatment with MET-targeted therapies. Downstream and bypass oncogenic signaling pathway activation cause therapeutic resistance to TKIs in cancers with other oncogenic drivers (9), but such mechanisms have not previously been described in METex14 NSCLC, in which second-site mutations in MET itself have taken center stage to date. We found that alterations (mutation or amplification) of certain RAS/RAF/MAPK pathway genes including NF1, KRAS, and NRAS alterations are enriched in METex14 NSCLC compared to NSCLCs with canonical EGFR-activating mutations, using a large, consecutively collected cfDNA clinical cohort described here. We additionally identify patients with METex14 NSCLC demonstrating RAS pathway co-alterations including KRAS mutation or copy number gain during MET TKI resistance, that were present upon both plasma cfDNA analysis and tumor-based DNA sequencing. This is a novel description of downstream MAPK pathway activation at resistance to MET TKI therapy, consistent with established roles of MAPK pathway hyperactivation in promoting resistance to EGFR, ALK, BRAF, and ROS1 targeted therapies via diverse mechanisms including KRAS amplification and KRAS, BRAF, and NF1 mutations (26-34).
The presence of acquired co-occurring downstream or bypass pathway alterations (e.g. KRAS copy number gain or mutation) may identify patients who will benefit from pathway-targeted rational polytherapy strategies. In our METex14 preclinical model system, KRAS overexpression induced MET TKI resistance that was overcome by the addition of the MEK inhibitor. The patient presented here who was similarly treated with this combination therapy showed evidence of molecular response but unfortunately experienced excess adverse effects when receiving this treatment. Though MEK inhibition was ineffective as treatment for activating KRAS mutations in the phase III SELECT-1 trial (35), the potential for activity against tumors bearing other genomic alterations leading to MAPK pathway activation, such as KRAS amplification or NF1 loss of function mutations warrant investigation. Future efforts to combine these agents will require attention to agent selection and dosing for tolerability and efficacy.
The plasma cfDNA analysis presented here from a large cohort of patients with METex14 NSCLC demonstrates co-occurring genomic alterations in the majority of samples, reminiscent of recent studies in patients with EGFR-mutated NSCLC and consistent with a smaller cohort of 13 patients previously described with advanced-stage METex14 NSCLC (11,36). While plasma cfDNA-based testing provides a minimally invasive tool to query the tumor genomic landscape and offers the potential to capture heterogeneity across multiple tumor sites, recent reports have highlighted the challenge of false positive tumor tissue or plasma cfDNA testing due to detection of gene alterations deriving from incorporation of hematopoietic cells undergoing clonal expansion (16,37). To account for this issue, we excluded 866 variants in 206 genes previously associated with clonal hematopoiesis (16) from our reported cfDNA cohort. Additionally, in the described cases with KRAS genetic alterations at MET TKI resistance, PBMC sequencing reduced the probability of clonal hematopoiesis as the origin of the described KRAS variants.
Clinically, while many MET second-site mutations acquired during type I MET TKI (e.g. crizotinib) treatment may be overcome by use of type II MET TKIs (e.g. cabozantinib) (6-8), the bypass pathway resistance mechanisms described here will likely require a different therapeutic strategy, such as pathway-targeted combination therapy. Overall, this study enhances the understanding of the role of co-occurring genetic alterations in METex14-mutated NSCLC, with implications for the improved molecular diagnosis, serial monitoring of resistance mechanisms during treatment, and development of new rational therapeutic strategies to combat MET TKI resistance and enhance clinical outcomes.
Methods
Patients
Patient samples were obtained and analyzed in accordance with IRB-approved protocols. The cfDNA analysis included 332 consecutive samples from 289 patients with advanced (stage IIIB/IV) non-small cell lung cancer with a METex14 mutation obtained between October 2015 and March 2018, and 1653 consecutively tested samples from 1489 patients with EGFR-mutated NSCLC (L858R and del19) obtained between April 2016 and May 2017.
Cell lines and reagents
Ba/F3 cells were purchased from ATCC (ATCC® HB-283™) and maintained in culture for a total of approximately 2-3 months in DMEM supplemented with 1 ng/mL IL-3 (Peprotech). Mycoplasma testing was not performed. Retrovirus was generated using TransIT-LT1 transfection reagent (Mirus). Cells were infected with filtered retrovirus, expressing either mCherry, human METWT or human METex14 in a pBABE-puro vector backbone as previously described (8) and selected in puromycin (2 μg/mL). Expression was confirmed by immunoblotting. KRAS-overexpressing cells were obtained by retroviral infection with a pBABE-hygro KRAS4B construct and selected in hygromycin (800 μg/mL). All drugs were purchased from Selleck Chemicals.
Transformation and cell proliferation assays
Transformation assay was performed by removing IL-3 through centrifugation and adding 50 ng/mL human HGF (Peprotech 100-39H). For proliferation assays cells were seeded in 96-well plates at 5,000 cells/well and the following day were exposed to crizotinib (Selleck Chemicals, #S1068) at 0 to 10 μM and/or trametinib (Selleck Chemicals, #S2673) at 0.01 μM. After 72 hours of drug exposure, CellTiter-Glo (Promega) reagent was added and luminescence was measured on a Spectramax spectrophotometer (Molecular Devices, Sunnyvale, CA, USA) according the manufacturer’s instructions. Data are presented as percentage of viable cells compared with control cells (vehicle treatment).
Immunoblotting
Cells were washed in PBS and lysed with 25 mM Tris-HCL (pH 7.6), 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate, 0.1% SDS supplemented with Halt Protease Inhibitor Cocktail (Thermo Fisher Scientific) and Halt Phosphatase Inhibitor Cocktail (Thermo Fisher Scientific). Lysates were separated in a 4%–15% SDS-PAGE gel and transferred onto a nitrocellulose membrane (Bio-Rad). Membranes were blocked with 5% fetal bovine serum (FBS) in Tris-buffered saline (TBS) containing 0.1% Tween and incubated with the appropriate antibodies. Detection was performed via ECL Prime (Amersham Biosciences). Antibodies against the following were obtained from Cell Signaling Technology (Danvers, MA, USA) and were used at a dilution of 1:1000: MET (#3148), p-MET Y1349 (#3121), pMEK S217/221 (#9121), ERK1/2 (#3493), p-ERK1/2 T202/Y204 (#9106), HSP90 (#4874), PARP (#9546), horseradish peroxidase (HRP)-conjugated anti-mouse (#7076) and HRP-conjugated anti-rabbit (#7074). The following antibody was obtained from EMD Millipore (Burlington, MA, USA): RAS (05-516, 1:2000 dilution). Detection was performed via ECL Prime (Amersham Biosciences).
Cell-free DNA analysis
Samples were shipped to a Clinical Laboratory Improvement Act (CLIA)-certified, College of American Pathologists–accredited laboratory (Guardant Health, Redwood City, CA). cfDNA was extracted from whole blood collected in 10-mL Streck tubes. After double centrifugation, 5–30 ng of cfDNA was isolated for digital sequencing of either a 70 or 73-gene panel (Supplementary Table 6) as previously described (15). Only those genes in common to these two panels were included in subsequent analysis. Nonsynonymous mutations were further processed with the R statistical computing program (version 3.3). Variants with unknown or neutral predicted functional significance were filtered prior to analysis as previously described (11). Mutations previously reported as associated with clonal hematopoiesis were also excluded (16). Residue numbering was standardized to MET UniProtKB-P08581.
Next Generation Sequencing
Tumor sample next generation sequencing (NGS) was performed in CLIA-approved laboratories. The Foundation One and Foundation ACT assays are commercially available assays which were used in the clinical standard-of-care setting. The UCSF500 assay sequences 479 cancer-associated genes to target 200X coverage, utilizing sequencing of a PBMC sample, target 100X coverage, as a control. The University of Florida GatorSeq NGS assay, utilized for both tumor tissue and PBMC analysis, performs sequencing of 76 cancer-associated genes with target 500X coverage (Supplemental Table S6). Germline mutations were subtracted utilizing sequencing of buccal swab samples to a target depth of 100X.
Droplet Digital PCR
Isolated genomic DNA extracted from FFPE was amplified using ddPCR Supermix for Probes (Bio-Rad) using KRAS and MET assays (PrimePCR ddPCR Mutation Assay, Bio-Rad, and custom-designed). DNA template (8 μL) was added to 10 μL of ddPCR Supermix for Probes (Bio-Rad) and 2 μL of the primer/ probe mixture. This 20-μL reaction mix was added to a DG8 cartridge together with 70 μL of Droplet Generation Oil for Probes (Bio-Rad) and used for droplet generation. Droplets were then transferred to a 96-well plate (Eppendorf) and then thermal cycled with the following conditions: 5 minutes at 95°C, 40 cycles of 94°C for 30 seconds, 55°C for 1 minute followed by 98°C for 10 minutes (ramp rate 2°C/second). Droplets were analyzed with the QX200 Droplet Reader (Bio-Rad) for fluorescent measurement of FAM and HEX probes. Gating was performed based on positive and negative controls, and mutant populations were identified. The ddPCR data were analyzed with QuantaSoft analysis software (Bio-Rad) to obtain fractional abundance of the mutant DNA alleles in the wild-type (WT)/normal background. The quantification of the target molecule was presented as number of total copies (mutant plus WT) per sample in each reaction. Fractional abundance is calculated as follows: F.A. % = (Nmut/(Nmut + Nwt)) × 100), where Nmut is number of mutant events and Nwt is number of WT events per reaction. Multiple replicates were performed for each sample. ddPCR analysis of normal control genomic DNA (gene fragment obtained from IDT) and no DNA template (water) controls was performed in parallel with all samples, including multiple replicates as contamination-free controls.
Statistical Analysis
Pairwise sample group comparisons for cfDNA analysis were carried out using a two-tailed Fisher’s exact t-test, with Benjamini-Hochberg correction for multiple comparisons, using a false discovery rate of <0.2%. For cell viability curves, comparisons were performed using the two-sided student t-test, with significance threshold of p-value < 0.05.
Author contributions
Conception and design: J. Rotow, C. Blakely, T. Bivona
Development of methodology: P. Gui
Acquisition of Data: J. Rotow, V. Raymond, R. Lanman, F. Kaye, N. Peled, F Fece de la Cruz, I Yeh, B. Bastian, P Starostik, K Newsom, V Olivas, B. Nadres, R. Corcoran
Analysis and interpretation of data: J. Rotow, P. Gui, W. Wu, E Collison, C. McCoach, A. Wolff, J. Fraser; Writing, review and or revision of the manuscript: J. Rotow, P. Gui, W. Wu, F. Fece de la Cruz, R. Corcoran, C. Blakely, T. Bivona
Administrative, technical, or material support: V. Raymond, R. Lanman