PhD Course “Clinical Pathophysiology”
Director: Prof. Fulvio Basolo
Doctorate Thesis
Development of a pharmacogenetic-based approach for
prevention of fluoropyrimidine-related toxicities and diagnosis
of resistance to targeted treatments in cancer patients
Candidate Tutor
Dott.ssa Marzia Del Re Chiar.mo Prof. Fulvio Basolo
Table of contents
Page
Preface – Pharmacogenetics as a branch of pharmacology 3
Section I
Chapter 1 – Prevention of fluoropyrimidine-related toxicities
Development of a pharmacogenetic-based approach for prevention of fluoropyrimidine-related toxicities
5
Section II
Development of a pharmacogenetic-based approach for diagnosis of resistance to targeted treatments in cancer patients
18 Chapter 1 – EGFR-mutated NSCLC
Contribution of KRAS mutations and c.2369C>T (p.T790M) EGFR to acquired resistance to EGFR-TKIs in EGFR mutant NSCLC: a study on circulating tumor DNA
Patients with NSCLC may display a low ratio of p.T790M vs. activating EGFR mutations in plasma at disease progression: implications for personalised treatment
19
Chapter 2 – ALK-mutated NSCLC
Detection of ALK and KRAS mutations in circulating tumor DNA of patients with advanced ALK-positive NSCLC with disease progression during crizotinib treatment
37 Chapter 3 – SCLC as a mechanism of resistance to EGFR-TKIs
Primary resistance to osimertinib due to SCLC transformation: issue of T790M determination on liquid re-biopsy
45 Chapter 4 – Circulating tumor nucleic acids to monitor immunotherapy response
PD-L1 mRNA expression in plasma-derived exosomes is associated with response to anti-PD-1 antibodies in melanoma and NSCLC.
60 Chapter 5 – Prostate cancer
The detection of androgen receptor splice variant 7 in plasma-derived exosomal RNA strongly predicts resistance to hormonal therapy in metastatic prostate cancer patients
69 Chapter 6 – Pancreatic cancer
Early changes in plasma DNA levels of mutant KRAS as a sensitive marker of response to chemotherapy in pancreatic cancer
80
Acknowledgments 93
Preface – Pharmacogenetics as a branch of pharmacology
The differences in response to medications both in the terms of lack of response and adverse reaction, and the hope to have a good treatment strategy for patients, are the call for a personalized medicine. To better understand of the relationship between human genetics and drug effects, pharmacogenetics has emerged as a branch of pharmacology, in order to identify genetic variants in candidate genes involved in drug metabolism, transport or molecular targets/pathways [1].
Pharmacogenetics has been defined as the study of variability in drug response due to heredity, while it was usually referred to genes determining drug metabolism, it is now used also for drug response [2]. Based on the NCI Dictionary of Cancer Terms, pharmacogenetics is “The study of how a person’s genes affect the way he or she respond to drugs. Pharmacogenetics is being used to learn ahead of time what the best drug or the best dose of a drug will be for a person”. An area of great interest for pharmacogenetics is in the field of anticancer therapy.
In oncology, genetic variants can be found either in the germline genome as germline variations or in the tumor genome as somatic mutations, both of which are the subjects of PGx research. A germline variation is any detectable and heritable variation in the lineage of germ cells, such as genetic polymorphisms in genes encoding drug-metabolizing enzymes; while somatic mutations are accidental changes in a genomic sequence of DNA, which may cause cancer. While germline variations could potentially predict drug efficacy and toxicities, somatic mutations are often used to optimize choice of chemotherapeutic agent to improve efficacy [2].
On these premises, the present work is divided into two parts, the first is aimed on the development a pharmacogenetic-based approach for prevention of
related toxicities on germinal DNA and the secondo ne on the diagnosis of resistance to targeted treatments in cancer patients.
Section I
Chapter 1 - Prevention of fluoropyrimidine-related toxicities
Development of a pharmacogenetic-based approach for
prevention of fluoropyrimidine-related toxicities
Data presented in this chapter were published as original articles in:
Del Re M, Michelucci A, Di Leo A, Cantore M, Bordonaro R, Simi P, Danesi R. Discovery of novel mutations in the dihydropyrimidine dehydrogenase gene associated with toxicity of fluoropyrimidines and viewpoint on preemptive pharmacogenetic screening in patients. EPMA J. 2015 Sep 2;6(1):17
Del Re M, Quaquarini E, Sottotetti F, Michelucci A, Palumbo R, Simi P, Danesi R, Bernardo A. Uncommon dihydropyrimidine dehydrogenase mutations and toxicity by fluoropyrimidines: a lethal case with a new variant. Pharmacogenomics. 2016;17(1):5-9
Data presented in this chapter were reviewed in:
Del Re M, Restante G, Di Paolo A, Crucitta S, Rofi E, Danesi R. Pharmacogenetics and Metabolism from Science to Implementation in Clinical Practice: The Example of Dihydropyrimidine Dehydrogenase. Curr Pharm Des. 2017;23(14):2028-2034 Danesi R, Del Re M, Ciccolini J, Schellens JHM, Schwab M, van Schaik RHN, van Kuilenburg ABP. Prevention of fluoropyrimidine toxicity: do we still have to try our patient's luck? Ann Oncol. 2017 Jan 1;28(1):183
Introduction
Fluoropyrimidine-based chemotherapy is the most widely used treatment for several solid tumors, including gastrointestinal, head and neck, pancreatic and breast cancers [3], and 5-fluorouracil (5-FU) and its prodrug capecitabine are the backbone of many combined regimens. However, despite their clinical benefit, fluoropyrimidines are associated with severe adverse drug reactions (ADRs), including gastrointestinal and hematological toxicities and hand-foot syndrome (HFS), which sometimes may also be life-threatening [4]. Toxicities require dose modifications and may limit treatment efficacy, because of dosing delay or reduction and/or treatment discontinuation. In this context, there is a critical need for the identification of predictive biomarkers for fluoropyrimidine-related toxicities, particularly in the adjuvant therapy setting [5]. Several enzymes are involved fluoropyrimidine metabolism, generating many intermediate metabolites, however, the rate-limiting step is dependent on dihydropyrimidine dehydrogenase (DPD). DPD metabolizes about the 80% of the administered fluoropyrimidines into the inactive 5-fluoro-5,6-dihydrouracil (5-FDHU) [6]. If DPD is not functional or has reduced activity, the amount of 5-FU increases, leading to 5-FU-related ADRs [6]. The major cause of DPD deficiency is the presence of mutations within the encoding gene DPYD, affecting splicing processes, transcription and enzyme activity [7] (Figure 1).
Figure 1. DPD deficiency and 5-FU toxicity [8]
Many DPYD variants have been discovered [8, 9], but most of them do not impair enzyme activity or their functional effect is unclear, with the exception of the splice site mutation in intron 14 (c.1905+1G>A) and the non-synonymous variant c.2846A>T (p.D949V), strongly associated with partial or complete loss of enzymatic activity and severe ADRs [10, 11] (Figure 2).
Figure 2. DPYD polymorphisms reported in literature [8]
Several attempts have been made to figure out the best approach to assess DPD deficiency and reduce the risk of toxicity [12]. However, the issue is still on debate and contrasting indications regarding the implementation of DPYD test in clinical practice have been issued [13-15]. For these reasons, this research was aimed at provide further evidences on the role of DPYD assessment by evaluating a large cohort of patients to discover which mutations should be tested to reduce the risk of ADRs and avoid unjustified costs of screening extremely rare DPYD genotypes.
Materials and Methods
Study design and patients
To evaluate the possible association of DPYD variants to treatment-related ADRs, a test group of 200 patients affected by gastrointestinal, pancreatic, head and neck and breast cancers, receiving fluoropyrimidine-based treatments was enrolled (cohort 1). All patients suffered from ADRs (CTCAE v.4 G≥3 hematological or G≥2 gastrointestinal ADRs). In this cohort, the pharmacogenetic analysis consisted on the sequencing of DPYD exons and flanking regions by standard Sanger approach, to identify all the variants potentially associated with ADRs. In addition, 982 patients with ADRs receiving fluoropyrimidine-based treatment for gastrointestinal, pancreatic, head and neck and breast cancers (cohort 2) were enrolled to confirm the presence of variants identified in cohort 1 and assess their frequency. Finally, the same DPYD variants selected in cohort 1 were also examined in a control population of 272 subjects (cohort 3) displaying optimal tolerability to treatment (no dose reduction, treatment delay or discontinuation). In cohorts 2 and 3, pharmacogenetic analysis was performed by the TaqMan® SNP Genotyping Assay (Life Technologies, Carlsbad, CA).
The study was approved by the Ethics Committee of Pisa University Hospital and conducted in accordance to the principles of the Declaration of Helsinki; all patients gave their signed informed consent before blood collection and DNA analysis.
Statistics
Qualitative characteristics of patients are described by frequency distributions, both absolute and relative, while quantitative characteristics are described by median and interquartile ranges (IQR). The association between DYPD variants and ADRs was evaluated by logistic regression, adjusting for age and gender, and a p-value less than 0.05 was considered significant. Data were analysed by the SPSS statistical software (IBM SPSS Statistics 20; IBM Corporation, NY, USA).
Results
A total of 1454 patients were enrolled in the study; 802 (55.2%) patients were female and 652 (44.8%) male, median age was 63 years (cohort 1 IQR: 11; cohort 2 IQR: 14; cohort 3 IQR: 10). The analysis of DPYD variants in cohort 1 was conducted in 200 patients (113 male [56.5%] and 87 female [43.5%]), suffering from fluoropyrimidine-related ADRs. The following variants were found: c.85T>C, c.496A>G, c.1601G>A, c.1627A>G, c.1896T>C, c.1905+1G>A, c.2194G>A and c.2846A>T. Table 1 reports characteristics of patients. The type and frequency of toxicities and genotypes are reported in table 2. At least one genetic variant potentially associated with enzyme deficiency was found in 87.1% of patients of this cohort (table 3). Gastrointestinal (G≥2) and hematological toxicities (G≥3) were present in 91.5% and 58% of patients, respectively. G≥2 hand-foot syndrome developed in 13% of patients.
Cohort 2 was composed by 982 patients (590 female [60.1%] and 392 male [39.9%]). Gastrointestinal (G≥2) and hematological toxicities (G≥3) were present in 69.7% and 46.6% of patients, respectively.
Table 1. Characteristics of patients of cohorts 1, 2 and 3. (1) Abbreviations listed as per NCI Thesaurus v. 16.08e (release 2016-08-29).
Table 2. Adverse drug reactions (ADRs) observed in patients of cohorts 1 and 2
The frequencies of the DPYD variants c.496A>G, c.1601G>A, c.1627A>G, c.1896T>C, c.1905+1G>A, c.2194G>A and c.2846A>T were comparable with cohort 1 and are reported in table 3. Due to the high frequency of the heterozygous genotype (48%), the c.85T>C was not further examined in cohorts 2 and 3. No significant correlation was found between each DPYD polymorphism and organ-specific toxicity, with the exception of the c.2194GA and AA genotypes, which were associated with higher risk of severe hand-foot syndrome and hematological ADRs, expecially when 5-FU or capecitabine were given in combination with myelotoxic drugs, i.e., epirubicin/cyclophosphamide and docetaxel/cisplatin (p=0.003).
Finally, 272 control patients (147 male [54%] and 125 female [46%], cohort 3) receiving
ADRs Cohort 1 Cohort 2
Gastrointestinal Grade ≥2 Grade ≥2
Nausea/Vomiting 21.5% 16%
Diarrhea 51% 39.7%
Stomatitis 19% 14%
Dermatological Grade ≥2 Grade ≥2
Hand-foot syndrome 13% 9.3%
Hematological Grade ≥3 Grade ≥3
Fever 4.5% 2.2% Leucopenia 16.5% 12.3% Neutropenia 21.5% 17.4% Febrile neutropenia 7% 4.7% Anemia 3.5% 4.2% Thrombocytopenia 5% 5.8%
interruption, were enrolled (tables 1 and 2) and the pattern of genotypes is reported in table 3. The strong association of c.1905+1G>A and c.2846A>T with toxicity was demonstrated by the absence of variant alleles in cohort 3 (table 3). The frequency of c.2194A allele in the population with ADRs (cohort 2, table 3) was significantly higher than in cohort 3 and associated with ADRs (OR=1.9; table 3). On the contrary, c.496A>G showed a borderline association with ADRs (OR=1.4), while c.1601G>A, c.1627A>G and c.1896T>C played no role in fluoropyrimidine toxicities (OR<1).
Discussion
The availability of a panel of DPYD mutations strongly associated with clinically significant DPD deficiency would be of significant value for the optimal management of patients administered fluoropyrimidines. Indeed, an extensive search of genetic variants of DPYD associated with enzyme deficiency and poor-metabolizer status has been performed and several genotypes were identified [7, 16]. Our study confirmed the well-known role of the c.1905+1G>A and c.2846A>T (p.D949V) mutations in DPD deficiency and fluoropyrimidine-associated severe toxicities. In agreement with the data of the present work, three meta-analyses confirmed the clinical relevance of c.1905+1G>A and c.2846A>T mutations as risk factors for the development of severe ADRs [10, 17, 18]. An additional meta-analysis also found an association between severe toxicities and the non-synonymous mutation c.1679T>G (p.I560S) as well as the non-synonymous variant c.1236G>A in complete linkage with HapB3 [19], a haplotype containing three novel intronic polymorphisms (IVS5+18G>A, IVS6+139G>A and IVS9-51T>G) [20, 21]. However, c.1679T>G and c.1236G>A/HapB3 were not found in the present study, suggesting a population-specific relevance. Several other variants have been associated with impaired enzyme activity and/or linked to fluoropyrimidine toxicities, although with weaker evidence. These variants have been defined as "probable reduced function" or "unclear function" based on the low level of evidence linking these genotypes to fluoropyrimidine toxicity or very low detection rate in the population. Such uncommon mutations are reported by the literature, e.g., c.257C>T, c.1850C>T [22] and c.2509-2510insC, c.1801G>C, and c.680G>A [23], but were not found in the present study, while the common non-synonymous variants c.85T>C (p.R29C) [24] and c.496A>G (p.M166V) [25] previously associated with the risk of severe toxicities alone or in association as haplotypes, were not liked to ADRs in the present study. Therefore, their role as markers of
toxicity remains unproven. In the present study we found a significant association of the non-synonymous variant c.2194G>A (p.V732I) with severe toxicity by fluoropyrimidines. The results of the present work add additional information on the debate on this pharmacogenetic marker. Despite the c.2194A allele seems to be relatively common, conflicting results have been reported concerning its influence on DPD enzyme activity. Initial reports seem not to attribute to c.2194G>A a significant role in fluoropyrimidine toxicity [26, 27]; the presence of this allele did not correlate with the reduced DPD catalytic activity in colorectal cancer patients. However, the small number of subjects carrying this variant and the even lower number of homozygous mutants would probably have affected such a conclusion. On the contrary, another study found a strong relationship of c.2194G>A with leukopenia (OR=8.17) and neutropenia (OR=2.78) [28] and an additional work also positively associated this variant with diarrhea [21]. A study found that c.2194G>A showed weak evidence for association with reduced DPD activity in the African-American cohort and c.2194GA heterozygous patients showed a 29% reduction in DPD activity compared to the wild-type, although the linkage with c.557A>G (p.Y186C) may have played a prominent role [29]. In the study by Schwab et al. [30] the role of c.2194G>A was not considered significant, but this result may have been affected by the number of carriers of this variant, which was not clear from the data reported. In addition to this, another study on the role of selected DPYD variants on treatment tolerability showed no association between G≥3 or greater toxicity and c.2194G>A, but this negative result may have been affected by the small subgroup of patients (95 subjects) [31]. Finally, a last publication in the field examined a cohort of 1545 patients who participated - from December 2005 to November 2009 - in the Pan-European Trials in Alimentary Tract Cancer (PETACC)-8 randomized phase 3 clinical trial and found a significant association of ADRs and c.2194G>A [32]. In more detail, statistically
significant associations were found between G≥3 ADRs by 5-fluorouracil and c.2194G>A variant (OR=1.7; P <0 .001). Moreover, G≥3 hematologic adverse events were associated with c.2194G>A (OR=1.9), and with G≥3 neutropenia (OR=1.8) [32], in agreement with the present study. The association of c.2194G>A with the occurrence of G≥3 AEs by 5-fluorouracil and overall hematologic adverse events was validated in an independent cohort of 339 patients with metastatic colorectal cancer receiving FOLFOX4 in the Fédération Francophone de Cancérologie Digestive 2000-05 phase 3 trial [32]. Although other mutations of DPYD may represent a risk for patients, their extremely low frequency does not suggest their inclusion in routine preemptive screening [13, 32, 33]. Interestingly, a meta-analysis conducted on 7 cohort studies, with a total of 946 colorectal cancer patients receiving 5-FU chemotherapy, found a significant association between the c.2194G>A polymorphism and the incidence of bone marrow suppression (p<0.001) and gastrointestinal ADRs (p<0.05) in colorectal cancer patients [17].
Despite the large number of published papers on DPD, the present study is one of the few addressing the issue of DPYD variants and treatment safety in a large population: 1182 patients suffering from severe toxicities and 272 patients with optimal tolerability to treatment. It is still a matter of debate when to screen subjects candidate to a fluoropyrimidine treatment: in some centers, patients are prospectively screened and dose reductions are made, if necessary. In other centers patients are only screened after an adverse event has occurred, thus abolishing the advantage of a preemptive genotyping to reduce the deleterious consequences of administering a fluoropyrimidine in a poor metabolizer. In conclusion, recommendations are available on which variants to test and which dose adjustement of the fluoropyrimidine should be adopted and include c.1236G>A/HapB3, c.1679T>G, c.1905+1G>A, c.2846A>T [13, 34]. The present article provides evidence that variants with previously unknown significance, including c.85T>C,
c.496A>G, c.1601G>A, c.1627A>G and c.1896T>C have no clinically meaningful significance and should not be analysed for preemptive purposes. On the contrary, c.2194G>A should also be examined, particularly in patients being administered a fluoropyrimidine in combination with myelotoxic drugs; a dose reduction in homozygous variant patients and a close monitoring of heterozygous subjects for ADRs are thus advisable.
Section II
Development of a pharmacogenetic-based approach for diagnosis of
resistance to targeted treatments in cancer patients
Preface
Intratumor heterogeneity may drive tumor adaptation and compromise the efficacy of personalized medicine approach. The heterogeneity degree both within a tumor (intratumor heterogeneity) and between tumors (intertumor heterogeneity) is only partially known. Such heterogeneity impacts on a personalised medicine approach, considering that the somatic mutational landscape may not be entirely represented by a single tumor biopsy. Moreover, intratumor heterogeneity may contribute to the acquisition of drug resistance and tumor progression by facilitating sub-clonal selection [35]. This phenomenon has been clearly demonstrated in several solid tumors, in which different treatments can differently act a clonal selection inducing acquired resistance to treatment.
The second part of this thesis, addresses how a pharmacogenetic-based approach using circulating tumor nucleic acids may be used for diagnosis of resistance to treatments in patients affected by lung, prostate and pancreatic cancer.
Chapter 1 – EGFR-mutated NSCLC
Contribution of KRAS mutations and c.2369C>T (p.T790M) EGFR to
acquired resistance to EGFR-TKIs in EGFR mutant NSCLC: a study on
circulating tumor DNA
Patients with NSCLC may display a low ratio of p.T790M vs. activating
EGFR mutations in plasma at disease progression: implications for
personalised treatment
Data presented in this chapter were published as original articles in:
Del Re M, Bordi P, Petrini I, Rofi E, Mazzoni F, Belluomini L, Vasile E, Restante G, Di Costanzo F, Falcone A, Frassoldati A, van Schaik RHN, Steendam CMJ, Chella A, Tiseo M, Morganti R, Danesi R. Patients with NSCLC may display a low ratio of p.T790M vs. activating EGFR mutations in plasma at disease progression: implications for personalised treatment. Oncotarget. 2017 Feb 21;8(8):13611-13619
Del Re M, Tiseo M, Bordi P, D'Incecco A, Camerini A, Petrini I, Lucchesi M, Inno A, Spada D, Vasile E, Citi V, Malpeli G, Testa E, Gori S, Falcone A, Amoroso D, Chella A, Cappuzzo F, Ardizzoni A, Scarpa A, Danesi R. Contribution of KRAS mutations and c.2369C>T (p.T790M) EGFR to acquired resistance to EGFR-TKIs in EGFR mutant NSCLC: a study on circulating tumor DNA. Oncotarget. 2017 Feb 21;8(8):13611-13619
Data presented in this chapter were reviewed in:
Bordi P, Del Re M, Danesi R, Tiseo M. Circulating DNA in diagnosis and monitoring EGFR gene mutations in advanced non-small cell lung cancer. Transl Lung Cancer Res. 2015 Oct;4(5):584-97
Del Re M, Rofi E, Restante G, Crucitta S, Arrigoni E, Fogli S, Di Maio M, Petrini I, Danesi R. Implications of KRAS mutations in acquired resistance to treatment in NSCLC. Oncotarget. 2017 Dec 21. doi: 10.18632/oncotarget.23553
Introduction
Activating mutations of epidermal growth factor receptor (actEGFR) predict sensitivity to tyrosine kinase inhibitors (TKI) in non-small cell lung cancer (NSCLC). Despite a very high response rate to EGFR-TKIs as first-line treatment (erlotinib, gefitinib or afatinib), tumors invariably progress after a median of 9 to 13 months [36-38]. The discovery of the molecular basis of the acquired resistance to TKIs [39] and its application to monitor treatment, may improve patients management by discontinuing ineffective treatments, towards most appropriate second line options. EGFR signaling is maintained in most cases developing secondary resistance [40] suggesting that additional molecular mechanisms can bypass EGFR-TKI inhibition reactivating tumor cells proliferation. In this context, heterogeneity is indeed a challenge for the development of selective pharmacologic treatments. Multiple evidences highlight that cells driving resistance in NSCLC are selected during treatment with EGFR-TKI [41], thus, several mechanisms of acquired resistance to EGFR-TKI have been described after progression, including p.T790M EGFR ( ~50% of patients) [41], MET (5–15%) [42] or HER2 (12%) [43] amplifications, PIK3CA (4.1%) [44], KRAS or BRAF (1%) [45] mutations or transformation into small cell histology (3%) [46].
The missense p.T790M is found in approximately 50-60% of NSCLCs after progression to EGFR-TKIs [47]; it is located within the gatekeeper residue of EGFR and increases the affinity of the receptor to ATP [48], strongly reducing the pharmacologic activity of first generation EGFR-TKIs [49]. Approximately 0.32% to 78.95% of patients with NSCLC harboring actEGFR display p.T790M before administration of EGFR-TKI, although this percentage is variable according to test sensitivity [50]. It is therefore hypothesized that the selective pressure of EGFR-TKIs may select and favor the growth of p.T790M sub-clones, leading to acquired resistance.
Although data on p.T790M at progression to first-line EGFR-TKI or pretreatment [51-53] are available, the information on p.T790M levels and activating EGFR mutation in patients progressing to first-generation EGFR-TKI are lacking.
Moreover, approximately 15–25% of patients with NSCLC have a KRAS mutation (mutKRAS), resulting in constitutive activation of the KRAS signaling pathways. Mutant KRAS has been found as a negative predictor of response to EGFR-TKIs in EGFR wild type (wtEGFR) NSCLC patients [54]. Since NSCLC heterogeneity can drive the therapeutic decisions, tissue availability is increasingly recognized as a crucial issue. Unfortunately, the location of the tumor and the risk of complications are serious limitations to re-biopsies in NSCLC [55]. Alternatively, the detection of somatic mutations in cell-free tumor DNA (cftDNA) released in plasma could be instrumental for a better understanding of the genetic modifications driven by the selective pressure of drug treatments and to monitor tumor dynamics [56].
The aims of this research have been:
1) to document the relationship between plasma levels of actEGFR mutations and p.T790M at the time of progression to first-line EGFR TKIs in patients with NSCLC and provide initial information on the dynamics of cftDNA changes upon treatment with osimertinib
2) to investigate the appearance of KRAS mutations during first/second-generation EGFR-TKI treatments and their contribution to drug resistance.
Materials and Methods
A total of 82 NSCLC patients with actEGFR were enrolled in this study and divided into two cohorts of 49 (Cohort I) and 33 (Cohort II) patients, respectively analysed for the two aims of the project.
The study was submitted and approved by the Ethics Committee of Pisa University Hospital and conducted in accordance to the principles of the Declaration of Helsinki; all patients gave their signed informed consent before blood collection and cftDNA analysis.
Cohort I included 49 NSCLC patients carrying the actEGFR and the p.T790M mutations in cftDNA at the progression of disease after TKI tratemnt. These patients were consecutively enrolled between June and October 2016 from a population of subjects with the following characteristics: 1) stage IIIb/IV disease carrying EGFR activating mutations in tumor tissue at the time of initial diagnosis; 2) treatment with first/second-generation EGFR-TKIs (gefitinib/erlotinib/afatinib), as per approved indication and 3) clinical and imaging evidence of disease progression, as per standard practice. The analysis of EGFR mutations in primary tumors was performed by standard diagnostic procedures in use in each centre participating to this study (i.e., EGFR TKI response®, Diatech, Jesi, Italy; Therascreen®, Qiagen, Valencia, CA, USA).
In order to detect EGFR mutations, one blood sample was taken in each patient at the time of disease progression; in 5 subjects, 2 to 4 additional blood drawings were obtained to monitor EGFR mutations during treatment.
Cohort II included 33 NSCLC patients with actEGFR, receiving EGFR-TKI (gefitinib or erlotinib) as per approved indication were included in this study. The analysis of EGFR mutations in primary tumors was performed by standard diagnostic procedures in use in each centre participating to this study (i.e., EGFR TKI response®, Diatech, Jesi, Italy; Therascreen®, Qiagen, Valencia, CA, USA). KRAS mutations were not examined at the time of diagnosis as per standard practice. Plasma and/or re-biopsy samples were taken at the time of disease progression.
Plasma collection and cftDNA extraction
Six ml of blood were collected in EDTA and centrifuged at 4°C for 10 min at 3000 rpm within two hours after blood drawing. Plasma samples were stored at −80°C until analysis. cftDNA was extracted using a QIAmp Circulating nucleic acid Kit (Qiagen®, Valencia, CA, USA) from 1 to 3 ml of plasma following the manufacturer’s protocol and the DNA was eluted in 100 µl of buffer.
Analysis of cftDNA
The investigational part of this study included the assessment of codons 12, 13 (p.G13D) and 61 of KRAS and NRAS genes, EGFR p.T790M, BRAF V600E in cftDNA. Other mutations potentially associated with EGFR-TKI resistance were not examined because of the limited amount of cftDNA available. The analysis of cftDNA was performed by digital droplet PCR (ddPCR, BioRad®, Hercules, CA, USA) and ddPCR Mutation Assay (BioRad®). The analytic procedure was unable to discriminate the nature of the KRAS mutation because the analysis was performed with a ddPCR KRAS Multiplex assay.
PCR reactions were assembled as per manufacturer’s protocols and the droplet reader (BioRad®) was used for fluorescence signal quantification.
In cohort II the concordance between KRAS mutations and EGFR p.T790M was assessed on pairwise cftDNA and tissue DNA of 8 patients who underwent re-biopsy for diagnostic purposes. DNA was extracted from formalin-fixed paraffin-embedded biopsies using the QIAmp DNA Mini Kit (Qiagen®) and analyzed using conventional diagnostics as reported above. As a positive control for KRAS, the cftDNA from 30 patients with known KRAS mutated pancreatic cancer was used, while the DNA extracted from plasma of 43 healthy blood donors was employed as negative control for KRAS mutations and EGFR p.T790M.
Statistical analysis
Progression-free survival (PFS) was calculated from the first day of first-line EGFR-TKI treatment to radiological evidence of disease progression according to RECIST criteria; deaths for other causes than disease progression were considered events [36]. Before inferential statistics, an exploration phase was performed using box plots and scatter plots. We assessed three main quantitative factors: 1) concentration of the activating mutations (ex19del, p.L858R), 2) levels of p.T790M, and 3) their ratio (resistant/activating), KRAS mutations. In order to verify if the quantitative data were normally distributed we used the Kolmogorov–Smirnov and Shapiro–Wilks tests. A comparison between the four factors and PFS (all patients relapsed to first-line EGFR-TKI) was performed by a nonparametric correlation analysis (Spearman’s rank correlation coefficient), whereas Kruskal–Wallis and Mann–Whitney (two-tailed) tests were used to perform comparisons among the factor associated to the different treatment lines. Finally, the Wilcoxon test (two-tailed) has been performed to compare the number of copies/ml of plasmatic EGFR activating and p.T790M mutations in cohort I. A p<0.05 was considered as statistically significant; all statistical analyses were performed using the SPSS version 24 software.
Results
The characteristics of patients of cohort I are summarized in Table 4. The median PFS of the first-line EGFR-TKI treatment was 21 months (range 6-57 months). The activating mutations detected in cftDNA at the time of first-line EGFR-TKI progression were ex19del (85.7%) and p.L858R (14.3%).
Table 4. Characteristics of patients in cohort I
The median values of ex19del, p.L858R and p.T790M were 3,550 (range 130-3,390,000), 420 (range 170-671,000) and 500 copies/mL (range 80-194,000) respectively.
Figure 3. Plasma levels of EGFR ex19del and p.L858R and p.T790M at disease progression to first/second-generation EGFR-TKI. The analysis showed the very low ratio of p.T790M vs. actEGFR.
The median value of the p.T790M/actEGFR ratio was 0.26 (range, 0.0004-0.9) showing, overall, a very low ratio at the time of progression to first/second-line EGFR-TKI (p<0,0001, Figure 3). The statistical analysis of PFS vs. the ratio of p.T790M/actEGFR activating mutations showed no significant difference (p=0.075; Spearman’s rank correlation coefficient=0.256) suggesting that previous EGFR-TKI does not influence the ratio p.T790M/actEGFR and that a high ratio is not required to obtain resistance to treatment (Figure 4). No differences in mutation amounts were seen between fast and slow progressing patients comparing months of PFS.
Figure 4. Absence of correlation between p.T790M/actEGFR ratio and PFS to first/second-generation EGFR-TKI. A high ratio is not required to obtain resistance to treatment, but also very low amounts drive the resistance.
Moreover, the amount of actEGFR and p.T790M in plasma was monitored during treatment with osimertinib in 5 patients (2 with complete response, 2 with partial response and 1 with stable disease at first-tumor evaluation after 12 weeks) in order to gather information about the dynamics of EGFR mutational pattern during treatment.
Figure 5. Decreasing of mutated EGFR alleles (ex19del, p. L858R, p.T790M) in plasma during treatment with osimertinib. The amount of EGFR mutant clones in plasma decreased according to tumor response.
0 500 1000 1500 2000 2500 3000 3500 4000 0 1 2 3 4 5 6 7 8 9 M u tan t c op ie s/ ml Months ex19del p.T790M Complete( response 0 1000 2000 3000 4000 5000 6000 0 1 2 3 4 5 6 7 M u tan t c op ie s/ ml Months ex19del p.T790M
Partial response Complete.response
0 5000 10000 15000 20000 25000 30000 35000 40000 0 1 2 3 4 5 6 7 M u tan t c op ie s/ ml Months ex19del p.T790M Complete(response 0 50 100 150 200 250 300 350 400 450 0 1 2 3 4 5 6 7 8 9 M utan t c op ie s/ ml Months p.L858R p.T790M Partial response 0 10000 20000 30000 40000 50000 60000 0 1 2 3 4 5 6 7 8 9 M u tan t c op ie s/ ml Months ex19del p.T790M Stable disease
The amount of EGFR mutant clones in plasma decreased in parallel (ex19del and p.T790M or p.L858R and p.T790M), and, in general, the reduction of both ex19del and p.T790M was more marked than in the only patient bearing both p.T790M and p.L858R (Figure 5).
Cohort II included 33 patients, 20 (60.6%) were female and 13 (39.4%) male. Median age was 62 years (range 41 – 75); 32 patients were affected by a stage IV disease, while one was a stage IIIB NSCLC. The clinical characteristics of patients are reported in Table 5. The frequency of actEGFR on tissue was as follows: the 60.6% of patients showed ex19del, 30.3% was carrier of the p.L858R, 6.1% displayed the p.L747P and 3% presented ex19ins 3%. The 81.8% of subjects received gefitinib and 18.2% erlotinib; treatment was administered as first-line in 69.7% (including 2 as maintenance), second-line in 18.2% and third or further lines in 12.1% of patients. The majority of them (66.7%) presented partial response (PR) to TKI treatment and only 1 patient showed complete response (CR) (Table 5). Stable and progressive diseases (SD, PD) were observed in 12.1% and 18.2%, respectively. Median time to progression (TTP) was 13.6 months (95% Confidence Interval, CI, range 8.0 – 19.2 months) and median overall survival (OS) was 40.2 months (95% CI range 25.8–54.7 months). Table 6 reports the description of patients with actEGFR in their primary tumors and the percentages of EGFR p.T790M and KRAS mutant alleles in cftDNA at the time of EGFR-TKI progression.
Table 6. Types of activating mutations of MUTEGFR in primary tumor and % of
p.T790MEGFR and MUTKRAS alleles in cftDNA. “-“ Indicates wild-type allele.
In 16 patients (48.5%), a mutation in KRAS codon 12 was detected in cftDNA (Figure 6). In addition to this, the p.T790M was present in the cftDNA of 24 patients (72.7%). Interestingly, both the KRAS and p.T790M mutations were present in 13 patients (39.4%), while 3 (9.1%) and 11 (33.3%) subjects displayed only KRAS or p.T790M, respectively
Figure 6. Occurrence of KRAS and p.T790M mutations in cftDNA of NSCLC patients treated with EGFR-TKI
The association between smoking habit and the occurrence of KRAS mutations in cftDNA was investigated. Regarding the 11 patients with smoking history, 2 (18.2%) presented a KRAS mutation and 9 (81.8%) were wild-type (WTKRAS). On the contrary, of 22 non-smoking subjects, 14 (63.3%) were MUTKRAS and 8 (36.4%) WTKRAS. Fisher’s extact test revealed that non-smoking habit and MUTKRAS were significantly associated (p = 0.026).
In 8 patients, paired re-biopsies and cftDNA were available. The 8 re-biopsies were performed in a different tumor site with respect to the initial diagnosis, because of several factors, i.e., anatomical accessibility, new or progressing lesions. The analysis of re-biopsies by standard methods and ddPCR demonstrated p.T790M in 4 (standard) vs. 2 (ddPCR) samples and MUTKRAS in none (standard) vs. 3 (ddPCR) specimens. p.T790M
and KRAS mutations were detected in 7 and 5 cftDNA specimens, respectively.
The analysis of KRAS status by ddPCR in the biopsies at diagnosis revealed the presence of KRAS mutations in 2 patients, who were WTKRAS by standard method. Therefore, the percentage of patients developing a KRAS mutation as a mechanism of acquired resistance is 42% (14 patients). In terms of TTP and OS, there was no difference between these
patients and the others (p=0.19 and p=0.13, respectively). The concordance between tissue of re-biopsies (standard methods) and plasma (ddPCR), calculated by combining positive and negative results, was 62.5% and 37.5% for p.T790M and KRAS mutations, respectively (Table 7). Moreover, three paired samples found positive for the p.T790M on re-biopsies (standard methods) and on cftDNA (ddPCR), were negative by ddPCR on primary tissue (Table 7). The analysis of survival data stratified according to KRAS status in cftDNA, showed that TTP and OS were not statistically different, despite a trend towards a better survival in WTKRAS vs. mutant KRAS patients, i.e., TTP 14.4 vs. 11.4 months (p = 0.97) and OS 40.2 vs. 35.0 months (p = 0.56), respectively. Finally, samples were analysed also for codons 13 (p.G13D) and 61 KRAS and NRAS, and BRAF (p.V600E) and all the samples were found wild-type.
Discussion
Detection of mutant alleles in oncogenic driver genes can be performed in tissue as well as plasma, and the presented data demonstrate that ddPCR is a sensitive technology suitable for cftDNA detection and analysis [57]. p.T790M can be detected in plasma at a median time of 2.2 months prior to disease progression, is a predictive factor of resistance and disease outcome [58], and shows a complex biological behaviour, since p.T790M status in patients may change both temporally and spatially among tumor sites at least in part due to the selective pressure from EGFR-TKI. Interestingly, p.T790M status of NSCLC varies after EGFR-TKI discontinuation and may change from positive to negative, thus justifying a potential re-challenge with EGFR-TKI. It is not surprising that after further EGFR-TKI administration, p.T790M status may change from negative to positive again [59]. Previous work already provided initial evidence of the different amounts of p.T790M vs. actEGFR in plasma samples in 9 patients [60]. In addition, a recent report suggested a possible clinical importance for detection of p.T790M at low levels in plasma samples [61]. Longitudinal cftDNA analysis revealed an increase in plasma actEGFR, and p.T790M announcing TKI resistance in some patients, whereas in others the activating mutation increased but p.T790M remained suppressed. These findings demonstrate the role of tumor heterogeneity when drugs targeting a single resistance mechanism are administered to patients [53]. The results obtained from Cohort I show that patients progressing to first/second-generation EGFR-TKIs displayed a low amount of p.T790M compared to the actEGFR, suggesting that p.T790M is probably present in a minority of the tumor cell population after TKI failure or both. Therefore, the unanswered questions are:
1) how p.T790M can drive tumor progression being so low within the tumor? 2) is there another mechanism of resistance in addition to p.T790M?
Our data indicate that, even in presence of low amount of p.T790M, tumors respond to the third-generation TKI osimertinib, which is mainly active against p.T790M. In patients monitored during osimertinib treatment, p.T790M reduction was marked and similar to the decline of plasma levels of EGFR activating mutations. The differences in concentration between them and the p.T790M may confirm that the tumor is heterogeneous, and it is composed by 1) wild-type clones, 2) cells carrying both the EGFR activating and p.T790M mutations, 3) clones with only the original EGFR activating mutation. A distinct population of cells with the p.T790M only is unlikely to be present. In addition to this, the presented results provide evidence that actEGFR and KRAS mutations can coexist after the selective pressure of EGFR-TKI treatment. In our study, EGFR p.T790M was more frequent in p.L858R mutant patients than in ex19del ones (80% vs. 67%); on the contrary, KRAS mutations in cftDNA were detected in 55% of patients with ex19del vs. 30% of patients with p.L858R. However, this cohort is too small to draw any conclusion. While the role of KRAS mutations in primary resistance to EGFR-TKIs in molecularly unselected NSCLC is quite well established [62, 63], its development and role in acquired resistance to EGFR-TKIs in EGFR mutant patients has not been explored in detail. Therefore, we addressed this issue and a sensitive ddPCR-based platform was employed to investigate the presence of KRAS mutant alleles besides the well-known EGFR p.T790M. Due to its high sensitivity, ddPCR is able to identify small amounts of mutant alleles, however, many methodological issues need to be addressed prospectically, particularly the threshold level of both KRAS and p.T790M mutations to be considered clinically relevant. Eight patients underwent re-biopsy after tumor progression during EGFR-TKI, allowing a comparison between tissue and cftDNA. The detection of mutations in cftDNA but not in re-biopsy, using both standard methods and ddPCR, could suggest the presence of heterogeneity within metastatic sites or the known lower
performance of ddPCR in the presence of paraffin. Nevertheless, the detection of mutations in both plasma and tissue by ddPCR, but not by standard methods, could be due to the higher sensitivity of ddPCR analysis. Two patients, initially diagnosed as wild-type KRAS by standard method, were re-analysed by ddPCR and were found KRAS mutant in the primary biopsy, suggesting that the KRAS mutant clones co-existed with actEGFR since the beginning, as also demonstrated in previous reports [64, 65]. In these patients, KRAS mutations cannot strictly be considered as a mechanism of resistance, however, it could be possible that EGFR-TKI treatment may have favored the expansion of mutant KRAS-positive clones. However, conclusions cannot be drawn as pre-treatment cftDNA was not available. Our observation is a pivotal evidence of the presence of KRAS mutations in cftDNA of tumors with EGFR mutant tumors resistant to EGFR-TKIs. It remains to be determined if the presence of EGFR p.T790M and KRAS mutations coexist in the same tumor cell or arise in different sub-clones. Targeting KRAS proteins is still a challenge [66]. Theoretically, combined treatment with KRAS and EGFR inhibitors can be administered to patients to prevent MUTKRAS-dependent resistance or restore sensitivity to EGFR-TKIs, as recently demonstrated co-targeting EGFR and MEK [67]. To date, EGFR p.T790M remains the most important predictor of efficacy of third generation EGFR-TKIs.
In conclusion, the present work demonstrates the feasibility of detecting very low amounts of mutant alleles of EGFR and KRAS in plasma by a sensitive analytical approach and that these mutations are associated with tumor progression – and response to osimertinib – even though it may be a minority with respect to activating EGFR mutations. Further studies are warranted to gain additional knowledge on the interaction between EGFR and KRAS mutations in TKI-resistant NSCLC, and to determine the clinical consequences to be connected to cftDNA outcomes in plasma. However, nowadays, is often required a
treatment adaptation based on pharmacogenetic data, and these results provide strong evidence supporting the usefulness of cftDNA as a good predictive biomarker.
Chapter 2 – ALK-mutated NSCLC
Detection of ALK and KRAS mutations in circulating tumor DNA of
patients with advanced ALK-positive NSCLC with disease progression
during crizotinib treatment
Data presented in this chapter were published as an original article in:
Bordi P, Tiseo M, Rofi E, Petrini I, Restante G, Danesi R, Del Re M. Detection of ALK and KRAS Mutations in Circulating Tumor DNA of Patients With Advanced ALK-Positive NSCLC With Disease Progression During Crizotinib Treatment. Clin Lung Cancer. 2017 Nov;18(6):692-697
Data presented in this chapter were reviewed in:
Bordi P, Del Re M, Danesi R, Tiseo M. Circulating DNA in diagnosis and monitoring EGFR gene mutations in advanced non-small cell lung cancer. Transl Lung Cancer Res. 2015 Oct;4(5):584-97
Bordi P, Del Re M, Tiseo M. Crizotinib Resensitization by Compound Mutation. N Engl J Med. 2016 May 5;374(18):1790.
Introduction
A small group of NSCLCs is characterized by the presence of the anaplastic lymphoma kinase (ALK) fusion oncogene, which defines a distinct clinico-pathologic subset of NSCLC [68]. The identification of ALK translocation is crucial because positive tumors are highly sensitive to ALK-TKIs. Crizotinib is the first TKI tested in NSCLC patients with ALK rearrangement and has shown higher efficacy than standard chemotherapy. However, after a median of 9 to 10 months, disease progression (PD) inevitably occurs because of acquired resistance [69]. Several distinct mechanisms of resistance have been reported in the literature: appearance of a secondary mutations within the ALK tyrosine kinase domain (i.e., p.G1269A, p.I1174N, p.L1196M), amplification of the ALK fusion gene (alone or in combination with a secondary resistance mutation), and activation of alternative signaling pathways [69]. These include abnormalities in the epidermal growth factor receptor (EGFR), tyrosine-protein kinase kit (KIT), and insulin-like growth factor-1 receptor pathways and suggest the potential need for combination therapies to overcome resistance. Moreover, also pharmacokinetic mechanisms (ie, brain penetration) could be responsible for PD [69]. Notably, patients resistant to crizotinib can be sensitive to other TKIs of recent development such as ceritinib, alectinib, and brigatinib [69-71]. Also these new generation ALK TKIs realize a selective pressure on neoplastic clones, inducing acquired resistance [72]. The spectrum of ALK mutations responsible for acquired resistance in ALK-rearranged NSCLC is heterogeneous and none of the ALK TKIs under investigation present the formal indication for a specific resistance mutation. Interestingly, each ALK TKI shows different potency with respect to the point mutation, suggesting a possible clinical algorithm mutation-based in the future [69]. When the tissue is not available, the analysis of circulating tumor DNA (cftDNA) can be a valid alternative for the determination of activating and resistance EGFR mutations. Therefore, we evaluated
acquired resistance mutations associated with ALK TKIs on cfDNA after the evidence of progression in patients treated with crizotinib.
Patients and Methods
Patients
Twenty consecutive patients were enrolled from 2 Italian centers (Parma and Pisa University Hospitals). All patients had an ALK-rearranged advanced adenocarcinoma and underwent PD after crizotinib treatment. The diagnosis of the ALK rearrangement had been performed according to clinical practice using fluorescence in situ hybridization. Plasma was collected at the time of PD during crizotinib treatment (PD samples) and was analyzed for the presence of the point mutations in ALK and KRAS genes. In case of a PD sample positive for ALK or KRAS mutations, additional samples were collected during the next line of therapy (chemotherapy or different ALK TKI) at the moment of the radiological revaluations (post-PD samples). This study was conducted in agreement with the Declaration of Helsinki and was approved by the institutional ethical review board of both institutions.
Plasma collection and cftDNA analysis
Six milliliters of blood were collected in EDTA tubes and centrifuged for 10 minutes at 3000 rpm within 2 hours after blood drawing. Plasma samples were stored at -80C until the analysis. CftDNA was extracted using a QIAmp Circulating nucleic acid Kit (Qiagen, Valencia, CA) from 3 mL of plasma following the manufacturer’s protocol. cftDNA was analyzed for ALK (p.L1196M, p.G1269A, and p.F1174L) and KRAS (codons 12 and 13) point mutations by a ddPCR (BioRad, Hercules, CA), according to manifacturer’s protocol.
Results
Patient characteristics are summarized in Table 8. The analysis of cftDNA after the evidence of PD was performed before the start of new systemic treatment in 18 patients; 17 started second-generation ALK TKI treatment (12 brigatinib, 4 ceritinib, and 1 alectinib), whereas 1 patient received chemotherapy with carboplatin-paclitaxel (Table 8). Plasma was collected before stereotactic radiotherapy in 1 patient (patient 14) who was treated with continued crizotinib therapy and before best supportive care in another one (patient 20).
Table 8. ALK and KRAS Mutations in Circulating DNA of ALK+ NSCLC [73].
Secondary mutations in the ALK gene were identified in 5 patients (Table 8). Three patients (patients 6, 11, and 17) were carriers of the ALK p.L1196M, p.G1269A, and
p.F1174L associated with the KRAS point mutation p.G12D; 1 patient was a carrier of the p.L1196M alone (patient 12), and another one of the double ALK mutations p.L1196M and p.G1269A (patient 2; Figure 7). Seven patients presented an isolated KRAS point mutation in codon 12.
Figure 7. Modifications of plasmatic levels of ALK and KRAS mutations expressed in terms of mutation copy per milliliter at the time of PD during crizotinib treatment and during treatment with brigatinib.
Post-PD samples were collected from 3 of 5 ALK-mutated patients and all showed plasmatic decreases of ALK mutations, together with tumor regression (2 objective responses and 1 stable disease according to Response Evaluation Criteria in Solid Tumors version 1.1 criteria) evidenced on computed tomography (CT) scan after 2 months of brigatinib treatment. Post-PD samples of patients 11 and 17, presented with ALK p.G1269A and p.F1174L, associated with KRAS p.G12D mutation in PD samples, were not collected because of rapid clinical worsening due to further PD. After PD during crizotinib treatment, 7 KRAS-positive patients received brigatinib, 1 ceritinib, 1 patient was treated with carboplatin-paclitaxel, and the remaining patient received best supportive
care only. In 3 KRAS-positive patients (patients 3, 7, and 8) treated with brigatinib, KRAS mutation was no longer detectable in plasma DNA alongside with tumor response. In patient 6, who developed ALK p.L1196M and KRAS p.G12D during crizotinib treatment (PD sample), both mutations disappeared during treatment with brigatinib and the patient achieved tumor response. Post-PD samples were not available in patients who did not attend the clinic because of drug toxicity (patient 4) or clinical deterioration (patients 1, 10, 11, 17, and 20). Figure 7 show the example of patient 2, treated with crizotinib for 13 months; at PD, CT scan revealed brain and liver metastases. The patient refused liver rebiopsy and only the cfDNA analysis was performed. At the time of PD there were 2 ALK mutations (p.L1196M and p.G1269A) in plasma. The patient started brigatinib treatment and after 2 months was reported a partial response of all the lesions. Accordingly, with clinical and radiological benefit, there was a reduction of the p.L1196M mutation and the disappearance of the p.G1269A, whereas the KRAS p.G12D was acquired in cftDNA. In the subsequent collection of samples the KRAS mutation disappeared and the radiological partial response was maintained.
Discussion
Considering our previous results with cfDNA in EGFR TKI-resistant NSCLC patients and growing evidence about different mutations in the ALK kinase domain as responsible for acquired resistance to ALK TKIs, we planned this study [74, 75]. Even if a next-generation sequencing approach after PD during crizotinib treatment would have been more informative, by exploring a wider range of possible resistance mechanisms, we decided to perform ddPCR, a low-cost procedure characterized by higher sensitivity, to optimize the chances of detection of p.L1196M, p.G1269A, and p.F1174L ALK mutations, because they were the most commonly associated with acquired resistance to crizotinib [76]. Five
of 20 patients (25%) presented an ALK mutation in cfDNA at the time of the PD during crizotinib treatment. There were 2 cases who presented with p.L1196M and 1 with p.F1174L mutations, 1 with p.G1269A, and a case with p.L1196M as well as p.G1269A. Several mutations of the kinase domain of ALK have been identified in tumor biopsies [76, 77]. In the article published by Gainor et al, an analysis of correlation genotype sensitivity of ALK mutations and crizotinib, ceritinib, brigatinib, alectinib, and lorlatinib was performed. The results showed that p.L1196M, p.G1269A, and p.F1174L/C need higher concentrations of crizotinib to inhibit cell phosphorylation (half maximal inhibitory concentration (IC50) of 339.0 nM, 117.0 nM, and 115.0 nM, respectively), with respect to the lower ones needed by new ALK- TKIs. This important information put the basis on a mutational-based algorithm to treat ALK-positive patients. The frequency of ALK mutations reported in this study is lower than previously reported and can be explained with the high number of patients who presented brain-limited PD, reasonably driven by a pharmacokinetic mechanism. However, our data are consistent with results recently reported by Gainor et al, who showed that only 20% of crizotinib-resistant patients carried ALK mutations, contrary to > 50% of patients who had PD during second-generation ALK TKI treatment [76]. As previously reported, the presence of a double ALK mutation (p.L1196M and p.G1269A) was observed (patient 2). In our case compound mutation suggests a possible heterogeneity of acquired resistance. The contribution of KRAS mutations in this contest is still un-certain, even if some evidence exists about KRAS mutations as a potential cause of ALK TKI resistance [78, 79]. Our findings are limited by the small number of patients and by the absence of comparison with tissue rebiopsy. However, we confirmed, to our knowledge, for the first-time in plasma, that ALK and KRAS mutations are associated with acquired resistance to crizotinib in ALK-positive
NSCLC. In particular, we showed that ALK mutations can be isolated in plasma and that their monitoring could serve as a response parameter.
Chapter 3 – SCLC as a mechanism of resistance to
EGFR-TKIs
Primary resistance to osimertinib due to SCLC transformation: issue of
T790M determination on liquid re-biopsy
Data presented in this chapter were published as an original article in:
Minari R, Bordi P, Del Re M, Facchinetti F, Mazzoni F, Barbieri F, Camerini A, Comin CE, Gnetti L, Azzoni C, Nizzoli R, Bortesi B, Rofi E, Petreni P, Campanini N, Rossi G, Danesi R, Tiseo M. Primary resistance to osimertinib due to SCLC transformation: Issue of T790M determination on liquid re-biopsy. Lung Cancer 2018;115:21-27
Data presented in this chapter were reviewed in:
Di Paolo A, Del Re M, Petrini I, Altavilla G, Danesi R. Recent advances in epigenomics in NSCLC: real-time detection and therapeutic implications. Epigenomics. 2016 Aug;8(8):1151-67
Introduction
Histological transformation towards small cell lung cancer (SCLC) and transition to an epithelial-mesenchymal (EMT) phenotype represent less common mechanisms of acquired resistance. Findings of p.T790M mutation and SCLC transformation allow proper therapies: third-generation EGFR-TKI or appropriate chemotherapy, respectively. Currently, p.T790M testing could be performed using circulating-free DNA (cftDNA) and, if plasma genotyping is positive for resistance mutation, it may obviate the need for an invasive tissue re-biopsy. If plasma genotyping for p.T790M is negative, tissue re-biopsy is recommended considering the sub-optimal sensitivity (about 70%) of plasma testing [80]. Response rates and progression-free survival to osimertinib are similar whether p.T790M is detected in tissue or in plasma samples [80]. Despite the high concordance between p.T790M on liquid biopsy with tissue sample, some information about other potential resistance mechanisms, such as histological transformation in SCLC and EMT may be lost, leading to potential osimertinib failure when cftDNA is analysed only. Here, we report our experience of five patients treated with osimertinib after p.T790M confirmation on liquid biopsy who present a primary resistance ultimately attributed to SCLC transformation.
Materials and methods
Patients
All patients reported in this series were enrolled in two different Institutions (University Hospital of Parma and Careggi Hospital of Firenze) to participate in the ASTRIS trial (NCT02474355), a real world treatment study testing the efficacy of osimertinib (80 mg os die) in advanced p.T790M-positive NSCLC that progressed to prior EGFR-TKI. Main inclusion criteria were: stage IIIB/IV NSCLC with T790M mutation, ECOG performance status 0–2, adequate bone marrow and organ function, absence of cardiac abnormalities at
ECG. Main exclusion criteria were: prior treatment with osimertinib, previous history of interstitial lung disease, uncontrolled systemic disease and symptomatic brain metastases. After progression to first-line EGFR-TKI, patients were considered eligible to osimertinib if p.T790M positive on tissue or plasma samples. All patients signed informed consent form before any trial procedure and the trial was properly approved by local Ethics Committee of both Institutions. Enrolled patients underwent ophthalmological evaluation at baseline, blood tests plus urine analysis and physical examination every six weeks and radiological revaluation every 12 weeks. Plasma samples for cftDNA analysis were collected from all patients before osimertinib initiation. Three out of five patients, referred to Parma University Hospital, were also enrolled in a local protocol of cftDNA monitoring in patients undergoing targeted therapies. In these patients, samples for cftDNA analysis were taken at baseline and at the same time-points of imaging revaluation.
Molecular analysis of EGFR mutational status
Baseline EGFR mutations were assessed as part of diagnostic procedure by validated methods including Sequenom (Diatech Pharmacogenetics®, Italy) or therascreen® EGFR RGQ PCR kit (Qiagen®, Valencia, CA, USA). The analysis of EGFR p.T790M was performed on cftDNA. Six ml of blood were collected in EDTA and centrifuged twice for 10 min at 2000 × g within one hour after blood drawing; plasma samples were stored at −80 °C until analysis. cftDNA was extracted using a QIAmp Circulating nucleic acid kit (Qiagen®, Valencia, CA, USA) from 1 to 3 ml of plasma following the manufacturer’s protocol. The analysis of EGFR activating and resistance mutations on cfDNA was performed by a digital droplet PCR (ddPCR) and ddPCR Mutation Assay (BioRad®, Hercules, CA, USA). In three out of five patients, analyses on cftDNA were performed also with therascreen® EGFR RGQ PCR kit (Qiagen®, Valencia, CA, USA), following
the manufacturer's instructions, as a comparative analysis. Next-Generation Sequencing (NGS) was performed with TruSight Tumor 26 genes on the MiSeq platform (Illumina®, San Diego, CA) only for the liver biopsy of case 4 at the time of osimertinib progression, in order to look for EGFR p.T790M mutation with a highly sensitive method. In the same NGS panel was included also p53 gene (but not RB1) useful for the aim of this study.
p53 and Rb1 immunohistochemistry
Expression of p53 and Rb1 was evaluated with immunohistochemistry (IHC) in case 2 and 5. With regard to the tissue sample from case 2, after fixation in formalin solution and inclusion in paraffin, sections of neoplastic tissue with a thickness 3–5 µm were stained with hematoxylin/eosin (H&E) for conventional evaluation. After deparaffinization and rehydratation, sections were treated with 3% hydrogen peroxidase for 5 min. For antigen retrieval, sections were treated with pH 9 Tris-EDTA buffer for 30 min in water-bath at 98 °C. The same method was used for case 5, after treatment of cytological sample decolorated from May-Grumwald Giemsa (MGG). Sections were stained with the following primary antibodies: anti-p53 (clone D0-7, Roche) and anti-Rb1 (clone Rb1, Dako, Agilent, Santa Clara, CA, USA). The sections were immunostained with the polymeric system Ultraview DAB Detection Kit (Ventana-Roche, Tucson, AZ, USA) in accordance with manufacturer specifications. Diaminobenzidine (DAB) was used for staining development and the sections were counterstained with haematoxylin. Negative controls consisted of substituting normal serum for the primary antibody.
Results
We report on a series of five patients (two females and three males) diagnosed with lung adenocarcinoma carrying an EGFR activating mutation that received first- or second-generation EGFR-TKI and, at resistance, were positive for EGFR p.T790M mutation in plasma samples. Nonetheless, all the patients progressed on osimertinib at first tumor assessment after 12 weeks or less, and tissue re-biopsies revealed a switch to SCLC histology. In four cases, re-biopsies were obtained after progression to osimertinib, while one patient was re-biopsied right after progression to first-line EGFR-TKI (i.e. before osimertinib administration). Clinico-biological characteristics representative of individual patients’ disease course are summarized in Table 9.
Case 1
A never smoker 69-year-old male was symptomatic for increasing back pain. A Magnetic Resonance Imaging (MRI) evidenced multiple bone lesions. After radiological documentation of lung, adrenal, bone and lymph node lesions, a bronchoscopy allowed the diagnosis of lung adenocarcinoma, positive for EGFR exon 19 deletion (ex19del). The patient received palliative bone radiotherapy and subsequently started afatinib 40 mg os die, later reduced to 30 mg os die due to grade 3 skin rash. Positron Emission Tomography (PET) scan performed after three months showed significant response and therapy was continued until progression of disease, for a total amount of 10 months. Liquid biopsy was performed showing ex19del and p.T790M (Table 9). Osimertinib was then administered, being optimally tolerated. However, after two months, osimertinib was permanently discontinued because of further skeletal, hepatic, pulmonary and nodal metastatic spread. A new bronchoscopy was performed and histological assessment was consistent with SCLC; molecular analysis did not reveal any genetic alterations on EGFR. Patient died few
weeks later without receiving further anticancer therapies due to rapid clinical deterioration.
Case 2
A 70-year-old never smoker woman reported dry cough and dyspnoea. Chest X-ray showed left pleural effusion and the patient was submitted to thoracentesis, negative for neoplastic cells. Computerized Tomography (CT) scan revealed a mass in left lower lobe, millimetric solid nodules spread in both lungs, pleural thickening and bone metastases. The patient underwent thoracoscopic biopsy, permitting diagnosis of lung adenocarcinoma positive for EGFR ex19del. She started afatinib 40 mg os die, with partial response as best response. After nine months of treatment, CT scan revealed disease progression due to increasing lung nodules. The research of resistance mutation was performed through liquid biopsy showing p.T790M (Table 9). The patient started osimertinib but her clinical conditions worsened because of bone pain. Restaging after two months of treatment showed disease progression in liver, lungs and bones. Liver re-biopsy was diagnostic for SCLC, with the persistence of EGFR ex19del while lacking p.T790M. Patient conditions rapidly declined and died in few weeks.
Case 3
The patient, a 58-year-old woman with a history of previous light smoker, referred impaired concentration, fatigue and peripheral tremors. CT scan evidenced multiple brain metastases associated to a lung mass at the right hilum. Diagnostic Endobronchial Ultrasound (EBUS) was performed and biopsy of the primary lesion resulted positive for lung adenocarcinoma carrying EGFR p.L858R activating mutation. Patient underwent whole brain radiotherapy and then started afatinib 40 mg os die. One month later, dose
