Mixed responses to targeted therapy driven by chromosomal instability through p53 dysfunction and genome doubling

Abstract

The phenomenon of mixed/heterogenous treatment responses to cancer therapies within an individual patient presents a challenging clinical scenario. Furthermore, the molecular basis of mixed intra-patient tumor responses remains unclear. Here, we show that patients with metastatic lung adenocarcinoma harbouring co-mutations of EGFR and TP53, are more likely to have mixed intra-patient tumor responses to EGFR tyrosine kinase inhibition (TKI), compared to those with an EGFR mutation alone. The combined presence of whole genome doubling (WGD) and TP53 co-mutations leads to increased genome instability and genomic copy number aberrations in genes implicated in EGFR TKI resistance. Using mouse models and an in vitro isogenic p53-mutant model system, we provide evidence that WGD provides diverse routes to drug resistance by increasing the probability of acquiring copy-number gains or losses relative to non-WGD cells. These data provide a molecular basis for mixed tumor responses to targeted therapy, within an individual patient, with implications for therapeutic strategies.

Introduction

Up to 50% of all never-smokers who develop lung adenocarcinoma (LUAD) harbor tumors with mutations in the epidermal growth factor receptor (EGFR)^1,2. EGFR mutations are predominantly clonal, making this an optimal therapeutic target. Unfortunately, only a minority of patients have a lasting treatment benefit for more than two years^3,4,5. The median progression-free survival for patients receiving EGFR tyrosine kinase inhibition (TKI) therapy with osimertinib is 18.9 months and 10.1 months for patients with EGFR mutation-positive metastatic non-small cell lung cancer (NSCLC) when receiving treatment in the first and second line, respectively^6,7. Primary resistance, with no objective treatment response, is seen in 20% and 29% of patients receiving osimertinib as first- and second-line treatment, respectively⁷. Acquired resistance is defined clinically as systemic progression as measured by RECIST (Response Evaluation Criteria in Solid Tumors^8,9) after a period of initial response to EGFR TKI therapy¹⁰. Common resistance mechanisms include secondary alterations in EGFR itself, such as the “gatekeeper” T790M mutation in response to erlotinib treatment¹¹, as well as TKI bypass via alternative signaling pathways and/or somatic copy-number alterations (SCNAs), including amplifications of genes such as MET, ERBB2, KRAS, NRAS, and BRAF¹². However, in ~30% of patients, the mechanisms of acquired resistance remain unknown^13,14.

Mixed treatment responses, also known as heterogeneous responses, where responding and non-responding metastases are detected within the same patient, have been observed with varying frequency in many cancer types. For example, 11% of patients with BRAF^V600E-mutant melanoma or thyroid cancer had mixed responses after treatment with a BRAF inhibitor¹⁵, whereas 56% of patients with renal clear cell cancer treated with anti-angiogenic tyrosine kinase inhibitors displayed mixed treatment responses¹⁶. In the CAIRO I/II studies of patients with colorectal cancer and liver metastasis, a mixed response to therapy was associated with poorer survival outcome¹⁷.

When measuring response to therapy using current clinical RECIST version 1.1 guidelines⁹, the sum of the diameters of all target lesions is used⁹. Within a cohort of patients classified as having a stable disease or partial response to treatment, there will be some patients with mixed responses i.e., responding lesions and, at the same time, progression of other lesions. Current RECIST reporting criteria do not consider such discordant radiological responses¹⁸, nor do they conform to the standard definitions of acquired resistance to therapy since both responding and resistant lesions occur within the same patient simultaneously.

Only a limited number of studies have explored the prevalence of mixed responses to TKI in LUAD^18,19, and to our knowledge, none have investigated the underlying mechanistic basis of acquired resistance in this context. Since a single progressing lesion might contribute to systemic re-seeding, disease progression, early treatment failure and drug discontinuation, understanding the mechanisms of clonal diversification and intra-patient mixed tumor response dynamics may improve patient screening and therapeutic strategies²⁰.

TP53 is mutated in around 40% of all patients with NSCLC, and TP53 pathway perturbations in EGFR-driven tumors are associated with shortened progression-free (PFS) and overall survival (OS), in the context of treatment with first-, second-, or third-generation EGFR inhibitors²⁰. It was recently suggested that loss of TP53 function, together with other genetic events, facilitates the acquisition of EGFR TKI resistance mutations²¹. Mechanistically, TP53 loss of function permits the tolerance of chromosomal instability (CIN) and is enriched in whole genome-doubled (WGD) tumors^22,23,24. Moreover, studies have demonstrated that WGD results in rapid propagation of CIN and acquisition of SCNAs^25,26,27,28.

We hypothesized that TP53 loss together with WGD permits the rapid onset of CIN and SCNA acquisition, leading to more diverse tumor genotypes and phenotypes, thereby contributing to the radiologically observed mixed tumor responses within patients with clonal actionable driver events. We investigated this hypothesis in multiple clinical cohorts of patients with clonal EGFR-activating mutations treated with EGFR TKI, in genetically engineered mouse models (GEMMs) driven by clonal EGFR activating mutations with or without Trp53 loss, and in isogenic cell lines to examine mechanisms of resistance and cellular evolution under therapeutic pressure using functional models, whole-exome DNA sequencing, and single-cell DNA sequencing.

Results

Mixed clinical responses to TKI therapy are prevalent in EGFR-driven lung adenocarcinoma

There is limited information available on mixed responses to targeted therapies and cytotoxic chemotherapy in NSCLC as most studies only report data required to meet RECISTv1.1 criteria for response¹⁸. We used the Reiter and Vogelstein¹⁵ defined response parameters to distinguish homogeneous from mixed tumor responses in both human and murine datasets. Unlike RECISTv1.1, which defines progressive disease as a 20% increase in total diameter calculated as the sum of the diameter of all measured lesions, Reiter et al. defined response in individual lesions. A lesion was considered to respond if it shrank by at least 30% in diameter and stable if it did not grow more than 10% or reduce in size by more than 30%. Progression was defined by at least a 10% increase in lesion diameter. A homogeneous objective response to therapy was defined as having at least one lesion with a greater than 30% reduction in size, in combination with other lesions being stable (less than 10% growth) or reducing in size. The appearance of new lesions was not considered in the Reiter et al. definition of a mixed response. However, in the context of lesions that meet the criteria for a radiological response, we included the appearance of one or more new lesions in the mixed response classification, even if all other lesions were responding. If lesions within the same patient were assigned to both the response and progression criteria or the patient developed a new lesion, the patient was classified as having a mixed response to therapy.

All assessments of response were performed at a single time point where imaging was available and, unless specified otherwise, were performed at the first response assessment following treatment (12 weeks post-treatment ± 2 weeks). As we assessed radiological response on a lesion-by-lesion basis early in the course of treatment, we were able to identify non-responding lesions before a patient’s overall tumor response reached the clinical definition of acquired resistance¹⁰. Therefore, to distinguish between clinical definitions of primary and acquired resistance, designed to standardize criteria for clinical trial enrollment, we refer to the growth of an established lesion as the “development of resistance”. Similarly, we apply the same nomenclature to our murine data and the genetic aberrations associated with the development of resistance¹⁹.

These radiological response parameters were first applied to analyse the European Organization for Research and Treatment of Cancer (EORTC) RECIST database¹⁹, which contains response assessments from patients in phase II and phase III clinical trials. The last imaging assessment occurring during the first 12 weeks of treatment was used for the analysis and compared to the baseline pretreatment imaging assessment. The dataset includes response data from 8,365 patients with lung cancer (NSCLC and SCLC). Of the 428 NSCLC patients treated with erlotinib, 237 patients had at least two target lesions as defined by RECISTv1.1. Of these patients, 31% (73/237) had at least one responding target lesion that reduced in size by 30% or more (Supplementary Fig. 1a). Within this group, 34% (25/73) had a mixed response to erlotinib treatment. The majority of patients within this group (21/25), had growth of at least one existing target lesion (Fig. 1a, b and Supplementary Table 1), whereas a minority of patients (4/25), had a mixed response to erlotinib due to the appearance of a new metastatic lesion or progression in a non-target lesion (Fig. 1b). To summarize, in all patients where a mixed response to erlotinib could be measured, i.e., two or more target lesions could be assessed, 25/237 patients had a mixed response….