For many decades the most common anti-cancer treatments have been the use of conventional chemotherapeutic agents. However their broad-based mechanisms (e.g., DNA alkylating agents) usually leads to severe systemic side effects. Today, molecular targeted therapies, which block specific molecules involved in cancer growth or progression, represent an integrative approach to cancer therapy that has already led to breakthrough clinical responses in specific types of cancers. This is particularly evident with several recently developed kinase inhibitors that target EGFR, BCR-ABL, HER2, ALK, VEGFR, mTOR, JAK2, and BRAF.
The discovery of signaling pathways associated with states of “oncogene addiction” has allowed scientifically guided drug discovery strategies to exploit specific tumor cell vulnerabilities, opening up a new paradigm of personalized cancer therapy. However, these are rarely curative, and suffer from the same major limitation associated with traditional chemotherapy drugs—the duration of any observed clinical benefit is invariably short-lived due to the relatively rapid acquisition of drug resistance.
Identifying the specific molecular mechanisms of resistance to chemotherapeutics has been very challenging. As a result, the discovery of second-generation chemotherapeutics that can effectively treat such acquired chemo-drug resistance has been limited. However, the mechanisms of acquired resistance to pathway-targeted drugs—for example, tyrosine kinase inhibitors (TKIs)—have been more tractable to some degree. The discovery has led to the development of follow-on drugs specifically designed to overcome acquired resistance.
As more mechanisms of acquired resistance are unraveled, there are further opportunities to develop new drugs that target the root cause of the resistance process. In conjunction with this, there is a pressing need for more clinically relevant and predictive preclinical models to address the high attrition rate of agents entering clinical trials, and also for the evaluation of new agents in models that replicate acquired resistance mechanisms.
Drug-naïve preclinical models and standard cell-derived xenograft (CDX) models may have genetic and phenotypic characteristics distinct from those patients in the clinic that have relapsed, therefore they would not predict clinical efficacy.
This article outlines the utility of models that are derived from both cell lines and patient-derived xenograft (PDX) models of acquired resistance that can be used to develop specific models of resistance to certain chemotherapeutics and targeted agents.
Generation of Acquired Resistance
PDX models are being increasingly used to improve and refine preclinical modeling and provide a more relevant heterogeneous system, in which human tumor and stromal cells are in close cooperation. Maintaining the human microenvironment in such models also sustains molecular, genetic, and histological heterogeneity of the original tumors. Collectively these important features help to ensure authentic responses to current targeted agents or chemotherapeutics, as well as sustain the properties of defined tumor subsets reflective of those seen in the clinic.
By repeatedly challenging heterogeneous PDX across several passages in vivo to specific clinical agents, it is possible to compare and contrast pre- and postresistant phenotype gene expression, pathway analysis, response to chemotherapeutic and targeted agents, as well as the effect of varying treatment regimens and dosing strategies to overcome resistance.
For example, LION102, which is a NSCLC ADC PDX with an activating L858R EGFR mutation and shows exquisite sensitivity to EGFRi, has been maintained under selected pressure by repeated exposure to EGFRi’s in vivo for several months at PRECOS to induce acquired resistance as observed in the clinic (Figure 1).
Similarly, CDX models of acquired resistance are also being developed. The relatively homogeneous NSCLC HCC827 cell line population, which harbors an activating EGFR gene (exon 19 deleted) has been exposed to treatment pressures in vitro resulting in variants that are resistant to EGRFi’s. (Figure 2). Genetic mechanisms of resistance to EGFR inhibitors in the clinic include secondary EGFR mutations (T790M in exon 20 which occurs in >50% of NSCLC patients), amplification of the gene encoding the MET kinase (5–10%), and mutations in the downstream-signaling lipid kinase, PIK3CA (<5%).
It is critical to understand the mechanism of resistance in a given patient’s tumor in order to identify the most appropriate follow-on therapy as part of a personalized medicine strategy (e.g., MET inhibitors for treating patients with acquired MET amplifications). Preclinical models with representative mechanisms of resistance in different subsets of cancer would therefore be invaluable.
Targeting Multiple Tumorigenic Pathways Simultaneously
Although mechanisms of drug resistance are being unraveled, they are not able to effectively account for all cases of drug resistance. Targeting multiple tumorigenic pathways simultaneously is a promising strategy. Critical to any drug discovery program is the knowledge that a candidate drug will be effective in a Phase II setting where acquired resistance to standard-of-care treatments or targeted treatments is already evident. Therefore understanding the mechanisms whereby the tumor cells may overcome treatment and acquire resistance becomes fundamental.
Once the mechanisms of acquired resistance in the preclinical setting are identified, strategies can be rapidly undertaken if targeted agents are already known or available to disrupt the specific mechanisms involved. Relevant PDX and CDX models of acquired resistance present more important opportunities in identifying key targeted combination studies and identifying new intellectual property around existing patents.
In the optimization of novel cancer therapeutics, strategies, and patient selection, preclinical models that closely replicate the human tumor heterogeneity are imperative. A comprehensive collection of well-characterized and subset-specific clinically derived tumor models characterized preresistance and postresistance provides a powerful screening platform to support the rapid pace of novel target identification/validation, biomarker, and drug evaluation studies.
These models can also be used to better understand the tumor microenvironment, EMT, and metastases processes. By implementing innovative approaches including resistance with human stroma, optical imaging, and 3D modeling, these models will be more efficient for oncology drug discovery.