Targeted therapies have altered the landscape of cancer treatment across an increasing number of cancer types. While these drugs have significantly benefited many cancer patients, a major ongoing challenge associated with targeted therapies is acquired drug resistance, which is believed to be a key factor in cancer-related deaths. In this post we explore how acquired drug resistance to targeted therapies can be tackled by using in vivo patient-derived xenograft (PDX) models, which are effective tools for this purpose as they are very clinically relevant and they adequately model the heterogeneity observed in drug-resistant tumors.
Mechanisms of Acquired Drug Resistance
Cancer cells acquire drug resistance to targeted therapies through various complex cellular and molecular mechanisms, which can be generally classified into three major categories:
- Genomic changes to the targeted driver oncogene (e.g., genetic mutations), such that targeted agents can no longer bind to their target.
- Bypassing the inhibited target (e.g., activation of the protein’s signaling pathway by a different route either downstream of the inhibitor or in parallel).
- Activating a completely different signaling pathway enhances cell survival.
Drugs that evade resistance or can re-elicit a drug response are urgently needed. While the problem is complex and multifaceted, clinically relevant in vivo PDX models are valuable tools that can be used to study acquired drug resistance so that predictive biomarkers of response can be developed, and novel targeted therapeutics or treatment strategies for preventing and/or overcoming acquired drug resistance can be tested.
Many Different Resistant Genotypes to a Single Drug
Acquired drug resistance is complicated by the fact that a single drug can induce many different resistant genotypes. For instance, the first mechanism mentioned above presents unique challenges in that several genomic alterations can occur following targeted drug treatment, each of which may require a different therapeutic approach or combination of therapies.
For instance, patients with gastrointestinal stromal tumors (GIST) harboring KIT gene mutations now have targeted therapy options that have revolutionized GIST treatment. Approximately 80% of patients with GIST exhibit an increased sensitivity to KIT inhibitors (due to a KIT gene mutation), and the KIT inhibitor imatinib mesylate (Gleevec) has dramatically increased the survival of GIST patients. However, acquired resistance to Gleevec develops quickly and within 20 months, 50% of tumors become drug resistant.
Gleevac resistance is linked to secondary, single-nucleotide KIT mutations across several exons. Identifying which exon has been mutated is critical for developing an effective treatment plan moving forward, as each acquired mutation genotype is known to respond differently to current second- and third-line drugs. For example, studies have shown that sunitinib is active against imatinib-resistant GIST cells with KIT ATP-binding pocket secondary mutations, while it is inactive against cells with KIT activation loop mutations.
Preclinical Models are Needed for a Range of Resistance Genotypes
There is no single existing preclinical model that can address all the complex questions that accompany acquired drug resistance, yet researchers must be able to develop drugs, or identify combinations of drugs, for a whole range of resistance genotypes and mechanisms.
While conventional cell-line-derived xenograft (CDX) models are useful for studying acquired drug resistance in some instances, CDXs are unfortunately not available to model all the relevant mutation genotypes or alternate pathways activated. Additionally, there is the chance of drift from original disease and resistance mechanism(s) when using CDXs.
As described below, PDX models are highly clinically relevant and have shown predictivity of patient response.
PDXs: Recapitulating Human Resistance Genotypes
PDXs are the gold standard for preclinical modeling of acquired drug resistance genotypes and mechanisms. This is because they are derived directly from primary tumor tissues, and they have been shown to highly recapitulate the histopathological and genetic profiles of original patient tumors, including the complex associations of both primary and secondary mutations seen in actual cancers. PDX models also retain original genetic and signaling alterations that drive resistance, and they provide abundant materials to support the necessary mechanistic and drug treatment strategy studies that are key to combatting acquired drug resistance.
Furthermore, PDX panels developed from a range of patients are now commercially available. Clinical diversity can be represented by using a PDX panel across a range of patient types–not only those with drug-resistant tumors, but also those that are still responsive to therapies. Additionally, PDX models generated from the same patient, pre- and post-treatment resistance, can be used for genomic and other -omics profiling technologies, to evaluate alterations associated with response and resistance. Investigational agents or existing drugs can then be tested on individual models with known genomic changes, or alternatively, across panels of models to assess a population response and identify how the new drugs relate to a broader range of mutations.
Since PDXs are highly predictive of patient response to treatment, efficacy data can be used to inform which single or combination drug treatments may be effective for specific patients, and can even guide clinical trial design and stratification. A recently published study by Yao et al. highlights the value of using PDXs to study acquired drug resistance. The investigators used acquired cetuximab resistance HNSCC PDX models to uncover resistance mechanisms. Whole exome sequencing and transcriptome sequencing demonstrated a range of patterns of clonal selection in acquired resistant PDXs, including the emergence of subclones with strongly activated RAS/MAPK.
The Value of Developing Your Own Model of Acquired Resistance During Preclinical Drug Development
To investigate a drug’s effect on a tumor with specific genomic features, researchers can develop their own models of resistance during preclinical drug investigation. Such models can be treated with consecutive drug cycles to identify the point at which resistance develops. Mechanisms of acquired resistance, such as new mutational events, overexpression, or other alterations, can be identified using genomic and other -omics technologies. Additionally, subsequent testing of other single and/or drug combinations can be done to determine which ones re-elicit a response. Given the highly predictive nature of PDXs, preclinical findings may also be relevant for the planning of clinical treatment regimens.
This is exactly what was done in the study by Yao et al. mentioned above. Cetuximab resistance was elicited, and the evolutionary trajectory of the acquired cetuximab mechanisms were then mapped. To do this, the PDXs models were treated with cetuximab and then genetically sequenced at multiple timepoints with the goal of revealing both pre-and post-treatment genomic features. A major finding was that during the acquisition of drug resistance to cetuximab, RAS signaling was activated (both mutational and transcriptional). Overall, the acquired resistance was determined to result from the acquisition of a diverse set of newly generated subclones, and the elimination of sensitive subclones such as those bearing amplification of EGFR, CCND1, and FGF3. Lastly, the investigators found that a combination therapy of cetuximab plus the RAC1/RAC3 dual-target inhibitor EHOP-016, successfully inhibited tumor growth in all cetuximab-acquired resistant mice.
Conclusion
Acquired resistance to targeted cancer therapies remains a major challenge. By using drug-resistant and highly clinically relevant PDX models during preclinical drug development, mechanisms of acquired resistance can be identified and new treatment strategies, including combination or next-generation therapies, can be tested.
