Genomic Heterogeneity Can Lead to the Selection of 'Incorrect' Targeted Inhibitors 

Get Permission

If we can understand the type of variability and the downstream pathways that are activated, we may be able to prevent the emergence of these drug-resistant clones.

—David B. Solit, MD

Genomic heterogeneity within tumors and among lesions varies widely, and “discordance among lesions could lead to the selection of the ‘incorrect’ targeted inhibitor,” according to David B. Solit, MD, of Memorial Sloan-Kettering Cancer Center, who spoke at the ASCO/American Association for Cancer Research (AACR) Joint Session during the 2013 ASCO Annual Meeting.1

In some cases, where a targeted inhibitor is given “in the wrong genetic context,” Dr. Solit explained, it can actually harm patients. For example, while BRAF mutation is known to predict response to vemurafenib (Zelboraf) in patients with melanoma, patients with BRAF wild-type melanoma do not respond. Not only might they not benefit, “they actually might see acceleration of their tumor growth, which could be very harmful,” Dr. Solit said.

Key Questions

“There are a lot of key questions for targeted therapy drug development that have to do with tumor heterogeneity, such as what samples should be profiled, is it adequate to profile the primary vs metastases, should we be performing more biopsies at time of progression, or relying on older samples that were collected at the time of diagnosis,” Dr. Solit stated. Another important question is whether tumor heterogeneity is “going to be an issue in one type of cancer more so than in another.” These questions are important, he said, “because it is already a reality in the clinic that we profile our patients molecularly and actually use that information to guide therapy.”

As an example, he cited patients with colorectal cancer who have KRAS mutations. “We don’t treat those patients with anti-EGFR therapies because they don’t respond,” Dr. Solit said. He referred to a 2008 study by Amado et al comparing panitumumab (Vectibix) monotherapy vs best supportive care in patients with metastatic colorectal cancer.2 “The patients who had KRAS mutations had no benefit. All the benefit of the panitumumab was in patients with KRAS wild-type tumors,” in terms of response rates and progression-free survival.

“This is an example of where, if we have the KRAS status wrong, because the tumor is heterogeneous, it is possible we are going to deny a patient effective therapy that they could otherwise have benefitted from,” Dr. Solit stated.

Discordance in KRAS Mutations

In laboratory work at Memorial Sloan-Kettering, Dr. Solit and his colleagues looked at the “issue of potential heterogeneity between primary and metastatic disease very carefully in colorectal tumors,” he said. That work included profiling more than 100 matched pairs of frozen primary and metastatic samples in addition to about 800 unmatched samples “to see what the heterogeneity is within tumors and among primary and metastatic pairs,” Dr. Solit explained. The results showed “incredible concordance” for the genes studied (including RAS, BRAF, PIK3CA, and TP53) and the copy number alterations.

On the other hand, the discordance in KRAS is likely due to multiple primary colorectal cancers in some of these patients. 

“The one example where we really are seeing dramatic discordance between primary and metastatic with things like BRAF and KRAS are these tumors in patients who have had a history of multiple primaries. I think we can make a clear clinical recommendation in those cases that if you have a colorectal cancer patient with a history of a polyp resection or a second primary, you probably should be biopsying the metastatic sites to determine KRAS and BRAF status,” Dr. Solit said.

He emphasized that findings are very different from tumor to tumor. “In colorectal cancer, we are seeing incredible concordance among these potential drivers used as predictors of drug response. In melanoma, we’ve only looked at a small number of tumors so far, and there has been a very disquieting amount of discordance among those lesions.”

Self-seeding Phenomenon

One possible explanation for “this incredible concordance in colorectal cancer between primary and metastatic cells,” Dr. Solit said, is the “self-seeding phenomenon,” initially demonstrated in breast cancer models and also seen in melanoma and colon cancer models. “It may be true that newer mutations are occurring in a metastasis, but it is possible that those mutations are now seeding back into the primary and into other tumors, thus minimizing the amount of discordance between primary and metastatic or different tumors than we otherwise would have seen or expected based upon a sequence of new mutations that were occurring over time,” he explained.

“It is possible that tumor seeding of primary and metastatic, which is ongoing, will actually minimize this issue in the clinic. But it is also possible that while it is a minimal issue in colorectal cancer, it is going to be more of an issue in other tumor types,” Dr. Solit said. Another consideration to take into account is that, so far, the investigators have been looking at untreated tumors. “Under the stress of a targeted agent or other therapy, we may see a more distinct selection for clones that will cause greater heterogeneity over time than we see in the primary situation,” Dr. Solit said.

Cetuximab Resistance

To address the issue of the selection of clones following exposure to drug therapy, Dr. Solit described a study looking at mechanisms of resistance to cetuximab (Erbitux). The investigators developed a resistant cell line and several different resistance models.

“If you look at those resistant models, in terms of the ability of the drug to inhibit the pathway, you will find that cetuximab fairly effectively inhibits EGFR activation in both the parental and resistant cell lines, but in the resistant cell lines we don’t see the same degree of ERK inhibition that we see in the parental cell line,” Dr. Solit said. “Notably, that inability to inhibit ERK is accompanied by activation of RAS, and it is known that the finding of a RAS mutation causes de novo resistance. We suspected that maybe these tumors with acquired resistance to these cells that were selected also had RAS mutations.” The investigators profiled both cell lines and found that “in both cases, as we treat with increasing passages of cetuximab, you find the emergence of these KRAS mutant clones,” Dr. Solit reported.

With deep-sequencing techniques, the investigators found that they were selecting out preexisting mutations under selective pressure with the drug. “But I don’t think that is the whole story. I think that de novo mutations are just as critical in this situation,” Dr. Solit added.

“I don’t think it is completely inevitable that all patients will develop resistance to targeted agents,” he concluded. “If we can understand the type of variability and the downstream pathways that are activated, we may be able to prevent the emergence of these drug-resistant clones.” ■

Disclosure: Dr. Solit has served in a consultant or advisory role for Abbott Laboratories, ArQule, AstraZeneca, Endo Pharmaceuticals, OrthoBiotech, Pfizer, Quintiles, and Roche. He has received research funding from AstraZeneca.


1. Solit DB: Genomic heterogeneity as a barrier to the development of targeted inhibitors. 2013 ASCO Annual Meeting. ASCO/AACR Joint Session. Presented June 2013.

2. Amado RG, Wolf M, Peeters, M, et al: Wild-type KRAS is required for panitumumab efficacy in patients with metastatic colorectal cancer. J Clin Oncol 26:1626–1634, 2008.