SAN FRANCISCO--As of yet, few molecularly targeted therapies provide more than a short reprieve for patients with advanced cancer, since resistance develops quickly even in patients with a strong initial response.
And although numerous investigators are working in tissue culture cells and model systems to uncover the mechanisms of resistance, Jeffrey A. Engelman, MD, PhD, and colleagues have taken their work right to patients, work described here at the AACR International Conference on Frontiers in Basic Cancer Research meeting.
The results are more complex than many researchers expected, suggesting that there are multiple cancer cell populations that need to be controlled, and therefore innovative treatment strategies will be required to substantially prolong patient lives.
“There is not a single targeted therapy out there in advanced cancers that induces responses for much more than a year,” said Dr. Engelman, Director of the Center for Thoracic Cancers at Massachusetts General Hospital and Assistant Professor of Medicine at Harvard Medical School. “Yet the oncology community talks about targeted agents as tremendous successes because well-selected patients do better on these agents than on standard chemotherapy. But even in lung cancer patients who have an EGFR mutation, at nine months half of them have progressed. And this is true for BRAF-mutant melanomas, EML4-ALK lung cancers, and HER2-amplified breast cancers.
“So the idea is coming clear that as we’ve opened up this one door and have seen these very nice responses to targeted therapies…there are about 10 more doors ahead of us that stand in our way of making this a real transformation for patients, where they can get years or decades of benefit from these therapies.”
To uncover molecular changes that drive resistance in patients, Dr. Engelman’s team grows cancer cells in the presence of drugs and treats xenograft animal models with the drugs. Once resistance arises they play detective to identify the key alterations, also examining patient biopsy samples taken at resistance.
“This has become a critical component of this resistance project,” Dr. Engelman said. “What we found is that each effort alone is unsatisfactory in terms of moving the field forward.”
Model systems are important functional tools to identify and test potential therapeutic strategies, but are not adequate by themselves. “We need the patient samples to see what is actually going on so that we prioritize things that are important and not try to treat things that happen in cells but that are never see in the clinic.”
Two Paradigms of Resistance to Crizotinib & Erlotinib
The team has characterized two paradigms of resistance to crizotinib and erlotinib: gatekeeper mutations and bypass mechanisms.
In the simplest scenario, a point mutation in the target gene interferes with drug binding and restores downstream signaling activity that promotes tumor growth. In the case of bypass mutations, changes in the cell allow downstream signaling to occur without reactivating the original target.
Both paradigms have been seen with other targeted agents, he noted, but that commonality, he thinks, points to the value of the models his team is using.
When the investigators look at the whole patient though, they find that classifying resistance becomes more difficult. In a recent study, the researchers examined what happened over time in 37 patients with EGFR-mutant lung cancer, comparing tumor samples removed prior to treatment with biopsy samples taken at resistance.
Each sample pair was tested for 60 different genetic alterations in 13 genes using a CLIA-approved multiplex pathology test, and for EGFR and MET expression using fluorescent in situ hybridization tests.
All tumors retained their original EGFR mutation, and the team was able to identify additional genetic or morphological changes in 70% of the patients. Eighteen (49%) of the patients had acquired a gatekeeper mutation in the EGFR gene, some with an amplification of the EGFR gene.
MET amplification, which is a bypass mechanism, occurred in two (5%) patients, and two patients had mutations in the PIK3CA gene, which was not predicted from cell or animal models. Additionally, the team saw strong epithelial-to-mesenchymal transition in two patients.
Perhaps most surprising, though, was that adenocarcinoma tumors in five (15%) patients transformed into small cell lung cancer. (One of these patients also developed a PIK3CA mutation.)
Alterations of Original Cancers
The newly arisen small cell lung cancers still carried the original EGFR mutation, leading Dr. Engelman et al to conclude that these were not new cancers but rather alterations of the original ones.
The small cell cancers in these patients expressed neuroendocrine markers typical of the disease and spread quickly through the body and responded to chemotherapy, as typical for this disease. It is not clear what drove the transformation or if that alteration is the same thing that drove resistance to erlotinib, but that would be the simplest hypothesis.
“We don’t understand the reason for the transformation,” Dr. Engelman said, “but it is quite a robust finding and we continue to see it as we acquire more cases.”
Once the team saw these data, their work was no longer just research, but was used to direct clinical care so patients could receive proper therapy.
“One thing to notice is that there are many different ways that cancer can become resistant to therapy,” Dr. Engelman said. “So the way we are going to overcome resistance is not a one-sized fits all approach, but clearly knowing the mechanism might guide our therapeutic strategy to put patients back into remission.”
Thinking about the tumor cell population in patients, Dr. Engelman noted that when patients with EGFR-mutant lung cancer develop resistance, about 25% of them will have a tumor “flare” when they are taken off the drug.
Therefore, even though the drug is no longer working the way one wants, it is doing something. Additionally, after a year or so of being off an EGFR inhibitor, many of these patients will have a second response when put back on the targeted agent.
Tese observations don’t “jive with idea that patients are simply developing [gatekeeper] mutations,” Dr. Engelman said.
To illustrate the complexity of resistance, Dr. Engelman described several patients’ responses over time:
The first patient had a strong initial response to erlotinib, but then became resistant. A biopsy indicated that the patient’s tumor had developed a T790M gatekeeper mutation in the EGFR gene, and the patient was thus taken off erlotinib and put on chemotherapy. When the patient was subsequently biopsied again, the T790M mutation was no longer observed and the patient responded to erlotinib again. “We don’t know what is going on, but at least we understand why the patients is sensitive again: We can’t find T790M.”
The second patient also had a nice early response to erlotinib but then became resistant. A biopsy showed that the tumor had developed a PIK3CA mutation and transformed into small-cell lung cancer. The clinicians switched the patient from erlotinib to chemotherapy. When the tumor was biopsied again after about a year on chemotherapy, the tumor had transformed back to adenocarcinoma and the PIK3CA mutation was no longer detectable. Subsequently, the patient had another response to erlotinib, although not as robust as the first time. At resistance, the small cell was back, with the PIK3CA mutation.
“It is not just that the cancer is sensitive or resistant, but the cancer has multiple populations in the tumor itself and we are seeing competing population dynamics,” Dr. Engelman said.
In other words, the tumor is heterogeneous, and as one cell population is controlled by a particular therapy, another population takes over. Then when treatment is switched to control that newly dominant population, the originally dominant one grows back because it was never completely killed off.
Dr. Engelman speculates that it is a regrowth of these residual cells that lead to a flare when some patients are taken off erlotinib. This shifting population dynamics means that a particular resistance mechanism might not be seen at any given time because it might be part of the minor cell population at that point in time, but that the population can rebound when no longer targeted by therapy.
“Who knows if this [hypothesis] is right or wrong,” Dr. Engelman said. “The data we’ve seen so far support some sort of competing populations, and it definitely adds complexity as we think about how to deal with resistance.”
Single Cell Line Can Generate Multiple Daughter Lines with Different Mechanisms of Action
Finally, even when the team is working in cell culture, they find that a single cell line can generate multiple daughter lines with different mechanisms of resistance.
If one shifts that concept to a patient, where different metastatic lesions are exposed to different microenvironments, it becomes clear that many patients won’t have a single mechanisms of resistance that will respond to a single therapeutic approach.
“We had some cases where during autopsy we saw multiple mechanisms of resistance in different tissues,” Dr. Engelman said. In one case, one lesion carried a gatekeeper T790M mutation, while another lesion from the same individual showed MET amplification. In another case, they found that some tumors had a small cell phenotype, while others remained adenocarcinomas with the T790M mutation.
“So you really get the sense that we are dealing with competing populations of sensitive and resistance cells, and multiple mechanisms of resistance,” he concluded.
“It really points to some difficulties and challenges as to how we develop therapeutics for these cancers if we want to really change it from a progression-free survival of one year to try to make it five years or ten years, something we would all be more proud of.”
During his presentation, Dr. Engelman did not elaborate at length on how to overcome the complexity he described, but he did briefly mention the possibility of rotating combinations of agents to keep multiple resistant populations in check.