Genetic mapping of brain metastases, in the laboratory of Priscilla Brastianos, MD, Director of the Central Nervous System Metastasis Center at Massachusetts General Hospital/Harvard Medical School, and Scott Carter, PhD, at the Harvard School of Public Health, is yielding findings that could someday drive targeted treatment.
Priscilla Brastianos, MD
Scott Carter, PhD
“Historically, we’ve had a limited understanding of how brain metastases genetically evolve from their primary tumor,” she said. “But research shows that brain metastases harbor distinct, clinically actionable genetic alterations compared to their primary tumors and extracranial metastases.”
If decisions for individualized treatment of brain metastases are based on biopsies of the primary tumor, as is the norm, opportunities may be missed for optimal treatment, Dr. Brastianos said at the 2020 Miami Breast Cancer Conference.1
Branched Evolution of Brain Metastasis and Primary Tumors
Dr. Brastianos and Dr. Carter performed whole-exome sequencing of 104 matched brain metastases, extracranial sites, primary tumors, and normal tissue.2 In all clonally related cancer samples, they observed branched evolution, where all metastatic and primary sites shared a common ancestor yet continued to evolve independently. “There were new mutations in the brain metastases,” she said.
“In 53% of cases, we found alterations in the brain metastases not detected in the matched primary tumor sample. In contrast, spatially and temporally separated brain metastasis sites were genetically homogeneous. Distal extracranial and regional lymph node metastases were highly divergent from brain metastases,” Dr. Brastianos noted.
The genomic analysis of brain metastases provides an opportunity to identify alterations that are potentially clinically informative—ones that would not be detected in clinically sampled primary tumors, regional lymph nodes, or extracranial metastases.
In a recent publication in Nature Genetics, Drs. Brastianos, Carter, and their teams performed a genomic analysis of 73 brain metastases from lung cancer compared to 503 primary lung cancers.3 In a unique case-control analysis, they showed that brain metastases harbor more frequent copy-number alterations compared to primary tumors in several genes, including CDKN2A/B, YAP1, MMP13, and MYC. They then validated these findings in patient-derived xenograft mouse models and showed that overexpression of YAP1, MYC, and MMP13 increased the incidence of brain metastases.
Are the genetic differences seen in brain metastases clinically significant? According to Dr. Brastianos’ study, in the brain metastases, alterations were associated with sensitivity to PI3K/AKT/mTOR, CDK, and HER2/EGFR inhibitors, suggesting the answer is “yes.”
For example, she described a patient whose situation is often seen in clinical practice. During trastuzumab treatment for a HER2-positive tumor, the patient developed a cerebellar tumor in the setting of stable extracranial disease. The cerebellar tumor harbored an EGFR mutation that was exclusive to the brain metastasis.
“This is the same mutation known to predict sensitivity to EGFR inhibitors in lung cancer, but in this case, we detected it in the breast cancer cerebellar tumor,” she explained. “If the EGFR mutation is exclusive to the cerebellar tumor, this probably represents a mutation that was a resistant subclone during trastuzumab therapy. Of note, when we looked across the entire cohort, more than half the cases had clinically actionable alterations in the brain metastasis that were not detected in the primary tumor biopsy.”
“This genetic divergence between primary metastatic samples could pose a major challenge to clinical decision-making in oncology,” Dr. Brastianos commented. If the targetable mutations found in brain metastases are not also found in the primary tumor, opportunities for optimal treatment could easily be lost by selecting treatment based on the genetic mutations found in the primary tumor, she added.
Homogeneity Within Brain Metastases
These findings led to the next question: Is there regional heterogeneity throughout the brain itself, or is a single brain metastasis sample representative of all central nervous system (CNS) disease? To answer this question, her team sequenced regionally, anatomically, and temporally distinct areas within the brain.
Dr. Brastianos described their findings through the example of a patient who underwent sequencing of a resected cerebellar tumor prior to whole-brain irradiation and a parietal lobe metastasis after whole-brain irradiation. These anatomically distinct regions of the brain metastases were found to share the same clinically actionable drivers. “They were more related to each other than to the primary tumor biopsy, so they’re relatively genomically homogeneous,” she concluded.
Although metastases within the CNS were found to be genetically homogeneous, it was not known whether extracranial sites mirror that picture. “Our next question was, to what extent do extracranial sites recapitulate genetic vulnerabilities in brain metastases?”
“It looks like extracranial disease is genetically divergent from the brain metastasis.”— Priscilla Brastianos, MD
Tweet this quote
To determine this, they examined regional lymph nodes and other metastatic sites from the same patient. The researchers found that the brain metastasis and extracranial sites shared a common ancestor, but the brain metastasis subsequently evolved to diverge significantly from both the primary tumor and the regional lymph node. “It looks like extracranial disease is genetically divergent from the brain metastasis,” she said, suggesting that genetic divergence can contribute to progressive brain metastases in the setting of stable disease.
Alterations in CDK, PI3K Pathways Often Observed
Will targeting molecular drivers in the CNS improve clinical outcomes? To inform this question and pave the way for clinical trials, Dr. Brastianos and Dr. Carter looked for commonalities across their sample of patients. They found a preponderance of alterations in the phosphoinositide 3-kinase (PI3K) and the cyclin-dependent kinase (CDK) pathways.
Alterations associated with sensitivity to PI3K inhibitors (PI3K/AKT/mTOR) were seen in 43% of cases. More than 50% of cases had CDK pathway alterations, suggesting that CDK inhibitors may also have a role in treating brain metastases.
To explore the clinical potential of the PI3K pathway as a target, the researchers looked at the efficacy of treatment with GDC-0084, a brain-penetrant, dual PI3K/mTOR inhibitor that has shown activity in a preclinical model of glioblastoma.4 They evaluated the compound in vivo in PIK3CA-mutant and PIK3CA wild-type breast cancer cell lines and in vivo in a breast cancer brain metastasis xenograft mouse model.
In vitro, GDC-0084 considerably decreased cell viability, induced apoptosis, and inhibited phosphorylation of Akt and p70 S6 kinase in a dose-dependent manner in the PIK3CA-mutant cell line—but not in the wild-type line. In vivo, treatment with GDC-0084 markedly inhibited the growth of PIK3CA-mutant brain tumors (with accompanying signaling changes) but again not PIK3CA wild-type tumors.
To clinically explore the idea that targeting alterations in the brain will lead to improved survival, Dr. Brastianos and Eva Galanis, MD, have initiated the national biomarker-driven Alliance A071701 trial in patients with a variety of cancer types and brain metastases. The primary endpoint is CNS response rate; circulating biomarkers and imaging will be part of the assessment. This trial currently has several cohorts, including a PI3K inhibitor, a CDK inhibitor, and an NTRK inhibitor.
Eva Galanis, MD
“We’re actually targeting patients based on the alterations that are found in the brain metastasis,” she said. This genomically guided brain metastasis trial is now underway at Alliance sites.
DISCLOSURE: Dr. Brastianos has consulted for Merck, Genentech, Angiochem, Eli Lilly, and ElevateBio; and has received institutional research funding from Merck, Pfizer, BMS, and Lilly.
1. Brastianos PK: Meet the faculty—Case discussions. 2020 Miami Breast Cancer Conference. Invited Lecture. Presented March 6, 2020.
2. Brastianos PK, Carter SL, Santagata S, et al: Genomic characterization of brain metastases reveals branched evolution and potential therapeutic targets. Cancer Discov 5:1164-1177, 2015.
3. Shih DJH, Nayyar N, Bihun I, et al: Genomic characterization of human brain metastases identifies drivers of metastatic lung adenocarcinoma. Nat Genet 52:371-377, 2020.
4. Ippen FM, Alvarez-Breckenridge CA, Kuter BM, et al: The dual PI3K/mTOR pathway inhibitor GDC-0084 achieves antitumor activity in PIK3CA-mutant breast cancer brain metastases. Clin Cancer Res 25:3374-3383, 2019.