StemCell Therapy

By Kevin E. Noonan --

The promise of an era of "personalized medicine" has been pursued for a generation, being one of the rationales for and purported benefits of the Human Genome Project. It has become such a sought-for goal that it has been used to drive policy: it is something that health care reform is banking on (literally), since by making medicine more individualized, success rates and fewer failed therapies are envisioned. It crept into the gene patenting debate, with Judge Bryson in his dissent crediting the ACLU's claim that gene patents will inhibit development of personalized medicine (they won't). But as others have noted (for example, Nicholas Wade, "A dissenting voice as the genome is sifted to fight disease," The New York Times, September 15, 2008), a promise is all it remains: achieving a personalized medicine future has proven (so far) to be much more daunting than its proponents believed (or told the rest of us to believe).

The results of a cancer study in the New England Journal of Medicine last week may have shed some light on what this has been so. The report, "Intratumor heterogeneity and branched evolution revealed by multiregion sequencing," a team of physicians and scientists from the UK and Harvard revealed that the genetics of human tumors is much more complicated than previously thought (Gerlinger et al., 2012, N. Engl. J. Med. 366: 883-92). These researchers obtained multiple biopsy samples from different regions of the same tumor (primary and metastatic) and performed multilocus genomic sequencing. The tumors were all a particular subtype of primary renal cell carcinoma, clear cell carcinoma (CCC) from patients that had been treated with everolimus (Zortress, an mTOR inhibitor) therapy before and after nephrectomy. Whole exome multiregion spatial DNA sequencing (see below) was performed on extracted tissue from fresh frozen samples as well as SNP analysis and mRNA expression profiling using gene arrays.

Exome Sequencing - Part I:

Exome Sequencing - Part 2

The results showed a significant amount of genetic heterogeneity that could be related to chemotherapeutic drug resistance and differential metastatic potential. In one CCC patient, nine regions of the primary tumor and three regions of metastatic tumors (as well as the germline DNA sequences) were assayed. A 2 bp deletion in the von Hippel-Lindau (VHL) tumor suppressor gene was found, a genetic characteristic of CCC. These analyses revealed 101 nonsynonymous point mutations and 32 instances of insertion or deletion (indels), with the assays showing a low false negative rate of detection. From a total of 128 mutations detected in the various samples, 40 were "ubiquitous" mutations (found in all samples), 59 were shared by "several but not all" regions, and 29 were unique to a particular region (called "private mutations"). Of the "shared" mutations, 31 were shared by most of the primary tumor samples, and 28 shared by most of the metastatic tumor samples. "The detection of private mutations suggested an ongoing regional clonal evolution," the authors concluded from this data.

From these results, the workers constructed a phylogenetic tree that revealed "branching rather than linear" tumor evolution. Deeper analysis showed that, in some regions, the primary tumor shared more mutation with the metastases than with other areas of the primary tumor, suggesting the existence of two "clonal populations of progenitor cells in this region." The study also compared these results with results from a "single" tumor biopsy study, which detected 70 somatic mutations (about 55% of the total detected using the multiregional approach). These figures were put into context by noting that only 31-34% of all mutations detected using the multiregional sampling/sequencing were detected in all regions. Finally, any major effect of the everolimus treatment on these results was discounted by finding that 67/71 mutations found after treatment were present in the tumor samples pretreatment, and that 64/66 chest wall metastasis mutations were found in post-treatment metastatic tumors. These results indicated to the researchers that "the two main branches of the phylogenetic tree were present before drug treatment" and that "60% of the mutations in pretreatment samples of the primary tumor and chest-wall metastases were not shared by both biopsy samples," i.e., evidence of clonal evolution that would have required reversion of somatic mutations during treatment (not very likely).

A conventional measure of tumor heterogeneity, ploidy analysis (i.e., how many chromosomes and chromosome fragments were present in the tumor cells) was also performed. While the primary tumor was predominantly diploid (i.e., facially "normal") there were two regions in the metastatic tumors that were subtetraploid (i.e., a few fewer than twice the [n]o limited by sample quality issues showed ubiquitous "allelic imbalance" on the short arm of the 3rd chromosome (3p) characterized by loss of heterozygosity at multiple allelic loci), including VHL and histone H3K36 methyltransferase SETD2. Even here, "tumor regions shared identical allelic-imbalance profiles, and heterogeneity of allelic imbalance within metastases, which is probably driven by aneuploidy, indicates that chromosomal aberrations contribute to genetic intratumor heterogeneity."

The study also compared the mutational status of genes known to be mutated in CCC, including VHL, SETD2, KDM5C, and mTOR. Only the VHL gene was ubiquitously mutated in all regions sampled, contrasted in the study by the mutational nature of SETD2: the metastases all showed a missense mutation while one primary region had a splice site mutation and the others showed a 2 bp frameshift deletion (which was also present in the region with the frameshift mutation). Convergent evolution was detected with regard to SETD2 histone methylation using functional assays; such convergent genetic evolution in tumor cells was also detected for the X chromosome-encoded histone methyltransferase KDM5C.

Another gene, mTOR, showed a missense mutation in the portion of the gene encoding a kinase domain; this mutation was found in all but one of the primary tumor regions tested. The researchers also reported that a currently used test for CCC, a 110-gene signature that assesses patient prognosis, displayed anomalous results: the metastases and one primary tumor sample showed the "good" prognostic pattern while all the other primary sites showed the "poor" prognostic pattern. The authors caution that "prognostic gene-expression signatures may not correctly predict outcomes if they are assessed from a single region of a heterogeneous tumor."

The workers performed similar analyses on three other patients. In one, patient 2, the researchers found 119 somatic mutations what also showed a branching pattern of clonal genetic evolution in this patient's tumor. Here, ~31-37 of the mutations were found ubiquitously (the lower number was obtained when the metastases were included). While no ploidy imbalance was detected in these tumor samples, allelic imbalance was found ubiquitously in all tested regions for 3p and on the long arm of chromosome 10 (10q). In addition to some of the 3p mutations found in patent 1's tumor, mutations were found for genes residing on 10q, including PTEN. Convergent evolution was also observed for the PTEN gene. Similar results were obtained and briefly noted for tumor samples obtained from patients 3 and 4 (patient 4's tumors showed allelic imbalance on chromosomes 5 (5q) and6 (6q)). However, "[t]hese early ubiquitous events were outnumbered by non-ubiquitous aberrations, indicating that the majority of chromosomal events occurred after tumors diverged, providing further evidence of branching evolution." Patient 4 also showed tumor heterogeneity in genes like SETD2 that had been detected in other tumor samples.

The authors summarized their results by noting that they had detected genetic heterogeneity in each tumor assayed, showing "spatial separated heterogeneous somatic mutations and chromosomal imbalances." These genetic lesions lead to phenotypic heterogeneity, with 63-69% of the mutations not detected in every tumor region sampled. Their detection of "ubiquitous alterations on the trunk of the tumor phylogenetic tree . . . may account for the benefits of cytoreductive nephrectomy" because it reduces the "reservoir" of primary tumors cells capable of genetic instability and failure to respond to more "conventional" regions of the tumor. Finally, the authors state that:

Genomics analyses from single tumor-biopsy specimens may underestimate the mutational burden of heterogeneous tumors. Intratumor heterogeneity may explain the difficulties encountered in the validation of oncology biomarkers owing to sampling bias, contribute to Darwinian selection of preexisting drug-resistant clones, and predict therapeutic resistance. Reconstructing tumor clonal architectures and the identification of common mutations located in the trunk of the phylogenetic tree may contribute to more robust biomarkers and therapeutic approaches.

These results illustrate a few things. First, the gene patenting debate per se is anachronistic and ten if not thirty years out of date. The complexity revealed by this study provides one reason why approaches tried thus far for implementing personalized medicine have not worked out as well as planned. This complexity suggests that it will take far more time to produce a worthwhile personalized medicine paradigm that fulfills all its unfulfilled promises and that the "gene age" will likely be long past by that time.

This very same complexity reinforces the risk in making any broad pronouncements against the patent-eligibility of "products to nature." With this level of complexity, the number of "false negatives" (and, presumably, false positives) may make it possible to identify diagnostic genetic markers for disease prognosis that can be protected without patents. As noted by the authors, identification of the "trunk" mutations (shared by the largest number of tumor samples) provide the best information on the tumor for treatment, prognosis and otherwise. The negative consequences of changing the incentives from disclosure (protected by patenting) and non-disclosure (protected, inter alia as a trade secret) has been discussed here before; this study points to ways that could be profitable for the company that develops the test at the cost of reaching the goal of personalized medical care. Because the alternative may be no personalized medical care at all, it behooves participants in the policy debate about gene patents, genetic diagnostic testing, and innovation to consider this study to be but the first in a long series demonstrating that, indeed, we are only at the beginning of the road when it comes to developing a robust personalized medicine system.

Images of exome sequencing (above) by SarahKusala, from the Wikipedia Commons (Part I & Part 2) under the Creative Commons Attribution 3.0 Unported license.

More here:
Patent Docs: Personalized Medicine

This entry was posted on August 4, 2016 at 9:36 am and is filed under Personalized Medicine. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed. |