For decades, researchers have viewed synonymous mutations as inconsequential quirks of the genome. Due to the way the genetic code is set up—where multiple three-base-pair codons can encode the same amino acid—mutations can arise that don’t change a protein’s amino acid sequence. Scientists have largely dismissed these anomalies as harmless oddities.
But like other historically underappreciated aspects of the genome, scientists are realizing that many “silent” mutations might not be so silent after all. Research suggests they’re often subject to selective pressure and could play a role in cancer, autism, and schizophrenia.
A study published online last week (February 12) in iScience adds to the mounting evidence that synonymous variants can have consequences. The authors describe a synonymous mutation in the gene BAP1 that was associated with a worse-than-expected prognosis in a kidney cancer patient. Their subsequent experiments suggest that the mutation has this effect by disrupting cells’ RNA splicing process, by which freshly transcribed messenger RNA (mRNA) is converted into digestible fragments ready to be translated into protein. Because the cancer patient lacked a second healthy copy of the gene, the silent mutation may have resulted in a complete loss of function of BAP1.
“To my knowledge, tying a specific synonymous mutation [to] a clinical outcome [in cancer] is a novelty,” remarks Fran Supek, a cancer geneticist at the Institute for Research and Biomedicine in Barcelona who wasn’t involved in the study. “I’m always glad to see that researchers are thinking a bit outside the box . . . and looking at understudied classes of genetic changes that may help us solve a certain number of patients with genetic diseases or with cancer.”
[There is] little awareness of synonymous mutations and their role in cancer. In general, they are ignored in sequencing studies as it has been felt that they are very unlikely to be drivers.—Thomas Mitchell, Wellcome Sanger Institute
While combing through The Cancer Genome Atlas (TCGA)—a public database of genomic samples from more than 11,000 patients around the world—Samuel Peña-Llopis and his colleagues discovered an entry from a patient with an unusual course of disease. The 73-year-old Caucasian woman had clear-cell renal cell carcinoma, the most common form of kidney cancer, with a mutation in PBRM1, a gene involved in chromatin remodeling.
Although PBRM1 mutations are normally associated with relatively good clinical outcomes in such patients—with a median survival of 117 months, according to TCGA data—the patient died only 56 months after diagnosis, says Peña-Llopis, a cancer geneticist specializing in kidney cancer and uveal melanoma with the German Cancer Consortium at the University Hospital Essen in Germany.
The team noticed that she also had a synonymous mutation in BAP1, which encodes an enzyme involved in regulating the degradation of proteins. The mutation changes a thymine to a guanine, which still results in the same amino acid, glycine, encoded both by GGT and GGG. Curiously, the patient also had very low abundance of BAP1 protein, in fact, it was on par with renal cell carcinoma patients who have nonsynonymous loss-of-function mutations in BAP1, which tend to be linked to severe outcomes. The team suspected that the silent BAP1 mutation might somehow affect the gene’s transformation into protein.
The path by which DNA turns into protein is a long and winding one. First, double-stranded DNA is teased apart and the strands are individually transcribed into single strings of pre-mRNA, a rough draft of the instructions needed to turn it into protein. Then it must be spliced, whereby various proteins bind to different sites across the pre-mRNA and cut out noncoding nucleotide sequences—introns—and fuse the coding parts—exons—together. Only then is the mRNA ready for other cellular machinists to translate into protein.
One way by which synonymous mutations can perturb this process, previous research suggested, is by altering the specific binding sites of RNA splicing proteins, which are required to properly integrate different exons. If they can’t bind—or the altered codon causes the wrong proteins to bind—they might end up skipping over important bits of genetic code—called “exon skipping”—which can result in a dysfunctional protein. Because the synonymous mutation located in BAP1’s exon 11 was close to a splice site critical for joining this exon to the next, “we thought that maybe the splicing system was affected,” Peña-Llopis recalls.
To find out, the team conducted a series of experiments with genetic constructs containing BAP1’s exon 11, into which they had inserted fluorescent proteins. They expressed the construct in a human cancer cell line. Based on the color that emerged under a microscope, they could tell if the exon was being integrated or skipped. They observed nearly 100 percent skipping when the construct contained the synonymous mutation, significantly more than when using the construct based on the unmutated version of BAP1.
If exon 11 is skipped, that likely causes a loss of BAP1 for that gene copy, Peña-Llopis explains. Because that exon has 185 base pairs—which is not a multiple of three—losing it will cause a shift of the three-base-pair reading frame that enzymes use for protein translation. That, in turn, would cause a codon further down the line to be misread as a stop codon, signaling the protein translation machinery to terminate. mRNA transcripts containing premature stop codons are typically degraded by the cell. In this particular patient, this likely led to a complete loss of BAP1 because she had lost her second copy due to a deletion of a small chromosome segment, which is common in that cancer subtype.
Synonymous mutations in kidney cancer patients
Back in the TCGA database, which includes nearly 500 clear-cell renal cell carcinoma patients, the team found another eight patients who had synonymous mutations in BAP1 exons near sites important for splicing, two of which were located inside the sites necessary for gluing exons 10 and 11 together. Considering that there are 32 splicing sites across the 17 exons that make up BAP1, finding two out of eight is a significant number, Peña-Llopis says, “suggesting that this is a hotspot for inactivation of BAP1.” However, the two patients had very different clinical outcomes, suggesting that other genetic alterations also play a role in the prognosis.
“I think this is an important finding,” remarks James Brugarolas, a physician-scientist and oncologist who directs the kidney cancer program at the University of Texas Medical Center. BAP1 is mutated in around 10–15 percent of all clear-cell renal cell carcinomas, mostly through nonsynonymous alterations. “The study provides relatively convincing evidence that . . . mutations that do not affect the protein sequence in BAP1 could be pathogenic driver mutations, leading to the inactivation of the tumor suppressor protein,” adds Brugarolas, who has collaborated with Peña-Llopis in the past but wasn’t involved in the new research.
Brugarolas says that the data would have been even more convincing had there been more RNA sequencing and immunohistochemistry data from the patients’ tumor available, which could yield more definitive evidence of exon skipping. And, of course, such findings can always be better supported by larger sample sizes and replication in independent datasets, Supek adds. That said, “I think the in vitro experiments that they [did] suggest that the mutation they’ve identified has the potential to alter splicing. And clearly, exon 11 escaping would result in a nonfunctional protein due to a premature stop codon. One could make a very convincing argument for that,” Brugarolas says.
The clinical relevance of the finding is not yet apparent. Targeting loss-of-function mutations in tumor suppressor genes such as BAP1 has lagged behind targeting gain-of-function mutations, for instance, in enzymes that control cell growth and function, Brugarolas says; it’s generally easier to inhibit misfit proteins than to correct something that has been already abolished by mutation. It’s also unclear if BAP1 mutations could be used as biomarkers to predict patients’ responsiveness to therapies. “How mutations in BAP1 [should] be leveraged for therapy remains unknown,” Brugarolas adds.
On the whole, the findings indicate that researchers should be paying more attention to synonymous mutations, notes Thomas Mitchell, a clinician-scientist focusing on kidney cancer at the Wellcome Sanger Institute in the UK. “[There is] little awareness of synonymous mutations and their role in cancer. In general, they are ignored in sequencing studies as it has been felt that they are very unlikely to be drivers.”
Nevertheless, larger studies in the past have predicted that synonymous mutations could have pathogenic effects. In 2014, Supek and colleagues estimated that around 6–8 percent of pathogenic single-nucleotide mutations in cancer genes are synonymous mutations. Taken together with other studies, research seems to converge on an estimate of 5 percent of all driver mutations being synonymous mutations—a “non-negligible amount” that may be quite significant for some individual cancer patients, Supek says. Exon skipping is one mechanism by which such mutations could have deleterious effects, “but it’s probably going to be different for every synonymous mutation.”
Understanding silent mutations, along with other overlooked genetic alterations, could help unlock the underlying causes of disease for many individual patients for whom mutations don’t clearly fall into the nonsynonymous bucket, and open the door to finding treatments. “The human genome is a very complex thing, and there are many ways in which it can break and result in disease,” Supek says.
J. Niersch et al., “A BAP1 synonymous mutation results in exon skipping, loss of function and worse patient prognosis,” iScience, doi:10.1016/j.isci.2021.102173.