Engaging cancer cells in a high stakes game of spot the imposter

From the Bradley Lab, Public Health Sciences and Basic Sciences Divisions

The year is 1995. A killer cyborg has arrived from the future to terminate John and Sarah Connor and, in doing so, destroy humanity’s only chance to survive the war to come. As Sarah engages the cyborg in a final showdown, the invader reveals a nifty trick, morphing itself into a perfect replica of Sarah. John stands by, ready to intervene and turn the tide of battle in Sarah’s favor, but there’s one big problem – he can’t tell who’s the human and who’s the cyborg.  -Plot of Terminator 2: Judgement Day.

Many of the diseases we face result from the presence of bad actors within our bodies, be they viruses (HIV, Covid-19), bacteria (tuberculosis, syphilis), fungi (ringworm), or parasites (malaria). Combating these bad actors often involves a general strategy that uses a combination of pharmaceutical agents and our own immunological defenses to selectively destroy them while sparing our own cells. Cancer, too, is such a disease, the cells of the tumor representing the bad actors we seek to eradicate. But cancer has a special feature that sets it apart from infectious diseases, and which makes it especially difficult to treat. While these other diseases result from the invasion of foreign entities, which are different in many ways from the cells of our own bodies, cancer cells are our cells, albeit perverted versions of them. Because the cancerous cells and the healthy cells in a body are so similar, it is extremely difficult for a drug, or for our immune systems, to tell them apart. This often leaves us with two bad options for treatment – hit the cancer hard and risk significant collateral damage to healthy tissues, or limit damage to healthy tissues and risk failing to kill the cancer cells. Recent major progress in cancer research has resulted from developing targeted treatments that improve the ability of therapeutic agents to recognize the subtle distinctions that exist between cancerous and healthy cells. Dr. Rob Bradley, a Professor in Fred Hutch’s Basic Sciences and Public Health Sciences Divisions and member of the Fred Hutch/UW Cancer Consortium, is putting a unique spin on this approach, exploiting the properties of cancer cells to enhance their differences.

A simplified view of genetics is that a gene provides the code to make a protein. In reality, many different proteins can often be made from one gene. This is because genes are composed of modular units (exons) connected via non-protein-coding linkers (introns). Depending on the context, as a transcript matures the machinery that cuts out the introns and splices together the exons can choose to include or exclude certain exons, ultimately generating related but distinct protein products.  Mutations in genes encoding splicing factors occur in many cancers, as splicing errors can in one fell swoop disrupt the function of myriad cancer-related genes. In a new paper published in Nature Biotechnology, Dr. Bradley’s lab, in collaboration with the labs of Fred Hutch colleague Dr. Hans-Peter Kiem and Dr. Omar Abdel-Wahab at Memorial Sloan Kettering Cancer Center, developed a way to transform splicing factor mutations from a cancer’s strength into its weakness. 

“Spliceosomal mutations alter splice site and exon recognition to cause dramatic mis-splicing of a restricted set of genes, while leaving most genes unaffected”, the authors write. Further, the fact that these mutations are common early drivers of cancer progression makes them particularly attractive therapeutic targets, although effective therapies against these mutations have not been identified. Rather than taking the more standard approach of identifying therapies that disrupt the function of the mutant protein, the group reasoned that if they could introduce a gene that was differentially spliced between cells with and without the spliceosomal mutation, they could trick the cancer into producing a protein that would allow the mutant cells to be targeted with existing therapies.

They focused their attention on cells with mutations in the SF3B1 gene, the most commonly mutated spliceosomal gene in cancer. First, the group searched the transcriptomes of patients with SF3B1 mutations to identify introns that were differentially spliced in these tumors. Of the ~1000 such introns found, they chose six of the strongest hits to examine further and found them to be consistently mis-spliced between cancers, cell types, and SF3B1 mutations. The group then used these introns as guides to develop short, synthetic introns with similar splicing properties. They then tested the function of their synthetic introns by inserting them between two exons that code for the fluorescent protein mEmerald, Introducing these constructs into wild-type and SF3B1 mutant cells, and measuring mEmerald fluorescence as an indicator of splicing efficiency. This experiment revealed strong mutation-dependent splicing.

The next step in the process involved connecting the synthetic introns to a gene that would, when translated, allow the mutant cell to be targeted by a therapeutic agent. “We selected the herpes simplex virus-thymidine kinase (HSV-TK) system, in which treatment of HSV-TK-expressing cells with the prodrug ganciclovir (GCV) causes cytotoxic metabolite production”, they wrote. They again placed their synthetic intron in the middle of the HSV-TK gene, inserted the construct into cells, and, incredibly, found that mutant, but not wild-type, cells became sensitized to GCV treatment.

The group then sought to optimize the function of their synthetic introns by performing a high-throughput screen to test mutation-dependent splicing activity on over eight thousand intron sequence variants. HSV-TK constructs containing the variant introns were placed into cells followed by treatment with GCV, and sequencing of the cells that survived treatment was used to determine which variants were lost due to GCV-dependent cell death. The results of this screen revealed not only which sequences were optimal for synthetic introns, but also the key features required for these introns to work as intended. Dr. Khrystyna North, former graduate student in the Bradley lab and co-lead author, points out another valuable aspect of this screening strategy, which is “to optimize [the synthetic intron] and tune the response to the desired therapeutic context allowing for further modification and applications of the technology in a wide variety of cancers”.

Finally, the authors asked whether their synthetic introns could mediate selective cell killing in in vivo cancer models. They introduced leukemia, melanoma, and breast cancer cells into mice that were either positive or negative for SF31B mutations, and found that GCV treatment selectively and effectively killed only the SF31B mutant cells. Lastly, they used lentiviral vectors to deliver the HSV-TK synthetic construct into established SF31B mutant tumors and found that this sensitized the tumors to GCV treatment.

Reflecting on the significance of this work, Dr. North says that these synthetic introns “can act as a “kill-switch” that directly responds to the internal state of the cell. This “switch” allows us to deliver therapies in a highly specific manner.” Moving forward, Dr. North is focused on how the technology can be adapted for use in patients. “Now that we have a method to develop and tune the response of synthetic introns, the next big question is about how to translate this into the clinic. This brings up further questions about how many and which other molecules can be delivered using the synthetic RNA approach? We focused on cells and cancers that have an SF3B1 mutation in the paper but we would like to know how to apply the method to cells with other splicing factor mutations, or even cancer cells that don’t have splicing factor mutations but still have abnormal splicing.”

synthetic intron cell killing
Left: representation of the spliced (top) and unspliced (bottom) HSV-TK gene containing the synthetic intron and their responses to GCV treatment. Right: Viability of wild-type (blue) and mutant (red) cells containing various versions of the HSV-TK gene upon different contrations of GCV treatment. Image provided by Khrystyna North.

This work was supported by the National Institutes of Health, the American Society of Hematology, the Leukemia and Lymphoma Society, the ARCS Foundation, the Washington Research Foundation, the Edward P. Evans Foundation, the Henry and Marilyn Taub Foundation, and the Fred Hutchinson Cancer Research Center.

Fred Hutch/UW Cancer Consortium members Rob Bradley and Hans-Peter Kiem contributed to this work.

North K, Benbarche S, Liu B, Pangallo J, Chen S, Stahl M, Bewersdorf JP, Stanley RF, Erickson C, Cho H, Pineda JMB, Thomas JD, Polaski JT, Belleville AE, Gabel AM, Udy DB, Humbert O, Kiem HP, Abdel-Wahab O, Bradley RK. Synthetic introns enable splicing factor mutation-dependent targeting of cancer cells. 2022. Nat Biotechnol. Online ahead of print.