The challenge of platinum resistance
Epithelial ovarian cancer (EOC) remains one of the most difficult gynaecological malignancies to treat. Most patients are diagnosed at an advanced stage, and although platinum-based chemotherapy can initially be highly effective, many patients eventually relapse with platinum-resistant disease. Understanding why ovarian cancer cells become resistant to platinum, and finding ways to prevent or overcome this resistance, has been a long-standing focus of our laboratory.
Rather than looking at platinum resistance only as a consequence of accumulating genetic alterations, we have become increasingly interested in the adaptive states that allow cancer cells to survive treatment. Changes in transcriptional programs and RNA processing can profoundly influence how cells respond to therapy, creating dependencies that may also represent therapeutic vulnerabilities.
From CDK12 to RNA metabolism
In our search for vulnerabilities associated with platinum response, Cyclin-Dependent Kinase 12 (CDK12) emerged as a candidate of interest. Although CDKs are traditionally associated with cell-cycle regulation, several members of this family also have important functions in transcription. Among them, CDK12 plays a crucial role in coordinating transcription with RNA processing by supporting transcriptional elongation and the expression of genes involved in the DNA damage response, while also contributing to splicing. CDK12 and its paralogue CDK13 are frequently altered in ovarian cancer, making them relevant candidates to explore in the context of platinum resistance. We therefore investigated the consequences of CDK12/CDK13 inhibition (CDK12i) in platinum-sensitive and platinum-resistant ovarian cancer models.
What emerged was broader transcriptional response than the expected reduction in the expression of DNA repair genes. CDK12i affected transcriptional programs, ribosome-related pathways and, importantly, RNA splicing. Altered splicing patterns were observed not only in parental and platinum-resistant cells, but also in models that we generated by selecting for adaptive resistance to CDK12i. These changes suggested that the consequences of CDK12i could not be understood solely by looking at gene expression. They pointed instead towards a more complex relationship between transcription, RNA processing and the final fate of the transcripts produced.
This raised a new question for us: how does CDK12 coordinate transcriptional elongation with the processing of these transcripts? And which RNA-binding proteins or splicing factors might cooperate with CDK12 to accomplish this?
Connecting transcription and RNA processing
Addressing these questions led us to SFPQ and p54nrb, two RNA-binding proteins involved in transcription and RNA processing. Importantly, their involvement was not the initial premise of the study. Rather, it emerged as we tried to understand how the effects of CDK12i could extend from transcriptional elongation to RNA processing.
We found that CDK12 functionally associates with the SFPQ/p54nrb complex and that this interaction contributes to the correct transcription and processing of long genes, including genes involved in the DNA damage response. We then became interested in how these proteins spatially organized within the nucleus. SFPQ and p54nrb participate in nuclear condensates, and CDK12i altered their localization and organization, suggesting that the response to CDK12i also involves changes in the nuclear environment in which transcription and RNA processing take place.
But what happens to the RNA itself? RNA-scope experiments provided an important piece of the puzzle. When the SFPQ/p54nrb system was disrupted, transcripts from long genes were retained in the nucleus rather than being efficiently processed. These findings connected several observations that had initially appeared as separate phenomena: CDK12 activity, RNA-binding proteins, transcriptional elongation and RNA processing. Together, they suggested that the ability of CDK12 to support DNA damage response programs depends not only on transcriptional elongation, but also on the coordinated processing and export of the resulting transcripts.
An unexpected relationship between resistance mechanisms
One of the most interesting observations came from a different experiment. We generated ovarian cancer cells resistant to CDK12i and asked how these cells had adapted to the treatment.
Interestingly, these cells did not appear significantly different from their parental counterparts when we looked at their global transcriptional programs. Their splicing patterns, however, were markedly altered.
Unexpectedly, the resistant models displayed a unique reciprocal sensitivity pattern: platinum-resistant cells remained vulnerable to CDK12i, while cells that had acquired resistance to CDK12i became highly sensitive to platinum. In our models, the two resistance states therefore appeared to involve different adaptations. This observation raised the possibility that CDK12i could have a broad therapeutic applicability either to overcome established platinum resistance and offer a strategy to support platinum rechallenge or to potentially limit the emergence of resistant disease when combined with platinum in the first-line setting.
From mechanism to therapeutic strategy
Most of the mechanistic work was performed using preclinical CDK12/13 inhibitors such as THZ531 and SR4835. To further support the therapeutic relevance of CDK12i, we then tested a compound with a more direct clinical translational impact: CT7439, a clinical-stage dual CDK12/13 inhibitor currently being evaluated in a Phase 1/2 clinical trial, thanks to the collaboration with Carrick Therapeutics.
Importantly, both SR4835 and CT7439 restored the sensitivity of platinum-resistant cells to platinum, and the synergistic combination of CDK12i with platinum also showed strong activity in vivo, reducing tumour burden and delaying the emergence of platinum-resistant disease in mouse models. These findings provided a link between the molecular mechanism we had uncovered and a compound already progressing through clinical development.
For us, this was an important aspect of the study: the goal was not simply to identify another molecular dependency, but to understand whether that dependency could eventually be exploited therapeutically. The clinical development of CT7439 provides an opportunity to explore this possibility further, including the potential for translating these findings into the clinical setting.
Behind the paper
Like many research stories, the final paper is much more linear than the path that led to it. We started from a relatively direct question: could targeting a transcriptional dependency make platinum-resistant ovarian cancer cells vulnerable again? The first results gave us a clear biological effect, but understanding why it happened proved considerably more challenging.
As we moved from the phenotype to the mechanism, the picture became progressively more complex. The effects of CDK12i were not limited to transcription or to the expression of DNA damage response genes. Changes in RNA processing and splicing pointed us towards SFPQ and p54nrb and, eventually, towards the organization and fate of RNA within the nucleus.
At the same time, the development of CDK12i-resistant models gave us an unexpected perspective on resistance itself. Their increased sensitivity to platinum suggested that different treatment pressures can lead cancer cells towards distinct adaptive states. Together, these findings suggested that platinum resistance and CDK12i resistance are not simply two independent states, but may reflect distinct cellular adaptations that can influence the response to subsequent treatments.
Looking ahead
There are still several questions we would like to answer. In particular, understanding how altered splicing contributes to tolerant adaptation to CDK12i, and whether specific RNA-processing changes can be used to identify or overcome resistance, will be important next steps.
More broadly, we hope that these findings will help clarify how transcriptional regulation and RNA processing cooperate to determine the response of ovarian cancer cells to therapy. The possibility of combining this mechanistic understanding with a clinical-stage CDK12i such as CT7439 provides a path towards testing whether this vulnerability can ultimately be translated into a useful strategy for patients with platinum-resistant disease.
A collaborative effort
Ilenia Pellarin and Valentina Rossi contributed as co-first authors, combining their work on the molecular mechanisms underlying platinum resistance and the role of RNA regulation in treatment response. The project was conducted under the scientific leadership of Gustavo Baldassarre, Scientific Director of CRO-Aviano, whose research focuses on the molecular mechanisms of cancer progression and drug resistance, including the therapeutic potential of CDKs.
This study was made possible by the contribution of colleagues within CRO Aviano, National Cancer Institute and collaborators from several institutions. We are grateful to our colleagues in Gynaecological Oncology, Medical Oncology and Pathology at CRO for their clinical expertise and contribution to the collection and characterization of surgical tumour specimens. We also thank our collaborators at SISSA, the University of Milan and Imperial College London, as well as Carrick Therapeutics, for their valuable contributions to the study and to the development and testing of CT7439.
Bringing together expertise in molecular and computational biology, preclinical models and tumour pathology was essential to move the study from mechanistic observations towards their potential therapeutic implications. We are grateful to everyone who contributed to this work.