Pan‐cyclin‐dependent kinase inhibition is a potential treatment for adenoid cystic carcinoma that downregulates the MYB::NFIB fusion and induces tumor regression

Adenoid cystic carcinoma (ACC) is an aggressive glandular cancer primarily affecting the major and minor salivary glands but may also occur in other anatomical locations such as the breast, prostate, lungs, and skin [1, 2]. ACC has a poor long-term prognosis due to frequent recurrences and distant metastases and is unresponsive to all so far tested systemic therapies. The MYB proto-oncogene, transcription factor (MYB), is a key oncogenic driver activated by gene fusions in ACC [3-7]. We have previously demonstrated that the canonical MYB::nuclear factor I B (NFIB) fusion (MYB::NFIB) is targetable by insulin-like growth factor 1 receptor/AKT serine-threonine kinase (IGF1R/AKT) inhibitors, demonstrating that it is indeed an actionable target in ACC [5]. To identify more effective therapeutic options for ACC patients, we conducted an in vitro drug screen using carefully validated MYB::NFIB-positive ACC cells and a custom-made library of more than 300 small-molecule inhibitors covering important druggable targets in precision oncology (Supplementary Materials). Notably, cyclin-dependent kinase (CDK) inhibitors emerged among the most effective agents that reduced the viability of ACC cells (Figure 1A, Supplementary Tables S1-S2). We further tested 8 different U.S. Food and Drug Administration (FDA)-approved/phase II–III CDK inhibitors, of which dinaciclib (a pan-CDK inhibitor) showed by far the highest potency (half maximal inhibitory concentration [IC50]: 10–13 nmol/L, P < 0.001) (Figure 1B, Supplementary Figure S1A). ACC cells were significantly less sensitive to U.S. FDA-approved CDK4/6 inhibitors, such as palbociclib and ribociclib, than to dinaciclib. The ACC cell line, UM-HACC-2A, showed similar sensitivity to dinaciclib as ACCX11 and ACC67 cells, whereas control pleomorphic adenoma (PA) cells were significantly less responsive (Supplementary Figure S1B). Additional top hits from our screen included histone deacetylase, proteasome, and IGF1R-phosphoinositide 3-kinase (PI3K)-AKT inhibitors, consistent with our previous findings [5, 8] and thus validating our screening approach. Dinaciclib caused a significant decrease in ACC spheroid formation at nanomolar concentrations (Figure 1C, Supplementary Figure S2A), suggesting that it inhibits cell division of immature stem-like ACC cells capable of tumor initiation. Flow cytometry analysis of dinaciclib-treated ACC cells revealed reduced S-phase entry and accumulation of cells in the G1 and G2/M phases, indicating cell cycle arrest at multiple checkpoints (Supplementary Figures S2B-C and S3). To test if CDK inhibition leads to apoptosis in ACC cells, we treated ACCX11 and ACC67 cells with CDK inhibitors at different concentrations for 24 h. Dinaciclib induced apoptosis at low concentrations, whereas Flavopiridol and AT7519 caused apoptosis at much higher concentrations; the other tested CDK inhibitors did not induce apoptosis (Figure 1D, Supplementary Figure S4A). Dinaciclib also induced apoptosis in UM-HACC-2A ACC cells but had no or very low effect on PA, mucoepidermoid carcinoma, and breast cancer cells (Supplementary Figure S4B-D), demonstrating the ACC-specific sensitivity of the drug. We then explored a potential MYB-dependence of the apoptotic response. Overexpression of wild-type MYB in ACC67 cells enhanced dinaciclib-induced apoptosis (Figure 1E). Conversely, MYB::NFIB knockdown in ACCX11 reduced the sensitivity to dinaciclib and cell proliferation without increasing apoptosis (Figure 1F, Supplementary Figure S5A-B), in line with previous findings in ACC67 cells [5]. ACCX11 cells, which had significantly higher MYB::NFIB expression, were more sensitive to dinaciclib than ACC67 cells (Figure 1D and 1G, Supplementary Figures S2C, S4A and S5C). These results suggest that high MYB levels sensitize ACC cells to pan-CDK inhibition and imply that they are “MYB-addicted”. RNA-seq analysis of dinaciclib-treated ACC cells revealed downregulation of genes involved in the cell cycle, DNA replication, and DNA repair, whereas upregulated genes were associated with programmed cell death, proteolysis, and autophagy (Supplementary Figure S6A-C). Gene set enrichment analysis showed significant downregulation of previously identified MYB signature genes [9] in dinaciclib-treated cells (Figure 1H), suggesting that dinaciclib perturbs MYB-driven gene expression. Notably, we also found that dinaciclib directly affected MYB::NFIB expression; both MYB::NFIB mRNA and protein levels were significantly downregulated in dinaciclib-treated cells (Figure 1I, Supplementary Figure S6D). Moreover, known MYB target genes involved in cell cycle regulation and apoptosis were also significantly decreased (Supplementary Figure S6E), indicating that dinaciclib disrupts the MYB::NFIB-driven transcriptional program by downregulating its expression. CDK9, a known dinaciclib target, regulates MYB expression in breast and colon cancer cells [10]. Western blotting of ACC cells showed decreased phosphorylation of RNA polymerase II serine 2 following dinaciclib treatment (Supplementary Figure S7A), confirming CDK9 as a target of dinaciclib also in ACC. CDK9 knockdown led to reduced MYB::NFIB expression but, unlike dinaciclib, did not induce apoptosis (Figure 1J-K, Supplementary Figure S7B-C). Similarly, treatment with the CDK9 inhibitor KB-0742 failed to induce apoptosis (Supplementary Figure S7D), suggesting that other CDK targets mediate cell death in ACC cells. Notably, all CDK genes known to be targeted by dinaciclib were overexpressed in primary ACCs compared to normal salivary gland tissue (Figure 1L). Individual knockdown of these genes revealed that silencing of CDK1, CDK2, CDK4, and CDK6 induced apoptosis in ACC cells but less effectively than dinaciclib (Figure 1M, Supplementary Figure S7E). Unexpectedly, knockdown of all tested CDK genes led to a significant decrease in MYB::NFIB expression at both the mRNA and protein levels, with CDK6 showing the most prominent effects (Figure 1N, Supplementary Figure S7F). CDK6 was also the most highly expressed CDK gene in ACC patient samples (n = 8) (Supplementary Figure S8). Taken together, these results suggest that several CDKs act in concert to regulate MYB::NFIB expression and that dinaciclib induces significant apoptosis in ACC cells by inhibiting multiple CDKs. Since both MYB::NFIB and CDK9 knockdown failed to induce apoptosis, our findings have implications for ongoing and completed clinical trials of single-agent MYB/CDK9 inhibitors in ACC (NCT06462183, NCT06297941, NCT04718675). To validate the in vivo efficacy of pan-CDK inhibition, we treated two ACC patient-derived xenograft (PDX) models with dinaciclib for 6 weeks. Both models showed significant tumor regression in response to dinaciclib with a maximum tumor volume reduction of 94% (ACCX11) and 27% (ACCX5M1), respectively, in individual mice (Figure 1O). Importantly, dinaciclib showed no significant effects on mouse weights (Supplementary Figure S9A) and previous clinical phase I/II trials including other cancer types have demonstrated that dinaciclib was generally well tolerated and safe (NCT00871663, NCT00732810). Notably, our study reports tumor regression with dinaciclib as monotherapy in PDX models, which has not been shown in other studies regardless of cancer type. Histopathological analysis of tumor tissues from dinaciclib-treated ACC PDX mice showed large, central areas with necrosis and focal infiltration of inflammatory cells compared with control tumors (Figure 1P). Moreover, Ki-67 and MYB::NFIB immunostainings showed a significant inhibition of tumor cell proliferation and a dramatic decrease in MYB::NFIB protein expression after dinaciclib treatment (Figure 1P, Supplementary Figure S9B). Collectively, these results demonstrate that dinaciclib treatment induces potent tumor regression in ACC PDX models through decreased tumor cell proliferation, increased cell death, and downregulation of MYB::NFIB expression. Our results indicate that ACCs are particularly vulnerable to combined inhibition of CDKs mediated by dinaciclib and we propose that high MYB levels is the underlying reason. This conclusion is based on our results from ACC cells with different levels of MYB::NFIB expression, forced overexpression of MYB, and dinaciclib treatment concurrent with MYB::NFIB knockdown (Figure 1D-G, Supplementary Figures S2C, S4A and S5C). A limitation of our study was that the mechanism behind the increased sensitivity of MYB::NFIB high ACC cells to dinaciclib and how CDKs regulate the expression of the fusion and apoptosis are still partly unknown. Both CDKs with critical functions in transcriptional regulation and cell cycle progression are involved in the regulation of MYB::NFIB. However, whether they directly or indirectly affect the expression and phosphorylation of the fusion protein is unknown. Recent studies have suggested that some CDKs may have dual functions, with both kinase-dependent and kinase-independent activities. This may explain the discrepancies in MYB::NFIB regulation/induction of apoptosis between CDK4/6 kinase inhibition and RNA interference-mediated depletion. Further mechanistic studies are needed to address these issues. Taken together, our study demonstrates that ACCs are highly sensitive to pan-CDK inhibition, resulting in decreased MYB::NFIB expression, induction of apoptosis, and tumor regression in ACC PDX models. Our findings reveal insights into the regulation of the MYB::NFIB fusion and advocate pan-CDK inhibition as a treatment strategy for ACC patients that should be further explored in clinical trials. Mattias K. Andersson and Göran Stenman conceptualized the study and managed the project. Junchi Huang, Peter Larsson, Maryam Kakay Afshari, Paloma Tejera Nevado, Tajana Tešan Tomić, André Fehr, Fredrik Jäwert, Göran Stenman, and Mattias K. Andersson took part in the acquisition, analysis, and interpretation of data. Mattias K. Andersson, Göran Stenman, and Peter Larsson acquired funding. Mattias K. Andersson drafted the manuscript, and all authors reviewed and edited the draft. All authors agreed to the submitted version of the manuscript. Resources for RNA-seq analysis were provided by SNIC through Uppsala Multidisciplinary Center for Advanced Computational Science (UPPMAX) under Project SNIC sens2017572. The authors thank Dr. Joakim Karlsson for bioinformatic assistance and Therese Carlsson for technical assistance. This study was supported by grants from the Swedish Cancer Society (24-3621), the Sjöberg Foundation (2020-119), the Wilhelm and Martina Lundgren Science Foundation (2025-SA-510), the Herbert and Karin Jacobsson Foundation (2024-forskning-293), the Lars Hierta Memorial Foundation (F02024-0432), the AG Fond (FB23-37), and the Swedish state under an agreement between the Swedish government and country councils, the ALF agreement (ALFGBG-1005807). The authors declare that they have no conflicts of interest. The study was approved by the regional ethics committee in Gothenburg (D-No: 178-08) and the Swedish Ethical Review Authority (D-No: 2025-03534-01). The requirement for informed consent was waived by the ethical review boards due to no or minimal risk for the patients and the use of patient material stripped of direct subject identifiers. The work was done in compliance with the Declaration of Helsinki. Animal experiments were approved by the Regional Animal Ethics Committee of Gothenburg (approval no. 273-2017) and were in accordance with Directive 2010/63/EU of the European Parliament and the European Council. RNA-seq data included in the present study are available for download from the Gene Expression Omnibus database (https://www.ncbi.nlm.nih.gov/geo/) using Accession No. GSE299929. Additional source data included in the study is available upon reasonable request from the corresponding author. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.