RAC‐ing up tumour stemness: Disabled homolog 2 interacting protein and triple‐negative breast cancer

This commentary frames the novelty of Xiong et al (Clinical and Translational Medicine, 2022) in the context of the field. Research most directly related to DAB2IP's function as a tumor suppressor are briefly summarized and referenced, particularly in reference to breast cancer and the findings by Xiong et al related to RAC1. The ability of triple-negative breast cancer (TNBC) cells to adapt to metastatic niches and chemotherapy through transcriptional reprogramming has enabled this disease to evade modern treatment strategies. Suppression of one master regulator has now been shown by Xiong et al. to directly enable TNBC cells to establish tumours in mice with fewer cells while also better surviving docetaxel chemotherapy in vivo: disabled homolog 2 interacting protein (DAB2IP).1 The oncogenic phenotype was mechanistically conferred by scaffolding functions of DAB2IP dysregulating RAC1 and β-catenin. DAB2IP regulates a multitude of oncogenic processes, including WNT/β-catenin,2 Ras through its Ras-GAP domain,3 RAF/MEK/ERK, AKT, JAK-STAT and TNFα/NF-κB pathways.4 While previous research has established DAB2IP as a β-catenin inhibitor through recruitment of PP2A to dephosphorylate and activate GSK3β,5 Xiong et al. provide strong evidence that DAB2IP also prevents β-catenin nuclear function via sequestration of RAC1, which is required for β-catenin nuclear translocation. Xiong et al. noted that previous investigations establishing the protein-protein interaction network for DAB2IP indicated broad regulation, implicating RAC1 as a direct interactor. RAC1 has been previously studied with DAB2IP and β-catenin signalling based on chemical inhibition of RAC1.6 Xiong et al. further added to the field by showing direct binding of DAB2IP with RAC1 through immunoprecipitation experiments in TNBC cells, which demonstrated dose-sensitive β-catenin release from RAC1 upon competitive DAB2IP binding. As a tumour suppressor, DAB2IP becomes inhibited during tumour development and progression. Suppression of either DAB2IP expression or DAB2IP function is accomplished through multiple means, all of which converge on observed TNBC genetics. First, as Xiong et al. importantly now also establish in a large independent Chinese TNBC population, DNA methylation commonly prohibits transcription of DAB2IP. Second, mutant p53 binds to DAB2IP scaffolding domains and ameliorates JNK activation while also activating DAB2IP-mediated NF-κB signalling and chemokine release.7 Mutation of p53 occurs in 71% of TNBC and 35% of TNBC primary tumours have missense mutant p53, which is often dominant-negative and expressed. Third, the chromosome region encoding DAB2IP has an allelic loss of over 40% of The Cancer Genome Atlas TNBC patients, as shown graphically by CAIRN analysis (Figure 1).8 Cis-chromosomal tumour suppressors involved in DNA repair, XPA and FANCC, as well as the mTORC1 inhibiting tumour suppressor TSC1, may exacerbate phenotypes associated with DAB2IP loss, as many tumours lose all these genes coincidentally via chromosome arm loss. The interactome of DAB2IP is ever-growing and an appreciation of how each interaction is regulated and contributes to tumour biology may enable a better understanding of molecular programs controlling metastasis, stemness and chemotherapy response. Roughly half of DAB2IP's interactome is selected for deletions in TNBC, whereas half is selected for gains and amplifications (Figure 2).9 RAC1, which appears at the top of the SWAN interactome figure and links with CUL1, is one of the few neutrally selected genes due to its numerous binding partners. Yet Xiong et al. clearly demonstrate DAB2IP inhibition of β-catenin occurs through RAC1 sequestration. Overexpression of DAB2IP restores cytoplasmic localization of β-catenin in TNBC spheroids with RAC1 inhibition. While the mechanistic interaction of DAB2IP with RAC1 is the main novelty of the work performed by Xiong et al., their publication strengthens the field through thorough demonstrations of in vivo relevance in a TNBC setting. Silencing DAB2IP conferred an ability to initiate suspended spheroids from single cells, correlating well with the observation that mammary fat pad injected cells were twice as likely to form tumours with just 100 cells injected or thrice as likely with 1000 cells injected, when DAB2IP was silenced. Tumour stemness is implicated in chemoresistance and Xiong et al. showed that under weekly docetaxel treatment in mice, TNBC cells with silenced DAB2IP grew faster than control tumours. Translation of these findings into patient care remains a significant challenge. Combatting transcriptional reprograming broad-scale using DNA-methyltransferase inhibitors such as 5-azacytidine and decitabine has been proposed, including by Xiong et al. Indeed, combining decitabine with docetaxel showed substantial and significant additivity in MDA-MB-231 xenografts. However, human trials containing DNA-methyltransferase inhibitors struggle due to toxicity limitations.10 While there remains optimism that toxicity can be circumvented using newer DNA-methyltransferase inhibitors, this has yet to be proven. Nonetheless, impactful basic science findings like those included in Xiong et al. allow for informed design of new targets, such as short peptides or small molecules which may strengthen the interaction between DAB2IP and RAC1, diminishing the oncogenic β-catenin signalling effects of reduced DAB2IP expression. The authors have nothing to report. The author declares no conflict of interest.