As noted earlier, CHD1 is homozygously deleted in 8~18% of prostate cancer, supporting the hypothesis that CHD1 is a tumor suppressor in prostate cancer. Earlier in vitro studies using siRNA showed that downregulation of CHD1 in nontumorigenic prostate epithelial cells promoted cell invasiveness and enhanced cell clonogenicity, but had no impact on cell growth (17, 18). To obtain the genetic evidence, our and other independent groups established prostate-specific Chd1 deletion genetically engineered mouse (GEM) models, in which conditional Chd1 alleles deleted by a Probasin (Pb) promoter-driven Cre recombinase (Pb-Cre; Chd1^L/L ) (21, 23, 27). Homozygous deletion of Chd1 in prostate glands showed no observed differences in cell proliferation, cell survival, androgen receptor (AR) expression, or glandular structure (21, 23, 27). No invasive adenocarcinoma was observed in mice up to 1 year of age, as characterized by well-maintained smooth muscle actin structures (21). This genetic evidence suggests that Chd1 loss alone is insufficient to drive tumorigenesis in the prostate.

Notably, CHD1-depleted tumors often harbor additional genetic alterations, including SPOP mutations and MAP3K7 deletion, but also show mutual exclusivity with PTEN loss or ERG translocation ( Figure 2C ) (14, 16, 17, 20, 25, 26, 62). CHD1 depletion reduced cell proliferation, invasiveness, and tumor growth of PTEN-deficient cancer cells (14, 20, 25, 26); while loss of MAP3K7 and CHD1 coordinates to promote aggressive prostate cancer (20, 61). These observations seem paradoxical at first glance; however, they established the context-dependent roles of CHD1 in prostate cancer ( Figure 3 ). Importantly, CHD1’s distinct roles in different contexts are largely mediated by the crosstalk with diverse signaling pathways, which will be introduced individually in the following subsections.

Prostate cancer is largely driven by androgen receptor (AR) signaling. Androgen deprivation therapy (ADT) and AR inhibition are the main strategies for prostate cancer treatment (78). Although CHD1 protein does not directly bind to AR (14, 16, 23), loss of CHD1 caused the transcriptome reprogramming of AR signaling and is strongly associated with the ERG translocation (14, 16, 22, 23). By performing the chromatin-bound interactome analysis, Augello et al. uncovered that CHD1 interacts with the cofactors of AR and other nuclear receptors (23). Chromatin immunoprecipitation (ChIP) sequencing showed that CHD1 colocalizes to gene enhancers enriched for AR and its cofactors, such as HOXB13, ETV1, and FOXA1 (23). Specifically, they found that CHD1 localizes to chromatin-containing canonical AR binding sites, but CHD1 loss causes AR to redistribute to HOXB13-enriched sites, which drives a unique AR transcriptome that contributes to the tumor formation (23).

In prostate cancer, the most common genetic rearrangement involves the fusion of the androgen-regulated gene TMPRSS2 with the ETS transcription factor ERG (79). The fusion joins the 5′-UTR of TMPRSS2 (21q22) with the 3′-end of ERG (21q22) and leads to the TMPRSS2:ERG mRNA fusion transcript, which is induced by androgen. Using whole exome sequencing, FISH, or confocal microscopy, several groups showed the mutual exclusivity of CHD1 deletion with ERG fusion in human prostate tumors (14, 16, 62). CHD1 deletion is also strongly associated with early PSA recurrence (14, 59). Using a doxorubicin/dihydrotestosterone-induced DNA double-strand breaks system, Burkhardt and colleagues showed that CHD1 depletion prevents the formation of ERG rearrangements. Mechanistically, they found that CHD1 is required to recruit AR to responsive promoters and regulates the expression of AR-responsive genes, such as NKX3-1, FOXO1, and PPARγ (14). Given that AR-dependent transcription is a prerequisite for ERG translocation, these studies concluded that a functional CHD1 supports AR signaling transcriptome and ERG fusion development in prostate cancer.

Lysine-specific demethylase 1 (KDM1A/LSD1) removes the mono- and di-methylation from H3K4 and H3K9, and plays an important role in regulating AR-dependent gene expression in prostate cancer (80, 81). A prior study by the Schule group reported that the LSD1 protein is modified by di-methylation at K114 (K114me2) (82). By solving the cocrystal structure, they identified CHD1 as an LSD1-K114me2 reader and uncovered that chromatin colocalization of CHD1 and LSD1-K114me2 drive AR-dependent transcription and TMPRSS2-ERG translocation (82). This structural study provides additional evidence and mechanistic insight into CHD1’s roles in modulating AR signaling and ERG fusions during prostate cancer evolution.

Transcriptomic and epigenetic profiling studies in ESCs and cancer cells showed that CHD1 is required for sustaining the opening of chromatin and the global transcription (10, 22, 23, 25, 27, 83). CHD1 deficiency causes the accumulation of heterochromatin, diminishing the pluripotency of ESCs (10). In prostate cancer, CHD1 co-localizes with H3K4me3 to the promoters of actively transcribed genes, while CHD1 depletion reduces H3K4me3 marked genes, alters the chromatin assembly across the genome, and reprograms the global transcription (22, 23, 25). Lineage plasticity of cancer cells has been proposed as a source of intratumoral heterogeneity and resistance to targeted anticancer treatments (84). In prostate cancer, the histological transformation from AR-dependent adenocarcinoma to AR-indifferent neuroendocrine or small-cell carcinoma is a well-known pathway of lineage plasticity, which might occur as a consequence of ADT (85, 86). In addition to the deregulation of AR signaling, CHD1 loss is linked to lineage plasticity by inducing a lineage-specific transcriptome (20, 22).

Cramer’s group initially proposed this hypothesis. They found MAP3K7 and CHD1 were significantly co-deleted in localized prostate tumors and combined loss correlated with poor disease-free survival of patients (20, 61). CHD1 knockdown reduced cell proliferation, impaired tumor growth, and prolonged the overall survival of mice in PTEN-deficient LNCaP-derived xenograft models. However, additional MAP3K7 loss completely rescued this effect and promoted prostate cancer progression (20, 61). Co-suppression of MAP3K7 and CHD1 induces androgen-independent growth and causes resistance to AR inhibitors, such as enzalutamide (61). Combining mouse prostate epithelial progenitor/stem cells (PrP/SC) and tissue recombination model, they found that CHD1-depleted PrP/SCs grafts are mostly benign, characterized by intact p63+ basal layer (20). This is consistent with the phenotypes observed in Pb-Cre; Chd1^L/L GEM model (21, 23, 27). In contrast, MAP3K7-depleted grafts displayed a mixture of benign, high-grade prostatic intraepithelial neoplasia (PIN), and carcinoma phenotypes. Strikingly, dual MAP3K7–CHD1 loss grafts displayed high-grade PIN and invasive carcinoma phenotypes (20). Compared to MAP3K7 or CHD1 depletion alone, dual depletion caused lineage switching, characterized by loss of AR and epithelial markers (CK5, p63, CK14, and CK18) along with the upregulation of neuroendocrine differentiation markers (SYP and Nestin) and mucin production (20). It remains unclear if MAP3K7/CHD1 double-depletion affects metastatic progression. Nevertheless, better understanding their interactions and underlying mechanisms might provide novel therapeutic strategies for MAP3K7/CHD1 loss prostate cancer.

Recently, Zhang et al. showed that CHD1 loss renders prostate cancer cells more resistant to AR inhibition via inducing lineage plasticity (22). They showed that loss of CHD1 induces the transcription factors of GR, BRN2, TBX2, and NR2F1, which are required to promote tumor heterogeneity and resistance to AR inhibitors in CHD1-deficient tumors (22). They also found that enzalutamide-resistant xenograft tumors with CHD1 depletion and high expression of those transcription factors, lost luminal lineage identities (AR, CK8, and CK18), but displayed increased basal markers (CK5 and p63) and epithelial to mesenchymal transition genes (SNAI2, TWIST1, SNAI1, and ZEB1) (22). These non-luminal lineage programs and plastic chromatin landscape induced by CHD1 loss may serve as mechanisms to enable heterogeneous subclones less dependent on AR.

Endogenous cell metabolism and environmental factors often cause DNA double-strand break (DSB). Homologous recombination (HR) and non-homologous end joining (NHEJ) are two major repair mechanisms in response to DSB (87). Using prostate cancer cell lines and GEM models, several studies reported that CHD1 loss causes defects in HR-mediated DNA damage repair (DDR) and increases sensitivity to DNA-damaging therapies (21, 24, 88, 89). Besides, recent studies in metastatic prostate cancer patients showed that CHD1 deletion is associated with HR deficiency-related mutational signatures (59, 90).

Mechanistically, CHD1 accumulates at the DNA damage sites, maintains the open status of chromatin, and co-localizes with γH2AX in response to DNA damage (24). CHD1 interacts with and recruits DDR factors, such as CtIP, 53BP1, RIF1, and KU70, to the DNA damage sites (21, 24). CtIP is a key player in HR by resecting DSB ends. CHD1 loss impairs the recruitment of CtIP to DNA damage sites and suppresses the initiation of HR (24, 88). As a key DDR protein, 53BP1 maintains the balance of repair pathway choices and genomic stability. Shenoy et al. found that CHD1 forms a complex with NHEJ components and negatively regulates the protein stability of 53BP1 (21). CHD1 loss stabilizes 53BP1 protein and causes the switch from HR to NHEJ pathway for DSB repair. Although AR signaling is known to regulate the expression of DDR-related genes and promotes NHEJ repair, the role of CHD1 in modulating DDR is independent of the AR pathway (21). CHD1 loss is also associated with chromosomal and genomic instability in prostate cancers (21, 22), and DDR defects may serve as one of the mechanisms.

Recurrent missense mutations in SPOP (speckle-type BTB/POZ protein) occur in 10-15% of localized prostate tumors and metastatic CRPC (60, 91–93). In occurrence with CHD1 deletion ( Figure 2C ), SPOP mutations define a distinct prostate cancer subtype, characterized by genomic instability, increased AR transcriptional activity, absence of ERG rearrangements, and increased DNA methylation (60, 91, 92, 94). SPOP protein is a substrate adaptor for the Cullin3-RING-based BCR E3 ligase complex (CUL3-SPOP), which mediates the ubiquitination and proteasomal degradation of target proteins. In prostate cancer, hotspot SPOP mutations are only observed in the MATH domain that is responsible for substrate recognition and recruitment. The mutant reduces the substrate-binding affinity and results in the aberrant accumulation of substrates (95).

Several oncogenic proteins in AR signaling were identified as substrates of CUL3-SPOP, such as AR (96), SRC3 (97), and ERG (98, 99). CUL3-SPOP complex mediates the ubiquitination-degradation of AR by binding to the ⁶⁴⁵ASSTT⁶⁴⁹ Motif in the hinge domain of AR. Prostate cancer-associated SPOP mutants (Y87C, Y87N, F102C, S119N, F125V, W131G, F133L, and F133V) fail to bind AR protein, thereby increasing the protein stability and activity of AR during tumorigenesis (96, 100). By establishing a tissue-specific SPOP-F133V overexpressing GEM model, Blattner and colleagues reported that SPOP mutation promotes prostate tumorigenesis through coordinate regulation of PI3K/mTOR and AR signaling (101). Clinical trials in men with metastatic prostate cancer found that SPOP mutations are associated with improved survival outcomes after ADT (93, 102, 103). Although it remains unclear whether SPOP mutations crosstalk with CHD1 loss when regulating AR signaling, a clinical study in metastatic CRPC showed that SPOP mutations and CHD1 loss are associated with a higher response rate to abiraterone (inhibitor of androgen biosynthesis) and a longer time on the abiraterone treatment (93).

In addition to modulating AR signaling, coordinate CHD1 deletion and SPOP mutations are also involved in DNA damage response. Phenocopying CHD1 loss, SPOP mutations also cause genomic instability and impaired HR DSB repair, as well as promote the sensitivity of prostate tumors to DNA-damaging therapeutic agents, such as PARP inhibitors (94, 104, 105). Mechanistically, SPOP is accumulated at DNA double-strand break sites, where it interacts with ATM kinase and plays an essential for DDR (94, 105). Depletion or mutations of SPOP inhibits HR and promotes NHEJ by downregulating DNA repair factors (RAD51, BRCA2, CHK1, and ATR), reducing RAD51 foci formation, and stabilizing 53BP1 (94, 106). Recent studies found that SPOP mutations and CHD1 deletion sensitize prostate cancer cells to DNA damage inducers and show synergistic effects on the DNA damage repair (59, 89). By generating prostate-specific Chd1 and/or Spop deletion GEM models, Zhu and colleagues found that co-deletion of Chd1 and Spop in the prostate synergistically induces the response to DNA DSBs, characterized by increased γH2AX staining (89). Besides, they showed that the combination of CHD1 depletion and SPOP mutations significantly augmented the DNA damage response and sensitized human prostate cells to DNA-damaging agents (89). Another study in AA men revealed that, compared to cases with either alteration alone, prostate tumors with both CHD1 deletion and SPOP mutations showed significantly higher levels of HR deficiency-associated signatures and large-scale structural rearrangements (59). These studies demonstrated the synergistic effects of CHD1 loss and SPOP mutations in modulating AR signaling and DDR pathways, providing insights into the molecular basis of their frequent co-occurrence in prostate cancers.

Tumor suppressor PTEN is frequently altered in prostate and other cancer types. As a dual lipid and protein phosphatase, PTEN dephosphorylates PIP3 and suppresses the activation of AKT, leading to a hyperactive PI3K signaling (107, 108). PTEN/AKT pathway is critical for cellular processes, such as metabolism and cell proliferation (109). Genetic deletion and mutations of PTEN occur in ~20% of localized prostate tumors and are further enriched in ~40% of CRPC with strong associations with metastatic disease and poor overall outcome (60, 110). In prior studies, we found that CHD1 deletions show a mutually exclusive pattern with PTEN loss in prostate tumors ( Figure 2C ), and CHD1 negatively correlates with PTEN expression at protein levels (25). Mechanistically, PTEN loss stabilizes CHD1 protein in cancer cells and prostate tumors by disrupting CHD1’s ubiquitination and degradation (25–27), as described above. Functionally, we identified CHD1 as a synthetic essential gene in cancers containing PTEN deficiency (25–27). CHD1 depletion significantly suppressed tumor growth in PTEN-deficient xenograft models (25), consistent with earlier observations in LNCaP xenograft tumors (20). However, CHD1 knockdown showed little effect on benign prostatic hyperplasia cells or PTEN-intact tumors (25, 26).

In GEM models, Pb-Cre-driven Pten loss (Pb-Cre; Pten^L/L ) in the prostate triggers non-lethal invasive tumors after a long latency (111). By crossing a Chd1 conditional knockout allele into this GEM model, Augello et al. reported that CHD1 loss promotes prostate tumor progression (23). The limitations of this study include the small animal cohort (n = 5), low frequency of tumor progression (1 in 5 mice), and lack of survival data. In contrast, we established prostate-specific Chd1 deletion in two well-established PTEN-deficient GEM models, Pb-Cre; Pten^L/L and Pb-Cre; Pten^L/L; Smad4^L/L (112), and then determined the impact of Chd1 deletion with much larger cohorts (n = 22 or 18) (27). In both models, we found that CHD1 depletion significantly delayed the development and progression of PTEN-deficient prostate tumors and prolonged the survival of mice, providing genetic evidence supporting the essential roles of CHD1 in the context of PTEN defects (27). Given that CHD1-null prostates are phenotypically normal (21, 23, 27), these studies revealed the therapeutic potential of targeting CHD1 in PTEN-deficient tumors with an acceptable therapeutic window. Despite these encouraging factors, it is worth noting that tumor progression was rarely observed in some Pten/Chd1 double-knockout mice. Although these cases appear to result from clonal expansion of prostate cancer cells undergoing incomplete Chd1 deletion, future study is needed to identify potential second-site suppression events that may underlie CHD1 bypass. It will also provide rational combinatorial strategies targeting CHD1 in PTEN-deficient tumors.

Tumor development and progression are largely driven by interactions between cancer cells, extracellular matrix, stromal cells, and immune cells in the tumor microenvironment (TME) (78). Prostate cancer has a TME characterized in part by a relative paucity of infiltrating T cells and a high proportion of immunosuppressive myeloid cells, including myeloid-derived suppressor cells (MDSCs) and tumor-associated macrophages (TAMs) (78, 113). MDSCs are a heterogeneous group of myeloid cells that play immunosuppressive roles via interaction with T and NK cells (114). Our prior studies using multiple GEM models demonstrated that CHD1 is involved in the inflammatory response and plays a crucial role in modulating the TME via promoting MDSC infiltration and suppressing tumor-infiltrating lymphocytes (TILs) (27). In PTEN-deficient prostate tumors, CHD1 deletion caused reduced MDSC infiltration and increased CD8+ T cells (27). Transcriptional and epigenetic profiling analyses revealed that CHD1 cooperates with NF-κB, the central player of inflammation, to regulate the transcription of inflammatory genes (25, 27). Besides, we identified IL-6 as a direct target of CHD1 and mediates the recruitment and activation of MDSC, which contributes to T cell suppression in the prostate tumors (27). In addition to NF-κB and IL-6/Stat3 signaling, CHD1 modulates several other TME-related pathways, such as inflammatory response, interferon alpha and gamma pathways, and angiogenesis (23, 27).

A recent immunogenicity study in localized prostate cancer provides additional evidence. Using multiplex immunofluorescence, Calagua and colleagues identified the genomic alterations associated with immunogenic (PD-L1 ≥5% and extensive TILs) and nonimmunogenic (PD-L1 negative and no TILs) tumor foci (115). They found that deep deletions of CHD1 are strongly associated with dendritic cell signatures and immunogenic phenotype, characterized by enriched T cell infiltration (115). The regulatory axis of CHD1/IL-6/MDSC may serve as one mechanism by which CHD1 loss drives immunogenicity. Besides, immunogenic localized prostate cancer shows high rates of genomic instability and variable tumor mutational burden (TMB) (115), suggesting chromatin instability and DDR defects induced by CHD1 loss may also contribute to immunogenic features.

As noted above, CHD1 plays a key role in DNA damage response and modulates the choice between HR and NHEJ DDR pathways. Several preclinical studies using prostate cancer cell lines, PDX models, and GEM models demonstrated that CHD1 loss leads to hypersensitivity to ionizing radiation (IR), PARP inhibition, and DNA-damaging agents, such as mitomycin C, carboplatin, irinotecan, and camptothecin (21, 24, 88–90).

By comparing the response of wildtype and Chd1-null (Pb-Cre; Chd1^L/L ) mice to a single dose of 10 Gy of IR, Shenoy and colleagues found that Chd1 deleted prostate tissues and ESCs are more sensitive to IR, as evidenced by increased γH2AX and phosphorylation of H2A and p53 (21). Similar phenotypes were also observed in CHD1-depleted prostate cell lines (21, 24). PARP (Poly-ADP-ribose polymerase) detects and initiates single-strand DNA breaks (SSB) DNA damage repair. Prior studies uncovered that PARP inhibitors have synthetic lethal effects in cells with HR defects, such as BRCA1 and BRCA2 loss (116). PARP inhibitors have been clinically tested in CRPC, and genetic alterations in DDR pathways are associated with better responses (117, 118). Preclinical studies showed that CHD1 loss-induced HR defects sensitize prostate tumors to PARP inhibitors, Olaparib and Talazoparib, both in vitro and in vivo (21, 24, 59, 88), suggesting CHD1 might be a predictive biomarker. The second-generation platinum agent, carboplatin, also showed a good response in a metastatic CRPC patient with homozygous CHD1 loss (21).

Notably, SPOP depletion also sensitizes cancer cells to IR and PARP inhibitors (94, 104, 105). A recent study demonstrated that SPOP mutations and CHD1 loss synergistically promote sensitivity to camptothecin, an inducer of double-strand breaks (89). Given that co-occurrence of SPOP mutations and CHD1 deletion define a distinct molecular subtype of prostate cancer, further studies are needed to assess if they have synergistic effects in response to DNA-damaging therapies. Their potential as biomarkers for predicting the response to radiotherapy, PARP inhibitors, and DNA-damaging agents in advanced prostate cancers remains to be determined.

In 1941, Huggins and Hodges reported that castration led to tumor regression in prostate cancer patients, first recognizing hormone responsiveness as a central feature of prostate cancer (119). Androgen deprivation by castration or agents that block the androgen pathway is the standard of care for prostate cancer. Resistance to ADT facilitates the development of CRPC with high rates of metastasis and mortality (120). Given the important role of CHD1 in AR signaling, preclinical and clinical studies have been conducted to determine the impact of CHD1 loss on response to antiandrogen therapies using different model systems (14, 16, 22, 23, 61).

Using an androgen-driven regrowth model, Augello et al. showed castrated Chd1-deficient mice (Pb-Cre; Chd1^L/L ) showed increased proliferation in regenerated epithelium upon androgen re-stimulation, suggesting Chd1 deletion may render the prostate tissue more dependent to androgen (23). However, Zhang et al. used AR-overexpressing LNCaP models and showed that CHD1 loss renders human prostate cancer cells more resistant to AR inhibitors in vitro and in vivo in castrated mice (22). They also found that low expression of CHD1 is associated with shorter clinical response to next-generation antiandrogen therapies (enzalutamide or abiraterone) in CRPC patients (22). Along the same line, Jillson and colleagues showed that co-suppression of MAP3K7 and CHD1 causes androgen-independent growth of prostate cancer cells and promotes resistance to AR inhibitor enzalutamide (61).

In prostate cancer patients, CHD1 loss was associated with a shorter time to PSA recurrence, suggesting its potential as a prognostic biomarker (14, 59, 61, 121). However, recent clinical trials in men with metastatic prostate cancer found that SPOP mutations are associated with improved survival outcomes after ADT (93, 102, 103). When considered as an individual variable, CHD1 loss is associated with a higher response rate to abiraterone (OR, 7.30, P= 0.08) and a longer time on abiraterone (HR, 0.50, P = 0.06) in metastatic CRPC patients (93). Prospective clinical trials are needed to validate the impact of CHD1 deletion on response to castration, abiraterone, enzalutamide, and other antiandrogen drugs in both hormone-sensitive and -resistant prostate cancers. Given the context-dependent role of CHD1 in prostate tumors, genes showing co-occurrence (SPOP or MAP3K7) or mutual exclusivity (ERG and PTEN) should also be considered as influence factors in these clinical studies.

Notably, the upregulation of transcription factors of GR, BRN2, TBX2, and NR2F1 was found to mediate the resistance to enzalutamide in CHD1-deficient prostate cancer, since inhibition of each factor re-sensitizes CHD1 loss prostate tumors to AR inhibitor (22). This offers new insights into synthetic lethal interactions with CHD1 and potential therapeutic vulnerabilities in prostate cancers containing CHD1 deficiency. Given that GR (Glucocorticoid Receptor) inhibition has been tested in clinical studies of CRPC (NCT02012296), future biomarker studies are needed to assess if GR inhibition is more effective in CRPC patients harboring CHD1 loss.

Our prior studies in xenograft and GEMM models established CHD1 as a synthetic essential gene and potential therapeutic target in prostate cancers containing PTEN defects (25–27). Several independent groups are dedicated to developing small-molecule inhibitors targeting CHD1, and the efficacies of top hits will be tested in cancer cell lines and diverse preclinical models. We expect that these drugs have better therapeutic effects on PTEN-deficient tumors but may have modest effects on PTEN-intact tumors. When some of them enter the early clinical phase, it is important to use PTEN as a biomarker for patient selection. Given that CHD1 inhibition sensitizes tumor cells to DNA-damaging agents, the combination of CHD1 inhibitors and DNA-damaging therapies should be tested in preclinical and clinical studies as well. It is also worth determining if CHD1 inhibitors synergize with AR or GR inhibitors in suppressing CRPC tumor growth and progression. However, caution should be taken when pharmacologically inhibiting CHD1 in prostate cancer with SPOP or MAP3K7 deletions, reasoning that CHD1 inhibition may play tumor-promoting roles in these contexts.

Combining high-throughput epigenetic screening and pan-cancer drug sensitivity analyses, we reported that CHD1 promotes the susceptibility of cancer cells to inhibitors targeting Aurora kinases (26). Aurora kinases are key players in mitotic control. Among three mammalian paralogues, Aurora A is required for centrosome maturation and mitotic spindle assembly (122–124). Several small-molecule inhibitors targeting Aurora kinases have been tested in clinical trials, and subsets of patients showed significant clinical benefits from the single agent or in combination with other agents (125–131).

In our recent study, we uncovered that CHD1 loss impaired the in vitro and in vivo efficacy of Aurora kinase inhibitors, while high expression of CHD1 is associated with increased sensitivity in a pan-cancer manner (26). Prior studies demonstrated that the activity of Aurora A is largely modulated by the autophosphorylation and interaction with the co-activator TPX2 (132–135). Mechanistic studies revealed that the regulatory axis of CHD1-KPNA2 suppressed the interaction between Aurora A and TPX2, thereby rendering cancer cells more vulnerable to Aurora A inhibition (26). Furthermore, our studies in GEM models, patient-derived organoids, and patient samples showed that PTEN defects are associated with a better response to Aurora A inhibition in advanced prostate cancer by inducing CHD1 protein stabilization (26). This study establishes the important role of CHD1 in modulating Aurora kinases and provides insights for using PTEN and CHD1 as predictive biomarkers to improve patient selections in clinical trials of Aurora A inhibitors.

Immunotherapy has shown only modest activity in advanced prostate cancer, partially due to low tumor mutation burden (TMB), lack of infiltrating T cells, and immunosuppressive TME (78, 113). Immune checkpoint inhibitors that target cytotoxic T-lymphocyte-associated protein 4 (CTLA-4), programmed death 1 (PD-1), and its ligand (PD-L1) display minimal or no activity as single agents or in combination with AR inhibitors in advanced prostate cancers (136–141). As noted above, CHD1 contributes to immunosuppressive TME by promoting MDSCs and suppressing tumor-killing T cells (27). Our recent studies in GEM and syngeneic models revealed that depletion of CHD1 reverses the immunosuppressive TME and sensitizes prostate tumors to the checkpoint immunotherapy (27). As a direct target gene of CHD1, IL-6 mediates the recruitment and activation of MDSCs in prostate tumors. Phenocopying CHD1 depletion, pharmacological inhibition of IL-6 and dual blockade of PD-1/CTLA-4 showed synergistic effects in preclinical models of PTEN-deficient prostate cancer (27). Notably, IL-6 inhibition was found to reduce immune-related adverse events in patients by de-coupling autoimmunity from antitumor immunity induced by immune checkpoint blockade (142). Further clinical studies are needed to test the above combinations in CRPC patients, particularly in PTEN-loss/CHD1-high tumors.