Prostate cancer (PCa) is the most common solid tumor malignancy in men. Activation of ETS transcription factors (TFs), ERG, ETV1, ETV4, and ETV5, are present in approximately 60% of PCa, underscoring their importance in prostate oncogenesis (Nelson et al., 2003). In human PCa, ETS TFs are most commonly activated through gene rearrangements that fuse the androgen responsive gene, TMPRSS2, to either ERG, ETV1, ETV4, or ETV5 (Sizemore et al., 2017). Beyond ETS TF fusions, little is known about the underlying molecular mechanisms that lead to increased expression of wildtype (WT) ETS factors, which confer aggressive malignant phenotypes and associate with poor clinical outcomes in fusion negative PCa patients (Baena et al., 2013).

Capicua (CIC) is a High-mobility group (HMG) box TF that silences ETV1, ETV4, and ETV5 through direct target gene repression (Kim et al., 2021; Simón-Carrasco et al., 2018). CIC is frequently altered in human cancer, where it functionally suppresses tumor growth and metastasis (Ahmad et al., 2019; Bettegowda et al., 2011; Choi et al., 2015; Dissanayake et al., 2011; Kim et al., 2018; Okimoto et al., 2017; Seim et al., 2017; Yoshiya et al., 2021). Notably, in PCa, CIC is commonly altered through genomic loss (homozygous and heterozygous deletion) in ~10% of PCa patients (Abida et al., 2019; Grasso et al., 2012; Hieronymus et al., 2014; Huang et al., 2017; Robinson et al., 2015; Cancer Genome Atlas Research Network, 2015) and inactivation of CIC de-represses ETV1, ETV4, and ETV5 transcription to promote tumor progression (Bettegowda et al., 2011; Choi et al., 2015; Kim et al., 2018; Okimoto et al., 2017). Leveraging mutational profiling data from multiple PCa cohorts, we previously observed concurrent loss of the ETS2 repressor factor (ERF) in CIC-deficient prostate tumors (Huang et al., 2017). Combinatorial loss is most commonly the result of focal deletions (homozygous and heterozygous) at the 19q13.2 locus, where CIC and ERF are physically adjacent (long and short isoforms of CIC are separated from ERF by approximately 15 and 30 kb, respectively) to one another in the genome. Since ERF is a transcriptional repressor that binds ETS DNA motifs (Bose et al., 2017), we hypothesized that in a fusion independent manner, CIC and ERF cooperate to mutually suppress ETS target genes in PCa.

Through an integrative genomic and functional analysis, we mechanistically show that CIC and ERF directly bind and co-repress a proximal ETV1 regulatory element limiting PCa progression. Concomitant loss of CIC and ERF de-represses ETV1-mediated transcriptional programs and confer ETV1 dependence in multiple PCa model systems. Thus, we reveal a fusion-independent mechanism to de-repress ETS-mediated PCa progression and potentially uncover a therapeutic approach to target CIC-ERF co-deleted PCa.

CIC is a TF that directly suppresses ETV1, ETV4, and ETV5 TF family members (Futran et al., 2015; Jiménez et al., 2012; Kim et al., 2021; Okimoto et al., 2017). CIC silences target genes through direct binding of a highly conserved DNA-binding motif (T[G/C]AAT[G/A]AA; Figure 1A; Ajuria et al., 2011; Futran et al., 2015; Jiménez et al., 2012). CIC is commonly altered in multiple human cancer subtypes where it suppresses tumor growth and metastasis (Kim et al., 2021). CIC is located on chromosome 19q13.2, directly adjacent to another transcriptional repressor, namely the ERF (Figure 1B). ERF binds and competes for ETS TF-binding sites (GGAA-motifs) and is frequently altered in human PCa, predominantly through focal deletions (Bose et al., 2017; Hou et al., 2020). We thus hypothesized that concurrent loss of CIC and ERF may de-repress an ETS-driven transcriptional program that drives PCa progression in a fusion-independent manner. To explore this, we first queried 15 PCa datasets curated on cBioPortal (Cerami et al., 2012; Gao et al., 2013) and identified a high cooccurrence rate (p<0.001, two-sided Fisher’s exact test [FET]) for CIC (10%) and ERF (12%) homozygous and heterozygous deletions (Figure 1C), suggesting that concurrent loss occurs through focal copy number change at the 19q13.2 locus. Through analysis of these clinically annotated specimens, we observed that the CIC-ERF co-deletion was present at an increased frequency in PCa with higher Gleason scores and later tumor stages when compared to CIC-ERF replete tumors (Figure 1D). In order to understand the association between CIC and/or ERF alterations in specific PCa cohorts, we stratified published datasets to identify patients that represent primary PCa (Fraser et al., 2017; Hieronymus et al., 2014; ‘ Cancer Genome Atlas Research Network, 2015) (PNAS 2014, n=272; Cell 2015, n=333; Nature 2017, n=477 primary PCas) and metastatic castrate resistant prostate cancer (mCRPC) (Abida et al., 2019; Grasso et al., 2012; Robinson et al., 2015) (Nature 2012, n=50; Cell 2015, n=150; PNAS 2019, n=429 mCRPCs). This analysis revealed enrichment of CIC and ERF alterations including the CIC-ERF co-deletion in mCRPC samples (Figure 1E, Figure 1—source data 1, Figure 1—figure supplement 1). Importantly, CIC-ERF co-deleted tumors clustered as a subgroup when compared to the more well-characterized molecular subsets including ERG, ETV1, ETV4, SPOP, and FOXA1 altered PCas, suggesting a distinct molecular subtype of PCa (Figure 1F). In order to explore clinical outcomes of patients harboring CIC-ERF co-deleted tumors, we performed a survival analysis using the aforementioned PCa datasets and observed significantly worse outcomes in patients who harbored the CIC-ERF co-deletion (p=0.001, disease-free survival [DFS] [ERF-CIC co-deletion 25 events/90 total; no ERF-CIC co-deletion 153 events/910 total] and p=0.01, progression-free survival [PFS] [ERF-CIC co-deletion 16 events/52 total; no ERF-CIC co-deletion 77 events/442 total]; Figure 1G). One limitation of this analysis is that individual studies included overlapping tumors from the same patient (i.e. TCGA PanCancer and Firehose Legacy cohorts). Despite this overlap, we utilized this insight to formulate a hypothesis that CIC and ERF are co-deleted with increasing frequency in mCRPC and that the CIC-ERF co-deletion is associated with worse clinical outcomes in PCa patients.

Figure 1 with 1 supplement see all

Download asset Open asset

Figure 1—source data 1

Prostate cancer studies identified in cBioPortal demonstrating the total number of patients, number of patients with shallow or deep deletions in Capicua (CIC)-ETS2 repressor factor (ERF), and the frequency of CIC-ERF alterations in each cohort.

Studies that analyzed predominantly primary prostate cancers (green) and metastatic castrate resistant prostate cancer (mCRPC; yellow) are highlighted.

https://cdn.elifesciences.org/articles/77072/elife-77072-fig1-data1-v1.xlsx

Download elife-77072-fig1-data1-v1.xlsx

Our initial findings provided rationale to explore the genetic and functional relationship between CIC and ERF. Independently, CIC and ERF have previously been reported to promote malignant phenotypes, including tumor growth and metastasis in multiple human cancer subsets (Bettegowda et al., 2011; Bose et al., 2017; Choi et al., 2015; Huang et al., 2017; Kawamura-Saito et al., 2006; Kim et al., 2018; Bunda et al., 2019; Okimoto et al., 2019; Okimoto et al., 2017; Simón-Carrasco et al., 2017; Wong et al., 2019; Yang et al., 2017). Since our clinical data indicated that combinatorial loss of CIC and ERF was associated with worse patient outcomes, we hypothesized that CIC and ERF loss may cooperate to enhance PCa progression. To investigate this, we engineered ERF, CIC, or both CIC and ERF-deficient immortalized prostate epithelial cells (PNT2) and performed a series of in vitro and in vivo experiments to test the combinatorial effect of the CIC-ERF co-deletion. Compared to single gene loss of CIC or ERF, genetic inactivation of both CIC and ERF increased colony formation (Figure 2A–B, Figure 2—figure supplement 1A) and spheroid formation (Figure 2C–D) in PNT2 cells. CIC-ERF co-deletion also enhanced malignant phenotypes including cellular viability, invasiveness, and migratory capacity in PNT2 (Figure 2E–G). Additionally, genetic silencing of both CIC and ERF increased the frequency of subcutaneous tumor xenograft formation in severe-combined immunodeficient (SCID) mice compared to control (Figure 2H, Figure 2—figure supplement 1B). Thus, our findings demonstrate that the combination of CIC and ERF loss augments the transformation of PNT2 and promotes malignant phenotypes.

Figure 2 with 1 supplement see all

Download asset Open asset

To assess the functional role of CIC and ERF in the context of human PCa progression, we leveraged two genetically annotated, androgen-insensitive PCa cell lines, DU-145 (moderate metastatic potential) and PC-3 (high-metastatic potential). DU-145 cells harbor a loss-of-function ERF mutation (ERF^A132S) (Huang et al., 2017) and express functional WT CIC, (Figure 3—figure supplement 1A). By comparison, PC-3 cells are deficient in CIC (homozygous deletion) and retain functional ERF (Figure 3—figure supplement 1B; Cerami et al., 2012; Gao et al., 2013; Seim et al., 2017). Thus, these cell line models provided isogenic systems to functionally interrogate the role of CIC and ERF in human PCa. Specifically, genetic reconstitution of ERF into ERF-deficient DU-145 cells decreased colony formation in both CIC proficient (parental cells) and CIC knockout (KO) conditions (Figure 3A–B, Figure 3—figure supplement 1C, D). While CIC loss did not enhance colony formation in ERF-deficient DU-145 cells, it significantly increased tumor cell viability, invasion, and migratory capacity compared to control (Figure 3C–D, Figure 3—figure supplement 1E). Importantly, we observed that ERF expression in CIC KO DU-145 cells rescued the CIC-mediated effects on viability and migration/invasion (Figure 3C–D, Figure 3—figure supplement 1E). These findings indicate that rescuing ERF can partially restore the functional effects of CIC loss in DU-145 cells. To further understand if ERF could suppress tumor growth in vivo, we reconstituted WT ERF into CIC proficient and deficient (CIC KO) DU-145 cells and generated subcutaneous xenografts in immunodeficient mice (NU/J). Consistent with our in vitro data, CIC KO increased the tumor growth rate in vivo and genetic reconstitution of WT ERF partially suppressed tumor growth compared to DU-145 CIC KO cells (Figure 3E). Moreover, genetic reconstitution of ERF into CIC proficient DU-145 cells suppressed the tumor growth rate in vivo (Figure 3E). We next used PC-3 cells, to further test how ERF and CIC functionally interact in the context of human PCa. We first noted that ERF overexpression (OE) or reconstitution of CIC alone decreased PC-3 colony formation, with combinatorial ERF OE and CIC rescue having the most significant reduction compared to parental PC-3 cells (Figure 3F–G, Figure 3—figure supplement 1F, G). Moreover, ERF and CIC expression had a similar impact on decreasing PC-3 viability, invasion, and migratory capacity (Figure 3H–I, Figure 3—figure supplement 1H). Similarly, genetic suppression of ERF resulted in a significant increase of colony formation, viability, and invasion compared to parental PC3 cells (Figure 3F–I, Figure 3—figure supplement 1I). Interestingly, OE of ERF in mice bearing CIC deficient PC-3 tumor xenografts decreased tumor growth compared to PC-3 parental and PC-3 cells expressing WT CIC (genetic rescue of CIC; Figure 3J). Since we consistently observed that ERF expression could partially rescue the effects of CIC loss in PCa, we hypothesized that WT CIC and ERF potentially cooperate to limit PCa progression.

Figure 3 with 1 supplement see all

Download asset Open asset

In order to mechanistically define how CIC and ERF (two TFs with known repressor function) were interacting to functionally regulate PCa, we performed chromatin immunoprecipitation followed by sequencing (ChIP-Seq) using a validated CIC antibody (Lin et al., 2020; Okimoto et al., 2019; Okimoto et al., 2017) in PNT2, and compared this to a publicly available ERF ChIP-Seq dataset in VCaP PCa cells (Bose et al., 2017), we were unsuccessful at pulling down ERF in PNT2 cells. This analysis identified 178 high-confidence (False-discovery rate (FDR) ≤0.05) CIC peaks that mapped to 130 annotated genes including known targets, ETV1, ETV4, and ETV5. Globally, CIC peaks were localized to distinct genomic regions including promoters (38.6%), untranslated regions (UTRs) (1.14%), introns (19.9%), and distal intergenic regions (38.64%). Interestingly, the distribution of ERF peaks was similar to CIC, with 32.8% in promoters, 1.2% in UTRs, 29.2% intronic, and 33.8% in distal intergenic regions (Figure 4A). Next, through a comparative ChIP-Seq analysis (Figure 4B), we identified 91 shared CIC and ERF target genes. Importantly, we focused on genes with shared CIC- and ERF-binding sites to potentially explain the functional cooperativity that we observed in our prior studies. In order to narrow down potential candidates, we performed Functional Clustering Analyses through the DAVID bioinformatics database (https://david.ncifcrf.gov) (Huang et al., 2009; Sherman et al., 2022) using the 91 shared CIC and ERF target genes (Supplementary file 1). Among these putative CIC and ERF targets, the PEA3 (ETV1, ETV4, and ETV5) TFs (known oncogenic drivers in PCa; Baena et al., 2013; Kim et al., 2016; Oh et al., 2012) were found to be the most highly enriched family. These findings suggested that CIC and ERF may co-regulate PEA3 family members through direct transcriptional control.

Figure 4 with 1 supplement see all

Download asset Open asset

Figure 4—source data 1

Full-length PCR gel images of ETV1 after Capicua (CIC) pull down in prostate epithelial cells (PNT2).

Cropped images and description shown in Figure 4E.

https://cdn.elifesciences.org/articles/77072/elife-77072-fig4-data1-v1.zip

Download elife-77072-fig4-data1-v1.zip

Figure 4—source data 2

Full-length PCR gel images of ETV1 after ETS2 repressor factor (ERF) pull down in prostate epithelial cells (PNT2).

Cropped images and description shown in Figure 4F.

https://cdn.elifesciences.org/articles/77072/elife-77072-fig4-data2-v1.zip

Download elife-77072-fig4-data2-v1.zip

Figure 4—source data 3

Full-length PCR gel images of ETV1 after ETS2 repressor factor (ERF) pull down in DU-145 cells.

Cropped images and description shown in Figure 4G.

https://cdn.elifesciences.org/articles/77072/elife-77072-fig4-data3-v1.zip

Download elife-77072-fig4-data3-v1.zip

In order to further identify how CIC and/or ERF impact PEA3 TF expression, we performed quantitative real time polymerase chain reaction (qRT-PCR) in PNT2 cells, assessing ETV1, ETV4, or ETV5 mRNA levels in response to genetic silencing of CIC and/or ERF. As expected, CIC KO and/or combinatorial CIC and ERF loss in PNT2 cells (ERF and CIC WT) consistently increased ETV1, ETV4, and ETV5 levels compared to control (Figure 4—figure supplement 1A-F). In contrast, genetic silencing of ERF consistently increased ETV1 mRNA expression, but not ETV4 or ETV5 (Figure 4C, Figure 4—figure supplement 1G-J). These findings indicated that ETV1 may be a shared transcriptional target of CIC and ERF in prostate cells. In order to test a potential proteomic interaction between CIC and ERF, we performed co-immunoprecipitation (Co-IP) experiments but did not observe binding when either ERF or CIC was pulled down in HEK293T cells (Figure 4—figure supplement 1K-L). This led us to hypothesize that CIC and ERF may potentially bind (but not interact) to distinct regions along the ETV1 regulatory element. To explore this, we first localized CIC (TGAATGGA) and ERF (GGAA) DNA binding motifs within the proximal upstream regulatory element of ETV1 and independently confirmed CIC and ERF occupancy of the ETV1 promoter through ChIP-PCR (Figure 4D–G, Figure 4—figure supplement 1M, P). To extend these findings into the context of PCa, we reconstituted ERF in ERF deficient DU-145 cells and this consistently decreased ETV1 expression (not ETV4 or ETV5) in both CIC proficient and CIC deficient settings (Figure 4H, Figure 4—figure supplement 1Q-T). Moreover, ERF Knockdown (KD) or ERF OE in CIC deficient PC-3 cells increased and decreased ETV1 mRNA expression, respectively (Figure 4I–J). Since ETV1 is a known target of CIC (Dissanayake et al., 2011; Jiménez et al., 2012), we focused on further validating ETV1 as a molecular target of ERF. To this end, we engineered a ETV1 luciferase based promoter assays and observed a decrease in luciferase activity following ERF expression in 293T cells (Figure 4K). These genetic tools further validate that ERF can suppress ETV1 expression through direct transcriptional silencing of the ETV1 promoter and identifies ETV1 as an ERF target.

Consistent with a repressor function, ERF loss was previously shown to transcriptionally associate with ETV1-regulated gene set signatures (Huang et al., 2017). Yet it remains unclear if ERF can directly regulate ETV1 and how combinatorial loss of CIC and ERF controls ETV1-mediated (or other ETS family members) transcriptional programs. To explore this, we generated a signature gene set of upregulated genes from our dual CIC and ERF deficient PNT2 cells (ERF KD +CIC KO) and projected the Cancer Genome Atlas PCa (TCGA-PRAD) dataset onto the transcriptional space of these signature gene sets using the ssGSEA module (Version 10.0.9) on GenePattern (Reich et al., 2006). We found that CIC and ERF loss were significantly associated with the ETV1-regulated gene set (Information coefficient (IC)=0.619, p=0.0009), but was anti-correlated with the TMPRSS2-ERG fusion signature gene set (Setlur et al., 2008; Figure 4L, IC = –0.45, p=0.0009). Thus, the enrichment of the ETV1-regulated gene set signature was shared between ERF loss alone (Huang et al., 2017) and CIC-ERF dual suppression (Figure 4L). In contrast, combinatorial CIC and ERF loss negatively correlated with the TMPRSS2-ERG fusion signature gene set, which was not consistent with our prior studies using ERF KD alone (Huang et al., 2017). These findings led us to hypothesize that the dual suppression of CIC and ERF may increase ETV1-mediated transcriptional programs in PCa. This was supported by two major lines of experimental and conceptual evidence including: (1) dual suppression of ETV1 expression and ETV1 mediated transcriptional output by CIC and ERF; and (2) the majority of tumors derived from PCa patients that harbor ERF deletions also contain deletions in CIC.

In order to demonstrate enhanced survival dependence on ETV1 in PCa cells, we genetically silenced CIC in ERF-deficient DU-145 cells and assessed drug sensitivity to an ETV1 inhibitor (BRD32048) (Pop et al., 2014). We validated the pharmacologic effect of ETV1 inhibition (BRD32048) through suppression of known downstream target genes (ATAD2 and ID2) using qPCR in LNCaP cells (Figure 5—figure supplement 1A-C). Since BRD32048 was previously shown to decrease invasiveness in ETV1-fusion positive PCa cells (LNCaP), but not significantly impact tumor cell viability, we unexpectedly observed that silencing CIC in DU-145 cells mildly enhanced sensitivity to BRD32048 (Figure 5A). To further confirm these findings and to mitigate potential off-target effects, we silenced ETV1 using siRNA in PC3 and DU-145 cells expressing Crispr-based sgRNA targeting CIC (Figure 5—figure supplement 1D-F). Consistent with our pharmacologic studies, the viability of DU-145 CIC KO cells was decreased upon genetic ETV1 inhibition (Figure 5B). Similarly, pharmacologic and genetic ETV1 inhibition decreased invasiveness of CIC deficient DU-145 cells (Figure 5C–D). These findings indicate that loss of CIC in ERF deficient PCa cells can potentially modulate the sensitivity to ETV1-directed therapies. To expand these findings, we also consistently observed a decrease in viability and invasiveness upon ETV1 inhibition (BRD32048) in PNT2 cells with dual ERF and CIC suppression and in PC-3 cells with ERF KD (Figure 5—figure supplement 2A-J). These in vitro findings indicate that ETV1 inhibition in CIC and ERF deficient prostate cells can suppress invasion and potentially limit viability. Further studies targeting ETV1 in patient derived specimens that harbor endogenous CIC-ERF co-deletions is warranted. Collectively, our data indicate that CIC and ERF may cooperate to silence ETV1 transcriptional programs, limiting ETV1-mediated PCa progression.

Figure 5 with 2 supplements see all

Download asset Open asset

Molecular and functional subclassification of human PCa has revealed a dependence on ETS family TFs including ERG, ETV1, ETV4, and ETV5 (Feng et al., 2014; Oh et al., 2012). The predominant mode of ETS activation in PCa is through chromosomal rearrangements that fuse ERG, ETV1, ETV4, and ETV5 to the androgen-regulated TMPRSS2 gene, leading to fusion oncoproteins that drive oncogenesis (Clark and Cooper, 2009; Feng et al., 2014; Helgeson et al., 2008; Tomlins et al., 2006; Tomlins et al., 2005). Interestingly, recent data indicate that ETV1, ETV4, and ETV5 are upregulated in a fusion-independent manner and are associated with poor clinical outcomes in PCa patients (Baena et al., 2013; Hermans et al., 2008). Our study focused on understanding the molecular mechanisms that drive fusion-independent upregulation of ETS family members and we reveal a new molecular subclass of PCa defined by a co-deletion of two TFs, CIC and ERF.

CIC is a TF that directly silences ETV1, ETV4, and ETV5 transcription through direct repression at proximal regulatory sites (Jiménez et al., 2012). We observed that in ~10–12% of human PCa, CIC and ERF are co-deleted through focal homozygous or heterozygous deletions. It has been recently shown that ERF competes for ETS DNA binding motifs and our studies identify cooperative regulation of key target genes between CIC and ERF. Specifically, through ChIP-Seq analysis coupled with a series of in vitro and in vivo studies, we identify a coordinated binding of CIC and ERF to the proximal ETV1 regulatory element that physically and functionally regulates ETV1 expression. Therefore, we reveal ETV1 as a novel ERF target gene. Rescuing ERF in CIC deficient PCa cells decreases ETV1 expression and limits malignant phenotypes including viability, migration and invasion.

The 19q13.2 locus contains CIC and ERF, which are physically adjacent and oriented in opposing directions. The long and short isoforms of CIC are separated from ERF by ~15 and 30 kb, respectively. Thus, future studies directed at defining the genome topology and epigenetic states, both within and around this highly conserved region are warranted. In particular, studies aimed at mapping topology associated domains and key histone marks can potentially reveal shared upstream regulatory elements including enhancers or superenhancers that may co-regulate CIC and ERF in concert. These findings could reveal non-genetic mechanisms to functionally regulate CIC and ERF expression in a coordinated fashion.

Beyond PCa, we and others have observed that CIC and ERF are co-deleted in a subset of stomach adenocarcinoma (LeBlanc et al., 2017; Okimoto et al., 2017). Thus, combined CIC and ERF loss are not entirely specific for PCa. We speculate that since prostate (Baena et al., 2013) and potentially stomach adenocarcinoma (Keld et al., 2011) are highly dependent on PEA3 transcriptional dysregulation to enhance tumor progression, the dual loss of CIC and ERF may, in part, represent an alternative mode to de-repress ETS mediated transcriptional programs in these cancer subsets.

The use of ETV1 inhibitors has been limited to preclinical studies (Pop et al., 2014). These studies have largely focused on direct targeting of ETV1 fusion oncoproteins in PCa (Pop et al., 2014). Our findings indicate that patients with CIC-ERF deficient PCa may have a survival dependence on ETV1 and that these patients may potentially benefit from ETV1 directed therapies. A limitation of this targeted approach in PCa patients is that we did not find a statistically significant difference in ETV1 transcript levels in CIC-ERF co-deleted tumors compared to CIC-ERF replete tumors in the TCGA-PRAD (n=455) dataset. One potential explanation is that ETV1 is commonly upregulated in human PCas through mechanisms beyond CIC-ERF loss. Thus, while our data support the upregulation and potential induced dependence on ETV1 in our CIC-ERF deficient systems, it remains unclear if this will translate beyond our cell-line based models into PCa patients that harbor CIC-ERF co-deletions. Thus, larger studies that aim to evaluate: (1) ETV1 mRNA and protein expression levels; (2) the biological significance of ETV1 function; and (3) the clinical application of ETV1 inhibitors in patients with endogenous CIC-ERF co-deleted tumors (compared to CIC-ERF WT tumors) is warranted in this subset of PCa. Collectively, we have uncovered a molecular subset of PCa defined by a co-deletion of CIC and ERF and further demonstrate a mechanism-based strategy to potentially limit tumor progression through ETV1 inhibition in this subset of human PCa.

All sequencing data including RNASeq and ChIP seq have been deposited in GEO under accession codes GSE216732 and GSE216318, respectively.

1. 1. Gupta N
   2. Song H
   3. Wu W
   4. Kriska Ponce R
   5. Lin Y
   6. Kim JW
   7. Small EJ
   8. Feng FY
   9. Huang FW
   10. Okimoto RA

   (2022) NCBI Gene Expression Omnibus

   ID GSE216732. Decoding the Protein Composition of Whole Nucleosomes with Nuc-MS.

   https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE216732
2. 1. Gupta N
   2. Song H
   3. Wu W
   4. Kriska Ponce R
   5. Lin Y
   6. Kim JW
   7. Small EJ
   8. Feng FY
   9. Huang FW
   10. Okimoto RA

   (2022) NCBI Gene Expression Omnibus

   ID GSE216318. The CIC-ERF co-deletion underlies fusion independent activation of ETS family member, ETV1, to drive prostate cancer progression.

   https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE216318

1. 1. Bose R

   (2017) NCBI Gene Expression Omnibus

   ID GSE83653. Loss of Function Mutations in ETS2 Repressor Factor (ERF) Reveal a Balance Between Positive and Negative ETS Factors Controlling Prostate Oncogenesis.

   https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE83653

1. 1. Abida W
   2. Cyrta J
   3. Heller G
   4. Prandi D
   5. Armenia J
   6. Coleman I
   7. Cieslik M
   8. Benelli M
   9. Robinson D
   10. Van Allen EM
   11. Sboner A
   12. Fedrizzi T
   13. Mosquera JM
   14. Robinson BD
   15. De Sarkar N
   16. Kunju LP
   17. Tomlins S
   18. Wu YM
   19. Nava Rodrigues D
   20. Loda M
   21. Gopalan A
   22. Reuter VE
   23. Pritchard CC
   24. Mateo J
   25. Bianchini D
   26. Miranda S
   27. Carreira S
   28. Rescigno P
   29. Filipenko J
   30. Vinson J
   31. Montgomery RB
   32. Beltran H
   33. Heath EI
   34. Scher HI
   35. Kantoff PW
   36. Taplin ME
   37. Schultz N
   38. deBono JS
   39. Demichelis F
   40. Nelson PS
   41. Rubin MA
   42. Chinnaiyan AM
   43. Sawyers CL

   (2019) Genomic correlates of clinical outcome in advanced prostate cancer

   PNAS 116:11428–11436.

   https://doi.org/10.1073/pnas.1902651116

   • PubMed
   • Google Scholar
2. 1. Ahmad ST
   2. Rogers AD
   3. Chen MJ
   4. Dixit R
   5. Adnani L
   6. Frankiw LS
   7. Lawn SO
   8. Blough MD
   9. Alshehri M
   10. Wu W
   11. Marra MA
   12. Robbins SM
   13. Cairncross JG
   14. Schuurmans C
   15. Chan JA

   (2019) Capicua regulates neural stem cell proliferation and lineage specification through control of ets factors

   Nature Communications 10:2000.

   https://doi.org/10.1038/s41467-019-09949-6

   • PubMed
   • Google Scholar
3. 1. Ajuria L
   2. Nieva C
   3. Winkler C
   4. Kuo D
   5. Samper N
   6. Andreu MJ
   7. Helman A
   8. González-Crespo S
   9. Paroush Z
   10. Courey AJ
   11. Jiménez G

   (2011) Capicua DNA-binding sites are general response elements for RTK signaling in Drosophila

   Development 138:915–924.

   https://doi.org/10.1242/dev.057729

   • PubMed
   • Google Scholar
4. 1. Baena E
   2. Shao Z
   3. Linn DE
   4. Glass K
   5. Hamblen MJ
   6. Fujiwara Y
   7. Kim J
   8. Nguyen M
   9. Zhang X
   10. Godinho FJ
   11. Bronson RT
   12. Mucci LA
   13. Loda M
   14. Yuan GC
   15. Orkin SH
   16. Li Z

   (2013) ETV1 directs androgen metabolism and confers aggressive prostate cancer in targeted mice and patients

   Genes & Development 27:683–698.

   https://doi.org/10.1101/gad.211011.112

   • PubMed
   • Google Scholar
5. 1. Bettegowda C
   2. Agrawal N
   3. Jiao Y
   4. Sausen M
   5. Wood LD
   6. Hruban RH
   7. Rodriguez FJ
   8. Cahill DP
   9. McLendon R
   10. Riggins G
   11. Velculescu VE
   12. Oba-Shinjo SM
   13. Marie SKN
   14. Vogelstein B
   15. Bigner D
   16. Yan H
   17. Papadopoulos N
   18. Kinzler KW

   (2011) Mutations in CIC and FUBP1 contribute to human oligodendroglioma

   Science 333:1453–1455.

   https://doi.org/10.1126/science.1210557

   • PubMed
   • Google Scholar
6. 1. Bose R
   2. Karthaus WR
   3. Armenia J
   4. Abida W
   5. Iaquinta PJ
   6. Zhang Z
   7. Wongvipat J
   8. Wasmuth EV
   9. Shah N
   10. Sullivan PS
   11. Doran MG
   12. Wang P
   13. Patruno A
   14. Zhao Y
   15. Zheng D
   16. Schultz N
   17. Sawyers CL
   18. International SU2C/PCF Prostate Cancer Dream Team

   (2017) ERF mutations reveal a balance of ETS factors controlling prostate oncogenesis

   Nature 546:671–675.

   https://doi.org/10.1038/nature22820

   • PubMed
   • Google Scholar
7. 1. Bunda S
   2. Heir P
   3. Metcalf J
   4. Li ASC
   5. Agnihotri S
   6. Pusch S
   7. Yasin M
   8. Li M
   9. Burrell K
   10. Mansouri S
   11. Singh O
   12. Wilson M
   13. Alamsahebpour A
   14. Nejad R
   15. Choi B
   16. Kim D
   17. von Deimling A
   18. Zadeh G
   19. Aldape K

   (2019) CIC protein instability contributes to tumorigenesis in glioblastoma

   Nature Communications 10:661.

   https://doi.org/10.1038/s41467-018-08087-9

   • PubMed
   • Google Scholar
8. 1. Cancer Genome Atlas Research Network

   (2015) The molecular taxonomy of primary prostate cancer

   Cell 163:1011–1025.

   https://doi.org/10.1016/j.cell.2015.10.025

   • PubMed
   • Google Scholar
9. 1. Cerami E
   2. Gao J
   3. Dogrusoz U
   4. Gross BE
   5. Sumer SO
   6. Aksoy BA
   7. Jacobsen A
   8. Byrne CJ
   9. Heuer ML
   10. Larsson E
   11. Antipin Y
   12. Reva B
   13. Goldberg AP
   14. Sander C
   15. Schultz N

   (2012) The cbio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data

   Cancer Discovery 2:401–404.

   https://doi.org/10.1158/2159-8290.CD-12-0095

   • PubMed
   • Google Scholar
10. 1. Choi N
    2. Park J
    3. Lee JS
    4. Yoe J
    5. Park GY
    6. Kim E
    7. Jeon H
    8. Cho YM
    9. Roh TY
    10. Lee Y

    (2015) MiR-93/mir-106b/mir-375-CIC-CRABP1: a novel regulatory axis in prostate cancer progression

    Oncotarget 6:23533–23547.

    https://doi.org/10.18632/oncotarget.4372

    • PubMed
    • Google Scholar
11. 1. Clark JP
    2. Cooper CS

    (2009) ETS gene fusions in prostate cancer

    Nature Reviews. Urology 6:429–439.

    https://doi.org/10.1038/nrurol.2009.127

    • PubMed
    • Google Scholar
12. 1. Dissanayake K
    2. Toth R
    3. Blakey J
    4. Olsson O
    5. Campbell DG
    6. Prescott AR
    7. MacKintosh C

    (2011) ERK/p90(rsk)/14-3-3 signalling has an impact on expression of PEA3 ets transcription factors via the transcriptional repressor capicúa

    The Biochemical Journal 433:515–525.

    https://doi.org/10.1042/BJ20101562

    • PubMed
    • Google Scholar
13. 1. Feng FY
    2. Brenner JC
    3. Hussain M
    4. Chinnaiyan AM

    (2014) Molecular pathways: targeting ETS gene fusions in cancer

    Clinical Cancer Research 20:4442–4448.

    https://doi.org/10.1158/1078-0432.CCR-13-0275

    • PubMed
    • Google Scholar
14. 1. Fraser M
    2. Sabelnykova VY
    3. Yamaguchi TN
    4. Heisler LE
    5. Livingstone J
    6. Huang V
    7. Shiah Y-J
    8. Yousif F
    9. Lin X
    10. Masella AP
    11. Fox NS
    12. Xie M
    13. Prokopec SD
    14. Berlin A
    15. Lalonde E
    16. Ahmed M
    17. Trudel D
    18. Luo X
    19. Beck TA
    20. Meng A
    21. Zhang J
    22. D’Costa A
    23. Denroche RE
    24. Kong H
    25. Espiritu SMG
    26. Chua MLK
    27. Wong A
    28. Chong T
    29. Sam M
    30. Johns J
    31. Timms L
    32. Buchner NB
    33. Orain M
    34. Picard V
    35. Hovington H
    36. Murison A
    37. Kron K
    38. Harding NJ
    39. P’ng C
    40. Houlahan KE
    41. Chu KC
    42. Lo B
    43. Nguyen F
    44. Li CH
    45. Sun RX
    46. de Borja R
    47. Cooper CI
    48. Hopkins JF
    49. Govind SK
    50. Fung C
    51. Waggott D
    52. Green J
    53. Haider S
    54. Chan-Seng-Yue MA
    55. Jung E
    56. Wang Z
    57. Bergeron A
    58. Dal Pra A
    59. Lacombe L
    60. Collins CC
    61. Sahinalp C
    62. Lupien M
    63. Fleshner NE
    64. He HH
    65. Fradet Y
    66. Tetu B
    67. van der Kwast T
    68. McPherson JD
    69. Bristow RG
    70. Boutros PC

    (2017) Genomic hallmarks of localized, non-indolent prostate cancer

    Nature 541:359–364.

    https://doi.org/10.1038/nature20788

    • PubMed
    • Google Scholar
15. 1. Futran AS
    2. Kyin S
    3. Shvartsman SY
    4. Link AJ

    (2015) Mapping the binding interface of ERK and transcriptional repressor capicua using photocrosslinking

    PNAS 112:8590–8595.

    https://doi.org/10.1073/pnas.1501373112

    • PubMed
    • Google Scholar
16. 1. Gao J
    2. Aksoy BA
    3. Dogrusoz U
    4. Dresdner G
    5. Gross B
    6. Sumer SO
    7. Sun Y
    8. Jacobsen A
    9. Sinha R
    10. Larsson E
    11. Cerami E
    12. Sander C
    13. Schultz N

    (2013) Integrative analysis of complex cancer genomics and clinical profiles using the cbioportal

    Science Signaling 6:pl1.

    https://doi.org/10.1126/scisignal.2004088

    • PubMed
    • Google Scholar
17. 1. Grasso CS
    2. Robinson DR
    3. Cao X
    4. Dhanasekaran SM
    5. Khan AP
    6. Quist MJ
    7. Jing X
    8. Lonigro RJ
    9. Brenner JC
    10. Asangani IA
    11. Ateeq B
    12. Chun SY
    13. Siddiqui J
    14. Sam L
    15. Anstett M
    16. Mehra R
    17. Prensner JR
    18. Palanisamy N
    19. Ryslik GA
    20. Vandin F
    21. Raphael BJ
    22. Kunju LP
    23. Rhodes DR
    24. Pienta KJ
    25. Chinnaiyan AM
    26. Tomlins SA

    (2012) The mutational landscape of lethal castration-resistant prostate cancer

    Nature 487:239–243.

    https://doi.org/10.1038/nature11125

    • PubMed
    • Google Scholar
18. 1. Helgeson BE
    2. Tomlins SA
    3. Shah N
    4. Laxman B
    5. Cao Q
    6. Prensner JR
    7. Cao X
    8. Singla N
    9. Montie JE
    10. Varambally S
    11. Mehra R
    12. Chinnaiyan AM

    (2008) Characterization of TMPRSS2:ETV5 and SLC45A3:ETV5 gene fusions in prostate cancer

    Cancer Research 68:73–80.

    https://doi.org/10.1158/0008-5472.CAN-07-5352

    • PubMed
    • Google Scholar
19. 1. Hermans KG
    2. van der Korput HA
    3. van Marion R
    4. van de Wijngaart DJ
    5. Ziel-van der Made A
    6. Dits NF
    7. Boormans JL
    8. van der Kwast TH
    9. van Dekken H
    10. Bangma CH
    11. Korsten H
    12. Kraaij R
    13. Jenster G
    14. Trapman J

    (2008) Truncated ETV1, fused to novel tissue-specific genes, and full-length ETV1 in prostate cancer

    Cancer Research 68:7541–7549.

    https://doi.org/10.1158/0008-5472.CAN-07-5930

    • PubMed
    • Google Scholar
20. 1. Hieronymus H
    2. Schultz N
    3. Gopalan A
    4. Carver BS
    5. Chang MT
    6. Xiao Y
    7. Heguy A
    8. Huberman K
    9. Bernstein M
    10. Assel M
    11. Murali R
    12. Vickers A
    13. Scardino PT
    14. Sander C
    15. Reuter V
    16. Taylor BS
    17. Sawyers CL

    (2014) Copy number alteration burden predicts prostate cancer relapse

    PNAS 111:11139–11144.

    https://doi.org/10.1073/pnas.1411446111

    • PubMed
    • Google Scholar
21. 1. Hou C
    2. McCown C
    3. Ivanov DN
    4. Tsodikov OV

    (2020) Structural insight into the DNA binding function of transcription factor ERF

    Biochemistry 47:4499–4506.

    https://doi.org/10.1021/acs.biochem.0c00774

    • PubMed
    • Google Scholar
22. 1. Huang DW
    2. Sherman BT
    3. Lempicki RA

    (2009) Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources

    Nature Protocols 4:44–57.

    https://doi.org/10.1038/nprot.2008.211

    • PubMed
    • Google Scholar
23. 1. Huang FW
    2. Mosquera JM
    3. Garofalo A
    4. Oh C
    5. Baco M
    6. Amin-Mansour A
    7. Rabasha B
    8. Bahl S
    9. Mullane SA
    10. Robinson BD
    11. Aldubayan S
    12. Khani F
    13. Karir B
    14. Kim E
    15. Chimene-Weiss J
    16. Hofree M
    17. Romanel A
    18. Osborne JR
    19. Kim JW
    20. Azabdaftari G
    21. Woloszynska-Read A
    22. Sfanos K
    23. De Marzo AM
    24. Demichelis F
    25. Gabriel S
    26. Van Allen EM
    27. Mesirov J
    28. Tamayo P
    29. Rubin MA
    30. Powell IJ
    31. Garraway LA

    (2017) Exome sequencing of african-american prostate cancer reveals loss-of-function erf mutations

    Cancer Discovery 7:973–983.

    https://doi.org/10.1158/2159-8290.CD-16-0960

    • PubMed
    • Google Scholar
24. 1. Jiménez G
    2. Shvartsman SY
    3. Paroush Z

    (2012) The capicua repressor--a general sensor of RTK signaling in development and disease

    Journal of Cell Science 125:1383–1391.

    https://doi.org/10.1242/jcs.092965

    • PubMed
    • Google Scholar
25. 1. Kawamura-Saito M
    2. Yamazaki Y
    3. Kaneko K
    4. Kawaguchi N
    5. Kanda H
    6. Mukai H
    7. Gotoh T
    8. Motoi T
    9. Fukayama M
    10. Aburatani H
    11. Takizawa T
    12. Nakamura T

    (2006) Fusion between CIC and DUX4 up-regulates PEA3 family genes in ewing-like sarcomas with t(4;19)(q35;q13) translocation

    Human Molecular Genetics 15:2125–2137.

    https://doi.org/10.1093/hmg/ddl136

    • PubMed
    • Google Scholar
26. 1. Keld R
    2. Guo B
    3. Downey P
    4. Cummins R
    5. Gulmann C
    6. Ang YS
    7. Sharrocks AD

    (2011) PEA3/ETV4-related transcription factors coupled with active ERK signalling are associated with poor prognosis in gastric adenocarcinoma

    British Journal of Cancer 105:124–130.

    https://doi.org/10.1038/bjc.2011.187

    • PubMed
    • Google Scholar
27. 1. Kim TD
    2. Jin F
    3. Shin S
    4. Oh S
    5. Lightfoot SA
    6. Grande JP
    7. Johnson AJ
    8. van Deursen JM
    9. Wren JD
    10. Janknecht R

    (2016) Histone demethylase JMJD2A drives prostate tumorigenesis through transcription factor ETV1

    The Journal of Clinical Investigation 126:706–720.

    https://doi.org/10.1172/JCI78132

    • Google Scholar
28. 1. Kim E
    2. Kim D
    3. Lee JS
    4. Yoe J
    5. Park J
    6. Kim CJ
    7. Jeong D
    8. Kim S
    9. Lee Y

    (2018) Capicua suppresses hepatocellular carcinoma progression by controlling the ETV4-MMP1 axis

    Hepatology 67:2287–2301.

    https://doi.org/10.1002/hep.29738

    • PubMed
    • Google Scholar
29. 1. Kim JW
    2. Ponce RK
    3. Okimoto RA

    (2021) Capicua in human cancer

    Trends in Cancer 7:77–86.

    https://doi.org/10.1016/j.trecan.2020.08.010

    • PubMed
    • Google Scholar
30. 1. LeBlanc VG
    2. Firme M
    3. Song J
    4. Chan SY
    5. Lee MH
    6. Yip S
    7. Chittaranjan S
    8. Marra MA

    (2017) Comparative transcriptome analysis of isogenic cell line models and primary cancers links capicua (CIC) loss to activation of the MAPK signalling cascade

    The Journal of Pathology 242:206–220.

    https://doi.org/10.1002/path.4894

    • PubMed
    • Google Scholar
31. 1. Lin YK
    2. Wu W
    3. Ponce RK
    4. Kim JW
    5. Okimoto RA

    (2020) Negative MAPK-ERK regulation sustains CIC-DUX4 oncoprotein expression in undifferentiated sarcoma

    PNAS 117:20776–20784.

    https://doi.org/10.1073/pnas.2009137117

    • PubMed
    • Google Scholar
32. 1. Nelson WG
    2. De Marzo AM
    3. Isaacs WB

    (2003) Prostate cancer

    The New England Journal of Medicine 349:366–381.

    https://doi.org/10.1056/NEJMra021562

    • PubMed
    • Google Scholar
33. 1. Oh S
    2. Shin S
    3. Janknecht R

    (2012) ETV1, 4 and 5: an oncogenic subfamily of ETS transcription factors

    Biochimica et Biophysica Acta 1826:1–12.

    https://doi.org/10.1016/j.bbcan.2012.02.002

    • PubMed
    • Google Scholar
34. 1. Okimoto RA
    2. Breitenbuecher F
    3. Olivas VR
    4. Wu W
    5. Gini B
    6. Hofree M
    7. Asthana S
    8. Hrustanovic G
    9. Flanagan J
    10. Tulpule A
    11. Blakely CM
    12. Haringsma HJ
    13. Simmons AD
    14. Gowen K
    15. Suh J
    16. Miller VA
    17. Ali S
    18. Schuler M
    19. Bivona TG

    (2017) Inactivation of capicua drives cancer metastasis

    Nature Genetics 49:87–96.

    https://doi.org/10.1038/ng.3728

    • PubMed
    • Google Scholar
35. 1. Okimoto RA
    2. Wu W
    3. Nanjo S
    4. Olivas V
    5. Lin YK
    6. Ponce RK
    7. Oyama R
    8. Kondo T
    9. Bivona TG

    (2019) CIC-DUX4 oncoprotein drives sarcoma metastasis and tumorigenesis via distinct regulatory programs

    The Journal of Clinical Investigation 129:3401–3406.

    https://doi.org/10.1172/JCI126366

    • PubMed
    • Google Scholar
36. 1. Pop MS
    2. Stransky N
    3. Garvie CW
    4. Theurillat JP
    5. Hartman EC
    6. Lewis TA
    7. Zhong C
    8. Culyba EK
    9. Lin F
    10. Daniels DS
    11. Pagliarini R
    12. Ronco L
    13. Koehler AN
    14. Garraway LA

    (2014) A small molecule that binds and inhibits the ETV1 transcription factor oncoprotein

    Molecular Cancer Therapeutics 13:1492–1502.

    https://doi.org/10.1158/1535-7163.MCT-13-0689

    • PubMed
    • Google Scholar
37. 1. Reich M
    2. Liefeld T
    3. Gould J
    4. Lerner J
    5. Tamayo P
    6. Mesirov JP

    (2006) GenePattern 2.0

    Nature Genetics 38:500–501.

    https://doi.org/10.1038/ng0506-500

    • PubMed
    • Google Scholar
38. 1. Robinson D
    2. Van Allen EM
    3. Schultz N
    4. Lonigro RJ
    5. Mosquera JM
    6. Montgomery B
    7. Taplin ME
    8. Pritchard CC
    9. Attard G
    10. Beltran H
    11. Abida W
    12. Bradley RK
    13. Vinson J
    14. Cao X
    15. Vats P
    16. Kunju LP
    17. Hussain M
    18. Feng FY
    19. Tomlins SA
    20. Cooney KA
    21. Smith DC
    22. Brennan C
    23. Siddiqui J
    24. Mehra R
    25. Chen Y
    26. Rathkopf DE
    27. Morris MJ
    28. Solomon SB
    29. Durack JC
    30. Reuter VE
    31. Gopalan A
    32. Gao J
    33. Loda M
    34. Lis RT
    35. Bowden M
    36. Balk SP
    37. Gaviola G
    38. Sougnez C
    39. Gupta M
    40. Mostaghel EA
    41. Cheng HH
    42. Mulcahy H
    43. True LD
    44. Plymate SR
    45. Dvinge H
    46. Ferraldeschi R
    47. Flohr P
    48. Miranda S
    49. Zafeiriou Z
    50. Tunariu N
    51. Mateo J
    52. Perez-Lopez R
    53. Demichelis F
    54. Robinson BD
    55. Schiffman M
    56. Nanus DM
    57. Tagawa ST
    58. Sigaras A
    59. Eng KW
    60. Elemento O
    61. Sboner A
    62. Heath EI
    63. Scher HI
    64. Pienta KJ
    65. Kantoff P
    66. de Bono JS
    67. Rubin MA
    68. Nelson PS
    69. Garraway LA
    70. Sawyers CL
    71. Chinnaiyan AM

    (2015) Integrative clinical genomics of advanced prostate cancer

    Cell 161:1215–1228.

    https://doi.org/10.1016/j.cell.2015.05.001

    • PubMed
    • Google Scholar
39. 1. Seim I
    2. Jeffery PL
    3. Thomas PB
    4. Nelson CC
    5. Chopin LK

    (2017) Whole-genome sequence of the metastatic PC3 and lncap human prostate cancer cell lines

    G3: Genes, Genomes, Genetics 7:1731–1741.

    https://doi.org/10.1534/g3.117.039909

    • PubMed
    • Google Scholar
40. 1. Setlur SR
    2. Mertz KD
    3. Hoshida Y
    4. Demichelis F
    5. Lupien M
    6. Perner S
    7. Sboner A
    8. Pawitan Y
    9. Andrén O
    10. Johnson LA
    11. Tang J
    12. Adami HO
    13. Calza S
    14. Chinnaiyan AM
    15. Rhodes D
    16. Tomlins S
    17. Fall K
    18. Mucci LA
    19. Kantoff PW
    20. Stampfer MJ
    21. Andersson SO
    22. Varenhorst E
    23. Johansson JE
    24. Brown M
    25. Golub TR
    26. Rubin MA

    (2008) Estrogen-dependent signaling in a molecularly distinct subclass of aggressive prostate cancer

    Journal of the National Cancer Institute 100:815–825.

    https://doi.org/10.1093/jnci/djn150

    • PubMed
    • Google Scholar
41. 1. Sherman BT
    2. Hao M
    3. Qiu J
    4. Jiao X
    5. Baseler MW
    6. Lane HC
    7. Imamichi T
    8. Chang W

    (2022) DAVID: a web server for functional enrichment analysis and functional annotation of gene lists (2021 update)

    Nucleic Acids Res 50:W216–W221.

    https://doi.org/10.1093/nar/gkac194

    • Google Scholar
42. 1. Simón-Carrasco L
    2. Graña O
    3. Salmón M
    4. Jacob HKC
    5. Gutierrez A
    6. Jiménez G
    7. Drosten M
    8. Barbacid M

    (2017) Inactivation of capicua in adult mice causes T-cell lymphoblastic lymphoma

    Genes & Development 31:1456–1468.

    https://doi.org/10.1101/gad.300244.117

    • PubMed
    • Google Scholar
43. 1. Simón-Carrasco L
    2. Jiménez G
    3. Barbacid M
    4. Drosten M

    (2018) The capicua tumor suppressor: a gatekeeper of ras signaling in development and cancer

    Cell Cycle 17:702–711.

    https://doi.org/10.1080/15384101.2018.1450029

    • PubMed
    • Google Scholar
44. 1. Sizemore GM
    2. Pitarresi JR
    3. Balakrishnan S
    4. Ostrowski MC

    (2017) The ETS family of oncogenic transcription factors in solid tumours

    Nature Reviews. Cancer 17:337–351.

    https://doi.org/10.1038/nrc.2017.20

    • PubMed
    • Google Scholar
45. 1. Tomlins SA
    2. Rhodes DR
    3. Perner S
    4. Dhanasekaran SM
    5. Mehra R
    6. Sun XW
    7. Varambally S
    8. Cao X
    9. Tchinda J
    10. Kuefer R
    11. Lee C
    12. Montie JE
    13. Shah RB
    14. Pienta KJ
    15. Rubin MA
    16. Chinnaiyan AM

    (2005) Recurrent fusion of TMPRSS2 and ETS transcription factor genes in prostate cancer

    Science 310:644–648.

    https://doi.org/10.1126/science.1117679

    • PubMed
    • Google Scholar
46. 1. Tomlins SA
    2. Mehra R
    3. Rhodes DR
    4. Smith LR
    5. Roulston D
    6. Helgeson BE
    7. Cao X
    8. Wei JT
    9. Rubin MA
    10. Shah RB
    11. Chinnaiyan AM

    (2006) TMPRSS2:etv4 gene fusions define a third molecular subtype of prostate cancer

    Cancer Research 66:3396–3400.

    https://doi.org/10.1158/0008-5472.CAN-06-0168

    • PubMed
    • Google Scholar
47. 1. Wong D
    2. Lounsbury K
    3. Lum A
    4. Song J
    5. Chan S
    6. LeBlanc V
    7. Chittaranjan S
    8. Marra M
    9. Yip S

    (2019) Transcriptomic analysis of CIC and ATXN1L reveal a functional relationship exploited by cancer

    Oncogene 38:273–290.

    https://doi.org/10.1038/s41388-018-0427-5

    • PubMed
    • Google Scholar
48. 1. Yang R
    2. Chen LH
    3. Hansen LJ
    4. Carpenter AB
    5. Moure CJ
    6. Liu H
    7. Pirozzi CJ
    8. Diplas BH
    9. Waitkus MS
    10. Greer PK
    11. Zhu H
    12. McLendon RE
    13. Bigner DD
    14. He Y
    15. Yan H

    (2017) cic loss promotes gliomagenesis via aberrant neural stem cell proliferation and differentiation

    Cancer Research 77:6097–6108.

    https://doi.org/10.1158/0008-5472.CAN-17-1018

    • PubMed
    • Google Scholar
49. 1. Yoshiya S
    2. Itoh S
    3. Yoshizumi T
    4. Yugawa K
    5. Kurihara T
    6. Toshima T
    7. Harada N
    8. Hashisako M
    9. Yonemasu H
    10. Fukuzawa K
    11. Oda Y
    12. Mori M

    (2021) Impact of capicua on pancreatic cancer progression

    Annals of Surgical Oncology 28:3198–3207.

    https://doi.org/10.1245/s10434-020-09339-z

    • PubMed
    • Google Scholar

Decision letter after peer review:

Thank you for submitting your article "The CIC-ERF co-deletion underlies fusion independent activation of ETS family member, ETV1, to drive prostate cancer progression" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Erica Golemis as the Senior Editor. The following individual involved in the review of your submission has agreed to reveal their identity: Stephen Yip (Reviewer #3).

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Essential revisions:

1. Can the authors confirm that the 15 cBioPortal datasets are really independent? To clarify this, it would be informative to have the numbers presented for each dataset in the associated Figure 1c legend. For example, the reader should then be able to see that the TCGA 2018 and TCGA numbers are quite small as most of these cases are also in TCGA 2015. Figure 1c is also hard to interpret given the data has been sorted (we assume) by CIC and ERF alterations. It would be more informative for the reader if it was sorted according to the study so we can see the distribution of alterations across each study. This then ties in with the later analysis looking at the difference in alterations between primary and metastatic prostate cancer datasets. It is also hard to interpret the stage data presented in Figure 1d, is it possible to condense the stages somehow? For example, condense T2A, T2B and T2C into one category (T2), etc.

2. Can the numbers of cases present in the primary and metastatic prostate cancer groups be specified in the text or in Figure 1e. Similarly, Can the number of cases present in each group of the survival analyses be specified in the text or in Figure 1g? Without any of the numbers requested above, it is difficult to determine the statistical power these analyses have.

3. The sample size used in each of the cell line assays should also be presented in the corresponding figure legends as (n=x). This includes bioreps of the cell culture experiments.

4. Supplementary Table 1 is missing.

5. The description of the ssGSEA work is very unclear both in the methods and Results sections. GSEA also stands for Gene Set Enrichment Analysis, not Gene Expression Analysis as stated in the results. It is unclear whether the authors used gene expression data from PNT2 cells or if they used data from the TCGA-PRAD dataset. If they used PNT2 gene expression data, how was this generated and how exactly was the TCGA-PRAD dataset involved? A more comprehensive and coherent description of these analyses is required in both the methods and Results sections. This includes the figure legend (4m) where there is no explanation for the abbreviations used.

6. The figures associated with the results describing the assessment of the ETV1 inhibitor in DU-145 cells actually show LNCap cells (Figure 4A-c). Please clarify.

7. The authors say they observe enhanced sensitivity to BRD32048 in DU-145 CIC KO cells, but this is not significant and should be noted. Why wasn't this work repeated in the other cell lines (PNT2 and PC-3, especially, given their contexts in prostate cancer metastatic potential) or in vivo? Or why not reconstitute ERF in DU-145 to see the effect of the inhibitor on CIC KO/KD alone? These experiments are essential to strengthen and generalize conclusions.

8. The supplementary figure (4d-f) also shows results from PC-3 cells, but this is not mentioned in the text.

9. The sub-figures (5b-5d) are in the wrong order, 5b shows invasiveness while 5c shows viability.

10. The methods section provides no information as to how the authors conducted their cell culture experiments using the ETV1 inhibitor (including technical/bioreps etc.), and as such, are not reproducible. Overall, the methods section is very disjointed and the methodology should be presented in the order that it is described in the results (aside from the cell line/culture information). This would also allow the authors to get rid of some repetition that is currently present and ensure there is still no missing methodology. In addition to some of the more specific methodology suggestions mentioned previously, can more information regarding the DAVID functional clustering analysis be added, including a reference/web link if available?

11. The authors should perform additional replicates to provide statistical significance for experiments currently lacking significance or temper their claims. For example, Figure 4l should probably be removed, unless additional replicates improve the data.

12. If ERF and CIC co-deletion function by increasing ETV1 transcription, then an increase in ETV1 mRNA levels in these tumors should be apparent in patient tumor datasets, such as those analyzed in Figure 1. This is a key missing analysis. This needs to be supplied.

13. A growth curve to measure proliferation rate changes after CIC and ERF manipulation would provide more rigor. Does ERF overexpression inhibit phenotypes in PNT2 cells?

14. Testing a few negative control genomic regions and showing no enrichment would alleviate the concern regarding controls for ChIP-PCR.

15. This phrase in the introduction needs references: "…and inactivation of CIC de-represses ETV1, ETV4, and ETV5 transcription to promote tumor progression."

16. It is important to investigate potential proteomic interactions between CIC and ERF using Western blot and immunoprecipitation. Whether they bind to one another and what happens when only CIC or ERF is not expressed. What happens to the level of CIC, or ERF, in the absence of the other protein and the consequence of this, if there is altered expression.

[Editors’ note: further revisions were suggested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "The CIC-ERF co-deletion underlies fusion independent activation of ETS family member, ETV1, to drive prostate cancer progression" for further consideration by eLife. Your revised article has been evaluated by Erica Golemis (Senior Editor) and a Reviewing Editor.

The manuscript has been improved but there are some remaining issues that need to be addressed, as outlined below:

In general, no further experiments are needed, and the remaining critiques can be addressed by writing changes. Reviewer 3 is satisfied with the current version of the article, and Reviewer 1 has points that are largely directed at clarity; please address these. However, Reviewer 2 has residual issues with the significance and interpretation of the data. Please address point 1 in the discussion; for point 2, we suggest you present alternative analyses of the data (perhaps in a new supplemental figure) that allow the comparisons suggested, and discuss the significance of the results obtained through this analysis.

Reviewer #1 (Recommendations for the authors):

The authors have addressed the majority of my concerns, however, I don't feel I can comment on whether they have adequately addressed the other reviewers' comments. I have just a few points below that require the authors' attention.

1. As suggested, the authors have included a Table (Supplementary Table 1) that presents the CIC-ERF data for the various studies. However, I'm not sure the authors understood our comment regarding study overlap. For example, they present two rows of data for TCGA samples (PanCancer Atlas and Firehose Legacy) but these studies essentially consist of the same samples (the second just has a few more samples included). Thus, it is somewhat misleading to present two primary predominant cohorts with lower CIC-ERF alterations, when this is really just a single cohort. The authors also need to determine how many samples overlap the two SU2C/PCF studies and the MSK studies. This also applies to the data presented in Supplementary Figure 1 and impacts the statement in the Results section that says "We analyzed over 6047 PCa tumors from 5839 patients" (line 99), as this is not the case given numbers overlap across the studies.

Figure 1C still appears to be sorted by something other than study. If it was sorted by study, the top bar in this figure would have, for example, all the MSK-IMPACT prostate samples grouped together, all the PRAD (MSKCC, 2020) samples grouped together, etc. So there would be a solid teal block, a solid pink block, etc. At the moment you can see that on the far left of the "Study" bar, it starts with a mix of green and purple study samples, and so on.

2. The first and second sentences of the abstract now essentially say the same thing. The second sentence can be removed.

3. Ensure all gene names are italicised.

4. There is an issue with one of the references (TCGA 2015 publication).

Reviewer #2 (Recommendations for the authors):

Overall the revision has not alleviated my concerns regarding the viability of the overall thesis that ERG and CIC co-deletion leading to activation of ETV1 is a driver of prostate cancer progression. There are two main reasons for this. First is the finding that there is no statistically significant difference in ETV1 expression between patient tumors with wild-type ERF and CIC, and co-deleted ERF and CIC. This seems to be a major problem. Second much of the critical data is underwhelming, does not reach statistical significance, and seems to be presented in ways intended to obscure the real results. Examples of this include the setting of every odd column to 1 and normalization of each even column to the prior odd column in Figures 5A, 5B, 5C, and 5S2B, C, E, and F – with no explanation that this is being done. This is also true of the ChIP control presented in the rebuttal, which includes an irrelevant input column that obscures the necessary comparison of IgG to anti-CIC and anti-ERF. It is also a problem that this control is presented in an entirely different way than the ChIP in Figure 4 that it is necessary to control for, and the control does not appear to be included in the revised manuscript.

Reviewer #3 (Recommendations for the authors):

I am satisfied with the authors' responses and the additional experiments performed. I am somewhat disappointed that the selective knockdown of CIC and ERF did not further elucidate the functional interactions between these two proteins but I appreciate the effort and the underlying complexity of the interactions which could also be affected by cell and developmental contexts.

Essential revisions:

1. Can the authors confirm that the 15 cBioPortal datasets are really independent? To clarify this, it would be informative to have the numbers presented for each dataset in the associated Figure 1c legend. For example, the reader should then be able to see that the TCGA 2018 and TCGA numbers are quite small as most of these cases are also in TCGA 2015. Figure 1c is also hard to interpret given the data has been sorted (we assume) by CIC and ERF alterations. It would be more informative for the reader if it was sorted according to the study so we can see the distribution of alterations across each study. This then ties in with the later analysis looking at the difference in alterations between primary and metastatic prostate cancer datasets. It is also hard to interpret the stage data presented in Figure 1d, is it possible to condense the stages somehow? For example, condense T2A, T2B and T2C into one category (T2), etc.

We agree with the reviewer and provide a detailed table with (1) the total number of cases in each study; (2) the type of CIC and ERF deletion (shallow or deep); and (3) the percentage of cases in each study with a CIC-ERF co-deletion. This analysis further suggests that the CIC-ERF co-deletion is enriched in studies that were predominantly composed of advanced prostate cancers (highlighted in yellow) compared to studies predominantly composed of primary prostate cancer (highlighted in green). We have included this table as Figure 1 – Table Supplement 1.

In order to illustrate the frequency of CIC-ERF co-deletions in each study we sorted according to the study as recommended by the reviewer. These data indicate that CIC-ERF co-deletions are observed at a higher frequency in studies that predominantly analyzed advanced prostate cancer relative to primary prostate cancers. We have incorporated this figure into our revised manuscript as Figure 1 – Supplementary Figure 1.

With respect to the staging data, we utilized the presorted selection in cBioPortal and were unfortunately, unable to further condense the sub-stages into a single staging category (ie. T2A, T2B, T2C into T2).

2. Can the numbers of cases present in the primary and metastatic prostate cancer groups be specified in the text or in Figure 1e.

We agree with the reviewer and have now incorporated detailed information on the number of cases in each of the primary and metastatic prostate cancer groups. We provide this information below and have incorporated these numbers into the text.

Primary prostate cancer cohorts:

PNAS 2014. Hieronymus et al.

272 primary prostate cancers

Cell 2015. TCGA

333 primary prostate cancers

Nature 2017. Fraser et al.

477 primary (localized, non-indolent) prostate cancer.

Metastatic prostate cancer cohorts:

Nature 2012. Grasso et al.

50 metastatic CRPCs and 11 treatment-naive, high-grade localized prostate cancers.

Cell 2015. Robinson et al.

150 mCRPC

PNAS 2019. Abida et al.

429 mCRPC

Similarly, Can the number of cases present in each group of the survival analyses be specified in the text or in Figure 1g? Without any of the numbers requested above, it is difficult to determine the statistical power these analyses have.

Again, we agree that adding these numbers are informative and have incorporated them into the text.

Disease-free survival (DFS):

Altered: 90 (# of total) 25 (# of events)

Non-altered 910 (# of total) 153 (# of events)

Progression-free survival (PFS):

Altered: 52 (# of total) 16 (# of events)

Non-altered: 442 (# of total) 77(# of events)

3. The sample size used in each of the cell line assays should also be presented in the corresponding figure legends as (n=x). This includes bioreps of the cell culture experiments.

We have now incorporated the sample size for each experiment in the corresponding figure legends.

4. Supplementary Table 1 is missing.

We apologize for this oversight and have now incorporated this table as Supplemental Table 2 in the revised version of our manuscript.

5. The description of the ssGSEA work is very unclear both in the methods and Results sections. GSEA also stands for Gene Set Enrichment Analysis, not Gene Expression Analysis as stated in the results. It is unclear whether the authors used gene expression data from PNT2 cells or if they used data from the TCGA-PRAD dataset. If they used PNT2 gene expression data, how was this generated and how exactly was the TCGA-PRAD dataset involved? A more comprehensive and coherent description of these analyses is required in both the methods and Results sections. This includes the figure legend (4m) where there is no explanation for the abbreviations used.

We agree with the Reviewer and have expanded the methods and Results sections to include additional details on how we performed the ssGSEA. We also apologize for the oversight, we have changed Gene Expression Analysis to Gene Set Enrichment Analysis.

6. The figures associated with the results describing the assessment of the ETV1 inhibitor in DU-145 cells actually show LNCap cells (Figure 4A-c). Please clarify.

We have added the explanation describing LNCaP cells for supplementary figure (4a-c)

7. The authors say they observe enhanced sensitivity to BRD32048 in DU-145 CIC KO cells, but this is not significant and should be noted. Why wasn't this work repeated in the other cell lines (PNT2 and PC-3, especially, given their contexts in prostate cancer metastatic potential) or in vivo? Or why not reconstitute ERF in DU-145 to see the effect of the inhibitor on CIC KO/KD alone? These experiments are essential to strengthen and generalize conclusions.

We thank the reviewer for this comment. We have now performed additional studies to address this concern. In particular, we genetically silenced CIC and ERF in PNT2 cells and observed enhanced sensitivity to the ETV1 inhibitor. Consistent with these results we also knocked down ERF in CIC deficient PC-3 cells and again noted an increase in ETV1 inhibitor sensitivity compared to scramble control. These data are incorporated into the revised manuscript as Figure 5 —figure supplement 1A-F.

Finally, we reconstituted ERF into ERF deficient DU145 cells with CIC KO. We did not observe a significant difference in ETV1 inhibitor sensitivity in this setting.

8. The supplementary figure (4d-f) also shows results from PC-3 cells, but this is not mentioned in the text.

We apologize for this oversight and have now mentioned PC-3 cells in the text.

9. The sub-figures (5b-5d) are in the wrong order, 5b shows invasiveness while 5c shows viability.

We apologize for this error. We have corrected the order of the sub-figure.

10. The methods section provides no information as to how the authors conducted their cell culture experiments using the ETV1 inhibitor (including technical/bioreps etc.), and as such, are not reproducible. Overall, the methods section is very disjointed and the methodology should be presented in the order that it is described in the results (aside from the cell line/culture information). This would also allow the authors to get rid of some repetition that is currently present and ensure there is still no missing methodology. In addition to some of the more specific methodology suggestions mentioned previously, can more information regarding the DAVID functional clustering analysis be added, including a reference/web link if available?

We have added a paragraph in the methods section describing the methodology using the ETV1 inhibitor. Overall, we have also restructured the methods section and the methodology is presented in the order as they appear in the Results section. We have also included a web link and references to DAVID in the manuscript text and methods sections.

11. The authors should perform additional replicates to provide statistical significance for experiments currently lacking significance or temper their claims. For example, Figure 4l should probably be removed, unless additional replicates improve the data.

We completely agree with the reviewer. We have removed figure 4l from the manuscript.

12. If ERF and CIC co-deletion function by increasing ETV1 transcription, then an increase in ETV1 mRNA levels in these tumors should be apparent in patient tumor datasets, such as those analyzed in Figure 1. This is a key missing analysis. This needs to be supplied.

We used the largest prostate cancer dataset (n=455) and observed a trend but not a statistically significant difference in ETV1 expression in tumors that harbored the CIC-ERF co-deletion compared to CIC-ERF replete tumors.

Author response image 1

Download asset Open asset

13. A growth curve to measure proliferation rate changes after CIC and ERF manipulation would provide more rigor. Does ERF overexpression inhibit phenotypes in PNT2 cells?

We assessed the growth rate (over 7 days) of PNT2 cells engineered to express shERF, sgCIC or combinatorial silencing of ERF (shERF) and CIC (sgCIC). We observed an increase in growth rate in PNT2 cells with dual ERF and CIC suppression compared to PNT2 control and either shERF or sgCIC alone.

Author response image 2

Download asset Open asset

We overexpressed ERF in PNT2 cells but did not observe a significant difference in viability (crystal violet or cell-titer glo assays) or in colony formation.

Author response image 3

Download asset Open asset

14. Testing a few negative control genomic regions and showing no enrichment would alleviate the concern regarding controls for ChIP-PCR.

We used two commercially validated human negative control q-pcr primer sets from Active Motif as an additional layer of control to address this concern. Using these primers targeting Gene Deserts on chromosome 12 and 4, we made comparisons to either IgG control or Input (2%).

Author response image 4

Download asset Open asset

15. This phrase in the introduction needs references: "…and inactivation of CIC de-represses ETV1, ETV4, and ETV5 transcription to promote tumor progression."

We have added the references for the indicated phrase.

16. It is important to investigate potential proteomic interactions between CIC and ERF using Western blot and immunoprecipitation. Whether they bind to one another and what happens when only CIC or ERF is not expressed. What happens to the level of CIC, or ERF, in the absence of the other protein and the consequence of this, if there is altered expression.

We agree with the reviewer and have performed co-immunoprecipitation experiments to address these concerns. Specifically, using HEK293T cells we expressed and pulled down GFP-tagged ERF, probing for CIC by western blot. CIC did not co-immunoprecipitate with GFP-tagged ERF under these conditions. The reciprocal experiment was also performed through expression of Myc-tagged CIC in HEK293T cells. ERF did not co-immunoprecipitate with CIC under these conditions. These data indicate that CIC and ERF do not interact in HEK293T cells. These studies have been incorporated to the revised manuscript as Figure 4 —figure supplement 1K-L.

We next performed ERF knockdown (KD) and ERF overexpression (OE) studies in HEK293T cells as requested by the Reviewer. Interestingly, we observed an increase in CIC expression following either ERF KD or ERF OE.

Author response image 5

Download asset Open asset

Moreover, we assessed ERF expression following CIC knockout or CIC overexpression in 293T cells. CIC expression increased using a single crispr-based sgRNA (sgCIC1) but this finding was not consistently observed with sgCIC2. CIC overexpression also increased ERF expression compared to control. Collectively, these findings indicate that the functional relationship between CIC and ERF may be complex and warrants further investigation into how these two proteins regulate one another. This will be an area of future interest and investigation.

Author response image 6

Download asset Open asset

[Editors’ note: further revisions were suggested prior to acceptance, as described below.]

Reviewer #1 (Recommendations for the authors):

The authors have addressed the majority of my concerns, however, I don't feel I can comment on whether they have adequately addressed the other reviewers' comments. I have just a few points below that require the authors' attention.

1. As suggested, the authors have included a Table (Supplementary Table 1) that presents the CIC-ERF data for the various studies. However, I'm not sure the authors understood our comment regarding study overlap. For example, they present two rows of data for TCGA samples (PanCancer Atlas and Firehose Legacy) but these studies essentially consist of the same samples (the second just has a few more samples included). Thus, it is somewhat misleading to present two primary predominant cohorts with lower CIC-ERF alterations, when this is really just a single cohort. The authors also need to determine how many samples overlap the two SU2C/PCF studies and the MSK studies. This also applies to the data presented in Supplementary Figure 1 and impacts the statement in the Results section that says "We analyzed over 6047 PCa tumors from 5839 patients" (line 99), as this is not the case given numbers overlap across the studies.

We agree with the reviewer. We have removed the statement “we analyzed over 6047 PCa tumors from 5839 patients”.

We also incorporated the following statement within the text to hopefully bring additional clarity. “One limitation of this analysis is that individual studies included overlapping tumors from the same patient (ie. TCGA PanCancer and Firehose Legacy cohorts).” Moreover, we emphasized in the text that we used this correlative analysis to generate a hypothesis to explore the functional interaction between CIC and ERF in the context of prostate cancer progression.

Figure 1C still appears to be sorted by something other than study. If it was sorted by study, the top bar in this figure would have, for example, all the MSK-IMPACT prostate samples grouped together, all the PRAD (MSKCC, 2020) samples grouped together, etc. So there would be a solid teal block, a solid pink block, etc. At the moment you can see that on the far left of the "Study" bar, it starts with a mix of green and purple study samples, and so on.

Figure 1C was sorted by gene alteration. For this analysis, we sorted by CIC and aligned this with ERF alterations. In response to the Reviewer, we previously incorporated the frequency of CIC and ERF co-deletions found in each cohort/study in Figure 1 – Supplementary Figure 1.

2. The first and second sentences of the abstract now essentially say the same thing. The second sentence can be removed.

We agree with the Reviewer and have removed the first sentence of the Abstract.

3. Ensure all gene names are italicised.

We have reviewed the manuscript and italicized all gene names.

4. There is an issue with one of the references (TCGA 2015 publication).

We thank the reviewer for identifying this issue. We have corrected the reference.

Reviewer #2 (Recommendations for the authors):

Overall the revision has not alleviated my concerns regarding the viability of the overall thesis that ERG and CIC co-deletion leading to activation of ETV1 is a driver of prostate cancer progression. There are two main reasons for this. First is the finding that there is no statistically significant difference in ETV1 expression between patient tumors with wild-type ERF and CIC, and co-deleted ERF and CIC. This seems to be a major problem.

We have expanded the Discussion section to include a paragraph on these findings.

“A limitation of this targeted approach in prostate cancer patients is that we did not find a statistically significant difference in ETV1 transcript levels in CIC-ERF co-deleted tumors compared to CIC-ERF replete tumors in the TCGA-PRAD (n=455) dataset. One potential explanation is that ETV1 is commonly upregulated in human prostate cancers through mechanisms beyond CIC-ERF loss. Thus, while our data support the upregulation and potential induced dependence on ETV1 in our CIC-ERF deficient systems, it remains unclear if this will translate beyond our cell-line based models into prostate cancer patients that harbor CIC-ERF co-deletions. Thus, larger studies that aim to evaluate: (1) ETV1 mRNA and protein expression levels; (2) the biological significance of ETV1 function; and (3) the clinical application of ETV1 inhibitors in patients with endogenous CIC-ERF co-deleted tumors (compared to CIC-ERF WT tumors) is warranted in this subset of PCa. Collectively, we have uncovered a molecular subset of PCa defined by a co-deletion of CIC and ERF and further demonstrate a mechanismbased strategy to potentially limit tumor progression through ETV1 inhibition in this subset of human PCa.”

Following ERG, ETV1 is the most frequently overexpressed ETS gene in prostate cancers. Prior studies (PMIDs: 16254181, 18594527, 21298110) have found both fusion dependent (ETV1-translocations) and fusion independent (increased WT ETV1 transcript levels) mechanisms that can lead to increased ETV1 mRNA expression in prostate cancers. Thus, increased ETV1 expression can be observed in other subsets of prostate cancer beyond CIC-ERF co-deleted tumors. We suspect that our analysis comparing CIC-ERF co-deletion to CIC-ERF WT tumors did not reach statistical significance due to the inability to identify other prostate tumors that have increased ETV1 expression from alternative mechanisms, beyond CIC-ERF co-deletions.

Second much of the critical data is underwhelming, does not reach statistical significance, and seems to be presented in ways intended to obscure the real results. Examples of this include the setting of every odd column to 1 and normalization of each even column to the prior odd column in Figures 5A, 5B, 5C, and 5S2B, C, E, and F – with no explanation that this is being done.

We have now modified the Figure title and clearly delineated the comparisons in the figure 5 legend. It now reads:

Figure 5. Combinatorial CIC and ERF loss can modulate ETV1 inhibitor sensitivity in prostate cells. (A) DU-145 cells were transfected with either siScramble (siSCM) or siCIC. After 48 hours, BRD32048 (ETV1 inhibitor) was added to both the transfected groups at the defined concentrations (0 μM, 25 μM, 50 μM). After 24 hours of BRD32048 treatment, cells were replated for crystal violet assay (0.4%) and images were taken and analyzed after 5 days (n=3). siCIC was compared to siSCRM conditions in each respective drug concentration. (B) Crystal violet viability assay (n=3). siETV1 groups were compared to siSCRM control groups +/- CIC expression (sgCtrl, sgCIC1, or sgCIC2). (C) DU-145 cells were transfected with either siScramble (siSCM) or siCIC. After 48 hours, BRD32048 was added to the transfected groups at defined concentrations (0 μM or 50 μM). Transwell invasion assays (n=3) were performed 24 hours after the addition of BRD32048. siCIC was compared to siSCRM in the 0 μM and 50 μM concentration groups. (D) Transwell invasion assays (n=3) comparing siETV1 to siSCRM control +/- CIC expression (sgCtrl, sgCIC1 or sgCIC2). P value = *p<0.05, **p<0.01 for all figures. Error bars represent SD.

Moreover, we now provide additional figures (Figure 5 —figure supplement 2D,E,I,J) to allow for comparisons between the PNT2 and PC-3 crystal violet assays in Figure 5 —figure supplement 2B, C, G, and H. We continue to observe a decrease in invasiveness and a milder, albeit consistent decrease in viability upon ETV1 inhibition in our engineered CIC-ERF deficient cells. We have further modified the text to highlight the shortcomings of our in vitro models and reserve claims on the potential impact on clinical translation until further studies are performed as we mention in the Discussion section.

This is also true of the ChIP control presented in the rebuttal, which includes an irrelevant input column that obscures the necessary comparison of IgG to anti-CIC and anti-ERF. It is also a problem that this control is presented in an entirely different way than the ChIP in Figure 4 that it is necessary to control for, and the control does not appear to be included in the revised manuscript.

We have included the ChIP control experiments into the revised manuscript as Figure 4

—figure supplement 1O-P. In addition, we quantified the ETV1 ChIP-PCR performed in Figure 4 using conventional densitometry (Image J), making comparison between CIC or ERF to IgG. This has been included in the revised manuscript as Figure 4 —figure supplement 1M-N.

Author details

1. Nehal Gupta

   Department of Medicine, University of California, San Francisco, United States

   Contribution

   Conceptualization, Data curation, Formal analysis, Investigation, Writing – original draft, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-8931-5759

2. Hanbing Song

   Department of Medicine, University of California, San Francisco, United States

   Contribution

   Data curation, Formal analysis, Writing – review and editing

   Competing interests

   No competing interests declared

3. Wei Wu

   Department of Medicine, University of California, San Francisco, United States

   Contribution

   Formal analysis, Writing – review and editing

   Competing interests

   No competing interests declared

4. Rovingaile K Ponce

   Department of Medicine, University of California, San Francisco, United States

   Contribution

   Data curation, Writing – review and editing

   Competing interests

   No competing interests declared

5. Yone K Lin

   Department of Medicine, University of California, San Francisco, United States

   Contribution

   Formal analysis, Writing – review and editing

   Competing interests

   No competing interests declared

6. Ji Won Kim

   Department of Medicine, University of California, San Francisco, United States

   Contribution

   Data curation, Writing – review and editing

   Competing interests

   No competing interests declared

7. Eric J Small

   1. Department of Medicine, University of California, San Francisco, United States
   2. Helen Diller Family Comprehensive Cancer Center, University of California, San Francisco, United States

   Contribution

   Supervision, Writing – original draft, Writing – review and editing

   Competing interests

   No competing interests declared

8. Felix Y Feng

   1. Department of Medicine, University of California, San Francisco, United States
   2. Helen Diller Family Comprehensive Cancer Center, University of California, San Francisco, United States
   3. Department of Radiation Oncology, University of California, San Francisco, United States

   Contribution

   Supervision, Funding acquisition, Writing – review and editing

   Competing interests

   No competing interests declared

9. Franklin W Huang

   1. Department of Medicine, University of California, San Francisco, United States
   2. Helen Diller Family Comprehensive Cancer Center, University of California, San Francisco, United States

   Contribution

   Data curation, Supervision, Writing – original draft, Writing – review and editing

   For correspondence

   franklin.huang@ucsf.edu

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-5447-0436

10. Ross A Okimoto

    1. Department of Medicine, University of California, San Francisco, United States
    2. Helen Diller Family Comprehensive Cancer Center, University of California, San Francisco, United States

    Contribution

    Conceptualization, Formal analysis, Supervision, Funding acquisition, Writing – original draft, Writing – review and editing

    For correspondence

    ross.okimoto@ucsf.edu

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0002-4467-8476

• 325

  Page views

• 38

  Downloads

• 0

  Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.