Sage Journals: Discover world-class research

Abstract

Background:

Transmembrane E3 ubiquitin ligase (RNF43) mutations are present in approximately 6–18% of colorectal cancers (CRC) and could enhance Wnt/β-catenin signaling, which is emerging as a promising therapeutic target. This study aims to investigate the clinical and molecular characteristics and potential heterogeneity of RNF43-mutant CRC.

Methods:

A total of 78 patients with RNF43-mutant CRC were enrolled from July 2013 to November 2022. Demographic data, clinical characteristics, treatment regimens used, and survival outcomes were collected and analyzed.

Results:

Our study uncovered that patients with RNF43 mutations in the N-terminal domain (NTD; n = 50) exhibited shorter overall survival (OS; median months, 50.80 versus not reached; p = 0.043) compared to those in the C-terminal domain (CTD; n = 17). Most RNF43 mutations in NTD had positive primary lymph node status, low tumor mutation burden (TMB-L), and correlated with proficient mismatch repair (pMMR)/microsatellite stable (MSS) status. By contrast, RNF43 mutations in CTD were significantly enriched in deficient MMR (dMMR)/microsatellite instability (MSI-H) tumors with high TMB (TMB-H). N-terminal RNF43-mutated tumors harbored a hotspot variant (RNF43 R117fs), which independently predicted a significantly worse outcome in pMMR/MSS CRC with a median OS of 18.9 months. Patients with RNF43 mutations and the BRAF V600E alterations demonstrated sensitivity to BRAF/EGFR inhibitors. Moreover, we observed that pMMR/MSS patients with RNF43 R117fs mutation had a higher incidence of stage IV, ⩾2 metastatic sites, low TMB, and none of them received PD-1/PD-L1 inhibitor therapy.

Conclusion:

Our findings provide the first evidence that RNF43 mutations in NTD and the R117fs variant correlate with a poorer prognosis in CRC patients, providing strategies for Wnt-targeted therapy to improve clinical efficacy.

Keywords

colorectal cancer predictive biomarker prognosis

Introduction

Colorectal cancer (CRC) is the third most commonly diagnosed cancer worldwide and the second leading cause of cancer-related mortality.¹ Previous studies suggest that most CRCs arise either via the adenoma-carcinoma sequence or the serrated neoplasia pathway.^2,3 Wnt/β-catenin signaling, the conventional pathway initiated by alterations in Wnt ligand-dependent genes (RNF43/ZNRF3/RSPO) or ligand-independent genes (APC), drives colorectal carcinogenesis.^4–6 RNF43, a transmembrane E3 ubiquitin ligase, acts as a feedback suppressor of the Wnt pathway by promoting the degradation of Frizzled (FZD) receptors⁷ via ubiquitination.

RNF43 is composed of two specific functional domains, the N-terminal domain (NTD) and the C-terminal domain (CTD). The NTD is mainly responsible for the inhibition of Wnt/β-catenin signaling and consists of an extracellular protease-associated (PA) domain that interacts with R-spondins (RSPO) or FZD, a single-pass transmembrane (TM) domain, and an intracellular RING finger (RING) domain that has ubiquitination functions.^7–10 The CTD of RNF43 contains the Disheveled-2 (DVL2) binding region (DIR), which is required for RNF43-mediated ubiquitination by binding with FZD receptors.^7,9–13 Nevertheless, the impact of DIR on oncogenesis remains highly controversial. There are two main hotspot subtypes of RNF43 mutations, namely G659fs and R117fs. In malignancies, mutations in the NTD of the RNF43 protein usually result in loss of function due to truncating events such as frameshift indels and nonsense mutations. One such variant is R117fs, which compromises the negative feedback regulation of the Wnt/β-catenin signaling pathway, rendering cancer cells susceptible to inhibition by Wnt ligands.^7,12,14

Previously, RNF43 mutations have been identified in several cancer types, including CRC, ovarian cancer, pancreatic cancer, and gastric cancer.^15–17 Approximately 6–18% of CRC patients harbor RNF43 mutations,¹⁷ which mainly served as a late event in the progression from serrated adenoma to malignancy.^3,18 Studies investigating the impact of RNF43 mutations in CRC have yielded conflicting results. Some studies have reported that RNF43 mutations are associated with poor outcomes and higher recurrence rates regardless of the MSI status, while others believe that RNF43 mutant tumors are associated with prolonged survival^19–22; therefore, the prognostic value of RNF43 remains to be determined. Interestingly, tumors with RNF43 mutations frequently exhibit a high frequency of BRAF V600E mutation, and the co-occurrence of these mutations is associated with worse survival outcomes. This subgroup of patients might benefit from anti-BRAF/EGFR therapy.^23–26 Nevertheless, it is worth noting that a phase Ib/II study of WNT974 + encorafenib + cetuximab in patients with BRAF V600E mutant metastatic CRC was recently discontinued due to bone-related toxicities.²⁷ Mutations in the NTD of RNF43 significantly enhance Wnt/β-catenin signaling and can be inhibited by Porcupine inhibitors. By contrast, mutations in the CTD, such as the G659fs variant, are commonly found in microsatellite instability (MSI) tumors and may respond to anti-PD-1/PD-L1 therapy.²⁸

Taken together, the above findings highlight the potential heterogeneity of RNF43 mutant tumors. Given that the prognostic implications of distinct genetic subgroups within RNF43 mutated CRC have not been elucidated, this study primarily aims to investigate the clinical, molecular, and potential prognostic characteristics of RNF43-positive CRC patients. Collectively, our findings will provide crucial insights into the role of RNF43 in Wnt/β-catenin signaling activation, allowing for the stratification of patients with distinct prognoses and the identification of potential therapeutic strategies.

Materials and methods

Patient and sample collection

We evaluated 78 patients with RNF43-mutant CRC between July 2013 and November 2022 at Sun Yat-sen University Cancer Center (SYSUCC, Guangzhou, China). Molecular alterations were detected in tumor tissue using Sanger sequencing or in tumor or plasma samples using next-generation sequencing (NGS). Variables including demographics (age, gender), clinical characteristics [histology pathologic differentiation, primary tumor location, Tumor-Node-Metastasis (TNM) stage, primary lymph node status, number of metastatic sites], treatment features (PD-1/PD-L1 inhibitors), and survival were also collected. Immunohistochemistry (IHC) was used to classify mismatch repair (MMR) as proficient MMR (pMMR) or deficient MMR (dMMR) by assessing four MMR proteins (MLH1, MSH2, MSH6, and PMS2). Microsatellite instability (MSI) status was determined via NGS in tumor tissue and classified as high MSI (MSI-H), microsatellite stable (MSS), and low MSI (MSI-L). Tumor mutational burden-high (TMB-H) was defined as the presence of at least 10 mutations per megabase (Mb),²⁹ and overall survival (OS) was defined as the time from disease diagnosis to death or end of follow-up.

Analysis of molecular features

We comparatively analyzed the genomic landscapes of RNF43-altered metastatic colorectal cancer (mCRC) patients from the Memorial Sloan Kettering Cancer Center (MSKCC, Cancer Cell 2018) using the cBioPortal database (https://www.cbioportal.org).

We stratified truncating and damaging missense mutations to NTD and CTD using codon 313 as a cutoff to demarcate the RING region. The in silico analyses of the missense variants utilized Polymorphism Phenotyping ver. 2 (PolyPhen-2, http://genetics.bwh. harvard. edu/pph2) to predict the functional effects of RNF43 protein.³⁰ The PolyPhen-2 software categorizes the investigated mutations as probably damaging (probability score ⩾ 0.85), possibly damaging (probability score between 0.16 and 0.85), or benign (probability score ⩽ 0.15).

Statistical analysis

The chi-square test or Fisher’s exact test was used to identify the differences in clinical characteristics, treatment features, and genomic alterations between groups. OS was examined using the Kaplan–Meier method and compared using the log-rank test. Univariate and multivariate Cox proportional hazard regression models were used to estimate the individual hazard ratio (HR). The HR with a 95% confidence interval (CI) was measured to estimate the hazard risk of individual factors. All statistical results were considered significant if the p value <0.05. Data were analyzed using SPSS software (version 24. 0, Armonk, NY:IBM Corp) and R software (version 4. 2. 0, R Foundation for Statistical Computing, https://www.r-project.org/, Vienna, Austria).

Results

Patient characteristics

Demographic and clinicopathological characteristics of the enrolled patients are summarized in Table 1. Of the 78 patients included, 49 (62.8%) were males and the median age at diagnosis was 49 years. Most patients had adenocarcinomas (71.8%) that were well or moderately differentiated (39.7%). Right-sided tumors were detected in 36 (46.2%) patients, and 61 (78.2%) patients had positive primary lymph node status. A total of 18 patients (23.1%) had more than one metastatic site, 41 (52.6%) were TNM stage IV, 30 (38.5%) had dMMR/MSI-H tumors, and 32 (41.0%) had a high tumor mutation burden (TMB-H). Meanwhile, 23 patients (29.5%) received immunotherapy and 7 patients were treated with a combination of vemurafenib, cetuximab, and irinotecan (VIC regimen). In terms of first-line regimen, 23 patients received Bevacizumab plus systemic chemotherapy, 8 patients were treated with cetuximab plus systemic chemotherapy, and 15 received standard chemotherapy only. In addition, we observed that among the included patients, 30 had concurrent KRAS mutations, 13 had BRAF V600E mutations, 24 had APC mutations, and 53 had TP53 mutations. Based on the estimation of mutation frequency, 15 patients exhibited RNF43 Gly659fs alterations, 10 had RNF43 Arg117fs alterations, and 3 patients presented with both.

Table 1.

Clinicopathological and molecular characteristics of 78 patients in RNF43-mutant CRC.

Characteristics	No. patients (%)
Characteristics	Overall (N = 78)
Age
Mean ± SD	49.28(±13.58)
Gender
Male	49 (62.8%)
Female	29 (37.2%)
Vital status
Death	26 (33.3%)
Alive	53 (66.7%)
OS
Median, CI	57.40 (48.76–64.78)
Histology
Adenocarcinoma	56 (71.8%)
Signet/mucinous	19 (24.4%)
Else	3 (3.8%)
Pathologic differentiation
Well or moderate	31 (39.7%)
Poor	32 (41.0%)
Else	15 (19.2%)
Tumor location
Left	41 (52.6%)
Right	36 (46.2%)
Multi-side	1 (1.3%)
Primary lymph node status
Positive	61 (78.2%)
Negative	11 (14.1%)
Unknown	6 (7.7%)
Number of organs involved
<2	59 (75.6%)
⩾2	18 (23.1%)
NA	1 (1.3%)
TNM stage
I–III	29 (37.2%)
IV	41 (52.6%)
NA	8 (10.3%)
MMR/MSI
dMMR/MSI-H	30 (38.5%)
pMMR/MSS	48 (61.5%)
TMB
H	32 (41.0%)
L	44 (56.4%)
NA	2 (2.6%)
RNF43
G659 mutant	15 (19.2%)
R117 mutant	10 (12.8%)
Else	53 (68.0%)
KRAS
Mutant	30 (38.5%)
Wild type	48 (61.5%)
BRAF
V600 mutant	13 (16.7%)
Else^a	65 (83.3%)
APC
Mutant	24 (30.8%)
Wild type	54 (69.2%)
TP53
Mutant	53 (67.9%)
Wild type	25 (32.1%)
Lynch
Yes	14 (17.9%)
No	64 (82.1%)
PD-1 during the disease	23 (29.5%)
VIC^b during the disease	7 (9.0%)
First-line therapy
Bevacizumab + chemotherapy	23 (29.5%)
Cetuximab + chemotherapy	8 (10.3%)
Chemotherapy	15 (19.2%)

BRAF else subgroup including wild types and else mutations except BRAF V600E.

VIC regimen involved a combination of vemurafenib, cetuximab, and irinotecan.

CRC, colorectal cancer; OS, overall survival; TMB-H, patients with high TMB; TMB-L, patients with low TMB.

The somatic mutational landscape in patients with RNF43 mutated CRC

In RNF43-mutant tumors with dMMR/MSI-H status, the most commonly mutated genes were MLL2 (80%), ARID1A (77%), RAD50 (63%), NOTCH3 (60%), and ASXL1 (60%) [Figure 1(a)]. By contrast, TP53 (83%), KRAS (29%), SMAD4 (23%), LRP1B (21%), and PIK3CA (21%) mutations were most frequently observed in the pMMR/MSS group [Figure 1(b)]. DMMR/MSI-H tumors exhibited more abundant alterations, and the majority of RNF43 mutations were multi-hit; meanwhile, pMMR/MSS tumors had more RNF43 truncating mutations (Figure 1). Furthermore, we created a visualization depicting the co-mutation landscape involving RNF43 in the MSKCC cohort. RNF43 showed a high frequency of concurrent alterations with MLL2 (68%), ARID1A (57%), FAT1 (53%), NOTCH3 (51%), and SPEN (51%) in the dMMR/MSI-H subgroup. However, in the pMMR/MSS subgroup, the most frequently co-occurring alterations with RNF43 were TP53 (77%), BRAF (48%), KRAS (32%), ARID2 (26%), and SMAD4 (26%), which is consistent with our findings (Supplemental Figure S1).

Figure 1.

The mutational landscape shows the high-frequency genomic alterations detected in RNF43-mutant colorectal cancers with [(a), n = 30] dMMR/MSI-H and [(b), n = 48] pMMR/MSS. The colors of the bars are indicative of the type of mutation, with gray = wild type.

The most frequent RNF43 mutation in our cohort was the p.Gly659fs variant in exon 9, which was observed in 18 out of 78 patients (23.08%). Meanwhile, the second most common mutation was the p.Arg117fs variant in exon 3, observed in 13 out of 78 patients (16.67%) (Supplemental Figure S2). All the p.Gly659fs variants were mutated in dMMR/MSI-H tumors. Moreover, the lollipop diagram illustrated that the majority of RNF43 mutations were truncating events (74%, 58/78) with a tendency to be located in the first half of the RNF43 protein, while missense mutations were more evenly distributed throughout the protein (Supplemental Figure S2).

The prognostic profiles of RNF43 mutations across distinct protein domains

Truncating mutations and damaging missense mutations could cause the loss of function in RNF43. We categorized them into the NTD and CTD based on the protein structure at amino acid 313, with the NTD containing the PA, TM, and RING domains (Supplemental Figure S2).

Notably, alterations in NTD were associated with a worse prognosis for OS (p = 0.043; median OS, 50.80 months) compared to CTD [Figure 2(a)]. In addition, as shown in Figure 2 and Table 2, most of the tumors in NTD tended to have positive primary lymph node status (NTD versus CTD, 82.0% versus 52.9%; p = 0.028). However, patients with C-terminal alterations preferred to have mutations in dMMR/MSI-H tumors (NTD versus CTD, 32.0% versus 70.6%; p = 0.009) and a higher proportion of high tumor mutation burden (TMB-H) (NTD versus CTD, 32.0% versus 70.6%; p = 0.004). No statistically significant differences between CTD and NTD were observed in other clinicopathological characteristics.

Figure 2.

(a) Kaplan–Meier analysis of OS in CRC patients with truncating and damaging missense mutations between distinct domains of RNF43 protein. (b) Kaplan–Meier analysis of OS in CRC patients with RNF43 mutations in different subgroups. Cox models were used to obtain HRs with 95% CIs, and a two-sided log-rank test was used for statistical comparisons without adjustment for multiplicity.

Table 2.

Differences in clinicopathological and molecular characteristics of RNF43-mutant CRC patients in distinct domains.

Characteristics	NTD (N = 50)	CTD (N = 17)	p-Value
Age	49.65 (±14. 94)	45.93 (±11. 22)	–
Gender
Male	33 (66.0%)	8 (47.1%)	0.249
Female	17 (34.0%)	9 (52.9%)
Histology
Adenocarcinoma	36 (72.0%)	13 (76.5%)	0.888
Signet/mucinous	11 (22.0%)	4 (23.5%)
Else	3 (6.0%)	0 (0.0%)
Pathologic differentiation
Well or moderate	20 (40.0%)	8 (47.1%)	0.375
Poor	19 (38.0%)	8 (47.1%)
Else	11 (22.0%)	1 (5.9%)
Tumor location
Right	23 (46.0%)	8 (47.1%)	1.000
Left	26 (52.0%)	9 (52.9%)
Else	1 (2.0%)	0 (0.0%)
Primary lymph node status
Positive	41 (82.0%)	9 (52.9%)	0.028
Negative	5 (10.0%)	6 (35.3%)
Unknown	4 (8.0%)	2 (11.8%)
Number of organs involved
<2	36 (72.0%)	14 (82.4%)	0.651
⩾2	13 (26.0%)	3 (17.6%)
NA	1 (2.0%)	0 (0.0%)
TNM stage
I–III	17 (34.0%)	10 (58.8%)	0.088
IV	25 (50.0%)	7 (41.2%)
NA	8 (16.0%)	0 (0.0%)
MMR/MSI
dMMR/MSI-H	16 (32.0%)	12 (70.6%)	0.009
pMMR/MSS	34 (68.0%)	5 (29.4%)
TMB
H	16 (32.0%)	12 (70.6%)	0.004
L	33 (66.0%)	4 (23.5%)
NA	1 (2.0%)	1 (5.9%)
KRAS
Mutant	16 (32.0%)	8 (47.1%)	0.380
Wild type	34 (68.0%)	9 (52.9%)
BRAF
V600 mutant	11 (22.0%)	1 (5.9%)	0.270
Else^a	39 (78.0%)	16 (94.1%)
APC
Mutant	12 (24.0%)	8 (47.1%)	0.123
Wild type	38 (76.0%)	9 (52.9%)
TP53
Mutant	35 (70.0%)	9 (52.9%)	0.243
Wild type	15 (30.0%)	8 (47.1%)
PD-1 during the disease
Yes	13 (26.0%)	7 (41.2%)	0.357
No	37 (74.0%)	10 (58.8%)

Bold text indicates significance at p < 0.05.

BRAF else subgroup including wild types and else mutations except BRAF V600E.

CRC, colorectal cancer; OS, overall survival; TMB-H, patients with high TMB; TMB-L, patients with low TMB.

The prognostic heterogeneity of RNF43 mutations in different subgroups

As illustrated in Supplemental Figure S2, the hotspot RNF43 mutation, p.Arg117fs variant, was located in the N-terminal region of the protein. We classified the tumor samples into five molecular subtypes: dMMR/MSI, pMMR/MSS R117, pMMR/MSS BRAF VIC (a combination regimen of vemurafenib, cetuximab, and irinotecan), pMMR/MSS BRAF ST (standard chemotherapy), and pMMR/MSS ELSE.

With a total of 78 patients, the median OS was 57.40 months. Among RNF43-mutant patients with BRAF V600E alteration in the pMMR/MSS subgroup, those treated with VIC therapy demonstrated a significantly better prognosis (p = 0.011; median OS, 47.67 months), whose survival curves resembled those of the pMMR/MSS ELSE subgroup. Notably, we observed that the pMMR/MSS R117 subgroup exhibited a notably shorter overall survival (p < 0.001; median OS, 18.9 months), with its survival curve comparable to that of the pMMR/MSS BRAF ST subgroup [p = 0.187; Figure 2(b)].

The results of the univariate Cox regression analysis revealed that TNM-stage, number of organs involved, APC and RNF43 subgroups were all significant prognostic factors for OS (Table 3). After adjusting for other clinicopathological factors, the number of organs involved and RNF43 subtypes remained independent factors for OS. Most importantly, the pMMR/MSS BRAF VIC subgroup displayed a similar outcome to the pMMR/MSS ELSE subgroup (HR: 1.00; 95% CI, 0.29–3.55; p = 0.992). On the other hand, the presence of RNF43 R117fs mutation was associated with a significantly poorer prognosis in pMMR/MSS patients (HR: 9.31; 95% CI, 2.19–39.48; p = 0.002; Table 3).

Table 3.

Univariate and multivariate Cox proportional hazards analysis of OS for patients in RNF43-mutant CRC.

Subgroup		Univariate Cox analysis			Multivariate Cox analysis
Subgroup		HR	95% CI	p-Value	HR	95% CI	p-Value
Histology (adeno versus signet/mucinous)		1.002	0.417–2.404	0.997
Differentiation (well or moderate versus poor)		1.196	0.514–2.784	0. 677
TNM (I–III versus IV)		2.821	1.106–7.194	0.030	2.579	0.194–34.349	0.473
Location (left versus right)		1.188	0.529–2.670	0.676
Primary lymph node status (negative versus positive)		3.239	0.746–14.071	0. 117
Number of organs involved (<2 versus ⩾2)		11.278	4.347–29.263	<0.001	3.906	1.137–13.424	0.031
KRAS (mutant versus WT)		0.458	0.183–1.148	0.096
APC (mutant versus WT)		0.169	0.040–0.717	0.016	0.545	0.107–2.786	0.466
TP53 (mutant versus WT)		1.120	0.459–2.734	0.804
Group	dMMR/MSI	0.176	0.039–0.799	0.024	0.062	0.006–0.694	0.024
	pMMR/MSS R117	12.488	3.515–44.367	<0.001	9.306	2.194–39.482	0.002
	pMMR/MSS BRAF VIC	1.483	0.470–4.683	0.502	1.006	0. 286–3.546	0.992
	pMMR/MSS BRAF ST	46.839	8.498–258.157	<0.001	28.129	3.795–208.496	0.001
	pMMR/MSS ELSE^a	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.
PD-1 (yes versus no)		0.265	0.079–0.887	0.031	2.357	0.476–11.676	0.294

Bold text indicates significance at p < 0.05.

pMMR/MSS ELSE subgroup including patients treated with standard chemotherapy without RNF43 R117 mutation and BRAF V600E mutation.

CRC, colorectal cancer; OS, overall survival; TMB-H, patients with high TMB; TMB-L, patients with low TMB; WT, wild type; VIC, target therapy (VIC regimen, a combination of vemurafenib, cetuximab, and irinotecan); ST, standard chemotherapy.

Clinical and molecular characteristics of RNF43 mutations in different subgroups

To further explore the clinical and molecular characteristics of RNF43 R117fs mutation, we assigned patients into three groups. In summary, no statistical differences were observed among the dMMR/MSI-H, pMMR/MSS-R117, and pMMR/MSS-ELSE subgroups when stratified by demographics (age and gender), primary tumor characteristics (histology, pathology differentiation, tumor location, and primary lymph node status), and molecular features (KRAS and BRAF V600E mutations) as shown in Table 4.

Table 4.

Differences of clinicopathological and molecular characteristics between subgroups in RNF43-mutant CRC patients. p₁ value, dMMR/MSI versus pMMR/MSS R117; p₂ value, pMMR/MSS R117 versus pMMR/MSS else.

Characteristics	No. Patients (%)
Characteristics	dMMR/MSI (N = 30)	pMMR/MSSR117 (N = 7)	pMMR/MSSelse^b (N = 41)	p-Value	p₁-Value	p₂-Value
Age	48.53 (±14.45)	49.29 (±14.23)	49.83 (±13.14)	–	–	–
Gender
Male	21 (70.0%)	3 (42.9%)	25 (61.0%)	0.345	0td .213	0.429
Female	9 (30.0%)	4 (57.1%)	16 (39.0%)
Histology
Adenocarcinoma	20 (66.7%)	4 (57.1%)	32 (78.0%)	0.478	0.728	0.289
Signet/mucinous	9 (30.0%)	3 (42.9%)	7 (17.1%)
Else	1 (3.3%)	0 (0.0%)	2 (4.9%)
Pathologic differentiation
Well or moderate	12 (40.0%)	3 (42.9%)	16 (39.0%)	0.927	0.871	0.668
Poor	12 (40.0%)	2 (28. 6%)	18 (43.9%)
Else	6 (20.0%)	2 (28.6%)	7 (17.1%)
Tumor location
Right	13 (43.3%)	5 (71.4%)	18 (43.9%)	0.408	0.515	0.237
Left	16 (53.3%)	2 (28.6%)	23 (56.1%)
Else	1 (3.3%)	0 (0.0%)	(0.0%)
Primary lymph node status
Positive	23 (76.7%)	7 (100.0%)	31 (75.6%)	0.850	0.769	0.738
Negative	4 (13.3%)	0 (0.0%)	7 (17.1%)
Unknown	3 (10.0%)	0 (0.0%)	3 (7.3%)
Number of organs involved
<2	28 (93.3%)	2 (28.6%)	29 (70.7%)	<0.001	<0.001	0.080
⩾2	1 (3.3%)	5 (71.4%)	12 (29.3%)
NA	1 (3.3%)	0 (0.0%)	0 (0.0%)
TNM stage
I–III	19 (63.3%)	0 (0.0%)	10 (24.4%)	0.001	0.006	0.458
IV	9 (30.0%)	6 (85.7%)	26 (63.4%)
NA	2 (6.7%)	1 (14.3%)	5 (12.2%)
TMB
H	29 (96.7%)	0 (0.0%)	3 (7.3%)	<0.001	<0.001	1.000
L	1 (3.3%)	7 (100.0%)	36 (87.8%)
NA	0 (0.0%)	0 (0.0%)	2 (4.9%)
KRAS
Mutant	16 (53.3%)	2 (28.6%)	12 (29.3%)	0.114	0.405	1.000
Wild type	14 (46.7%)	5 (71.4%)	29 (70.7%)
BRAF
V600 Mutant	2 (6.7%)	1 (14.3%)	10 (24.4%)	0.158	0.477	1.000
Else^a	28 (93.3%)	6 (85.7%)	31 (75.6%)
APC
Mutant	16 (53.3%)	0 (0.0%)	8 (19.5%)	0.002	0.012	0.583
Wild type	14 (46.7%)	7 (100.0%)	33 (80.5%)
TP53
Mutant	14 (46.7%)	6 (85.7%)	33 (80.5%)	0.006	0.097	1.000
Wild type	16 (53.3%)	1 (14.3%)	8 (19.5%)
Lynch
Yes	14 (46.7%)	0 (0.0%)	0 (0.0%)	<0.001	0.031	–
No	17 (53.3%)	7 (100.0%)	41 (100.0%)
PD-1
Yes	19 (63.3%)	0 (0%)	4 (9.8%)	<0.001	0.003	1.000
No	11 (36.7%)	7 (100%)	37 (90.2%)

Bold text indicates significance at p < 0.05.

BRAF else subgroup including wild types and else mutations except BRAF V600E.

pMMR/MSS else including patients without dMMR/MSI status and RNF43 R117fs mutation.

CRC, colorectal cancer; OS, overall survival; TMB-H, patients with high TMB; TMB-L, patients with low TMB; WT, wild type.

Nevertheless, our analysis revealed that patients in the pMMR/MSS-R117 subgroup had a significantly higher rate of metastasis to different organs (⩾2 organs involved) (71.4% versus 3.3%; p < 0.001), a higher proportion of tumors at TNM stages IV (85.7% versus 30.0%; p = 0.006), and a lower tumor mutational burden (100.0% versus 3.3%, p < 0.001) compared to the dMMR/MSI-H subgroup. In particular, we observed that R117fs variants were concurrent exclusivity with APC mutations (53.3% versus 0.0%; p = 0.012) in pMMR/MSS tumors compared to the dMMR/MSI-H subgroup (Table 4).

Discussion

In this retrospective study, we selected 78 RNF43 mutant patients from a cohort of more than 1000 colorectal cancer (CRC) patients, making it one of the largest series of RNF43-mut CRCs reported to date. Our dataset suggests that RNF43 R117fs mutation is strongly associated with a poor survival outcome and could serve as an independent prognostic factor in patients with pMMR/MSS tumors.

With a median age at diagnosis of 49 years, RNF43-mutant patients tend to be younger than the majority of CRC cases, which needed more concerns. The nature of the interaction between RNF43 and Frizzled (FZD) receptors, specifically whether it occurs via the DVL-interaction region (DIR) or direct binding with the PA domain, remains a subject of debate. Besides, it has been suggested that some truncating mutations in the DIR domain may confer gain-of-function properties.^7,8,10,12,13 Hence, we opted to utilize the end of the RING region as a cutoff point to examine the impact of truncating and deleterious missense mutations, which was in line with previous investigations.^7–10

Recent studies have shown that N-terminal truncating mutations of RNF43 are more efficient in enhancing Wnt/β-catenin (canonical WNT pathway) signaling activity compared to C-terminal mutations, with the majority of these mutations being loss of function.^13,18,31 Importantly, our data suggest that the C-terminal region (CTD) of RNF43 mutations is associated with a better OS, with most patients having dMMR/MSI-H and TMB-H, which could partly explain the better prognosis. Similar to previous reports, we observed that the most frequent mutations in our study were Gly659fs and Arg117fs, which were located in the C-terminal and N-terminal regions, respectively.¹⁷ All Gly659fs mutations observed in our study were identified in dMMR/MSI-H tumors; moreover, the majority of these patients received immune checkpoint inhibitor (ICI) therapy and achieved a more favorable clinical outcome. While the RNF43 R117fs mutation was predominantly observed in pMMR/MSS tumors, suggesting its potential as a biomarker in pMMR/MSS patients. Previous studies by Elez et al. and our research have both indicated that RNF43-mutant tumors co-occurring with BRAF V600E alterations exhibit sensitivity to anti-BRAF/EGFR therapy.^25,26 As a result, we categorized patients into five subgroups: dMMR/MSI, pMMR/MSS R117, pMMR/MSS BRAF VIC (a combination regimen of vemurafenib, cetuximab, and irinotecan), pMMR/MSS BRAF ST (standard chemotherapy), and pMMR/MSS ELSE. Interestingly, BRAF V600E mutant patients who received the VIC regimen demonstrated extended OS in contrast to previous reports,^32,33 suggesting that RNF43 mutation accompanied by BRAF V600E alteration represents a distinct subtype with predictive value for BRAF/EGFR inhibitor treatments. Importantly and innovatively, patients belonging to the pMMR/MSS R117 subtype exhibited a much poorer outcome, which was also found to be an independent prognostic factor. The observed difference in outcome between patients belonging to the pMMR/MSS R117 subtype and other subtypes could potentially be attributed to several factors, including the presence of more than two metastatic sites, a higher TNM stage, lower tumor mutational burden, and a lower likelihood of receiving ICI therapy.²⁹

In agreement with our findings, a previous study found a significant difference in the location of RNF43 mutations between MSI and MSS cancers, with the RNF43 G659fs mutation being frequently associated with colorectal tumors having an MSI phenotype and a favorable prognosis.¹⁷ RNF43 G659fs mutation has been reported to have equal activity with wild type, which indicates that it might be a passenger mutation or a secondary mutation effect triggered by MSI status.^13,34 On the other hand, the characteristics of RNF43 R117fs mutation, which has been reported to positively regulate Wnt signaling, are still largely unknown.¹² Li et al. previously uncovered that among a panel of RNF43-mutant CRC cell lines, only HCT116, which carries a homozygous R117fs mutation, exhibited reduced pDVL2 levels and a consistent growth inhibition following Wnt-secretion blockade.¹³ Numerous studies have highlighted the importance of RNF43 in immune modulation, with implications for ICI therapy response.^22,35 Notably, our previous work demonstrated the predictive value of RNF43 mutations in anti-PD-1/PD-L1 combination therapy for BRAF V600E mCRC patients.²⁵

Co-mutations in APC, KRAS, TP53, and SMAD4 are frequently observed in MSS-CRCs, while MSI-H CRCs harbor a higher frequency of BRAF and RNF43 alterations.³⁶ It has been observed that RNF43 mutations are exclusively associated with APC mutations and frequently co-occur with BRAF V600E mutations.²³ Likewise, all the tumors in our pMMR/MSS R117 group exhibited exclusivity with APC mutation. Tumorigenesis is a complex process that involves multiple factors, including genetic mutations in various pathways. Studies by Bert Vogelstein et al. and Yaeger et al. have demonstrated a potential trend in which right-sided colorectal tumors are associated with MSI-H status, BRAF-RNF43 mutations, and poor prognosis, while left-sided tumors are linked to the MSS phenotype, APC-KRAS mutations, and better outcomes.^2,36–38 Moreover, Vogelstein et al. also surmise that APC mutations may serve as an early rate-limiting event in the conventional pathway, while RNF43 mutations are thought to occur in the later stages of the serrated pathway.^18,23,39,40 These studies suggest that the development of CRC may involve different mechanisms of activation for the WNT and RAS-RAF-MEK-ERK signaling pathways.^24,41 Both Elez et al. and our previous studies have demonstrated that RNF43 mutations can improve the prognosis of patients with BRAF V600E mutant mCRC who are treated with anti-BRAF/EGFR therapy, providing further evidence for the co-activation of the WNT and MAPK signaling pathways.^25,26 Other studies have shown that RNF43 mutations can downregulate P53 expression and promote tumorigenesis through a multi-step process involving the WNT-RAS-P53 signaling axis.^8,20 In addition, RNF43 G659fs mutation can increase PI3K signaling by promoting p85 ubiquitination.²⁸ Taken together, these findings suggest the existence of a potential cross-talk between the WNT/MAPK/TGF-β/PI3K/P53 pathway, warranting further investigation.

The activity of N-terminal alterations, specifically the R117fs mutation, appears to be reliant on the activation of WNT signaling, which can be targeted using blocking antibodies.^9,13,31 Porcupine inhibitors (PORCi) have been shown to have anti-tumor effects by inhibiting the secretion of Wnt-ligands, and RNF43 loss-of-function mutations have been observed to be sensitive to PORCi treatment.^14,42 Moreover, blocking the WNT signaling pathway has been shown to activate the immune system by stimulating T cells and dendritic cells.^43,44 Therefore, targeting RNF43 mutations may provide a viable strategy for anti-cancer drugs, such as LGK974 and ETC159, either alone or in combination with ICIs (ClinicalTrials. gov Identifier: NCT01351103).^42,45,46

Nevertheless, the current study has a limited sample size, highlighting the need for further research using a larger cohort that also includes RNF43 wild type. Moreover, the underlying mechanisms of RNF43 mutations on tumor progression and therapy response are complex and require further elucidation. Therefore, additional investigations are necessary to fully understand the implications of RNF43 mutations in CRC.

In conclusion, our study sheds light on the prognostic differences between the genetic subgroups of RNF43 mutations in CRC. Our analysis provides comprehensive insights into RNF43-mutated CRC, including its association with clinical, molecular, and prognostic features. We verified that RNF43 got a predictive value in response to BRAF/EGFR inhibitors in BRAF V600E tumors. Particularly, we found that the RNF43 R117fs mutation was associated with a poorer prognosis and could serve as a potential biomarker in RNF43-mutant CRC. Importantly, our findings could be utilized to stratify patients with CRC and guide treatment decisions.

Supplemental Material

sj-docx-1-tam-10.1177_17588359231220600 – Supplemental material for Clinical and molecular characteristics of RNF43 mutations as promising prognostic biomarkers in colorectal cancer

Supplemental material, sj-docx-1-tam-10.1177_17588359231220600 for Clinical and molecular characteristics of RNF43 mutations as promising prognostic biomarkers in colorectal cancer by Zi-Yao Huang, Lei Wen, Liu-Fang Ye, Yu-Ting Lu, William Pat Fong, Ren-Jing Zhang, Si-Xian Wu, Zhi-Gang Chen, Yan-Yu Cai, Rui-Hua Xu, Yu-Hong Li, Zi-Ming Du and De-Shen Wang in Therapeutic Advances in Medical Oncology

Supplemental Material

sj-eps-2-tam-10.1177_17588359231220600 – Supplemental material for Clinical and molecular characteristics of RNF43 mutations as promising prognostic biomarkers in colorectal cancer

Supplemental material, sj-eps-2-tam-10.1177_17588359231220600 for Clinical and molecular characteristics of RNF43 mutations as promising prognostic biomarkers in colorectal cancer by Zi-Yao Huang, Lei Wen, Liu-Fang Ye, Yu-Ting Lu, William Pat Fong, Ren-Jing Zhang, Si-Xian Wu, Zhi-Gang Chen, Yan-Yu Cai, Rui-Hua Xu, Yu-Hong Li, Zi-Ming Du and De-Shen Wang in Therapeutic Advances in Medical Oncology

Supplemental Material

sj-eps-3-tam-10.1177_17588359231220600 – Supplemental material for Clinical and molecular characteristics of RNF43 mutations as promising prognostic biomarkers in colorectal cancer

Supplemental material, sj-eps-3-tam-10.1177_17588359231220600 for Clinical and molecular characteristics of RNF43 mutations as promising prognostic biomarkers in colorectal cancer by Zi-Yao Huang, Lei Wen, Liu-Fang Ye, Yu-Ting Lu, William Pat Fong, Ren-Jing Zhang, Si-Xian Wu, Zhi-Gang Chen, Yan-Yu Cai, Rui-Hua Xu, Yu-Hong Li, Zi-Ming Du and De-Shen Wang in Therapeutic Advances in Medical Oncology

Footnotes

Acknowledgements

The authors sincerely thank all the patients involved in this study and are grateful to the funders who supported this work.

Declarations

ORCID iDs

Yu-Hong Li

De-Shen Wang

Supplemental material

Supplemental material for this article is available online.

References

Siegel

Miller

Jemal

. Cancer statistics, 2020. CA Cancer J Clin 2020; 70: 7–30.

Vogelstein

Fearon

Hamilton

, et al. Genetic alterations during colorectal-tumor development. N Engl J Med 1988; 319: 525–532.

Leggett

Whitehall

. Role of the serrated pathway in colorectal cancer pathogenesis. Gastroenterology 2010; 138: 2088–2100.

Kinzler

Vogelstein

. Lessons from hereditary colorectal cancer. Cell 1996; 87: 159–170.

Nusse

Clevers

. Wnt/β-Catenin signaling, disease, and emerging therapeutic modalities. Cell 2017; 169: 985–999.

Klaus

Birchmeier

. Wnt signalling and its impact on development and cancer. Nat Rev Cancer 2008; 8: 387–398.

Tsukiyama

Fukui

Terai

, et al. Molecular role of RNF43 in canonical and noncanonical wnt signaling. Mol Cell Biol 2015; 35: 2007–2023.

Spit

Fenderico

Jordens

, et al. RNF43 truncations trap CK1 to drive niche-independent self-renewal in cancer. EMBO J 2020; 39: e103932.

Yusoff

PAM

Woutersen

DTJ

, et al. The functional landscape of patient-derived RNF43 mutations predicts sensitivity to wnt inhibition. Cancer Res 2020; 80: 5619–5632.

10.

Jiang

Charlat

Zamponi

, et al. Dishevelled promotes Wnt receptor degradation through recruitment of ZNRF3/RNF43 E3 ubiquitin ligases. Mol Cell 2015; 58: 522–533.

11.

Koo

Spit

Jordens

, et al. Tumour suppressor RNF43 is a stem-cell E3 ligase that induces endocytosis of Wnt receptors. Nature 2012; 488: 665–669.

12.

Cho

Sul

Kim

, et al. RNF43 R117fs mutant positively regulates Wnt/β-catenin signaling by failing to internalize FZD expressed on the cell surface. Sci Rep 2022; 12: 7013.

13.

Lavrijsen

Bakker

, et al. Commonly observed RNF43 mutations retain functionality in attenuating Wnt/β-catenin signaling and unlikely confer Wnt-dependency onto colorectal cancers. Oncogene 2020; 39: 3458–3472.

14.

Jiang

Hao

Growney

, et al. Inactivating mutations of RNF43 confer Wnt dependency in pancreatic ductal adenocarcinoma. Proc Natl Acad Sci USA 2013; 110: 12649–12654.

15.

Wang

Yuen

, et al. Whole-genome sequencing and comprehensive molecular profiling identify new driver mutations in gastric cancer. Nat Genet 2014; 46: 573–582.

16.

Waddell

Pajic

Patch

, et al. Whole genomes redefine the mutational landscape of pancreatic cancer. Nature 2015; 518: 495–501.

17.

Giannakis

Hodis

Jasmine Mu

, et al. RNF43 is frequently mutated in colorectal and endometrial cancers. Nat Genet 2014; 46: 1264–1266.

18.

Bugter

Fenderico

Maurice

. Mutations and mechanisms of WNT pathway tumour suppressors in cancer. Nat Rev Cancer 2021; 21: 5–21.

19.

Fennell

Kane

Liu

, et al. APC mutation marks an aggressive subtype of BRAF mutant colorectal cancers. Cancers 2020; 12: 1171.

20.

Tsukiyama

Zou

Kim

, et al. A phospho-switch controls RNF43-mediated degradation of Wnt receptors to suppress tumorigenesis. Nat Commun 2020; 11: 4586.

21.

Eto

Miyake

Nosho

, et al. Impact of loss-of-function mutations at the RNF43 locus on colorectal cancer development and progression. J Pathol 2018; 245: 445–455.

22.

Seeber

Battaglin

Zimmer

, et al. Comprehensive analysis of R-Spondin fusions and RNF43 mutations implicate novel therapeutic options in colorectal cancer. Clin Cancer Res 2022; 28: 1863–1870.

23.

Matsumoto

Shimada

Nakano

, et al. RNF43 mutation is associated with aggressive tumor biology along with BRAF V600E mutation in right-sided colorectal cancer. Oncol Rep 2020; 43: 1853–1862.

24.

Yan

HHN

Lai

JCW

, et al. RNF43 germline and somatic mutation in serrated neoplasia pathway and its association with BRAF mutation. Gut 2017; 66: 1645–1656.

25.

Huang

Chen

, et al. Monitoring tumour resistance to the BRAF inhibitor combination regimen in colorectal cancer patients via circulating tumour DNA. Drug Resistance Updates 2022; 65: 100883.

26.

Elez

Ros

Fernández

, et al. RNF43 mutations predict response to anti-BRAF/EGFR combinatory therapies in BRAF^V600E metastatic colorectal cancer. Nat Med 2022; 28: 2162-2170. 2022/09/14. DOI: 10.1038/s41591-022-01976-z.

27.

Tabernero

Van Cutsem

Garralda

, et al. A Phase Ib/II Study of WNT974 + Encorafenib + cetuximab in patients with BRAF V600E-Mutant KRAS wild-type metastatic colorectal cancer. Oncologist 2023; 28: 230–238.

28.

Fang

Ford-Roshon

Russo

, et al. RNF43 G659fs is an oncogenic colorectal cancer mutation and sensitizes tumor cells to PI3K/mTOR inhibition. Nat Commun 2022; 13: 3181.

29.

Marabelle

Fakih

Lopez

, et al. Association of tumour mutational burden with outcomes in patients with advanced solid tumours treated with pembrolizumab: prospective biomarker analysis of the multicohort, open-label, phase 2 KEYNOTE-158 study. Lancet Oncol 2020; 21: 1353–1365.

30.

Adzhubei

Schmidt

Peshkin

, et al. A method and server for predicting damaging missense mutations. Nature methods 2010; 7: 248–249.

31.

Yamamoto

Oshima

Wang

, et al. Characterization of RNF43 frameshift mutations that drive Wnt ligand- and R-spondin-dependent colon cancer. J Pathol 2022; 257: 39–52.

32.

Kopetz

Guthrie

Morris

, et al. Randomized trial of irinotecan and cetuximab with or without vemurafenib in BRAF-mutant metastatic colorectal cancer (SWOG S1406). J Clin Oncol 2021; 39: 285–294.

33.

Kopetz

Grothey

Yaeger

, et al. Encorafenib, binimetinib, and cetuximab in BRAF V600E-mutated colorectal cancer. N Engl J Med 2019; 381: 1632–1643.

34.

Park

, et al. The most common RNF43 mutant G659Vfs*41 is fully functional in inhibiting Wnt signaling and unlikely to play a role in tumorigenesis. Sci Rep 2019; 9: 18557.

35.

Hosein

Dangol

Okumura

, et al. Loss of Rnf43 accelerates kras-mediated neoplasia and remodels the tumor immune microenvironment in pancreatic adenocarcinoma. Gastroenterology 2022; 162: 1303–1318.e1318.

36.

Yaeger

Chatila

Lipsyc

, et al. Clinical sequencing defines the genomic landscape of metastatic colorectal cancer. Cancer Cell 2018; 33: 125–136.e123.

37.

Fearon

Vogelstein

. A genetic model for colorectal tumorigenesis. Cell 1990; 61: 759–767.

38.

Walther

Johnstone

Swanton

, et al. Genetic prognostic and predictive markers in colorectal cancer. Nat Rev Cancer 2009; 9: 489–499.

39.

McCarthy

Serra

Chetty

. Traditional serrated adenoma: an overview of pathology and emphasis on molecular pathogenesis. BMJ Open Gastroenterol 2019; 6: e000317.

40.

Tsai

Liau

Yuan

, et al. RNF43 Is an early and specific mutated gene in the serrated pathway, with increased frequency in traditional serrated adenoma and its associated malignancy. Am J Surg Pathol 2016; 40: 1352–1359.

41.

Kleeman

Koelzer

Jones

, et al. Exploiting differential Wnt target gene expression to generate a molecular biomarker for colorectal cancer stratification. Gut 2020; 69: 1092–1103.

42.

Shah

Panchal

Patel

. Porcupine inhibitors: novel and emerging anti-cancer therapeutics targeting the Wnt signaling pathway. Pharmacol Res 2021; 167: 105532.

43.

Suryawanshi

Manoharan

Hong

, et al. Canonical wnt signaling in dendritic cells regulates Th1/Th17 responses and suppresses autoimmune neuroinflammation. J Immunol 2015; 194: 3295–3304.

44.

Gattinoni

Zhong

Palmer

, et al. Wnt signaling arrests effector T cell differentiation and generates CD8+ memory stem cells. Nat Med 2009; 15: 808–813.

45.

Rodon

Argilés

Connolly

, et al. Phase 1 study of single-agent WNT974, a first-in-class Porcupine inhibitor, in patients with advanced solid tumours. Brit J Cancer 2021; 125: 28–37.

46.

DeVito

Sturdivant

Thievanthiran

, et al. Pharmacological Wnt ligand inhibition overcomes key tumor-mediated resistance pathways to anti-PD-1 immunotherapy. Cell Rep 2021; 35: 109071.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.04 MB

1.91 MB

0.76 MB