Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2021 Dec;11(12):e601.
doi: 10.1002/ctm2.601.

Multi-omics consensus ensemble refines the classification of muscle-invasive bladder cancer with stratified prognosis, tumour microenvironment and distinct sensitivity to frontline therapies

Xiaofan Lu  1 , Jialin Meng  2 , Liwen Su  1 , Liyun Jiang  1   3 , Haitao Wang  4 , Junkai Zhu  1 , Mengjia Huang  1 , Wenxuan Cheng  1 , Li Xu  1 , Xinjia Ruan  1 , Shuyuan Yeh  5 , Chaozhao Liang  2 , Fangrong Yan  1
Affiliations

Multi-omics consensus ensemble refines the classification of muscle-invasive bladder cancer with stratified prognosis, tumour microenvironment and distinct sensitivity to frontline therapies

Xiaofan Lu et al. Clin Transl Med. 2021 Dec.
No abstract available

PubMed Disclaimer

Conflict of interest statement

The authors have no conflict of interest.

Figures

FIGURE 1
FIGURE 1
The multi‐omics consensus ensemble identifies four molecular subtypes of MIBC. (A) Design overview. This study enrolled a total of 1097 muscle‐invasive bladder cancer (MIBC) cases and identified four MIBC subtypes under multi‐omics framework that stratify prognosis, tumour microenvironment and distinct sensitivity to frontline therapies. An R package 'refineMIBC' was provided for MIBC refinement in clinical setting. (B) Comprehensive heatmap showing the molecular landscape of four integrative consensus subtypes (iCSs) of MIBC (n = 396). Other previously defined gene expression‐based MIBC subtypes were annotated at the top of the heatmap, including prediction analysis of microarrays‐based (PAM), one nearest neighbour (oneNN) prediction model‐based, Lund, TCGA and consensus molecular subtypes (CMS). (C) Pie charts showing the proportion of other gene expression‐based MIBC subtypes in the current iCS. Kaplan–Meier curves of (D) progression‐free survival and (E) overall survival with log‐rank test for 396 MIBC patients stratified by iCS
FIGURE 2
FIGURE 2
Molecular landscape of four MIBC iCSs. (A) Heatmap showing profiles of regulon activity for 23 regulator (top panel), and potential regulators that are relevant with chromatin remodelling (middle panel); Similar patterns were shared by iBS1 and iBS4, but iBS1 differed with high activity of GATA6, FGFR1 and ESR1 for regulon activity, while iBS4 was distinctly associated with high activity of TP63. A total of 296 unique differentially methylated promoters derived from each iCS versus adjacent normal samples (bottom panel); the iLS3 (265 vs. 45 in iLS2) and iBS4 (191 vs. 26 in iBS1) had more hypermethylated promoters (296 unique loci) than the 21 normal samples had. (B) Genomic alteration landscape according to iCS. Samples are sorted in descending order according to the cumulative contribution of APOBEC‐relevant mutational signatures (i.e., SBS2 + SBS13) within each iCS. TMB, relative contribution of four mutational signatures, selected differentially mutated genes (>5%) and broad‐level CNAs (>20%), and selected genes located within Chr9p21.3 are shown from the top to the bottom panels. Of note, iBS1 harboured more mutations of TP53 (74%; p < 0.001), RB1 (39%; p < 0.001) and KMT2A (18%; p = 0.033) than others, while iBS4 was enriched in mutations of NFE2L2 (16%; p = 0.001; p = 0.005 compared to iBS1 [3%]) and TRANK1 (10%; p = 0.06; p = 0.001 compared to iBS1 [0%]); KIAA0947 (11%; p = 0.005), MED12 (11%; p = 0.005), COL6A6 (13%; p = 0.008), and ARID2 (14%; p = 0.01) were mutated more frequently in iLS2, whereas iLS3 was enriched in mutations of FGFR3 (34%; p < 0.001), STAG2 (22%; p = 0.006), and SPEN (11%; p = 0.05). The proportion of iCS in each alteration is presented in the right bar charts. The distributions of TMB and APOBEC contributions are shown in (C) and (D), respectively. (E) Distribution of fraction genome altered (FGA) and fraction genome gain/loss (FGA/FGG). Bar charts are presented as the mean ± standard error of the mean. Statistical p values were calculated by Kruskal–Wallis rank sum test for multiple comparison
FIGURE 3
FIGURE 3
Differential immune profile across MIBC subtypes and its association with genomic alteration and immunotherapeutic response. (A) Heatmap showing the immune profile in the MIBC‐TCGA cohort, with the top panel showing the expression of genes involved in immune checkpoint targets and the bottom panel showing the enrichment level of 24 microenvironment cell types. The immune enrichment score, stromal enrichment score and DNA methylation of tumour‐infiltrating lymphocytes (MeTILs) were annotated at the top of the heatmap. (B) Immune profile heatmap for the IMvigor210 cohort with annotations for immune enrichment score, stromal enrichment score, immune phenotype, immune cell (IC) level, and best confirmed overall response, including four categories, namely, complete response (CR), partial response (PR), stable disease (SD), and progressive disease (PD). Barplots showing the distribution of (C) CR and PD, (D) immune phenotype and (E) IC levels in iCS of the IMvigor210 cohort. (F) Subclass analysis revealed that the iBS1 subtype could be more sensitive to the anti‐PD‐L1 and anti‐PD1 agents (both, Bonferroni‐corrected p = 0.001) using two reference cohorts in which patients received immunotherapy. (G) Heatmap showing the correlation between Chr4 copy number deletion and expression profiles of immune markers across ICSs in the MIBC‐TCGA cohort. A positive correlation indicates that the deletion of Chr4 may trigger the downregulation of the relevant immune gene expression. (H) Positive correlation between downregulation of expression of immune regulatory genes and Chr4 deletion in immune‐cold iBS4 of the MIBC‐TCGA cohort; statistical p value was calculated by Spearman rank‐order correlation coefficient
FIGURE 4
FIGURE 4
Oncogenic pathways, replication stress and targeted inhibitors in MIBC as well as predictive performance of random forest basal‐classifier and its application in pan‐carcinoma investigation. Dysfunctional oncogenic pathways quantified by single‐sample gene set enrichment analysis are presented in (A) a heatmap and (B) a boxplot for the MIBC‐TCGA cohort. Relevant mutations involved in several oncogenic pathways are annotated at the top. The cell cycle oncogenic pathway was significantly activated in basal‐inflamed/noninflamed MIBC; the luminal‐desert and basal‐noninflamed subtypes showed relatively higher activation of the oncogenic NRF2 pathway; the luminal‐excluded subtype was highly enriched in the WNT pathway, while the luminal‐desert subtype showed the lowest enrichment in angiogenesis genes but the highest activation of the TGF‐β pathway. (C) Enrichment heatmap showing the signalling pathways (Reactome database) that are involved in DNA maintenance as well as activated cell cycle regulation in DNA damage response and replication stress. Two replication stress (RS) subtypes were identified for the MIBC‐TCGA cohort. Subtypes with (D) high replication stress (RS‐High) or (E) basal‐like MIBC (iBS and iBS4) were inferred to be much more sensitive to both ATR (i.e., VE‐822, AZD6739 and VE‐821) and WEE1 (i.e., Wee1 inhibitor and MK‐1775) inhibitors by applying a ridge regression model using 727 human cancer cell lines. Drug sensitivity was measured as ln(IC50), and the lower the value was, the more sensitive the patient would be to the treatment. (F) Distribution of COX2 expression between immune‐hot (i.e., iBS1 and iLS2) and immune‐cold (i.e., iLS2 and iBS4) phenotypes of MIBC‐TCGA. (G) ROC curve showing predictive performance (area under the curve [AUC] and accuracy [ACC]) when using the basal‐classifier to refine basal‐like MIBC into basal‐inflamed and basal‐noninflamed subtypes. The prognostic value of the basal‐classifier in refining previously identified basal‐like subtypes of MIBC is presented in (H) for the TCGA‐basal/squamous subtype, (I) for the PAM‐basal subtype, (J) for the oneNN‐basal subtype, (K) for the Lund‐SCCL subtype, and (L) for the CMS‐Ba/Sq subtype. (M) Diagonal heatmap showing global immunological divergence across 22 basal‐like carcinomas with a total of 2459 cases. The upper triangle of each heatmap cell represents the average expression of the immune checkpoint/cell in the predicted basal‐inflamed subtype given a specific tumour type, while the lower triangle represents the predicted basal‐noninflamed subtype
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Choi W, Porten S, Kim S, et al. Identification of distinct basal and luminal subtypes of muscle‐invasive bladder cancer with different sensitivities to frontline chemotherapy. Cancer Cell. 2014;25(2):152‐165. - PMC - PubMed
    1. Damrauer JS, Hoadley KA, Chism DD, et al. Intrinsic subtypes of high‐grade bladder cancer reflect the hallmarks of breast cancer biology. Proc Natl Acad Sci U S A. 2014;111(8):3110‐3115. - PMC - PubMed
    1. Kamoun A, de Reynies A, Allory Y, et al. A consensus molecular classification of muscle‐invasive bladder cancer. Eur Urol. 2020;77(4):420‐433. - PMC - PubMed
    1. Robertson AG, Kim J, Al‐Ahmadie H, et al. Comprehensive molecular characterization of muscle‐invasive bladder cancer. Cell. 2018;174(4):1033. - PMC - PubMed
    1. Lu X, Meng J, Zhou Y, et al. MOVICS: an R package for multi‐omics integration and visualization in cancer subtyping. Bioinformatics. 2020;36:5539‐5541. - PubMed

Publication types

MeSH terms

Substances