Abstract
Background
Bladder cancer (BLCA), the 10th most common cancer worldwide, presents
a worsening prognosis as the disease progresses. Reliable tools for
predicting BLCA prognosis and treatment efficacy remain urgently
needed.
Methods
Expression profiles of lactylation related genes were analyzed
utilizing the Cancer Genome Atlas (TCGA) database and BLCA data from
the [37]GSE13507 dataset. Two distinct clusters were identified through
unsupervised clustering analysis. Lactylation associated gene
signatures were established and subsequently validated using training
cohort and different validation cohorts. Immune cell infiltration
patterns and drug response profiles were systematically evaluated.
Parallel analyses of lactylation related genes were conducted at the
single-cell resolution. A series of in vivo and in vitro experiments
were subsequently performed to validate the findings.
Results
We examined the mRNA expression profiles of 22 lactylation related
genes in BLCA tissues. Through comprehensive analysis, we identified
two distinct lactylation clusters that exhibited significantly
different clinical outcomes and tumor immune microenvironment
characteristics. Building upon these findings, we subsequently
stratified patients into two molecular subtypes according to the
lactylation clusters and established a robust genetic signature for
predicting survival outcomes in BLCA patients. The lactylation risk
score showed a strong connection with survival outcomes and correlated
with the tumor microenvironment (TME) immunosignature and predicted
immunotherapy efficacy. DHCR7 emerged as a pivotal prognostic gene from
the nine gene model, prompting subsequent focused analyses. Single-cell
analysis confirmed that DHCR7 reached peak expression in tumor
epithelial cells, whereas TCGA data and single-cell data demonstrated
strong associations between DHCR7 and diverse immune-cell populations.
For the first time, we identified that knockdown of DHCR7 enhances the
efficacy of both cisplatin chemotherapy and immunotherapy, highlighting
DHCR7 as a key player in cisplatin resistance and its influence on
immunotherapy effectiveness in BLCA. These findings offer valuable
insights into potential combined therapeutic strategies.
Conclusions
We developed a robust lactylation risk prediction model for accurately
forecasting BLCA prognosis and identified DHCR7 as a pivotal biomarker
involved in cisplatin resistance and influencing immunotherapy efficacy
in BLCA.
Keywords: bladder cancer, lactylation, chemotherapy, Dhcr7,
immunotherapy, prognostic signature, tumor immune microenvironment
Introduction
Bladder cancer is the 10th most common malignancy globally and the
second most prevalent urological tumor ([38]1, [39]2). At diagnosis,
70–75% of patients present with non-muscle invasive bladder cancer
(NMIBC), 20–25% with muscle-invasive bladder cancer (MIBC), and 5% with
metastatic disease ([40]3). NMIBC management typically involves
endoscopic resection and risk based intravesical adjuvant therapy. In
contrast, MIBC treatment involves more aggressive approaches, including
surgery, radiation, and chemotherapy. Advanced BLCA treatment relies on
systemic therapy ([41]3). Despite multiple treatment options and
incremental advances in BLCA management, many patients face tumor
recurrence even after standardized therapy. Limited sensitivity to
current treatments contributes to persistently high overall mortality
([42]4–[43]9). Therefore, identifying novel prognostic biomarkers and
predictive tools, along with understanding factors that influence
treatment efficacy, particularly for cisplatin-based chemotherapy and
immunotherapy, is crucial for advancing personalized and precise
treatment strategies for BLCA.
Enhanced glycolysis, one of the most critical metabolic changes during
hypoxia, leads to increased lactate production in cancer and
participates in various cellular processes ([44]10). Lactate, an
abundant tumor metabolite, arises from the Warburg effect within the
tumor microenvironment ([45]11). Lactylation involves the covalent
attachment of a lactate molecule to a protein through a chemical
reaction between lactate and a lysine residue on the protein ([46]12).
Lactylation modifications play a crucial role in various biological
activities, such as tumorigenesis ([47]13, [48]14), tumor progression
([49]15), macrophage polarization ([50]16), and drug resistance
([51]17, [52]18). At present, the relationship between lactylation and
BLCA progression, prognosis, immunotherapy, tumor immune
microenvironment, and drug resistance remains unclear. Additionally,
predictive models for evaluating the prognostic significance of
lactylation related genes in BLCA are still lacking. Consequently,
exploring the pathological processes, potential biological functions,
and effective predictive models of lactylation could provide new
strategies for the diagnosis and treatment of BLCA.
In this study, we utilized bulk RNA transcriptome and single-cell RNA
sequencing data, integrating various algorithms such as Consensus
Clustering, immune infiltration analysis, enrichment analysis, and
predictive modeling of lactylation related genes. The stability and
reliability of the prognostic model were validated in external cohorts
to comprehensively examine the expression patterns of lactylation
related genes in BLCA. Among the modeled genes, DHCR7 was identified as
significantly overexpressed in BLCA, strongly associated with
prognosis, and implicated in regulating the tumor immune
microenvironment. Subsequently, a series of in vivo and in vitro
experiments revealed that DHCR7 knockdown enhances the efficacy of both
cisplatin chemotherapy and immunotherapy.
Materials and methods
Bulk RNA-seq data acquisition and preprocessing
We retrieved BLCA transcriptomic data and clinical profiles from The
Cancer Genome Atlas database
([53]https://www.cancer.gov/ccg/research/genome-sequencing/tcga), our
analysis included 412 tumor samples and 19 normal samples. For the
IMvigor210 dataset, RNA-seq and clinical information were obtained
using the R package IMvigor210CoreBiologies ([54]19). In addition,
RNA-seq data and relevant clinical details from BLCA cohorts
([55]GSE13507, [56]GSE19423, [57]GSE32894, [58]GSE48075, [59]GSE48276)
were obtained from the Gene Expression Omnibus (GEO) database.
[60]GSE13507 contains microarray data of 165 primary BLCA samples, 23
recurrent NMIBC tissues, 58 adjacent tissues and 9 normal bladder
samples ([61]20). [62]GSE19423 included microarray gene expression
analysis of 48 patients with primary pT1 BLCA who received BCG
immunotherapy ([63]21). [64]GSE32894 contains the gene expression
profile data of 308 cases of urothelial carcinoma ([65]22).
[66]GSE48075 included 142 cases of primary BLCA, among which 73 cases
were MIBC ([67]23). [68]GSE48276 contains the gene expression profile
analysis of 116 cases of urothelial carcinoma ([69]24). Six
cisplatin-related datasets ([70]GSE165767, [71]GSE235066, [72]GSE15372,
[73]GSE77515, [74]GSE33482, [75]GSE45553) were obtained from the GEO
database. [76]GSE165767 presents the gene expression difference map of
BLCA cell line T24 with or without cisplatin treatment ([77]25). The
[78]GSE235066 dataset collected transcriptome sequencing data of RT112
and 5637 BLCA cell lines treated with or without cisplatin ([79]26).
[80]GSE15372, [81]GSE33482 and [82]GSE45553 contain microarrays of
expression of normal and cisplatin-resistant ovarian cancer cell lines
([83]27, [84]28). The transcriptome sequencing data of breast cancer
cells treated with cisplatin were obtained in the [85]GSE77515 dataset
([86]29). In the GEO dataset, genes were matched to probes based on
platform annotations. For genes corresponding to multiple probes, the
maximum expression value was used. In these high throughput
experiments, we used R software for corresponding processing, using the
Combat function from the sva R package to remove batch effects
([87]30). The lactylation associated genes were identified in previous
studies ([88]31). The above datasets were used for the construction and
validation of the prediction model and for the analysis of cisplatin
sensitivities related genes.
Single-cell RNA sequencing analysis
Three single-cell sequencing datasets ([89]GSE135337, [90]GSE130001,
[91]GSE129845) were obtained from the GEO database, comprising 9 tumor
samples and 4 normal tissue samples. Standardized single-cell RNA
sequencing (scRNA-seq) data from these BLCA patients were analyzed
using the R package Seurat ([92]32). The following cells were excluded:
1) mitochondrial gene expression exceeding 10%; 2) fewer than 200
feature genes; and 3) more than 4000 feature genes. To eliminate batch
effects, data integration was performed using the Harmony R package
([93]33). Five algorithms—AUCell, UCell, singscore, ssgsea, and
AddModuleScore were applied to single-cell data for enrichment score.
CellCall ([94]34), a toolkit that utilizes KEGG pathway-based
ligand-receptor-transcription factor (L-R-TF) axis datasets, was
employed to infer intercellular communication networks and internal
regulatory signals by integrating intra and intercellular signals.
Using CellCall R package, we further elucidated specific pathways
between DHCR7 high and low expression group cells.
Consensus clustering and differential gene expression
Unsupervised cluster analysis was conducted using the R package
ConsensusClusterPlus to identify unique patterns of genes associated
with lactylation. Expression profiling data of lactylation related
genes were used to classify patients for further analysis. To ensure
classification reliability, 100 replications were performed.
Model development
To construct a model based on lactylation related genes, we conducted a
series of analyses, including cluster differential analysis, univariate
Cox regression, and least absolute shrinkage and selection operator
(LASSO) regression with ten-fold cross-validation using the glmnet R
package ([95]35). For LASSO regression, we selected lambda.min to
prevent overfitting. A final set of nine genes (DHCR7, P4HB, CD109,
FADS1, HOXC8, CLDN5, TMC7, KRT4, ADIRF) were identified to construct a
prognostic formula termed ‘Riskcore’ Risk
[MATH:
score=
∑i=1
nCoei·Exp{i
} :MATH]
, where
[MATH:
Coei
:MATH]
represents the coefficients of the genes and
[MATH:
Exp{i} :MATH]
represents the relative expression of genes in the cohort.
Kaplan-Meier survival analysis and tumor mutational burden
Kaplan-Meier (K-M) analysis was conducted to compare survival between
high and low groups. The predictive accuracy of Riskscore for 1-, 3-,
and 5-year survival was evaluated using ROC curves generated by the
timeROC R package ([96]36). We also investigated the expression levels
of immune checkpoint genes and tumor mutational burden (TMB) to
evaluate their potential as predictive markers for immunotherapy
response. BLCA mutation data was retrieved from the TCGA database, and
tumor mutational burden was calculated using the maftools R package
([97]37).
Functional enrichment analysis
Functional enrichment analyses were conducted using the R package
clusterProfiler ([98]38), focusing on Gene Ontology (GO) and Kyoto
Encyclopedia of Genes and Genomes (KEGG) pathways. Genomic Variation
Analysis (GSVA) was conducted using the GSVA R package to compare
pathway activation between different groups ([99]39). Gene set
enrichment analysis (GSEA) was performed using the R software package.
To obtain a comprehensive overview of the proteins translated by each
group of differential genes, we utilized the Proteomaps database
([100]40).
Analysis of tumor microenvironment cell infiltration
The relative abundance of various cells in the tumor was determined
using the CIBERSORT and ssGSEA (Single Sample Gene Set Enrichment
Analysis) methods, implemented through the CIBERSORT and GSVA R
packages, respectively. Using the ESTIMATE tool ([101]41), this study
analyzed BLCA gene expression data to estimate stromal content, tumor
purity, and immune cell infiltration in cancerous tissues, predicting
immune scores, stromal scores, and tumor purity in BLCA.
Prediction of drug sensitivity
Drug sensitivity analysis was performed using data from the Genomics of
Drug Sensitivity in Cancer 2 (GDSC2) database
([102]https://www.cancerrxgene.org/). The relationship between DHCR7
expression and drug sensitivity was analyzed using the oncoPredict R
package ([103]42).
Cell culture
The T24(#SC0113), J82(#SC0116), and MB49(#SC0512) BLCA cell lines were
procured from Yuchi Biology (Shanghai, China). They were cultured in
DMEM or RPMI-1640 media (Gibco, USA) at 37°C in 5% CO2, respectively.
These media contained 10% fetal bovine serum (AC03L055, Shanghai Lifei
Lab Biotech, China) and 1% penicillin-streptomycin. All cells were
subjected to STR authentication. Additionally, Mycoplasma contamination
was checked regularly using a Mycoplasma detection kit (Biotool,
Houston, TX).
Antibodies and reagents
β-actin (#20536-1-AP, Proteintech, 1:6000 dilution), cleaved
caspase-3(#19677-1-AP, Proteintech,1:1000 dilution), DHCR7(#ab103296,
Abcam, 1:1000 dilution). The following chemicals and reagents were
used: water for injection (WFI) for Cell. Culturesodium (ThermoFisher,
A1287301) lactate (L-lactate, #867-56-1, MedChemExpress), Oxamate
(#565-73-1, MedChemExpress) and Cisplatin (#S1166, Selleck).
Plasmids and transfection
The short hairpin RNAs (shRNAs) of DHCR7 was obtained from GeneCopoeia
(Guangzhou, China). The sequences as followed: human:
5′-GATCCCCTGACTTCTGCCATAAGTTCTCGAGAACTTATGGCAGAAGTCAGGGTTTTTG -3′;
mouse: 5′-GATCCCACAGATTTCTGCCAGGTTACTCGAGTAACCTGGCAGAAATCTGTGGTTTTTG
-3′ For gene knockdown experiments, cells were cultured in plates or
dishes to undergo starvation treatment with serum-free Opti-MEM medium
(Gibco, USA) for 12 hours. Then transfected with 2 μg/ml of indicator
vector, after 72 hours, cells are collected for further experiments.
All transfections were performed using Lipofectamine 2000 (Invitrogen,
America) according to the manufacturer’s instructions. After puromycin
selection, we obtained cells stably transfected with the indicator
plasmids.
Cell proliferation assay
The cells were seeded in 96-well plates; approximately 10^4 cells were
seeded per well. After culturing for 24 h at 37°C in 5% CO2, the cells
were divided into several groups with different treatments. Each group
had at least 3 repetitions. Ten microliters of CCK-8 reagent (#C0037,
Beyotime, China) were added to each well and incubated for 1 h under
the above conditions. The absorbance at 450 nm was measured by a
microplate reader. The CCK-8 assay was applied to measure the half
maximal inhibitory concentration (IC50) of cisplatin after treatment
with a serial dose of cisplatin for 24 h in T24 and J82 cells.
Apoptosis assay
Caspase-3 activity and Annexin V-FITC/PI assays were used to assess
cell apoptosis. The caspase-3 activity assay was performed using the
Caspase-3 Assay Kit (ab39401, Abcam) according to the manufacturer’s
protocol. For the Annexin V-FITC/PI assay, cells were stained with
Annexin V-FITC and PI using the Annexin V-FITC Apoptosis Detection Kit
(A21102, Vazyme), following the manufacturer’s instructions. After
staining, cells were incubated at room temperature for 15 minutes and
analyzed using a flow cytometer. Data analysis was performed with
FlowJo software.
Western blot
Cells were lysed with RIPA buffer containing protease and phosphatase
inhibitors (#P0013, Beyotime, China). Protein concentration was
determined using a Micro BCA Protein Assay Kit (Sigma-Aldrich). Equal
amounts of protein were resolved via sodium dodecyl
sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred
onto polyvinylidene fluoride (PVDF) membranes with 0.45 μm pores
(Millipore, Bedford, MA, USA). The membranes were blocked with 5%
skimmed milk and incubated with primary antibodies overnight at 4°C.
The next day, membranes were incubated with secondary antibodies and
visualized using enhanced chemiluminescence (ECL) reagent
(Sigma-Aldrich).
RT-qPCR
Total RNA (1 μg) was extracted using Trizol reagent (#AG21102, Accurate
Biotechnology, Hunan, China) according to the manufacturer’s
instructions. RNA concentration and quality were assessed using a
NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific). cDNA was
then synthesized using a gDNA-free reagents kit (#AG11728, Accurate
Biotechnology, Hunan, China). Reverse transcription quantitative PCR
(RT-qPCR) was performed using SYBR Premix ExTaq (Cat. No. AG11701,
Accurate Biotechnology, Hunan, China). The primer sequences are
provided in [104]Supplementary Table 1 . GAPDH was used as the internal
control for sample normalization. The results are presented as
expression levels relative to the control group, which was set to 1.
Mice study
All animal procedures were approved by the Ethics Committee of the
Second Xiangya Hospital, Central South University (Approval No.
20241117). Six-week-old C57BL/6 mice were purchased from Shulaibao
Biotechnology (Wuhan, China). MB49 cells (1×10^7in 100 µl 1×PBS)
infected with shControl or shDhcr7 lentivirus were injected s.c. into
the right flank of mice. After the xenografts reached a size of
approximately 50 mm^3, mice carrying similar types of tumors were
randomized into different groups and treated with anti-PD-1(BioXcell,
Clone RMP1-14)/IgG (BioXcell, Clone 2A3) (200 μg, i.p., given at days
0, 3, 6). Mouse was euthanized when it meets the end-point standard
required by ethics committee. And the tumor was collected for
immune-fluorescence.
Statistical analysis
All data analyses in the bioinformatics section were conducted using R
software (version 4.3.2).
Comparisons between two independent groups were conducted using the
two-tailed Wilcoxon test for the raw letter portion unless otherwise
specified. All basic experimental data were analyzed using GraphPad
Prism V.10, and differences between groups were analyzed using
Student’s two-tailed t-test. Spearman correlation analysis was used to
evaluate the relationships between variables. Survival differences were
assessed using K-M survival curves with log-rank tests. For clarity in
presentation, p-values > 0.05 were labeled as ‘ns’, p-values < 0.05 as
‘*’, p-values < 0.01 as ‘**’, and p-values < 0.001 as ‘‘***”. The
experimental data were collected from three independent experiments and
expressed as mean ± SD. P values below 0.05 indicate statistical
significance.
Results
Identification of lactylation clusters and differential gene expression in
BLCA
In this study, we examined the expression profiles of lactylation
related genes in both normal and tumor samples. Our analysis revealed
that genes such as ARID3A, CCNA2, and DDX39A were consistently and
significantly overexpressed in cancer tissues ([105] Figure 1A ). High
expression of most lactylation related genes correlates with poorer
prognosis ([106] Supplementary Figures S1A–O ). Using the
ConsensusClusterPlus R package, we performed unsupervised clustering of
BLCA patients based on lactylation related genes, resulting in the
identification of two distinct patient clusters ([107] Figures 1B, C ).
We observed a significant difference in immune cell infiltration
between the two patient clusters. Notably, Cluster B exhibited higher
levels of infiltrating immune cells, including natural killer T cells
([108] Figure 1D ). Patients in Cluster B also demonstrated a more
favorable prognosis ([109] Figure 1E ), suggesting that enhanced immune
cell presence within the tumor microenvironment may contribute to
improved clinical outcomes. In light of the survival differences
observed between the two clusters, we conducted a pathway enrichment
analysis using the GSVA algorithm. This analysis revealed distinct
pathway profiles for each cluster. Specifically, Cluster A exhibited
higher enrichment scores for cell cycle, ubiquitin-regulated protein
degradation, and RNA degradation pathways. The significant enrichment
of these pathways may explain the poorer prognosis associated with
Cluster A ([110] Figure 1F ).
Figure 1.
[111]Panel A displays box plots comparing gene expression levels
between normal and tumor samples. Panel B shows a consensus matrix
heatmap for k=2 clustering. Panel C is a scatter plot showing PCA with
two clusters labeled A and B. Panel D presents box plots of immune cell
infiltrations for clusters A and B. Panel E is a Kaplan-Meier survival
curve demonstrating differences in survival between the two clusters,
with a p-value of 0.007. Panel F is a heatmap of KEGG pathways, with
rows as pathways and columns as samples colored by cluster assignment.
[112]Open in a new tab
Identification of lactylation clusters and differential gene expression
in BLCA. (A) Expression differences of lactylation related genes
between cancer and normal tissues. (B) Consensus matrix showing
clustering results with the number of clusters (k) set to 2. (C) PCA
plot illustrating the two clusters. (D) Immune cell expression
differences between the clusters. (E) K-M curves showing overall
survival differences between the two clusters. (F) Heatmap showing KEGG
enrichment analysis differences between clusters for 25 pathways. The
symbols *, **, and *** represent P < 0.05, P < 0.01, and P < 0.001,
respectively.
Identification of gene subtypes influenced by BLCA lactylation clusters
Subsequently, we employed the “limma” package to identify 1,250
differentially expressed genes (DEGs) associated with the lactylation
clusters. This enabled us to further explore the distinct biological
behaviors exhibited by each cluster. We performed functional enrichment
analyses, including GO and KEGG pathway evaluations, on the DEGs
associated with the lactylation clusters ([113] Figures 2A, B ). From
these differentially expressed genes, we identified 692 genes with
significant prognostic associations (p < 0.05) through one-way Cox
regression analysis. Patients were grouped into two genetic subtypes
based on these 692 prognoses associated genes ([114] Figures 2C, D ).
We observed a significant difference in prognosis between the two gene
clusters ([115] Figure 2E ). Interestingly, RACGAP1 was the only gene
that showed a significant difference between the two gene clusters
([116] Figure 2F ).
Figure 2.
[117]Group of six panels showing various data visualizations: A) Bar
chart of chromosomal and molecular processes with counts and associated
significance values. B) Bar chart of biological pathways and diseases
with counts and significance values. C) Heatmap showing consensus
matrix for k=2 clusters. D) Line chart depicting consensus cumulative
distribution function (CDF) for different cluster numbers. E)
Kaplan-Meier survival curves comparing two gene clusters with a
significant p-value. F) Box plots showing gene expression levels for
multiple genes across two clusters.
[118]Open in a new tab
Identification of gene subtypes influenced by BLCA lactylation
clusters. (A) Gene Ontology (GO) analysis highlighting enriched
biological processes (BP), cellular components (CC), and molecular
functions (MF) between clusters. (B) KEGG enrichment analysis of
differentially expressed genes between clusters. (C, D) Consensus
matrix illustrating clustering results with the number of clusters (k)
set to 2. (E) Kaplan Meier curves of overall survival differences
between two clusters. (F) Expression differences of lactylation related
genes among gene clusters. The symbols *represent P < 0.05.
Development and validation of a lactylation related gene signature
All patients were randomly divided into training and validation sets at
a 1:1 ratio. From the initial 1,250 differentially expressed genes, 692
were identified as preliminary prognostic genes through one way cox
regression analysis. Using the LASSO algorithm, this set was further
narrowed down to a final selection of nine genes ([119] Figures 3A, B
). The final prognostic risk score was derived from nine gene
signatures associated with patient prognosis. Prognostic scores were
calculated from the expression levels of these genes using the
following formula:
Riskscore=DHCR7*0.240+P4HB*0.604+CD109*0.128+FADS1*0.216+HOXC8*0.210+CL
DN5*0.296+TMC7*(-0.250) +KRT4*0.103+ADIRF*(-0.041). Patients were
classified into high or low score groups based on the median Risk score
value. To illustrate the relationships among clustering, Risk score
subgroups, and survival status in BLCA, we utilized Sankey diagrams
([120] Figure 3C ). Cluster A exhibited a consistently poorer prognosis
and a higher likelihood of falling into the high-risk group, indicating
that patients in cluster A are more prone to adverse outcomes. Patients
belonging to the lactylation or gene cluster A exhibited higher risk
scores, which aligns with the patterns observed in the Sankey diagram
([121] Figures 3D, E ). Additionally, we observed that high risk
patients exhibited higher expression levels of lactylation related
genes ([122] Figure 3F ). The KM survival curve clearly shows that
patients in the high score group have a significantly worse prognosis
(P < 0.05) ([123] Figures 3G–I ). Risk scores demonstrated a worsening
trend across clinicopathologic subgroups, with higher risk score groups
correlating with poorer prognosis ([124] Supplementary Figures S3A–N ).
High risk patients exhibit an increased risk of death ([125]
Supplementary Figures S4A–C ). In the train patient cohort, we assessed
the predictive performance of our prognostic model by generating
receiver operating characteristic (ROC) curves for 1-, 3-, and 5-year
overall survival (OS) ([126] Figure 3J ). The area under the curve
(AUC) values were 0.77, 0.82, and 0.83, respectively, highlighting the
model’s strong predictive accuracy, particularly for long-term
survival. Furthermore, the model demonstrated high predictive
performance in both the validation and overall cohorts ([127]
Figures 3K, L ). Additionally, we validated the accuracy of the model
in three independent external validation datasets ([128]GSE13507,
[129]GSE32894, [130]GSE48075) achieving highly satisfactory results
with 5-year AUC values of 0.71, 0.89, and 0.69, respectively. In the
immunotherapy cohort, we observed that high-risk patients had a worse
prognosis following immunotherapy. Similar trends were noted within
both the responding and non-responding subgroups ([131] Supplementary
Figures S2H–J ). Besides, elevated expression levels of PDL1, PD1, and
CTLA4 were observed in the high risk group ([132] Supplementary Figures
S3O–Q ). Univariate and multivariate Cox regression analyses,
incorporating other clinicopathologic factors, demonstrated that the
Risk core is an independent risk factor ([133] Figures 3M, N ). To
investigate the biological basis of the prognostic differences between
high and low risk groups, we conducted further analyses. GSEA
enrichment analysis of differences between high and low risk groups
revealed the top five pathways, with significant activation of the cell
cycle pathway in the high-risk group ([134] Supplementary Figure S3L ).
The lactylation risk score correlated with tumor microenvironment
cells, with a high score indicating increased M2 macrophage
infiltration and reduced CD8+ T cells ([135] Supplementary Figures
S4D–I ). This indicates that the high risk group likely contributes to
tumor progression by activating cell cycle pathways and inhibiting the
tumor immune microenvironment.
Figure 3.
[136]Twelve-panel image showing data analysis results. A) Plot of
partial likelihood deviance against Log of lambda values. B)
Coefficient plot with lines converging at zero. C) Sankey diagram
mapping clusters to risk levels and survival status. D, E) Box plots
comparing risk scores for Gene and Lact Clusters, with significant
p-values. F) Box plots of gene expression levels separated by risk.
G-I) Kaplan-Meier survival curves showing significant differences
between high and low-risk groups. J-L) ROC curves evaluating model
performance in train, test, and all data sets. M, N) Forest plots of
hazard ratios for various clinical variables with p-values.
[137]Open in a new tab
Development and validation of a lactylation-related gene signature. (A,
B) LASSO regression analysis using the minimal lambda value. (C) Sankey
diagram linking clusters, risk score groups, and BLCA survival status.
(D, E) Risk score differences between clusters. (F) Lactylation related
genes expression differences between risk groups. (G-I) Patients with
high lactylation risk scores showed worse prognosis in the train,
validation and overall groups. (J-L) The ROC curves indicate higher
model effectiveness. (M, N) Univariate and multivariate Cox regression
analyses of risk scores and clinical features in the integrated cohort.
The symbols *, **, and *** represent P < 0.05, P < 0.01, and P < 0.001,
respectively.
DHCR7 as an important prognostic gene
High risk patients exhibit higher stromal scores and immune scores
([138] Figure 4A ). Given that the high-score group showed increased
immune cell infiltration, we next analyzed the correlation between the
genes in the constructed model and various immune cell types ([139]
Figure 4B ). Each gene in the model demonstrates a strong correlation
with various immune cell types. Recognizing the poorer prognosis of
high-risk patients, we further analyzed the relationship between each
gene in the model and patient outcomes. A univariate Cox regression
forest plot illustrates the prognostic relevance of all genes in the
model ([140] Figure 4C ). For instance, DHCR7, P4HB, CD109, and FADS1
are identified as risk factor genes, whereas TMC7 and ADIRF serve as
protective genes. All these genes exhibit p-values below 0.05. In the
TCGA cohort and [141]GSE13507 cohort, DHCR7 and P4HB consistently
exhibited high expression in tumor tissues ([142] Figures 4D, E ). In
the [143]GSE19423 cohort, only high DHCR7 expression was linked to
poorer prognosis ([144] Figures 4F, G ). In the [145]GSE48276 and
[146]GSE48075 cohorts, patients with high DHCR7 expression exhibited
poorer prognoses ([147] Figures 4H, I ). Building on these findings,
the analysis will now focus on DHCR7.
Figure 4.
[148]Multiple panels display a variety of data visualizations related
to gene expression and survival analysis. Panel A shows violin plots
comparing StromalScore, ImmuneScore, and ESTIMATEScore with different
risk levels. Panel B is a heatmap correlating immune cell types with
gene expressions. Panel C presents a forest plot of genes with their
hazard ratios, confidence intervals, and p-values. Panels D and E are
box plots showing gene expression levels in normal and tumor samples.
Panels F, G, H, and I contain Kaplan-Meier plots showing survival
probabilities associated with various genes and risk categories.
[149]Open in a new tab
DHCR7 as an important prognostic gene. (A) Tumor microenvironment score
differences between high and low risk lactylation groups. Green for
high risk group, blue for low risk group. (B) Correlation analysis
between model genes and various immune cells. Different colors
represent different correlation coefficients. (C) Univariate Cox
regression analysis of genes in the model. (D, E) Expression
differences of model genes between normal and tumor tissues [(D)
[150]GSE13507 cohort, (E) TCGA cohort]. (F, G) In the [151]GSE19423
cohort, only DHCR7 showed a significant association with prognosis. (H,
I) In the [152]GSE48276 and [153]GSE48075 cohort, K-M survival curves
comparing DHCR7 High and Low expression groups. The symbols *, **, and
*** represent P < 0.05, P < 0.01, and P < 0.001, respectively.
Multi dimensional analysis of DHCR7 in BLCA
Samples categorized into high and low expression groups based on the
median DHCR7 value, with absolute logFC > 0.5, were analyzed for
enrichment using Proteomaps. Up-regulated genes show significant
enrichment in lipid metabolism, transcription factors, cell cycle,
ubiquitination, and chromosome-associated pathways. In contrast,
down-regulated genes exhibit significant enrichment in the immune
system, signaling molecules, and signal transduction pathways ([154]
Figure 5A ). However, using the WGCNA method, a correlation analysis of
the DHCR7 gene was conducted. The yellow gene module was selected for
GO and KEGG enrichment analyses, revealing significant associations
between DHCR7 and the P53-related pathway, DNA repair, fatty acid
synthesis, and glutathione metabolism. These findings suggest that
DHCR7 may contribute to tumor progression through these pathways ([155]
Figures 5B–D ). Supplementary figures illustrate the WGCNA analysis
process ([156] Supplementary Figures S5A–C ). We conducted a waterfall
plot analysis of gene mutation data. The analysis revealed that TP53
mutations ranked first, with patients exhibiting high DHCR7 expression
showing a higher TP53 mutation rate. TTN mutations followed in
frequency ([157] Figures 5E, F ). Boxplot demonstrated significant
differences in HLA family gene expression between the DHCR7 high and
low expression groups ([158] Figure 5G ). Specifically, HLA-F and
HLA-DOB showed significantly higher expression in the DHCR7 low
expression group, indicating potential immune regulatory roles
associated with DHCR7 expression levels. BLCA response to immunologic
and antibody-drug conjugate (ADC) therapeutic drugs. Next, the
correlation between DHCR7 and individual immune checkpoints and ADC
target were examined. Boxplot showed that patients with high DHCR7
expression had elevated levels of PVR, NECTIN4, and TDO2, while BTLA
and CD209 exhibited reduced expression ([159] Figure 5H ). TMB plays a
pivotal role in tumor biology. Analyzing its relationship with DHCR7,
we found that K-M survival analysis stratified patients into four
distinct prognostic groups based on combined DHCR7 expression and TMB
levels ([160] Figure 5I ). Patients with high DHCR7 expression and low
TMB demonstrated the poorest survival outcomes, while those with low
DHCR7 expression and high TMB had the best outcomes, highlighting the
prognostic significance of integrating DHCR7 expression and TMB in
BLCA. Based on this observation, we hypothesized that DHCR7 could play
a role in BLCA immunotherapy.
Figure 5.
[161]This composite image contains multiple panels. Panel A shows a
colorful tree map categorizing biological functions related to the
immune system. Panel B features two line graphs of scale independence
and mean connectivity versus soft threshold. Panel C displays a heatmap
of module-trait relationships. Panel D presents a bubble chart for GO
and KEGG annotation with varying colors and sizes indicating
significance. Panels E and F depict bar graphs of mutation data in
samples. Panel G shows box plots of gene expression across subtypes.
Panel H is another series of box plots for comparison of gene
expression levels. The final panel displays a Kaplan-Meier survival
curve.
[162]Open in a new tab
Multi dimensional analysis of DHCR7 in cancer. (A) Proteomaps
enrichment of DHCR7 high and low expression gene group. (B) Selection
of soft-thresholding power in WGCNA. (C) Module-trait relationships for
DHCR7 expression groups. (D) GO and KEGG enrichment of yellow module
genes. (E, F) Mutation landscape in DHCR7 high and low expression
groups. (G) Differential expression of HLA-related genes between DHCR7
high- and low-expression groups. (H) Immune checkpoints and ADC targets
show differential expression between DHCR7 expression groups. (I)
Combined survival analysis of DHCR7 expression and tumor mutation
burden (TMB). The symbols, *, **, and *** represent P < 0.05, P < 0.01,
and P < 0.001, respectively.
Single-cell transcriptomics reveals DHCR7 modulation of tumor
microenvironment
The advent of single-cell technology has significantly advanced our
understanding of the tumor immune microenvironment. After applying
quality control and dimensionality reduction, we analyzed 55,951 cells,
which were further classified and annotated into eight distinct
clusters ([163] Figures 6A, B ). The supplementary figure presents the
manually annotated result map for single-cell analysis ([164]
Supplementary Figure S6A ). Differential gene expression across
distinct cell types ([165] Figure 6C ). Notably, the proportions of
these cell clusters differed markedly between normal and tumor tissues,
highlighting the heterogeneity of the tumor immune microenvironment
([166] Figure 6D ). Heatmap showing enrichment analysis results for
intercellular Hallmark gene sets in cancer tissues ([167] Supplementary
Figure S6B ). We observed elevated glycolysis levels in epithelial
cells and fibroblasts within tumor tissues. We applied five algorithms
to analyze the enrichment of lactylation associated genes. Overall,
epithelial cells exhibited the highest lactylation levels ([168]
Figure 6E ). Across tissues, most cells in tumor samples showed
elevated lactylation levels, with the exception of smooth muscle cells,
which displayed lower levels ([169] Figure 6F ). These findings suggest
that lactylation plays a critical role in tumorigenesis and
development. At the bulk level, DHCR7 exhibited high expression in
tumor tissues. Additionally, we also observed that DHCR7 was
significantly overexpressed in tumor epithelial cells ([170] Figure 6G
). The analysis above highlighted the significant role of DHCR7 in
BLCA. To further investigate its interaction with other cells in the
tumor microenvironment, tumor epithelial cells were classified into
DHCR7 positive and DHCR7 negative groups based on DHCR7 expression. The
higher level of lactylation score in the DHCR7 positive expression
group also suggests a potential relationship between DHCR7 expression
and lactylation ([171] Figure 6H ). To further explore the relationship
between DHCR7 epithelial positive cells and other cell types, we
performed cellular communication analysis in tumor samples. The
analysis revealed that DHCR7 positive epithelial cells interacted with
various other cells through distinct pathways, driven by different
transcription factors ([172] Figures 6I–P , [173]Supplementary Figures
S6C, D ). A potential relationship between DHCR7 positive epithelial
cells and other cells was observed at the single-cell level. To further
investigate, K-M survival analyses were performed to assess the
combined prognostic impact of DHCR7 expression and various immune cell
subpopulations, including CD8+ T cells, NK cells, macrophages, and CD4+
T cells. Among these, The DHCR7 low + T cells CD8+ high group
consistently displayed the best prognosis, highlighting a synergistic
effect of reduced DHCR7 expression and elevated CD8+ T cell
infiltration on survival. Conversely, the DHCR7 high + T cells CD8+ low
group showed the worst outcomes, emphasizing the pivotal influence of
DHCR7 expression and immune cell composition on patient prognosis.
These results highlight the potential interplay between DHCR7
expression and immune cell mediated tumor microenvironment in shaping
patient outcomes ([174] Supplementary Figures S6E–O ).
Figure 6.
[175]Scientific data visualization panel showing various graphs and
charts related to cell types and gene expression. The panel includes
UMAP plots for cell clusters, a bar chart for cell type proportions,
bubble plots depicting expression levels, a box plot comparing
lethality scores, and several flowcharts representing signaling
pathways and interactions between various genes and cell types. Each
element highlights different aspects of cellular and molecular data,
with color codes and annotations for clarity.
[176]Open in a new tab
Single-cell transcriptomics reveals DHCR7 modulation of tumor
microenvironment. (A, B) UMAP plot showing the distribution of 8 main
cell types in the integrated dataset. (C) Top 5 genes for each cell
type. (D) Cellular composition ratio chart. (E, F) Analysis of
lactylation enrichment score differences. (G) Dot plot of DHCR7
expression across cells and tissue types. (H) Lactylation score
differences between DHCR7 groups. (I, J) Cellular communication between
epithelial cells and other cells in the DHCR7-positive or negative
group. (K-P) Cellular communication in the DHCR7 positive group:
DHCR7-positive epithelial cells on the left, other cells in the center,
and transcription factors on the right. [(K) Macrophages Cells, (L)
Monocytes Cells, (M) T Cells, (N) DHCR7 positive epithelial, (O) Smooth
Muscle Cells, (P) Plasma Cells] The symbols ns, *, and *** represent
not significant, P < 0.05, and P < 0.001, respectively.
DHCR7 inhibition augments response to anti-PD1 therapy
Immune checkpoint inhibitors have extended survival in advanced BLCA
patients, yet their efficacy remains constrained, with many patients
developing resistance to immunotherapy ([177]43). This underscores the
urgent need to identify the underlying causes of immunotherapy
resistance. Based on the following findings: Patients with high
lactylation risk scores exhibited poorer responses to immunotherapy. In
TCGA data, genes in the low DHCR7 expression group showed activation of
the immune system in Proteomaps. Patients with low DHCR7 expression and
high TMB demonstrated better prognoses. At the single-cell level,
interactions between DHCR7-positive epithelial cells and T cells were
observed. Additionally, TCGA data revealed that patients with low DHCR7
expression and high CD8+ T cells infiltration had improved survival
outcomes. DHCR7 is not only related to lactoylation, but also
participates in cholesterol synthesis. Previous studies have linked
elevated cholesterol levels to increased PD-L1 expression, facilitating
immune escape ([178]44). Another recent study revealed that DHCR7
significantly influences the tumor microenvironment, with high
cholesterol levels contributing to CD8+ T cell exhaustion ([179]45).
Based on the above result, we hypothesized that DHCR7 might influence
BLCA immunotherapy and that its knockdown could enhance sensitivity to
anti-PD-1 treatment. To test this, we conducted in vivo experiments and
confirmed that Dhcr7 knockdown significantly suppressed tumor growth.
When combined with anti-PD-1 treatment, tumor suppression was further
enhanced ([180] Figures 7A–D ). Immunofluorescence analysis showed that
Dhcr7 knockdown markedly elevated CD3 expression and the combination
with anti-PD-1 treatment amplified this effect ([181] Figures 7E, F ).
These findings suggest that DHCR7 influences tumor immunotherapy
sensitivity by modulating the tumor immune microenvironment.
Figure 7.
[182]Schematic diagram and data on tumor progression and immune
response in a study involving MB49 cells and immunotherapy. (A)
Timeline of antibody administration and tumor volume measurement. (B)
Tumor samples from different treatment groups. (C) Graph showing tumor
volume over time with various treatments. (D) Bar graph comparing tumor
mass across treatments. (E) Immunofluorescence images showing DAPI and
CD3 staining. (F) Bar graph illustrating CD3 expression levels.
Treatments include shControl+IgG, shControl+αPD-1, shDhcr7+IgG, and
shDhcr7+αPD-1, with significant differences noted.
[183]Open in a new tab
DHCR7 inhibition enhances sensitivity to anti-PD-1 therapy. (A-D) MB49
cells were infected with lentivirus vectors expressing control or Dhcr7
shRNAs. These cells were infected with shControl or shDhcr7 and
subcutaneously injected into the right dorsal flank of C57BL/6 mice.
Mice with subcutaneous MB49 tumors (n = 6/group) were treated with
anti‐PD‐1 (200 µg) or nonspecific IgG three times as shown in the
schematic diagram (A). The image of the tumor is shown in panel (B).
The tumor growth curve was demonstrated in panel (C). Tumor volumes are
shown in panel (D). (E, F) At the end of treatment, the tumors excised
from the mice were dissociated and tumor cells were harvested for
immunofluorescence staining. The symbols **, and *** represent P <
0.01, and P < 0.001, respectively.
Role of DHCR7 in cisplatin sensitivity
We analyzed the differences between DHCR7 positive and DHCR7 negative
epithelial cells in tumor tissues and identified highly expressed genes
for GO and KEGG enrichment analyses. Interestingly, the analysis
revealed enrichment in pathways related to cisplatin resistance. This
finding led us to hypothesize that DHCR7 may contribute to cisplatin
resistance in BLCA ([184] Figure 8A ). Using the oncoPredict R package,
we further analyzed DHCR7 and found that patients with high DHCR7
expression had higher IC50 values ([185] Figure 8B ), indicating
reduced sensitivity to cisplatin. In BLCA data ([186]GSE165767), we
also observed the involvement of DHCR7 in cisplatin resistance ([187]
Figure 8C ). Additionally, cisplatin showed greater sensitivity
compared to carboplatin in one dataset, further complicating the
relationship between DHCR7 expression and drug response ([188]
Figure 8D ). To investigate whether this phenomenon extends beyond
BLCA, we analyzed other tumors and found that DHCR7 was highly
expressed in cisplatin-resistant groups across various cancers ([189]
Figures 8E–I ). These findings strongly suggest that DHCR7 is likely
involved in cisplatin resistance in BLCA and potentially in other
tumors. Next, we aim to experimentally validate the role of DHCR7 in
cisplatin resistance in BLCA and determine whether its involvement in
resistance is associated with lactylation.
Figure 8.
[190]Panel A displays a bubble plot of GO and KEGG annotation with
differently colored and sized bubbles representing various biological
processes and pathways. Panels B and D show box plots comparing DHCR7
expression with cisplatin sensitivity, indicating significant
differences. Panels C, E, F, G, H, and I feature volcano plots from
different studies emphasizing the differential expression of DHCR7,
with distinct color gradients representing statistical significance
levels.
[191]Open in a new tab
Role of DHCR7 in cisplatin sensitivity. (A) GO and KEGG enrichment of
differential genes between epithelial DHCR7 groups at the single-cell
level in BLCA. (B) Cisplatin IC50 values across DHCR7 groups in BLCA.
(C) Volcano plot of cisplatin resistance in BLCA ([192]GSE165767). (D)
DHCR7 expression changes across chemotherapeutic agents in BLCA
([193]GSE235066). (E-I) DHCR7 involvement in cisplatin-resistant
volcano plots of other cancers. [(E) Ovarian Cancer, (F) Breast Cancer,
(G) Ovarian Cancer, (H) Ovarian Cancer, (I) Breast Cancer].
Elevated lactylation levels reduce sensitivity to cisplatin therapy in BLCA
Our results demonstrated that lactate treatment of BLCA cells increased
the median inhibitory concentration (IC50) and reduced their
sensitivity to cisplatin ([194] Figure 9A ). Conversely, treatment with
Oxamate resulted in a reduction in IC50 ([195] Supplementary Figure S7A
). Indicating a role for lactate in modulating cisplatin resistance.
Colony formation and CCK-8 assays demonstrated that lactate promotes
tumor progression and attenuated the sensitivity of BLCA cells to
cisplatin ([196] Figures 9B, C ). Oxamate treatment led to decreased
cell proliferation ([197] Supplementary Figure S7B ). As expected,
lactate reduces the pro-apoptotic effect of cisplatin on BLCA cells
([198] Figures 9D, E ), whereas oxamate enhanced this effect ([199]
Supplementary Figures S7C, D ). Then, how does lactete regulate DHCR7
and affect the sensitivity to cisplatin treatment? Our results showed
that lactate upregulated DHCR7 expression at both protein and mRNA
levels ([200] Figures 9F, G ). Conversely, oxamate inhibited the
expression of DHCR7 ([201] Supplementary Figures S7E, F ). The IC50
curve showed that DHCR7 knockdown significantly increased the
sensitivity of BLCA cells to cisplatin and reversed the effect of
lactate on cisplatin sensitivity ([202] Figure 9H ). DHCR7 knockdown
increased cisplatin sensitivity in cell proliferation ([203]
Figures 9I, J ). Similarly, down-regulation of DHCR7 promotes apoptosis
and enhances the sensitivity of BLCA cells to cisplatin ([204]
Figures 9K, L ). In conclusion, lactylation-associated DHCR7 reduces
cisplatin sensitivity in patients with BLCA.
Figure 9.
[205]The image contains multiple panels depicting experimental data on
cell proliferation, caspase-3 activity, colony formation, protein
expression, and mRNA levels. Panels A and H show dose-response curves
for different treatments in T24 and J82 cells. Panel B and I illustrate
cell proliferation over time with various treatments. Panel C and J
present images and graphs of colony formation assays. Panel D and L
depict caspase-3 activity in treated cells. Panels E, F, and K show
Western blots for protein expression. Panel G presents bar graphs of
mRNA levels. Statistical significance is indicated with asterisks.
[206]Open in a new tab
Elevated lactylation levels reduce sensitivity to cisplatin therapy in
BLCA. (A) T24 and J82 cells were treated with lactate (10 mM) for 24
hours, followed by cisplatin at varying doses for another 24 hours.
CCK-8 assays were performed to measure cisplatin IC50 values. (B) T24
and J82 cells were treated with lactate (10 mM) for 24 hours, followed
by collection for the CCK-8 assay. Data are presented as the mean ± SD
from three independent experiments. (C) T24 cells were treated with or
without lactate (10 mM) for 24 hours, followed by treatment with or
without cisplatin (1.0 nM) for an additional 24 hours. The cells were
then collected for a colony formation assay. Data are presented as the
mean ± SD from three independent experiments. (D, E) T24 and J82 cells
were treated with lactate for 24 hours, followed by treatment with or
without cisplatin for an additional 24 hours. The cells were then
collected for caspase-3 activity assay (D) and Western blot analysis
(E). (F, G) T24 and J82 cells were treated with varying doses of
lactate for 24 hours. The cells were then collected for Western blot
analysis (F) and qPCR analysis (G) to measure DHCR7 expression levels.
(H) T24 and J82 cells were transfected with the specified shRNA for 72
hours. After puromycin selection, the cells were treated with varying
doses of cisplatin for 24 hours and then subjected to CCK-8 assays to
determine cisplatin IC50 values. (I) T24 and J82 cells were transfected
with the specified shRNA for 72 hours. After puromycin selection, the
cells were treated with lactate for an additional 48 hours, followed by
treatment with or without cisplatin. The cells were then subjected to
CCK-8 assays. (J) T24 cells were transfected with the specified shRNA
for 72 hours. After puromycin selection, the cells were treated with or
without cisplatin and lactate, and then subjected to colony formation
or CCK-8 assays. Data are presented as the mean ± SD from three
independent experiments. (K, L) T24 cells were transfected with the
specified shRNA for 72 hours. After puromycin selection, the cells were
treated with or without cisplatin and lactate. The cells were then
collected for Western blot analysis (K) and caspase-3 activity assay
(L). Data are presented as the mean ± SD from three independent
experiments. The symbols *, **, and *** represent P < 0.05, P < 0.01,
and P < 0.001, respectively.
Discussion
Significant progress has been achieved in the diagnosis and treatment
of BLCA. However, the survival rate remains unsatisfactory, hindered by
challenges such as metastasis, recurrence, and drug resistance. Recent
studies have identified lactylation modification as a critical factor
in cancer progression, drug-resistant and metastasis ([207]17, [208]46,
[209]47). Li et al. revealed the association between histone
lactylation and cisplatin resistance in BLCA through single-cell
transcriptome analysis. Their research showed that cisplatin-resistant
BCa cells exhibit intracellular lactate accumulation and increased
levels of histone H3K18 lactylation (H3K18la). H3K18la activates the
expression of YBX1 and YY1, inducing cisplatin resistance in BCa
([210]48). Similarly, Deng et al. demonstrated that H3K18la regulates
PRKN-mediated mitophagy, promotes M2 macrophage polarization, and
facilitates immune escape in BCa ([211]49). Xie et al. showed that
CircXRN2 inhibits H3K18la-driven tumor progression by activating the
Hippo signaling pathway in human BLCA ([212]50). These findings suggest
that histone lactylation may play an important role in BLCA
progression. Beyond histone lactylation, recent studies have indicated
that numerous non-histone proteins can also undergo lactylation
modification and play critical roles in BLCA progression. A study by
Xing et al. confirmed that increased lactylation levels of YTHDC1
suppress the sensitivity of BLCA to enfortumab vedotin treatment
([213]51). Jin et al. revealed that mannose inhibits PKM2 lactylation,
induces pyroptosis in BLCA, and activates anti-tumor immune responses
([214]52). Given the pivotal role of lactylation in BLCA development,
this study aims to investigate the biological significance of
lactylation-related genes in BLCA, identify potential therapeutic
targets, and provide novel insights for BLCA treatment.
Our results showed that the lactylation risk score characterization
model performed well in the BLCA cohort and three independent external
GEO validation sets and showed similar effectiveness in other tumors
([215]53, [216]54). Cluster A patients exhibited higher risk scores,
poorer prognoses, and distinct differences in the immune
microenvironment. The prognostic utility of risk score was further
validated in an immunotherapy cohort. Analysis of single-cell data
revealed that lactylation related genes play a significant role in the
tumor microenvironment. As previously reported, these genes may
contribute to tumor progression and immunotherapeutic response by
regulating the tumor immune microenvironment ([217]55). Our study
revealed that lactylation risk score associated DHCR7 knockdown
enhances the response to BLCA immunotherapy, aligning closely with
findings from a recent study in Glioblastoma multiforme ([218]45). The
primary function of DHCR7 involves regulating genes associated with
cholesterol synthesis. At the single-cell level, DHCR7 showed high
expression in macrophages, T cells, and endothelial cells.
Additionally, DHCR7-positive epithelial cells interacted with various
other cells through distinct pathways. Dong et al. found that the
elevated cholesterol level of macrophages in glioma can promote the
growth of tumor cells and inhibit the anti-tumor effect of CD8+ T cells
([219]45). In our study, prognostic analysis combining DHCR7 with
different immune cells revealed that DHCR7 might play a role in
modulating the tumor immune microenvironment. Notably, patients with
low DHCR7 expression and high CD8-positive T cell infiltration
exhibited the best prognosis, while those with low DHCR7 expression and
low M2 macrophage infiltration also had favorable outcomes. Another
study indicates that DHCR7 contributes to M2 macrophage polarization in
hepatocellular carcinoma, promoting tumor growth and metastasis
([220]56). These findings suggest that lactylation associated genes may
contribute to tumor progression not only through tumor cells but also
by altering the tumor immune microenvironment via interactions with
other cells.
We identified DHCR7 as significantly overexpressed in tumors and
involved in cisplatin resistance. Zeng et al. indicated that DHCR7 is
involved in the progression of BLCA through the cholesterol pathway
([221]57). Kanmalar et al. confirmed that increased cholesterol
synthesis plays an important role in the sensitivity of cisplatin
treatment for BLCA ([222]58). In addition, sequencing data revealed
that DHCR7 knockdown leads to the downregulation of ENO2, a key
glycolysis gene ([223]59, [224]60). Previous studies have shown that
histone lactylation promoting YTHDF2 expression ([225]13), and the
DHCR7 is regulated by YTHDF2 ([226]57). This suggests that DHCR7 may be
indirectly influenced by lactylation through multiple pathways.
Elevated DHCR7 may, in turn, upregulate ENO2 and ENO2 contributes to
lactate accumulation, forming a vicious cycle that drives BLCA
progression. At the single-cell level, DHCR7 positive epithelial cells
displayed higher lactylation scores compared to DHCR7 negative cells.
In vitro experiments further demonstrated that lactate increased DHCR7
expression, supporting our hypothesis. However, a more detailed
exploration of the underlying mechanisms requires further investigation
in future studies.
Some studies suggest that chemotherapy and immunotherapy yield better
outcomes in BLCA ([227]61–[228]63); however, certain patients still
experience suboptimal responses. Our study offers new perspectives on
BLCA treatment. In patients with markedly elevated DHCR7, DHCR7
inhibitate effectively enhance the efficacy of combined chemotherapy
and immunotherapy. Despite the robustness of our predictive model,
several limitations should be acknowledged. First, this study primarily
relies on retrospective data, necessitating validation through
prospective cohort studies. Second, while we demonstrated that DHCR7
mediates cisplatin resistance in BLCA, the precise molecular mechanisms
require further investigation. Thirdly, while the tumor immune
microenvironment of BLCA remains highly complex, our study has only
provided an initial investigation into the immunotherapeutic role of
DHCR7. Further research is required to elucidate the detailed
mechanisms underlying DHCR7’s involvement in the BLCA immune
microenvironment.
Conclusion
Our study offers valuable insights into the expression patterns and
roles of lactylation related genes in BLCA. Lactylation risk score
emerges as an effective predictive tool, capable of forecasting
prognosis and immunotherapy efficacy, serving as a guide for precision
medicine. Additionally, we found that DHCR7 mediates cisplatin
resistance in BLCA, and its knockdown significantly enhances
immunotherapy efficacy. This study offers a novel therapeutic approach
for BLCA patients with limited response to cisplatin-based combination
immunotherapy.
Acknowledgments