Abstract

Background

   Colon adenocarcinoma (COAD) is a prevalent malignancy worldwide, yet,
   its underlying pathogenesis and genetic characteristics are still
   unclear. Previous studies have suggested that NADH dehydrogenase 1
   alpha subcomplex subunit 4-like 2 (NDUFA4L2) may affect tumor
   progression across various cancers. However, this effect on COAD has
   rarely been reported. Thus, this study investigated NDUFA4L2's
   prognostic and diagnostic relevance and explored its potential
   connection with immune cell infiltration in COAD.

Methods

   To achieve this, RNA sequencing data from Cancer Genome Atlas (TCGA)
   was analyzed to assess NDUFA4L2's prognostic value in COAD, and factors
   relevant to the prognosis of COAD, including NDUFA4L2, were scrutinized
   using Kaplan-Meier analyses as well as univariate and multivariate Cox
   regression. A nomogram model was created to project prognosis based on
   the results of multivariate Cox analysis. Furthermore, gene set
   enrichment analysis (GSEA) was employed to pinpoint key
   NDUFA4L2-related pathways, and single-sample GSEA (ssGSEA) on TCGA data
   was employed to investigate the connections of NDUFA4L2 with cancer
   immune infiltrations.

Results

   Our findings revealed significant associations of high NDUFA4L2
   expression with poor overall survival, progression-free interval, and
   disease-specific survival of COAD patients. GSEA indicated close links
   of NDUFA4L2 with several signaling pathways implicated in
   tumorigenesis, including extracellular matrix receptor interaction, the
   intestinal immune network for immunoglobulin A production, natural
   killer (NK) cell-mediated cytotoxicity, pathways in cancer, cell
   adhesion molecules, mitogen-activated protein kinase signaling pathway,
   Hedgehog signaling pathway, transforming growth factor beta signaling
   pathway, and chemokine signaling pathway. Additionally, ssGSEA
   identified a positive link between increased NDUFA4L2 expression and
   higher infiltration degree of various immune cells, such as immature
   dendritic cells, macrophages, NK cells and dendritic cells.

Conclusions

   Collectively, our findings demonstrate the association of increased
   NDUFA4L2 expression with adverse prognosis and heightened immune cell
   infiltration in COAD patients.

   Keywords: COAD, NDUFA4L2, Tumor immune infiltration, Prognosis

Highlights

     * •
       NDUFA4L2 expression is elevated in COAD tumor tissues.
     * •
       NDUFA4L2 expression is significantly augmented in SW620 and HT-29
       cell lines relative to the normal CCD-18o cell line.
     * •
       High NDUFA4L2 expression is connected with unfavorable prognosis
       and augmented immune cell infiltration in COAD patients

1. Introduction

   Colon adenocarcinoma (COAD) ranks among the primary causes of
   cancer-related death [[37]1]. The current standard therapeutic
   strategies for COAD encompass surgical resection, chemotherapy, and
   radiotherapy [[38]2]. Despite the effectiveness of targeted drugs and
   immunotherapies, the 5-year overall survival (OS) for COAD is
   persistently low, primarily due to late-stage diagnosis, aggressive
   progression and frequent metastasis [[39][3], [40][4], [41][5]].
   Therefore, there is a compelling demand to inspect the molecular
   mechanisms of COAD and recognize novel biomarkers for survival
   assessment and targeted therapy.

   NDUFA4L2, an increasingly expressed electron transport chain complex I
   subunit under hypoxic conditions has garnered attention. It can inhibit
   the activity of complex I, leading to reduced mitochondria oxygen
   species (ROS) levels under a hypoxic environment [[42]6]. However, the
   functions and roles of NDUFA4L2 in cancer development information are
   not fully understood. Studies have shown that NDUFA4L2 knockdown cells
   can survive under hypoxic conditions while exhibiting increased ROS
   accumulation in mitochondria, suggesting its ability to suppress ROS
   production and mitigate, oxidative stress in cancer cells [[43]7].
   Additionally, NDUFA4L2 expression has been associated with Warburg-like
   shift from mitochondrial respiration towards glycolysis and increased
   lysosomal diameter and decreased lysosomes in RCC4 cells, indicating
   that NDUFA4L2 regulates mitochondrial and lysosomal activities in clear
   renal cell carcinoma (ccRCC) cells. Furthermore, elevated NDUFA4L2
   transcription and translation have been observed in colorectal
   carcinoma (CRC) [[44]8]. However, NDUFA4L2 expression and its potential
   prognostic relevance in COAD have not been extensively elaborated.

   Therefore, this study utilized COAD datasets from Human Protein Atlas
   (HPA), Cancer Genome Atlas (TCGA), and Gene Expression Omnibus (GEO) to
   explore the correlation between NDUFA4L2 levels and prognosis, as well
   as clinicopathological features in COAD patients. Additionally, this
   study also investigated the potential functions of NDUFA4L2, developed
   a nomogram to predict patients’ prognosis, and assessed the impact of
   NDUFA4L2 on immune infiltration in COAD. Our findings revealed the
   significant role of NDUFA4L2 in COAD, highlighting the underlying
   mechanisms behind its increased expression and potential correlation
   with tumor-immune infiltrations.

2. Methods

2.1. Data sourcing and manipulation

   Clinical information and level 3 HTSeq-FPKM data 41 normal and 480 COAD
   samples were retrieved from TCGA ([45]https://portal.gdc.cancer.gov/).
   After excluding those without clinical information, FPKM data were
   transfigured to the transcripts per million reads (TPM) to ensure
   consistency and subsequently analyze following the TCGA publication
   guidelines. Additionally, [46]GSE37364 and [47]GSE33133 datasets from
   the GEO database ([48]https://www.ncbi.nlm.nih.gov/geo/) were utilized
   to certify NDUFA4L2 expression.

2.2. NDUFA4L2 over-expression in COAD patients

   NDUFA4L2 expression in non-paired and paired tissues of COAD patients
   was compared using immunohistochemical (IHC) staining with HPA078265
   antibody based on HPA ([49]https://www.proteinatlas.org) datasets.
   Additionally, the diagnostic utility of NDUFA4L2 was assessed by
   employing the R package (1.18.0) through the receiver operating
   characteristic (ROC) curve analysis.

2.3. Differentially expressed gene (DEG) analysis

   HTseq-Counts were categorized into low- and high-NDUFA4L2 groups, with
   the cutoff being the median NDUFA4L2 level. DEGs in COAD were
   identified in accordance with the adjusted P < 0.05 and |log[2]-fold
   change (FC)| > 1.5 and analyzed by employing DESeq2 package in R
   version 1.36.0 [[50]9].

2.4. Gene set enrichment analysis (GSEA)

   GSEA was implemented to assess the functional and pathway disparities
   between two groups exhibiting varying NDUFA4L2 levels using R package
   clusterProfiler (version 4.4.4) [[51]10,[52]11] by referencing to
   c5.all.v2022.1.Hs.symbols.gmt (Gene ontology) and
   c2.cp.kegg.v7.5.1.symbols.gmt (KEGG Pathway Database) from the MSigDB
   Collections ([53]https://www.gseamsigdb.org/gsea/msigdb/index.jsp),
   with 1000 permutations set for each analysis. Gene sets were considered
   substantially enriched if they had a false discovery rate (FDR) below
   0.25 and an adjusted P-value below 0.05.

2.5. Nomogram construction and evaluation

   The rms package in R version 6.3–0 was utilized to construct a
   survival-predicting nomogram at 1-, 3-, and 5-year intervals,
   incorporating all independent prognostic factors identified via
   multivariate analysis. Their performance was ascertained using the
   calibration curve and consistency index (C-index).

2.6. Immune infiltration analysis

   To explore the link between NDUFA4L2 and marker genes of 24 distinct
   immune cell types, an immune infiltration study was executed using the
   ssGSEA [[54]12] approach in the GSVA package of R version 1.46.0
   [[55]13]. Immune cell infiltration in the low- and high-NDUFA4L2 groups
   was examined using the Wilcoxon rank sum test, its connection with
   NDUFA4L2 level was assessed using Spearman correlation.

2.7. Cell culture

   CCD-18Co, SW620 and HT-29 cell lines were from Otwo Biotech Inc.
   (Shenzhen, China) and cultured in McCoy's 5a (Procell) and DMEM
   (Biosharp) with 10 % fetal bovine serum (Sijiqing) and 1 %
   penicillin/streptomycin (Sangon Bioteck), respectively.

2.8. Real-time quantitative PCR (RT-qPCR) analysis

   Total RNA was isolated using RNAeasy™ Animal RNA Isolation Kit
   (Beyotime Biotechnology) and converted to cDNA using BeyoRT™ III First
   Strand cDNA Synthesis Kit (Beyotime Biotechnology). qPCR was executed
   using BeyoFast™ SYBR Green qPCR Mix (2X) (Beyotime Biotechnology) on a
   QuantStudio apparatus (Applied Biosystems) with the following
   amplification program: an initial denaturation step 95°C for 2 min,
   followed by 40 cycles of 15s at 95 °C for denature and 30s at 60°C for
   annealing/extension. Primers 5′-ATGATCGGCTTAATCTGCCTG-3′ and
   5′-TCCGGGTTGTTCTTTCTGTCC-3′ were used for NDUFA4L2, and primers
   5′-GGAGCGAGATCCCTCCAAAAT-3′ and 5′-GGCTGTTGTCATACTTCTCATGG -3′ were for
   GAPDH.

2.9. Statistical analysis

   Data were visualized using the ggplot 2 package in R version 3.3.6.
   Differences in NDUFA4L2 expression levels between COAD patients and
   normal controls were compared using the Wilcoxon rank sum test and
   Wilcoxon signed rank test. The association of clinicopathological
   characteristics with NDUFA4L2 levels was scrutinized using the
   Kruskal-Wallis test, followed by Dunn's post hoc test and logistic
   regression. The impact of NDUFA4L2 expression on COAD prognosis was
   determined using Kaplan-Meier method and validated using the OSppc web
   server [[56]14]. NDUFA4L2's prediction ability for 5-, 3-, and 1-year
   OS was predicted using the time-dependent analysis of the ROC curve,
   and the impacts of NDUFA4L2 level and clinicopathological
   characteristics on OS were scrutinized using the univariate and
   multivariate Cox regression analyzes. Statistical significance in
   RT-qPCR was evaluated using one-way analysis of variance (ANOVA).
   Statistical analyses were fulfilled using R (4.2.1), and a significance
   threshold of P < 0.05 (two-tailed) was applied to determine
   significance.

3. Results

3.1. Clinical characteristics of the included COAD patients

   [57]Fig. 1 shows the study's flow chart. The clinical information of
   478 COAD patients was collected, including residual tumor, gender, N
   stage, M stage, T stage, carcinoembryonic antigen level, history of
   colon polyps, pathologic stage, age, disease-specific survival (DSS),
   OS, colon polyps present, and progression-free interval (PFI)
   ([58]Table 1). The correlation analysis revealed that NDUFA4L2
   exhibited a trending link with OS (P = 0.035), DSS (P = 0.011), T stage
   (P = 0.030), N stage (P = 0.008), pathologic stage (P = 0.049), and PFI
   (P = 0.004), as determined by the chi-square test. However, NDUFA4L2
   expression was not correlated with other clinicopathologic
   characteristics.

Fig. 1.

   [59]Fig. 1
   [60]Open in a new tab

   The flow chart of this study.

Table 1.

   Clinicopathological features of patients with COAD.
   Characteristics Low expression of NDUFA4L2 High expression of NDUFA4L2
   P value
   n 239 239
   Gender, n (%) 0.067
   Female 103 (21.5 %) 123 (25.7 %)
   Male 136 (28.5 %) 116 (24.3 %)
   T stage, n (%) 0.030
   T1 7 (1.5 %) 4 (0.8 %)
   T2 49 (10.3 %) 34 (7.1 %)
   T3 161 (33.8 %) 162 (34 %)
   T4 21 (4.4 %) 39 (8.2 %)
   N stage, n (%) 0.008
   N0 151 (31.6 %) 133 (27.8 %)
   N1 58 (12.1 %) 50 (10.5 %)
   N2 30 (6.3 %) 56 (11.7 %)
   M stage, n (%) 0.133
   M0 178 (42.9 %) 171 (41.2 %)
   M1 27 (6.5 %) 39 (9.4 %)
   Pathologic stage, n (%) 0.049
   Stage I 51 (10.9 %) 30 (6.4 %)
   Stage II 92 (19.7 %) 95 (20.3 %)
   Stage III 64 (13.7 %) 69 (14.8 %)
   Stage IV 27 (5.8 %) 39 (8.4 %)
   Residual tumor, n (%) 0.138
   R0 178 (47.6 %) 168 (44.9 %)
   R1 1 (0.3 %) 3 (0.8 %)
   R2 8 (2.1 %) 16 (4.3 %)
   CEA level, n (%) 0.882
   ≤ 5 97 (32 %) 99 (32.7 %)
   >5 52 (17.2 %) 55 (18.2 %)
   History of colon polyps, n (%) 0.918
   No 135 (33.1 %) 127 (31.1 %)
   Yes 76 (18.6 %) 70 (17.2 %)
   Colon polyps present, n (%) 0.080
   No 78 (31.3 %) 84 (33.7 %)
   Yes 52 (20.9 %) 35 (14.1 %)
   OS event, n (%) 0.035
   Alive 197 (41.2 %) 178 (37.2 %)
   Dead 42 (8.8 %) 61 (12.8 %)
   DSS event, n (%) 0.011
   No 205 (44.4 %) 193 (41.8 %)
   Yes 22 (4.8 %) 42 (9.1 %)
   PFI event, n (%) 0.004
   No 189 (39.5 %) 161 (33.7 %)
   Yes 50 (10.5 %) 78 (16.3 %)
   Age, median (IQR) 69 (60, 75.5) 69 (57, 78) 0.935
   [61]Open in a new tab

3.2. NDUFA4L2 is upregulated in COAD

   NDUFA4L2 exhibited higher levels in 480 COAD samples relative to 41
   normal samples (P < 0.001) ([62]Fig. 2A). NDUFA4L2 level was
   dramatically augmented in tumor tissues among the 41 paired samples
   (P < 0.001) ([63]Fig. 2B). Consistent results were noticed in
   additional analyses using [64]GSE37364 ([65]Fig. 2D) and [66]GSE33133
   datasets ([67]Fig. 1E), supporting the validity of the TCGA data. To
   assess NDUFA4L2 protein levels, IHC staining results obtained from the
   HPA were analyzed. These results demonstrated the absence of NDUFA4L2
   IHC staining in normal colon tissues but revealed medium intensity in
   COAD tissues ([68]Fig. 2C). Furthermore, NDUFA4L2's diagnostic
   potential in COAD was evaluated using ROC curve analysis, indicating a
   reasonably accurate predictive capability (area under the ROC curve
   [AUC] = 0.876, CI = 0.836–0.916) ([69]Fig. 2F). Subsequently, 286 DEGs
   were uncovered in high- and low- NDUFA4L2 groups, based on its median
   level, using adjusted P < 0.05 and |log[2]-fold change (FC)| > 1.5 as
   the cutoff, including 74 upregulated and 212 downregulated. The
   distribution of DEGs was visualized in the volcano plot shown in
   [70]Fig. 2G. The correlation between NDUFA4L2 and the top 10 DEGs was
   illustrated in a heatmap ([71]Fig. 2H). Pearson correlation analysis
   demonstrated a positive correlation of NDUFA4L2 level with HIF-1α level
   ([72]Fig. 2I).

Fig. 2.

   [73]Fig. 2
   [74]Open in a new tab

   Differential expression levels of NDUFA4L2 in COAD from the TCGA data.
   (A) Expression levels of NDUFA4L2 in COAD and normal samples. (B)
   Expression levels of NDUFA4L2 in paired tumor and normal samples. (C)
   The protein expression of NDUFA4L2 in COAD and normal colon tissues.
   (D) Expression levels of NDUFA4L2 in [75]GSE37364 (E) Expression levels
   of NDUFA4L2 in GSE 33133 (F) Diagnostic ROC curve of NDUFA4L2. (G) The
   volcano plot of differentially expressed mRNAs. (H) The heat map of the
   10 genes correlated to NDUFA4L2. (I) Positive correlation trend between
   NDUFA4L2 and HIF-1α mRNA expression in the TCGA data. *P < 0.05,
   **P < 0.01, ***P < 0.001.

3.3. Association between NDUFA4L2 expression and clinical features

   The connection between clinicopathological characteristics and NDUFA4L2
   expression was revealed using the Kruskal-Wallis test, followed by
   Dunn's t post hoc test and logistic regression, which demonstrated
   dramatic correlations of NDUFA4L2 expression with T stage, N stage,
   pathologic stage ([76]Fig. 3A–C), OS, DSS, and PFI events ([77]Fig.
   3D–F). Logistic regression analysis found dramatic correlations of
   NDUFA4L2 expression with the T stage (OR = 1.628, 95 % CI: 1.033–2.588,
   P = 0.037) and N stage (OR = 2.006, 95 % CI: 1.245–3.280, P = 0.005)
   ([78]Table 2).

Fig. 3.

   [79]Fig. 3
   [80]Open in a new tab

   Association between NDUFA4L2 expression and clinicopathological
   characteristics, including (A) T stage, (B) N stage, (C) pathologic
   stage, (D) OS event, (E) DSS event, and (F) PFI event. *P < 0.05,
   **P < 0.01, ***P < 0.001.

Table 2.

   Univariate and multivariate Cox regression analyses for OS.
   Characteristics Total(N) Univariate analysis
     __________________________________________________________________

   Multivariate analysis
     __________________________________________________________________

   Hazard ratio (95 % CI) P value Hazard ratio (95 % CI) P value
   Gender 477 0.626
   Female 226 Reference
   Male 251 1.101 (0.746–1.625) 0.627
   Age 477 0.024
   ≤ 65 194 Reference Reference
   >65 283 1.610 (1.052–2.463) 0.028 2.179 (1.332–3.564) 0.002
   T stage 476 < 0.001
   T1&T2 94 Reference Reference
   T3&T4 382 3.072 (1.423–6.631) 0.004 3.311 (1.013–10.817) 0.047
   N stage 477 < 0.001
   N0&N1 391 Reference Reference
   N2 86 3.419 (2.286–5.113) < 0.001 1.529 (0.893–2.619) 0.122
   M stage 414 < 0.001
   M0 348 Reference Reference
   M1 66 4.193 (2.683–6.554) < 0.001 2.322 (1.363–3.955) 0.002
   Pathologic stage 466 < 0.001
   Stage I&Stage II 267 Reference Reference
   Stage II &Stage IV 199 2.947 (1.942–4.471) < 0.001 1.685 (0.927–3.064)
   0.087
   NDUFA4L2 477 0.008
   Low 238 Reference Reference
   High 239 1.691 (1.141–2.507) 0.009 1.646 (1.059–2.560) 0.027
   [81]Open in a new tab

3.4. High NDUFA4L2 expression is connected with adverse outcomes in COAD

   Evaluation of NDUFA4L2's prognostic value in COAD patients through
   Kaplan–Meier analysis signaled that augmented NDUFA4L2 level was linked
   to unfavorable OS (HR = 1.69, 95 % CI: 1.14–2.51, P = 0.009).
   Consistent results were discerned for DSS (HR = 2.07, 95 % CI:
   1.23–3.46, P = 0.006) and PFI (HR = 1.90, 95 % CI: 1.33–2.91,
   P < 0.001) ([82]Fig. 4A–C). Additionally, Kaplan–Meier survival plots
   in [83]GSE175336 and [84]GSE29623 datasets also revealed that high
   NDUFA4L2 expression group had poorer OS ([85]Fig. 4D and E). Assessment
   of the accuracy of NDUFA4L2 expression in projecting 5-, 3-, and 1-year
   OS using time-dependent ROC analysis yielded AUC values of 0.561,
   0.612, and 0.659 respectively, illustrating NDUFA4L2 being a reliable
   prognostic indicator for assessing OS of COAD patients ([86]Fig. 4F).

Fig. 4.

   [87]Fig. 4
   [88]Open in a new tab

   Prognostic value of NDUFA4L2 for patients with COAD. (A–C) The
   prognostic value of NDFA4L2 in OS, DSS, and PFI of COAD. (D, E) The
   prognostic value of NDFA4L2 in OS of COAD in [89]GSE17536 and
   [90]GSE29623 (F) Time-dependent ROC curves of the predictive value of
   NDFA4L2 for OS.

3.5. Cox regression analyses and prognostic nomogram

   Cox regression analysis to assess the independent prognostic factors
   for OS revealed that age (P = 0.028), pathologic stage (P < 0.001), T
   stage (P = 0.004), N stage (P < 0.001), M stage (P < 0.001), and
   NDUFA4L2 expression (P = 0.009) were correlated with OS ([91]Table 3).
   Among them, age (P = 0.002), T stage (P = 0.047), NDUFA4L2 expression
   (P = 0.027), and M stage (P = 0.002) were independent prognostic
   factors for OS ([92]Fig. 5A). Additionally, analysis of NDUFA4L2 levels
   in relation to the risk score and survival time revealed a clear
   association of higher risk scores with worse prognosis ([93]Fig. 5B).
   Furthermore, a predictive nomogram was developed to estimate the 1-,
   3-, and 5-year OS according to the multivariate Cox analysis ([94]Fig.
   5C). The survival model C-index was 0.745 (0.716–0.774), and the 5-,
   3-, and 1-year OS calibration curves signified a satisfactory
   prediction effect of the nomogram ([95]Fig. 5D).

Table 3.

   Association between NDUFA4L2 expression and clinical pathological
   characteristics (logistic regression).
   Characteristics Total(N) Odds Ratio (OR) P value
   T stage (T3&T4 vs. T1&T2) 477 1.628 (1.033–2.588) 0.037
   N stage (N2 vs. N0&N1) 478 2.006 (1.245–3.280) 0.005
   M stage (M1 vs. M0) 415 1.397 (0.824–2.393) 0.218
   Pathologic stage (Stage III&Stage IV vs. Stage I&Stage II) 467 1.311
   (0.908–1.895) 0.149
   Residual tumor (R1&R2 vs. R0) 374 2.211 (0.998–5.258) 0.058
   CEA level (>5 vs.≤5) 303 1.015 (0.634–1.628) 0.949
   History of colon polyps (YES vs. NO) 408 0.979 (0.652–1.468) 0.918
   Colon polyps present (YES vs. NO) 249 0.657 (0.385–1.110) 0.118
   [96]Open in a new tab

Fig. 5.

   [97]Fig. 5
   [98]Open in a new tab

   Prognostic prediction model of NDUFA4L2 in COAD. (A) Multivariate Cox
   regression visualization by the forest plot. (B) NDUFA4L2 expression
   distribution and survival status. (C) Nomogram for 1-, 3-, and 5-year
   OS of patients with COAD. (D) Calibration plots for 1-, 3-, and 5-year
   OS prediction.

3.6. Function and pathway enrichment analysis by GSEA

   A GSEA was conducted using the MSigDB collection to identify relevant
   biological functions associated with different NDUFA4L2 expression
   levels in COAD. The top 20 gene ontology (GO) terms that positively
   correlated with high NDUFA4L2 levels and the top 20 GO terms that
   negatively correlated with high levels of NDUFA4L2 based on their
   normalized enrichment scores are presented in [99]Fig. 6A and B,
   respectively. GSEA indicated several key signaling pathways related to
   tumorigenesis were positively connected with high NDUFA4L2 expression
   phenotype, including antigen processing and presentation, chemokine
   signaling, the intestinal immune network for immunoglobulin A (IgA)
   production, NK cell-mediated cytotoxicity, leukocyte transendothelial
   migration, the Hedgehog signaling pathway, cytokine–cytokine receptor
   interaction, cell adhesion molecules (CAMs), extracellular matrix (ECM)
   receptor interaction, pathways in cancer, the mitogen-activated protein
   kinase (MAPK) signaling pathway, and the transforming growth factor
   beta (TGF-β) signaling pathway ([100]Fig. 6C and D). These findings
   suggest that NDUFA4L2 may contribute to COAD progression via various
   cancer-related signaling pathways in COAD.

Fig. 6.

   [101]Fig. 6
   [102]Open in a new tab

   Enrichment analysis of NDUFA4L2 in COAD. (A, B) GSEA of DEGs in low-
   and high- NDUFA4L2-expression samples. (C, D) GSEA of DEGs in low- and
   high- NDUFA4L2-expression samples.

3.7. Connection between NDUFA4L2 level and immune infiltration

   ssGSEA showed a positive connection of NDUFA4L2 level with various
   immune cell populations, including NK cells, dendritic cells (DCs), NK
   CD56dim cells, immature dendritic cells (iDCs), macrophages, Th1 cells,
   cytotoxic cells, neutrophils, mast cells, Treg, T follicular helper
   cells, activated DCs, NK CD56bright cells, eosinophils, T cells,
   plasmacytoid DCs, CD8 T cells, T gamma delta, B cells, and Th2 cells
   while negatively correlated with T central memory cells ([103]Fig. 7A).
   Examination using Spearman correlation showed a weakly (R < 0.5) but
   significantly (P < 0.05) linear correlation of NDUFA4L2 level to the
   infiltration level, which was quantified by the ssGSEA score, of NK
   cells (R = 0.496), iDCs (R = 0.402), DCs (R = 0.398), and macrophages
   (R = 0.398), among 24 immune cells ([104]Fig. 7C). Consistently,
   Wilcoxon rank sum test revealed significantly higher infiltration
   levels of NK cells, iDCs, DCs, and macrophages in high NDUFA4L2 group
   (P < 0.001) ([105]Fig. 7B), indicating that NDUFA4L2 is essential in
   the immune infiltration of COAD. Moreover, the correlations among the
   percentages of 24 tumor-infiltrating immune cell subpopulations were
   evaluated and illustrated in a heat map ([106]Fig. 7D).

Fig. 7.

   [107]Fig. 7
   [108]Open in a new tab

   NDUFA4L2 expression and immune infiltration. (A) Correlation between
   NDFA4L2 expression and the relative abundances of 24 immune cells. (B)
   Comparison of the different immune cells infiltration levels under
   high- and low- NDUFA4L-expression conditions. (C) The positive
   correlation between NDUFA4L2 expression and NK cells, iDCs, DCs and
   macrophages. (D) Heatmap of 24 immune infiltration cells in COAD.
   *P < 0.05, **P < 0.01, ***P < 0.001.

3.8. Validation of the expression of NDUFA4L2 mRNA by RT-qPCR

   As depicted in [109]Fig. 8, using RT-qPCR, NDUFA4L2 level was
   significantly augmented in SW620 and HT-29 cell lines relative to the
   normal CCD-18o cell line (P < 0.001).

Fig. 8.

   [110]Fig. 8
   [111]Open in a new tab

   Validation of the expression of NDUFA4L2 mRNA in CCD-18Co, SW620 and
   HT-29 cell lines by RT-qPCR analysis. *P < 0.05, **P < 0.01,
   ***P < 0.001.

4. Discussion

   Currently, there is limited available information on NDUFA4L2's role in
   diseases. However, some studies have shed light on its involvement in
   certain contexts. For instance, in cardiac stem cell research, it has
   been demonstrated that hypoxia can enhance their survival in a
   serum-free medium by activating the HIF-1α/NDUFA4L2 pathway. This
   activation suppresses ROS production in a mitochondrial
   function-dependent manner. Additionally, NDUFA4L2 has been shown to
   accelerate pulmonary artery smooth muscle cell proliferation by
   modulating HIF-1α and its downstream p38-5-lipoxygenase signaling
   pathways. This, in turn contributes to pulmonary vascular remodeling
   and pulmonary arterial hypertension [[112]15,[113]16]. Although there
   have been reports on NDUFA4L2 expression and function in tumors, the
   available literature is sparse and lacks comprehensive substantiation.
   The relationship of NDUFA4L2 expression with its role in COAD stay
   largely unknown. Therefore, investigated NDUFA4L2's potential role in
   COAD. Analyzing the TCGA dataset uncovered the overexpression of
   NDUFA4L2 in COAD tissues, which exhibited a significant correlation
   with a poorer OS. Consistent with these findings, our analysis of the
   HPA database revealed higher NDUFA4L2 expression in COAD tissues.
   Through further examination of the TCGA dataset, we validated that
   NDUFA4L2 expression acts as an independent prognostic factor for COAD,
   providing support to recent reports and underscoring the potential
   significance of NDUFA4L2 expression in the prognosis of COAD [[114]8].
   Furthermore, our analyses revealed the association between NDUFA4L2
   expression and several critical tumorigenesis-related signaling
   pathways. Further a positive correlation of NDUFA4L2 to infiltration
   levels of iDCs and NK cells was observed. These novel insights into the
   potential involvement of NDUFA4L2 in immune cell infiltration further
   highlight its promising application as a biomarker for cancer therapy
   and prognosis.

   NDUFA4L2 functions as an inhibitory subunit of the mitochondrial
   oxidative respiratory chain complex I, exerting essential control over
   metabolic reprogramming and oxidative stress in malignant tumors
   [[115]7,[116]17]. Research has shown aberrant NDUFA4L2 expression in
   various cancer types, including ccRCC, glioblastoma, hepatocellular
   carcinoma (HCC), and CRC. Additionally, a high NDUFA4L2 expression
   level has been linked to poor prognosis [[117]8,[118][18], [119][19],
   [120][20], [121][21]]. NDUFA4L2 is highly augmented in HER2-positive
   breast cancer with trastuzumab resistance. Augmentation of NDUFA4L2
   level leads to an increase in the Warburg effect and aerobic glycolysis
   while reducing oxygen consumption and suppressing ROS production
   [[122]22]. Furthermore, mitochondrial NDUFA4L2 is regulated by HIF-1α,
   which is involved in the adaption to hypoxia by decreasing ROS
   production, oxygen consumption, and complex I activity in mitochondria
   [[123]7]. In HCC cells, NDUFA4L2 is modulated by HIF-1α, and its
   inactivation results in elevated mitochondrial activity, ROS
   accumulation, oxygen consumption, and ultimately apoptosis [[124]23].
   NDUFA4L2, induced by HIF-1α，improve osteosarcoma cell survival,
   metastasis, and epithelial-mesenchymal transition by inhibiting ROS
   generation. Additionally, as a target of HIF-1α in ccRCC, NDUFA4L2
   downregulation exhibits significant antiproliferative effects
   [[125]6,[126]24]. NDUFA4L2 level is enhanced in both human non-small
   cell lung cancer cell lines and tissues under hypoxic conditions and is
   modulated by HIF-1α and its silencing significantly increases
   mitochondrial ROS production and inhibits NSCLC viability [[127]25].

   Our study revealed a positive connection between HIF-1α and NDUFA4L2
   levels, supporting the notion that HIF-1α induces NDUFA4L2 expression,
   thereby suggesting the involvement of HIF-1α/NDUFA4L2 pathway in COAD
   progression. However, conflicting findings have been reported in other
   studies. For instance, NFIB has been found to induce sorafenib
   resistance in HCC cells by promoting NDUFA4L2 transcription, which
   decreases ROS production. Additionally, enhanced glycolysis promotes
   bladder cancer progression, metastasis, and gemcitabine resistance by
   modulating NDUFA4L2 expression through NXPH4 [[128]20,[129]26].
   Furthermore, Pearson correlation analysis of clinical ccRCC samples
   revealed an reverse correlation between NDUFA4L2 level and HIF-1α
   expression but a positive correlation with ELK1 level, indicating that
   ELK1 may potentially induce NDUFA4L2 expression [[130]18]. Thus,
   investigations are required to further evaluate NDUFA4L2's role in COAD
   progression.

   Due to the lack available data on NDUFA4L2 function, its potential
   involvement in various biological pathways was explored using
   functional annotation. The analysis revealed that NDUFA4L2 might be
   implicated in cytokine-cytokine receptor interaction, antigen
   processing and presentation, the chemokine signaling pathway, NK cell
   mediated-cytotoxicity, the intestinal immune network for IgA
   production, leukocyte transendothelial migration, CAMs, ECM receptor
   interaction, pathways in cancer, the MAPK signaling pathway, the TGF-β
   signaling pathway, and the Hedgehog signaling pathway. Several
   chemokines, including CCL2, CCL3, CCL5, CXCL1, CXCL2, and CXCL8, have
   been linked to the risk and progression of colon cancer [[131]27].
   Inhibiting the intestinal immune network for IgA production signaling
   pathway suppresses HCC proliferation and migration [[132]28]. The cell
   surface glycoproteins CAMs are entangled in cell-cell and cell-ECM
   interactions [[133]29]. In cancer, adhesion to the ECM is essential for
   cell migration, and various CAMs play significant roles in mediating
   these interactions. The ECM serves as the natural niche for cells, and
   CAM-mediated interactions with surface ligands activate key signals
   involved in cell proliferation, differentiation, and dissemination
   [[134]30]. Numerous studies have demonstrated the critical involvement
   of the ECM receptor interaction signaling pathway in CRC development
   and metastasis [[135][31], [136][32], [137][33]]. TGF-β, a secreted
   multifunctional cytokine, activates its type I and type II membrane
   receptors and downstream SMAD [[138]34]. Dysregulated TGF-β signal
   transduction leads to various tumors, including esophageal cancer, CRC,
   gastric cancer, HCC, and pancreatic cancer, [[139]35]. Mutations in
   TGF-β or SMAD can result in an aberrant TGF-β signaling pathway,
   especially CRC metastasis [[140]36]. MAPK signaling pathway impacts
   tumor cell proliferation, apoptosis, and metastasis. Activation of ERK2
   and downregulation of p38 MAPK have been observed in primary tumors,
   promoting CRC metastasis to the liver and lung, respectively [[141]37].
   However, more investigations are necessary to unravel the potential
   regulatory mechanisms of NDUFA4L2 in COAD.

   Our investigation on the links between NDUFA4L2 and immune infiltration
   levels in COAD revealed a strong connection between NDUFA4L2 levels and
   various immune cells, including NK cells, iDCs, DCs, and macrophages.
   Tumors with high levels of NDUFA4L2 showed a significant infiltration
   of these immune cells. Immune cells, stromal cells, cancer cells, and
   ECM within the tumor microenvironment (TME) collectively influence
   immune evasion, tolerance, and tumor growth [[142]38]. Within the tumor
   microenvironment (TME), tumors and tumor-associated cells/factors exert
   control over DC function, enabling tumor cells to evade immune
   recognition and suppression. Tumor-infiltrated DCs frequently exhibit
   reduced or deficient function, leading to immunosuppression
   [[143]39,[144]40]. Infiltrated NK cells in the TME also affect
   immunotherapy, and targeting NK cells can enhance antitumor immune
   responses. It has been demonstrated that infiltrated NK cells exhibit
   impaired function and phenotype leading to their dysfunction or
   exhaustion in the TME [[145][41], [146][42], [147][43]]. DCs, as highly
   specialized antigen-presenting cells, is critical in connecting innate
   and adaptive immune responses [[148]44]. Macrophages, key players
   within the innate immune system, are crucial in promoting tumor
   metastasis and growth by accelerating angiogenesis, immunosuppression,
   and cancer cell proliferation and producing numerous inflammatory
   cytokines, chemokines, and growth factors [[149]45,[150]46]. In
   summary, our results suggest that NDUFA4L2 is crucial role in
   recruiting and regulating immune infiltrating cells in COAD.

   While this work enhanced our understanding of NDUFA4L2 in COAD, it is
   not without limitations. First, additional research with large sample
   size and prospective design is necessary to mitigate selection and
   recall biases and generate more robust results. Second, the mRNA and
   protein expression of NDUFA4L2 should be validated at cellular and
   clinical levels. Lastly, further in vitro and animal tests are required
   to confirm the underlying mechanism through which NDUFA4L2 contributes
   to COAD.

5. Conclusions

   This study revealed that NDUFA4L2 expression is elevated in COAD tumor
   tissues, and high NDUFA4L2 expression is connected with unfavorable
   prognosis and augmented immune cell infiltration in COAD patients
   suggesting that NDUFA4L2 shows promise as a prognostic biomarker and
   target for immunotherapy in COAD. However, further experimental
   validation is needed to elucidate the precise biological effects of
   NDUFA4L2.

Funding

   This work was supported by Education Department of Heilongjiang
   Province (2019-KYYWF-1227).

Data availability statement

   Data included in article/supp. material/referenced in article.

Ethics declarations

   Not applicable.

CRediT authorship contribution statement

   Qingbu Mei: Writing – original draft. Ping Chen: Formal analysis. Ying
   Lv: Formal analysis. Lihong Zheng: Validation. Dan Liu: Validation.
   Minglong Zhang: Resources, Software, Visualization. Wanquan Liu:
   Conceptualization, Supervision, Writing – review & editing. Penghui Li:
   Software, Visualization.

Declaration of competing interest

   The authors declare that they have no known competing financial
   interests or personal relationships that could have appeared to
   influence the work reported in this paper.

Acknowledgement