A nuclear isoform of hydroxyacyl-COA dehydrogenase inhibits tumor progression in colorectal cancer

Summary

Alternate translation initiation expands protein diversity, which is adopted by cancers for progression. However, the majority of these alternative initiation events in cancers remain largely undefined. In this study, we demonstrate that the human hydroxyacyl-CoA dehydrogenase (HADH) gene, catalyzing the third step of the mitochondrial β-oxidation cascade, has two alternative translation start codons. The translation from the downstream start codon produces a short isoform (HADH-S) primarily localized in the nucleus. HADH-S is downregulated in colorectal cancer. Overexpression of HADH-S inhibits colorectal tumorigenesis in vitro and in vivo. Mechanistically, HADH-S antagonizes protein arginine methyltransferase 5 (PRMT5)-mediated histone arginine methylation, thereby facilitating gene transcription. Notably, HADH promotes the transcription of the Wnt suppressor AXIN2. These findings collectively identify a novel nuclear isoform of HADH that exerts inhibitory effects on colorectal tumorigenesis.

Graphical abstract

Highlights

• •Discovery of a novel short isoform of the human HADH gene
• •The short HADH isoform is mainly localized in the nucleus
• •HADH short isoform suppresses colorectal cancer progression
• •HADH short isoform regulates gene transcription by inhibiting PRMT5 activity

Cancer; Biochemistry

Introduction

The canonical initiation of mRNA translation begins at AUG start codons, which are identified by ribosome small subunits scanning the 5′ untranslated region (5′ UTR).¹ Accurate selection of the translation initiation site (TIS) on mRNAs is critical for synthesizing specific proteins. In certain instances, a single mRNA may contain multiple start codons, leading to the initiation of translation through a process termed alternative translation initiation.² In-frame alternative translation initiation, either from upstream TISs (uTISs) or downstream TISs (dTISs), produces extended or truncated protein isoforms, which enhance protein diversity and regulates gene function.^3,4,5,6,7

Cancers frequently adopt alternative translation initiation mechanisms to drive progression. For example, the tumor suppressor gene phosphatase and tensin homolog (PTEN) contains multiple translation initiation sites, giving rise to distinct isoforms with varying localization and function.^8,9,10,11 Technologies that focus on initiation-site sequencing, such as ribosome profiling (Ribo-seq) coupled with initiating ribosome inhibitors such as harringtonine or lactimidomycin, enable the mapping of translation initiation sites on a transcriptomic scale.^3,4,5,12 Data from these approaches reveal that alternative initiation is far more prevalent than initially anticipated, yet most of these alterations in cancers remain poorly characterized.^6,7

HADH belongs to the 3-hydroxyacyl-CoA dehydrogenase family and primarily operates within the mitochondrial matrix, catalyzing the third step in the β-oxidation cascade: the conversion of 3-hydroxyacyl-CoA to 3-ketoacyl-CoA.¹³ Its enzymatic activity is notably highest for medium- and short-chain fatty acids.¹³ Acting as a crucial enzyme in fatty acid metabolism, HADH is of particular importance for organs highly reliant on mitochondrial oxidative phosphorylation, such as the liver, heart, and kidney.¹⁴ Mutations within the HADH gene lead to familial hyper-insulinemic hypoglycemia.¹⁵ Evidence also suggests a role of HADH in cancer development. Dysregulated expression of the HADH gene is closely associated with the prognosis of various human cancers, including gastric cancer and kidney cancer.^16,17 Furthermore, in vitro studies indicate that the downregulation of HADH promotes gastric cancer progression.¹⁸ Nevertheless, the precise function of HADH in the development of colorectal cancer remains largely unexplored.

In our study, we demonstrated that the downstream initiation of HADH mRNA generates a nuclear isoform that effectively impedes colorectal cancer (CRC) tumorigenesis. Furthermore, we established that nuclear HADH suppresses PRMT5 activity, thereby regulating gene transcription.

Results

The short isoform of hydroxyacyl-CoA dehydrogenase is mainly localized in the nucleus

A recent study has identified two alternative translation start codons in the murine Hadh gene.¹⁹ By analyzing translation initiation sequencing data,⁵ we discovered a potential dTIS in human HADH (Figure S1). The alignment of human HADH protein sequence with the mouse HADH protein sequence showed high sequence conservation (Figure 1A). Unlike murine Hadh dTIS, which employs UUG as the initiator, human HADH dTIS utilizes AUG as the initiator.⁵ Translation initiation from the annotated TIS (aTIS) yields a HADH precursor bearing a 12-amino-acid mitochondrial signal peptide. Subsequent cleavage of this signal peptide results in the mature HADH localized in the mitochondria. Notably, the dTIS is situated within the mitochondrial signal peptide and within the same reading frame. Translation from the dTIS generates a short isoform of HADH (HADH-S), exhibiting a truncation of 8 amino acids. This truncation in the mitochondrial signal might influence its localization within the mitochondria (Figure 1B).

Figure 1.

Alternative translation initiation of human HADH mRNA produces a nuclear isoform

(A) Protein alignment comparing HADH from Homo sapiens (human) and Mus musculus (mouse). Conserved amino acids are denoted by asterisks. The canonical ATG start site (orange) and the downstream initial site (green) are highlighted.

(B) Schematic representation of engineered HADH plasmids with a C-terminal FLAG tag.

(C) Immunoblotting analysis of FLAG-tagged HADH, HADH-L, and HADH-S in RKO cells.

(D) Immunofluorescence analysis of the subcellular localization of FLAG-tagged HADH, HADH-L, and HADH-S in RKO cells. Scale bars, 30 μm

(E) Immunoblotting assessment of nuclear, cytosolic, and mitochondrial fractions in RKO cells expressing the plasmids indicated in (B).

To investigate the subcellular localization of HADH, we expressed HADH with an FLAG tag in RKO cells. Immunofluorescence staining revealed that wild-type HADH was localized in both mitochondria and nucleus (Figures 1C and 1D). To differentiate the short isoform (HADH-S) and the long isoform (HADH-L), we introduced mutations into either the aTIS (AUG→AUA, mut1) or the dTIS (AUG→AUA, mut2) (Figure 1B). Mutation of aTIS prevented the expression of HADH-L, while mutation of dTIS hindered HADH-S expression. The data showed that HADH-L is localized predominantly within the mitochondria, whereas HADH-S primarily resided in the nucleus (Figure 1D). Additional cellular fractionation experiments validated the immunofluorescence results (Figure 1E). Collectively, these findings affirm that HADH undergoes translation from the downstream AUG start codon, producing a nuclear isoform of HADH.

The nuclear isoform of hydroxyacyl-CoA dehydrogenase is reduced in colorectal cancers

To confirm the results of the overexpression experiments, we used the HADH antibody to detect endogenous HADH in colorectal cancer cells. Apart from its mitochondrial distribution, a majority of HADH displayed nuclear localization in both RKO and HCT116 cells (Figures 2A and 2B). Clinically, the analysis of the TCGA database revealed decreased HADH expression in higher-grade colorectal cancer tissues (Figure 2C). In addition, higher expression of HADH was associated with better patient survival, as evidenced by Kaplan-Meier survival curves (Figure 2D, n = 68, p = 0.013). To further validate the clinical relevance of HADH in CRC, we conducted an immunohistochemistry analysis of HADH expression in a CRC tissue array. We observed nuclear staining of HADH in epithelial cells (Figure 2E), implying the expression of HADH-S isoform. Moreover, nuclear staining of HADH was decreased in CRC tissues compared to adjacent tissues (Figures 2F and 2G). Collectively, these findings suggest that nuclear HADH in CRC is associated with a tumor-suppressive effect.

Figure 2.

HADH-S is decreased in human colorectal cancer tissue

(A) Immunofluorescence analysis of endogenous HADH expression using anti-HADH antibody in RKO and HCT116 cells. Scale bars, 10 μm.

(B) Immunoblotting assessment of nuclear, cytosolic, and mitochondrial fractions of RKO and HCT116 cells.

(C) The expression of HADH in different grade CRC tumors from the TCGA database.

(D) Kaplan-Meier curves depict the overall survival rate in patients with colon cancer with high or low expression of HADH from the TCGA database.

(E) Representative IHC staining of HADH in colorectal cancer tissues and para-cancer tissues. Scale bars, 50 μm. The red arrows indicate nuclear staining of HADH.

(F) Comparison of HADH IRS in CRC tissues and paired para-cancer tissues (mean ± SD; two-tailed one-way analysis of variance).

(G) Nuclear staining of HADH in CRC tissues and paired para-cancer tissues (mean ± SD; ∗p < 0.05; two-tailed t test).

Hydroxyacyl-CoA dehydrogenase-S inhibits colorectal tumorigenesis in vitro and in vivo

Given the above clinical indications, we proceeded to investigate the potential role of HADH in colorectal tumorigenesis. Firstly, we stably expressed HADH-WT, HADH-S, or HADH-L in CRC cells and determined cell proliferation using direct cell counting and EdU incorporation. We observed a significant inhibition of cell proliferation in both RKO cells and HCT116 cells overexpressing HADH-S compared to control cells, while HADH-L had minimal impact on cell proliferation (Figures 3A–3E). On the contrary, knockdown of HADH promoted cell proliferation and Edu incorporation (Figures 3F–3H). The suppression of cell proliferation induced by HADH-S overexpression was further corroborated by a colony-forming assay (Figure 3I).

Figure 3.

HADH-S inhibits the tumorigenesis of colorectal cancer

(A) The proliferation assay of RKO cells stably expressing HADH, HADH-L, or HADH-S. Cell number at day 4 was statistically analyzed.

(B) EdU incorporation assay in RKO expressing HADH, HADH-L, or HADH-S. Scale bars, 20 μm. The analysis of EdU incorporation data is shown in the right panel.

(C) Immunoblotting analysis of FLAG-tagged HADH, HADH-L, and HADH-S in HCT116 cells.

(D) The proliferation assay of HCT116 cells expressing HADH, HADH-L, or HADH-S. Cell number at day 4 was statistically analyzed.

(E) EdU incorporation assay in HCT116 expressing HADH, HADH-L, or HADH-S.

(F) Immunoblotting analysis of HADH knockdown in HCT116 cells.

(G) The proliferation assay of HCT116 cells with or without HADH knockdown. Cell number at day 4 was statistically analyzed.

(H) EdU incorporation assay in HCT116 with or without HADH knockdown.

(I) Colony formation assay of HCT116 cells expressing HADH, HADH-L, or HADH-S. The analysis is shown in the right panel.

(J) The volume of xenograft tumors from mice implanted with the indicated HCT116 cells.

(K) The images of xenograft tumors from mice implanted with the indicated HCT116 cells. The analysis of tumor weight is shown in the right panel. All the data are presented as mean ± SD. For panels G and H, a two-tailed t test was used. For panels A, B, D, E, I, J, and K, one-way ANOVA was used to compare multiple groups.∗p < 0.05, ∗∗p < 0.01, ns not significant.

Next, we sought to ascertain whether the expression of HADH-S could impede tumor growth in the xenograft mouse model. Subcutaneous injection of HCT116 cells expressing vector, HADH-L, or HADH-S into the flanks of nude mice was performed, and tumor growth was monitored. Consistent with our in vitro results, cells expressing HADH-S but not HADH-L developed smaller tumors as compared with the vector control group (Figures 3J and 3K), indicating that HADH-S inhibits colorectal cancer progression in vivo.

Hydroxyacyl-CoA dehydrogenase-S regulates gene transcription in colon cancer cells

The nuclear localization implies the role of HADH-S in gene transcription. Therefore, we conducted genome-wide mapping of HADH-S binding using chromatin immunoprecipitation-sequencing (ChIP-seq) in RKO cells. The ChIP-seq analysis revealed a pronounced enrichment of HADH-S in the vicinity of transcription start sites (TSSs) (Figures 4A, S2A). High-quality peaks from the HADH-S ChIP data identified a total of 91,348 peaks (q < 0.01, using the input DNA sample as a control; Figure 4B, Table S1). Notably, more than 55% of HADH-S-binding sequences were localized within gene promoters (Figures 4B and 4C). Motif analysis using the MEME Suite detected an enrichment of CT-rich sequences in these HADH-S-binding regions (present in 6,304 out of 91,348 sequences) (Figures 4D, S2). To ascertain whether HADH-S recognizes this binding motif and affects gene transcription, a 12 × CT motif was cloned into the pGL3 vector, and the luciferase assay was performed. HADH-S, but not HADH-L, significantly enhanced the luciferase activity of the CT motif (Figure 4E). Electrophoretic mobility shift assay showed that HADH binds to the CT motif (Figure 4F). These findings imply HADH-S’s ability to recognize CT-rich sequences and promote gene transcription.

Figure 4.

HADH-S regulates gene transcription in RKO cells

(A) Genomic distribution analysis of HADH-S binding regions around transcription start sites (TSSs) using ChIP-seq signal profile (top panel) and heatmap (bottom panel), normalized by the reads per genome coverage (RPGC).

(B) Bar graph illustrates the distribution of HADH-S-bound chromatin fragments.

(C) Distribution of HADH-S binding sites relative to TSS.

(D) Consensus motif of HADH-binding sites.

(E) Schematic representation of the luciferase construct containing the potential HADH-S binding sequence (Top panel). Luciferase activity of the construct, as indicated in the top panel, in RKO cells (mean ± SD; ∗∗p < 0.01; two-tailed t test).

(F) Electrophoretic mobility shift assay (EMSA) showing HADH binding to the CT motif.

(G) Venn diagram demonstrates the overlap of genes with HADH-S binding sites in the promoter region from chip-seq data and genes expressed in cells with or without HADH-S overexpression.

(H) Violin plot depicts log₂-fold change of cells with or without HADH-S overexpression for non-targets (light pink) and potential HADH-S targets (grayish-green) (∗∗∗∗p < 0.0001, Mann-Whitney test).

(I) Volcano plot depicts the mRNA changes after HADH-S overexpression. The significantly upregulated (red) and downregulated (blue) genes are highlighted (Mann-Whitney test).

To explore the target genes of HADH-S in CRC cells, we conducted RNA-seq in RKO cells. Genes with peaks in the promoter region in the ChIP-seq data were considered potential HADH-S targets. We identified 8,052 potential HADH-S-targeted genes (Figure 4G, Table S2). Notably, HADH-S-targeted genes exhibited a substantial increase at the mRNA level (RNA-seq) following HADH-S overexpression compared to non-HADH-S-targeted genes (Figure 4H), indicating a positive regulatory role of HADH-S in gene expression. Among the 8,052 potential HADH-S target genes, 65 genes showed significant up-regulation (Figure 4I). KEGG analysis unveiled that these up-regulated genes are associated with metabolic pathways, and many pathways in cancer (Figure S2). We focused on these upregulated genes for further analysis.

Hydroxyacyl-CoA dehydrogenase-S promotes the transcription of AXIN2

From the up-regulated HADH-S target genes, we selected eight top genes for validation, and consistently observed induced expression of AXIN2 in HADH-S-expressing cells. We then investigated HADH-S regulating AXIN2 expression in CRC. Our findings revealed a substantial increase at both the mRNA level and protein level of AXIN2 in HADH-S cells, whereas HADH-L cells have little effect (Figures 5A and 5B). Moreover, luciferase assays demonstrated that the AXIN2 promoter was activated by HADH-S (Figure 5C). Independent ChIP-qPCR experiments unveiled a dramatic enrichment of HADH-S at the promoter region but not the coding region of the AXIN2 gene (Figures 5D and 5E). These data suggest that AXIN2 is a direct target of HADH-S in CRC cells.

Figure 5.

AXIN2 is a direct target of HADH-S

(A and B) The mRNA (A) and protein (B) levels of AXIN2 expression in RKO cells expressing HADH-S or HADH-L. The activity of the wild-type AXIN2 promoter or the CT-mutated promoter in RKO cells.

(C) Luciferase activity of wild-type or CT-mutant AXIN2 promoter in RKO cells transfected with vector or HADH-S expression plasmid.

(D) Genome browser tracks illustrate HADH-S ChIP-seq and input reads density at the AXIN2 peak region (the red rectangle) and non-peak region (the orange rectangle).

(E) ChIP-qPCR analysis of HADH-S binding to AXIN2 loci, with IgG serving as the control.

(F) Immunofluorescence detection of β-catenin in RKO cells expressing HADH-S or HADH-L treated with or without Wnt3a conditioned medium for 12 h. Scale bars, 10 μm.

(G) β-catenin/TCF4 reporter activity in control, HADH-S, or HADH-L cells treated with Wnt3a conditioned medium for 12 h.

(H) β-catenin/TCF4 reporter activity in control or HADH-S expressing SW480 cells.

(I) Relative mRNA levels of MYC and CCND1 in SW480 cells with or without HADH-S expression.

(J) Immunoblotting analysis of HADH-S expressing RKO cells with or without AXIN2 knockdown.

(K) The proliferation assay of the indicated cells. All the data are presented as mean ± SD. For panel H, a two-tailed t test was used. For panels A, G, and K, one-way ANOVA was used. For panels C, E, and I, two-way ANOVA was used. ∗p < 0.05, ∗∗p < 0.01, ns not significant.

AXIN2 is a scaffold protein within the β-catenin destruction complex, which blocks β-catenin nuclear translocation and activation.^20,21 Therefore, we determined whether HADH-S affects Wnt/β-catenin signaling in CRC. Under normal growth conditions, the β-catenin signaling was low in RKO cells, since RKO cells have intact APC/β-catenin molecules.²² When treated with Wnt-3a conditioned medium, β-catenin was activated and accumulated in the nucleus of RKO cells. The nuclear β-catenin was reduced in HADH-S cells upon Wnt treatment, indicating the inhibition of β-catenin (Figure 5F). Consistently, using the TOP^Flash/FOP^Flash luciferase reporter assay, we detected a significant reduction of β-catenin transcription activity in HADH-S expressing cells (Figure 5G). HADH-S also reduced β-catenin transcription activity in SW480 cells, bearing APC mutation and catenin activation (Figure 5H). Moreover, HADH-S also significantly down-regulated β-catenin target genes MYC and CCND1 in SW480 cells (Figure 5I). To ascertain the functional relationship between AXIN2 and HADH-S in CRC, we conducted AXIN2 knockdown experiments in HADH-S cells (Figure 5J). The knockdown of AXIN2 abolished the inhibitory effect of HADH-S on CRC cell proliferation (Figure 5K).

Hydroxyacyl-CoA dehydrogenase-S antagonizes protein arginine methyltransferase 5-mediated histone arginine methylation

To explore the mechanism through which HADH-S facilitates AXIN2 transcription, we immunoprecipitated the interactome of FLAG-tagged HADH-S, followed by mass spectrometry analysis. Among the proteins interacting with HADH-S, PRMT5 exhibited high abundance (Table S3). The interaction between HADH-S and endogenous or HA-tagged PRMT5 was validated by independent Co-IP experiments (Figure 6A). We further demonstrated that the middle 100 amino acids of HADH are essential for its interaction with PRMT5 (Figures 6B and 6C). Deletion of the middle 100 amino acids abolished HADH-regulated AXIN2 expression and cell proliferation, suggesting that HADH’s function relies on its PRMT5 interaction (Figures 6D and 6E). PRMT5 is known to catalyze the symmetric di-methylation of arginine residues (Rme2s) in histones H3 and H4, which are commonly linked with transcriptional repression.²³ Specifically, H4R3me2s is associated with coupling histone and DNA methylation in gene silencing.²⁴ We hypothesized that HADH-S might modulate PRMT5’s histone methylation function. First, we studied the regulation of PRMT5 on AXIN2 expression. ChIP experiments revealed a direct binding of PRMT5 with the AXIN2 promoter region (Figure 6F). Moreover, knockdown of PRMT5 in RKO cells resulted in increased AXIN2 expression, phenocopying HADH-S overexpression (Figures 6G and 6H). Although the expression of HADH-S led to a slight inhibition of total H4R3me2s levels (Figure 6I), we observed a significant decrease in the H4R3me2s level at the AXIN2 promoter region in HADH-S cells (Figure 6J). Collectively, these findings suggest that HADH-S promotes AXIN2 expression by counteracting PRMT5-mediated histone arginine methylation.

Figure 6.

HADH-S antagonizes PRMT5-mediated histone arginine methylation

(A) Co-immunoprecipitation analysis of HADH-S interaction with PRMT5. HAHD-S cell lysates were subjected to immunoprecipitation using IgG or anti-FLAG antibody, followed by immunoblotting analysis.

(B) Schematic of deletion mutants of the HADH protein.

(C) Co-immunoprecipitation analysis of the interaction between deletion mutants of HADH-S and PRMT5.

(D) The mRNA level of AXIN2 in RKO cells expressing HADH or its deletion mutant.

(E) The proliferation assay of RKO cells expressing HADH or its deletion mutant.

(F) ChIP-qPCR analysis of PRMT5 binding to AXIN2 loci, with IgG serving as the control.

(G and H) The mRNA (G) and protein (H) levels of AXIN2 expression after PRMT5 knockdown.

(I) Immunoblotting analysis of H3R4me2s levels in control or HADH cells.

(J) ChIP-qPCR analysis of H3R4me2s levels at AXIN2 promoter region.

(K) The working model of HADH-inhibited colorectal cancer tumorigenesis. Alternative translation initiation of HADH produces a nuclear isoform that upregulates AXIN2 expression and suppresses tumorigenesis. All the data are presented as mean ± SD. For panels D, E, and G, one-way ANOVA was used. For panels F and J, two-way ANOVA was used. ∗p < 0.05, ∗∗p < 0.01, ns not significant.

Discussion

Hydroxyacyl-CoA dehydrogenase (HADH) is crucial for fatty acid β-oxidation and is intricately linked to tumorigenesis. The correlation of low HADH expression in gastric cancer, CRC, and renal cancer with poor prognosis emphasizes the need to unravel the intricate mechanisms behind these observations. Our data indicated that the translation initiation of HADH is altered in colorectal cancer to produce a truncated isoform HADH-S with nuclear localization, aligning with our prior identification of a similar isoform in mice.¹⁹ Strikingly, the nuclear isoform but not the mitochondrial isoform inhibited tumor growth in cultured cells and mouse models, underscoring a tumor suppressive role of this nuclear isoform in CRC (Figure 6K). Thus, altered translation initiation of HADH provides a potential approach for the therapeutic intervention of colorectal cancer.

Our findings unveil a pivotal transcriptional regulatory role for nuclear HADH in cancer suppression. Specifically, HADH binds to CT-rich sequences within the transcription start site (TSS) region, facilitating the expression of the tumor suppressor gene AXIN2. The transcriptional activation of AXIN2 by HADH, as evidenced by our ChIP-seq and RNA-seq analyses, provides mechanistic insights into the tumor-suppressive network orchestrated by HADH. AXIN2 acts as a negative regulator of the Wnt/β-catenin signaling pathway,^25,26 playing a crucial role in maintaining pathway equilibrium and curtailing tumor progression. In addition, AXIN2’s involvement in modulating tumor growth and metastasis processes further underscores its significance in suppressing cancer progression.^27,28,29 Thus, the HADH-S/AXIN2 regulatory axis offers a potential avenue for therapeutic interventions aimed at harnessing the anti-cancer properties of HADH and AXIN2 in CRC.

Furthermore, our study uncovers a nuanced interaction between HADH and PRMT5. PRMT5-catalyzed histone arginine methylation is a crucial post-translational modification in gene transcription silencing.²³ Upon binding to PRMT5, HADH decreases H4R3me2s modification at the AXIN2 promoter region, phenocopying the consequences observed upon PRMT5 inhibition. These data further support the notion that HADH-S enhances AXIN2 mRNA expression by inhibiting PRMT5 activity. Although we currently do not have direct biochemical evidence showing how HADH-S inhibits PRMT5 enzymatic activity, our co-immunoprecipitation and mass spectrometry results indicate that HADH-S physically interacts with PRMT5. One plausible explanation is that HADH-S binding interferes with PRMT5’s ability to access histone substrates or alters the assembly of PRMT5-containing complexes, thereby reducing its local catalytic activity at the AXIN2 promoter.

Transcriptomic-scale translation sequencing has revealed pervasive and dynamic translation initiation selection.^3,4,5,12 Studies have confirmed alternative translation initiation in numerous oncogenes, such as c-Myc,³⁰ and tumor suppressor genes, such as PTEN.^8,9,10,11 These alternative initiation events enable the precise regulation of protein activities and tumor progression. Intriguingly, most of these alternative initiations use non-AUG codons or codons in the sub-optimal Kozak context, raising a key question governing the mechanism of translation initiation site selection by the ribosome small subunit. One potential explanation is that the presence of regulatory elements or secondary structures, such as hairpins or internal ribosome entry sites (IRESs), dictates start codon selection.^30,31 Furthermore, non-canonical translation initiation factors such as eIF2A³² and eIF4G2³³ are frequently upregulated in cancers. These initiation factors may direct the ribosome to initiate translation, disobeying the canonical scanning model.

In conclusion, our study demonstrates a nuclear isoform of HADH as a novel tumor suppressor in CRC. This nuclear HADH emerges as a pivotal transcriptional factor impacting PRMT5-mediated epigenetic modifications and the Wnt signaling pathway. These findings pave the way for future investigations into the underlying molecular mechanisms and therapeutic implications of HADH in cancer biology.

Limitations of the study

Despite the valuable insights provided by this study, two main limitations should be acknowledged. First, the sample size and diversity of colorectal cancer (CRC) models used in both in vitro and in vivo experiments may limit the generalizability of the findings. Additional validation using a larger cohort of patient-derived samples and diverse genetic backgrounds would strengthen the conclusions. Second, while the study identifies and characterizes the short isoform HADH-S, the regulation of alternative translation initiation and the molecular mechanisms controlling the selection of the downstream start codon remain insufficiently explored.

Resource availability

Lead contact

Requests for further information, resources, and reagents should be directed to and will be fulfilled by the lead contact, Zhanghui Chen, at (zjcell@126.com).

Materials availability

All plasmids and cell lines generated by this study are available upon request from the lead contact with a completed Materials Transfer Agreement (MTA).

Data and code availability

• •The ChIP-seq and RNA-seq datasets generated in the current study have been deposited to Gene Expression Omnibus: GSE253660. (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE253660). secure token: clebwsyqtzetbwr.
• •This article does not report original code.
• •Any additional information required to reanalyze the data reported in this article is available from the lead contact upon request.

Acknowledgments

This work was supported by the National Natural Science Foundation of China (82170053 and 82370099 to Z.C., 82372727 and 82073110 to X.W.), the Zhejiang Provincial Natural Science Foundation of China (LZ23H160003), and the Zhanjiang Central Hospital Funding.

Author contributions

X.G. and Z.C. designed the experiments. Y.G. performed most of the experiments. L.Y., J.H., and Y.H. performed the experiments. X.G. and Y.G. wrote the article. X.G. and Z.C. administered the project. All the authors discussed.

Declaration of interests

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 article.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
FLAG tag	Cell Signaling Technology	Cat#14793
HA tag	Cell Signaling Technology	Cat#3724
HADH	Proteintech	Cat#19828-1-AP RRID:AB_10667408
Lamin B	Cell Signaling Technology	Cat#13435 RRID:AB_10896336
Tubulin	Proteintech	Cat#11224-1-AP RRID:AB_2210206
COX5A	Abclonal	Cat#A6437 RRID:AB_2767039
TOM20	Abclonal	Cat#A19403 RRID:AB_2862646
Chemicals, peptides, and recombinant proteins
Lipofectamine 3000	Invitrogen	Cat#L3000015
Puromycin	Selleck	Cat#S7417
Fetal bovine serum (FBS)	GIBCO	Cat#26140079
Dulbecco’s modified Eagle’s medium (DMEM)	GIBCO	Cat#11965092
Annexin V and Propidium Iodide	eBioscience	Cat#88-8007-72
Proteinase K	NEB	Cat#P8107S
Mito-Tracker Red CMXRos	Beyotime	Cat#C1439S
Critical commercial assays
BD Pharmingen FITC Annexin V Apoptosis	BD Biosciences	Cat#556547
EdU detection kit	Beyotime	Cat#C0075L
Nuclear and Cytoplasmic Protein Extraction Kit	Beyotime	Cat#P0027
Mitochondria Fractionation kit	Beyotime	Cat#ELS-ALX-850-276-KI01
Dual-Luciferase® Reporter Assay System	Promega	Cat#E1910
EZ-ChIP kit	Merck	Cat#17-371
Universal DNAseq Library Prep Kit	KAITAI-BIO	Cat#AT4106
Deposited data
ChIP-seq	This paper	Gene Expression Omnibus: GSE253660
RNA-seq	This paper	Gene Expression Omnibus: GSE253660
Experimental models: Cell lines
HCT116	ATCC	Cat#CCL-247
RKO	ATCC	Cat#CRL-2577
HEK293T	ATCC	Cat#CRL-3216
Oligonucleotides
HADH qPCR forward primer	This paper	ACTCGGGTTTGGGCTTTTCT
HADH qPCR reverse primer	This paper	CCACCCATCCACGATGAACT
AXIN2 qPCR forward primer	This paper	TAACCCCTCAGAGCGATGGA
AXIN2 qPCR reverse primer	This paper	AGTTCCTCTCAGCAATCGGC
Recombinant DNA
pCDH3.1-FLAG-HADH-L	This paper	N/A
pCDH3.1-FLAG-HADH-S	This paper	N/A
CT-repeat-reporter	This paper	N/A
Software and algorithms
GraphPad Prism 8	GraphPad Software	https://www.graphpad.com/
ImageJ	National Institutes of Health (NIH)	https://imagej.nih.gov/ij/
DNAMAN	Lynnon Corporation	https://www.lynnon.com/dnaman.html
Flowjo	BD Biosciences	https://www.flowjo.com/
Integrative Genomics Viewer	Broad Institute	https://software.broadinstitute.org/software/igv/download
Other
Flow cytometer	BD Biosciences	FACSVerse
Real-time PCR	Roche	LightCycler 480

Experimental model and study participant details

Cell culture

Human colorectal cancer cell line HCT116 and RKO were procured from the ATCC and maintained in DMEM medium (Gibco) supplemented with 10% fetal bovine serum at 37°C in a 5% CO₂ environment. All experiments were performed with mycoplasma-free cells. All human cell lines have been authenticated using STR profiling within the last three years has been included.

Mouse tumor xenografts

Six-week-old female Balb/c-nude mice were utilized for the tumor xenograft model. HCT116 cells expressing vector, HADH-L, and HADH-S, at a concentration of 2 × 10⁶ cells per mouse, were subcutaneously injected into the axillary region. Mice were randomly divided into four groups, with five mice in each group. After injection, tumor diameter measurements were conducted three times per week, alongside the monitoring of any alterations in mouse body weight. Upon reaching a tumor diameter of 1 cm after four weeks post-injection, mice were euthanized. The tumors were then extracted, weighed, and measured for volume. The excised tumor tissues were carefully preserved for subsequent investigations.

Mice were maintained in pathogen-free conditions at the Animal Center of Zhejiang University. All animal studies were performed in compliance with the Guide for the Care and Use of Laboratory Animals by the Medical Experimental Animal Care Commission of Zhejiang University(approval number: ZJU20240214). All animal studies used the protocol that has been approved by the Medical Experimental Animal Care Commission of Zhejiang University.

Method details

Plasmids construction

The HADH gene was amplified from human RKO cDNA. The resulting fragments were then inserted with a C-terminal FLAG-tag into the lentivirus vector pCDH-CMV-MCS-EF1-Puro. aTIS or dTIS was mutated to AUA using a mutagenesis kit. For the construction of a luciferase reporter, a potential HADH-S binding motif (12×CT sequence) was inserted into the upstream region of the minimal promoter of the pGL3-enhancer vector. For knockdown experiments, shRNA oligos targeting PRMT5 or AXIN2 were cloned into the lentiviral vector pLKO.1.

Cell fractionation

Nuclear, cytoplasmic, and mitochondrial proteins were prepared using the nuclear fractionation kit and mitochondria fractionation kit (Beyotime).

Immunoblotting

Whole cell lysates or cell fractions were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto PVDF membranes (Millipore). After blocking with 5% non-fat dry milk, membranes were incubated with primary antibodies, followed by secondary antibody incubation. Membranes were visualized using chemiluminescence (GE Technology). Anti-Flag-tag (14793, Cell Signaling Technology), Anti-HA-tag (3724, Cell Signaling Technology), anti-HADH (19828-1-AP, Proteintech), anti-Lamin B (13435, Cell Signaling Technology), anti-Tubulin (11224-1-AP, Proteintech) and anti-COX5A (A6437, Abclonal) were used.

Immunofluorescence

Cells were treated with Mito-Tracker Red CMXRos (Beyotime), fixed with 4% paraformaldehyde (PFA), permeabilized with 0.1% Triton X-100, and blocked using 5% goat serum. Staining for C-terminal FLAG-tagged proteins or endogenous HADH was carried out using anti-FLAG (Cell Signaling Technology) or anti-HADH antibody, followed by incubation with secondary antibodies. The nucleus was visualized using Hoechst (ThermoFisher Scientific). Fluorescent images were captured using a Nikon A1R confocal microscope.

Immunohistochemistry

The colorectal cancer tissue arrays were purchased from Shanghai Outdo Biotech Company (HCOA160CS01-M-128). Ethical consent was granted from the Ethical Committee Review Board of Shanghai Outdo Biotech Company. Following deparaffinization and hydration of tissue array sections, heat-induced antigen retrieval was performed using the citrate buffer solution. Subsequently, the sections were treated with a 3% H₂O₂ solution to quench endogenous peroxidase activity and 5% goat serum to block nonspecific binding. The sections were incubated with anti-HADH (Proteintech, 1:100) antibodies overnight at 4 C, followed by incubation with HRP-conjugated secondary antibodies. Staining was achieved using diaminobenzidine as the enzyme substrate, and hematoxylin served as the counterstain. The slides were digitally scanned using an Olympus digital slicing scanner VS200.

Cell proliferation and colony formation assays

To access cell proliferation, HCT116 cells expressing the vector, HADH-L or HADH-S were seeded in a 12-well plate at a concentration of 1×10⁵ cells per well, followed by direct cell counting. The EdU incorporation assay was carried out using an EdU incorporation kit (Beyotime). For the colony formation assay, cells were seeded into 6-well dishes (3×10³ cells/well) for approximately 3 weeks, and subsequent staining with 1mg/ml Nitrotetrazolium Blue chloride (NBT) was conducted. Colonies were observed and enumerated under a microscope.

Chromatin immunoprecipitation (ChIP-seq)

ChIP was performed using the EZ-ChIP kit (Merck). Briefly, RKO cells were fixed with 1% formaldehyde for 15 min at room temperature and quenched with 0.125 M glycine/PBS for 5 min. Cells were harvested and lysed in SDS lysis buffer (1% SDS, 10 mM EDTA, 50 mM Tris pH 8.1). The chromatin was fragmented using a non-contact sonicator to achieve fragments of approximately 300-500 bp. 5% of sonicated chromatin was used as input. The rest chromatin solutions were immunoprecipitated overnight at 4 °C using antibody-conjugated protein-A beads. Following sequential washing with low salt buffer, high salt buffer, LiCl buffer, and TE buffer, the beads were eluted with elution buffer. The immunoprecipitates were de-crosslinked and treated with proteinase K (NEB) and RNAse A. The released DNA was purified using a gel purification kit and was then subjected to DNA library construction or qPCR analysis. Library construction was performed using Universal DNAseq Library Prep Kit (KAITAI-BIO, AT4106). Library sequencing was performed on NovaSeq 6000 with paired-end 2 × 150 bp read length.

RNA-seq

Total RNA was isolated from both control cells and cells overexpressing HADH-S utilizing the TRIzol reagent. RNA samples were then submitted to mRNA library construction and sequencing by using an Illumina NovaSeq 6000.

Sequencing data analysis

Illumina sequencing reads were initially processed using fastp (version 0.23.2) for adaptor removal and quality trimming.³⁴ The processed reads were aligned to the human reference genome (gencode.v41) using Bwa-mem2 (version 2.2.1) with default parameters. SAMtools (version 1.9) was applied to eliminate PCR duplications.³⁵ Peak calling and annotation were performed utilizing MACS2 (version 1.4.2) along with the ChIPseeker package.³⁶ Visualization of identified peaks with specific genomic regions was achieved through the Integrative Genomics Viewer (IGV, version 2.11.2).³⁷ Heatmaps and spectra depicting ChIP-seq signals around transcription start sites (TSS) were generated using deepTools (version 3.5.1). Additionally, motif analysis was conducted using MEME.

Luciferase report assay

The luciferase reporter containing the potential HADH-S binding motif was co-transfected with pRL-TK into RKO cells expressing the vector, HADH-L or HADH-S. After 24 h of transfection, cells were harvested and lysed. Firefly and Rinella luciferase activities were measured using the Dual-Luciferase Assay Kit (Promega). The luciferase activities were expressed as normalized relative light units to the Rinella internal control.

To study the β-catenin activity, cells were co-transfected with TOP^Flash or FOP^Flash firefly luciferase reporter plasmid with pRL-TK as an internal control. The activities of the TOP^Flash and the FOP^Flash reporter constructs were normalized to the Rinella internal control to obtain FOP and TOP activities. For FOP, the three replicates were averaged to yield the mean FOP activity for each group. TOP activity was then normalized to the corresponding mean FOP to calculate TOP/FOP activity. Finally, TOP/FOP values were normalized to the control (Con) group.

Real-time quantitative RT-PCR

Total RNA was extracted from cells using RNAiso plus (Takara) and subjected to reverse transcription with M-MLV Reverse Transcriptase (AGbio). Real-time PCR was carried out using the SYBR Green I master mix on a LightCycler 480 real-time PCR system (Roche). Normalization of all data was performed relative to the expression levels of ACTB or GAPDH.

Co-immunoprecipitation

Cells lysis was performed using ice-cold lysis buffer (50 mM Tris-pH 7.4, 150 mM NaCl, 0.5% NP40) supplemented with protease the inhibitor cocktail. Lysates were incubated overnight at 4 °C with anti-FLAG M2 magnetic beads (Sigma). The immunoprecipitated proteins were efficiently eluted from the beads and subjected to Western blotting or mass spectrometry.

Quantification and statistical analysis

Number of biological replicates, repetitions of the experiments, and statistical significance were described in figure legends. Data are presented as mean ± SD.

Statistical analyses were conducted using Student’s t-test and one-way ANOVA to compare differences between experimental and control groups. All data were analyzed by GraphPad Prism 8. A two-tailed p-value of ∗p ≤ _0.05, ∗∗p ≤ _0.01 was considered statistically significant.

Published: January 10, 2026