P4HA2 mediates hypoxia-induced collagen remodeling to promote tumor progression in head and neck squamous cell carcinoma

Abstract

Purpose

This study investigated the clinical significance, regulatory mechanisms, and functional roles of Prolyl 4-Hydroxylase Subunit Alpha 2 (P4HA2) in head and neck squamous cell carcinoma (HNSCC), focusing on its involvement in hypoxia-driven extracellular matrix (ECM) remodeling and tumor progression.

Methods

Immunohistochemical (IHC) analysis was performed on 108 clinical HNSCC specimens to evaluate the expression of P4HA2 and Hypoxia-Inducible Factor-1 Alpha (HIF-1α) and their association with patient prognosis. These clinical findings were further validated using bioinformatic data from public cohorts. The transcriptional regulation of P4HA2 by HIF-1α under hypoxic conditions was investigated via chromatin immunoprecipitation (ChIP). Functional assays, including assessments of P4HA2’s dependency on its essential cofactor vitamin C, were conducted in HNSCC cell lines to evaluate cell proliferation, migration, and invasion in vitro. The role of P4HA2 in tumor growth and microenvironmental remodeling in vivo was examined using xenograft mouse models.

Results

P4HA2 and HIF-1α were upregulated in HNSCC and associated with poor prognosis. Hypoxia-induced HIF-1α directly activated P4HA2 transcription. P4HA2 promoted collagen deposition and ECM remodeling, facilitating HNSCC cell proliferation, migration, and invasion, with its collagen-producing effects enhanced by vitamin C. In vivo, P4HA2 overexpression accelerated tumor growth, accompanied by enhanced HIF-1α signaling and collagen production.

Conclusion

P4HA2 is a key effector of the hypoxia/HIF-1α axis, driving collagen remodeling to promote HNSCC progression, suggesting its potential as a therapeutic target.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s13402-026-01268-5.

Introduction

Head and Neck Squamous Cell Carcinoma (HNSCC) is a malignant tumor originating from the epithelial lining of the oral cavity, larynx, pharynx, nasal cavity, paranasal sinuses, and salivary glands. It represents the most common type of head and neck cancer, accounting for over 90% of all cases in this region [1]. The etiology of HNSCC is multifactorial, with strong associations with tobacco use, alcohol consumption, and betel quid chewing [2]. Despite advances in treatment modalities such as surgery, radiotherapy, chemotherapy, and targeted therapies, the 5-year survival rate remains approximately 50%, due largely to rapid tumor proliferation, local invasion, and distant metastasis [3–5]. Notably, the quality of life of HNSCC survivors is severely compromised; functional and aesthetic impairments not only cause psychological distress but also contribute to an elevated risk of suicide among these patients [6, 7].

Tumor metastasis represents a major therapeutic challenge and is a critical determinant of prognosis. Data from the American Cancer Society indicate a pronounced decline in 5-year survival rates as the disease progresses from localized to distantly metastatic stages [8]. Therefore, a deeper understanding of the mechanisms underlying HNSCC development and metastasis is urgently needed to reveal disease progression and provide a theoretical foundation for novel therapeutic strategies aimed at improving survival and quality of life.

During metastasis, tumor cells remodel the extracellular matrix (ECM) to breach the basement membrane and invade adjacent tissues. This process involves collagen fiber cross-linking and linearization, forming migration tracks for tumor cells [9, 10]. Collagen-modifying enzymes play essential roles in this context by altering collagen structure and properties to establish a microenvironment conducive to tumor growth and dissemination [11, 12]. Thus, investigating the mechanisms of ECM remodeling in HNSCC is of great therapeutic relevance.

The prolyl 4-hydroxylase (P4H) family, key enzymes in collagen biosynthesis, contribute significantly to ECM remodeling and tumor microenvironment modulation. Recent studies indicate that P4HA1 reduces prolyl hydroxylation of Hypoxia-Inducible Factor-1 Alpha (HIF-1α), enhancing its stability and activating the P4HA/HIF-1 axis, which promotes cancer stemness and chemoresistance in triple-negative breast cancer [13]. Among P4H members, Prolyl 4-Hydroxylase Subunit Alpha 2 (P4HA2) has been implicated in facilitating collagen deposition and driving tumor progression and metastasis in breast cancer [14]. Beyond its role in collagen regulation, P4HA2 influences a spectrum of oncogenic processes. For example, Cao et al. reported that P4HA2 knockdown suppresses proliferation, migration, and invasion of cervical cancer cells and inhibits tumor growth in xenograft models [15].

Pharmacologically, targeting P4HA2 represents a promising therapeutic strategy. Aspirin, for instance, has been shown to disrupt P4HA2-driven hepatocellular carcinoma growth and collagen deposition [16], whereas budesonide inhibits P4HA2-mediated hydroxylation to suppress triple-negative breast cancer metastasis [17]. Notably, the functional impact of P4HA2 appears to be context dependent. In prostate cancer, it suppresses YAP1 expression and thereby restrains cell migration and invasion [18], highlighting its tissue-specific regulatory behavior.

The tumor-specific upregulation of P4HA2 is potentially driven by hypoxic conditions, a hallmark of rapidly growing malignancies characterized by inadequate vascular supply [19, 20]. Under hypoxia, HIF-1α transcriptionally upregulates P4HA2 expression [21], which in turn facilitates ECM remodeling through collagen cross-linking and matrix stiffening [22]. These biomechanical modifications not only provide structural guidance for tumor cell invasion but also enhance integrin-mediated signaling pathways, thereby further promoting metastatic behavior [23].

Based on the emerging role of P4HA2 as a multifunctional effector in malignant progression [24] and its strong association with poor prognosis in HNSCC [25], this study aims to systematically investigate how P4HA2, under hypoxic conditions, drives HNSCC aggressiveness through the HIF-1α/COL1A1 axis.

Materials and methods

Samples collection and characteristics

A total of 108 patients with HNSCC who underwent surgical resection in the Department of Oral and Maxillofacial-Head and Neck Surgery at the First Affiliated Hospital of Fujian Medical University between January 2012 and December 2016 were retrospectively enrolled in this study. The inclusion criteria were as follows: (1) diagnosis of primary HNSCC; (2) underwent standard surgical resection including primary tumor excision and cervical lymph node dissection; and (3) availability of complete follow-up data. Exclusion criteria included: (1) incomplete clinical or pathological records; and (2) mortality attributable to non-HNSCC causes.

Tumor histopathological grading was assessed according to WHO 2016 classification system, and clinical staging was determined based on the 8th edition of the AJCC TNM staging system. Follow-up was conducted through telephone interviews and electronic medical record review from outpatient clinics. The date of initial diagnosis was defined as the start of follow-up, with death due to HNSCC as the endpoint event.

Immunohistochemistry (IHC) and evaluation

A tissue microarray (TMA) was constructed by transferring core biopsies from representative tumor and adjacent normal tissues into a recipient paraffin block with a 7 × 6 array. For immunohistochemical analysis, TMA sections were subjected to standard procedures including baking, deparaffinization, antigen retrieval with pH 9.0 EDTA buffer, and blocking of endogenous peroxidase. The sections were then incubated with primary antibodies against P4HA2 (Abcam, cat# ab233197; 1:900), HIF-1α (Abcam, cat# ab179483; 1:200), p53 (MXB Biotechnologies, cat# MAB-0674; 1:200) or p16 (MXB Biotechnologies, cat# MAB-0673; 1:300) at 37 °C for 2 h, followed by an HRP-conjugated secondary antibody and DAB visualization. Finally, the sections were counterstained with hematoxylin, dehydrated, and mounted.

Staining for P4HA2 was evaluated semi-quantitatively based on the proportion of positive cells (scored 0–4) and staining intensity (scored 0–3). The final score (proportion × intensity) was categorized as negative (–), weak (+), moderate (++), or strong (+++), with five high-power fields assessed per sample. HIF-1α expression was evaluated using the H‑Score method. H-Score = Σ (proportion of positive cells × staining intensity), yielding a scale from 0 to 300. A score > 100 was classified as high expression, and ≤ 100 as low expression. For p16, positivity was defined as diffuse and strong nuclear and cytoplasmic staining in ≥ 70% of tumor cells. p53 expression was categorized as low (scattered weak‑to‑moderate nuclear staining) or high (diffuse strong nuclear staining or complete absence of staining, null pattern).

Cell culture

Human HNSCC cell lines HN6 (provided by Prof. Wantao Chen, Shanghai Jiao Tong University, China) and CAL27 (purchased from ATCC, USA) were cultured in high-glucose DMEM (HyClone, USA) supplemented with 10% FBS (Gibco, USA) at 37 °C in a humidified atmosphere containing 5% CO₂. For hypoxia experiments, cells were incubated in a triple-gas incubator (Heal Force, China) under conditions of 3% O₂, 5% CO₂, and balanced N₂. For vitamin C treatment, cells were supplemented with L‑ascorbic acid (Sigma‑Aldrich, USA) at a final concentration of 100 μm for 48 h. The culture medium was refreshed every 24 h to maintain ascorbate activity.

CAL27 and HN6 cells were selected because they represent HPV-negative oral squamous cell carcinoma models with altered p53 backgrounds, consistent with the predominant molecular characteristics observed in our clinical cohort [26].

siRNA transient transfection

siRNAs targeting HIF-1α and P4HA2 (Table S1) were transfected using Lipofectamine RNAiMAX (Thermo Fisher Scientific, USA) according to the manufacturer’s instructions. CAL27 and HN6 cells were seeded in 6-well plates, transfected at 30–50% confluence with siRNA-lipid complexes in Opti-MEM (Gibco, USA), and incubated under normoxia or hypoxia (3% O₂) for 24–48 h.

Plasmid transfection

The P4HA2 overexpression plasmid was purchased from Aibive Biotechnology (Changsha, China). Plasmids were extracted from E. coli cultures using a plasmid purification kit (Tiangen Biotech, China). Cells were transfected with the P4HA2 plasmid or an empty vector control using Lipofectamine 2000 (Thermo Fisher Scientific, USA) in Opti-MEM (Gibco, USA). The culture medium was replaced after 6–12 h, and cells were harvested after 24–48 h.

shRNA-mediated stable knockdown

To establish stable P4HA2 knockdown cell lines, lentiviral shRNA targeting P4HA2 (shP4HA2) and a non-targeting control shRNA (shNC) were used (Table S2). CAL27 and HN6 cells were seeded in 6-well plates and transduced with lentiviral particles in the presence of polybrene (5 µg/mL). After 24 h, the medium was replaced with fresh culture medium containing puromycin (2 µg/mL) for selection. Stable knockdown cells were maintained in puromycin-containing medium for 7 days.

Protein extraction and quantification

Cellular proteins were lysed in RIPA buffer (Beyotime, China) containing 1× PMSF (Beyotime, China) and protease inhibitor cocktail (MedChemExpress, China), scraped from plates, and centrifuged (12000 × g, 10 min, 4 °C) to collect supernatants. Tissue proteins were extracted by homogenizing samples in RIPA buffer on ice using a handheld homogenizer (Labgic, China), followed by centrifugation (12000 × g, 15 min, 4 °C). Protein concentration was determined using a BCA assay kit (Beyotime, China) with BSA (BioFroxx, Germany) as the standard. All samples were adjusted to 1–2 µg/µL, denatured in 5× SDS loading buffer (Beyotime, China) at 100 °C for 10 min, and stored at − 20 °C.

Western blotting

Proteins (20–40 µg) were separated by SDS-PAGE using precast gels (GenScript, USA) and transferred to PVDF membranes (Merck KGaA, Germany). Membranes were blocked with 5% BSA (BioFroxx, Germany) and incubated with primary antibodies against HIF-1α (Abcam, UK; cat# ab179483; 1:1000), P4HA2 (Abcam, UK; cat# ab233197; 1:1000), COL1A1 (Abcam, UK; cat# ab255809; 1:1000), GAPDH (Wuhan Sanying, China; cat# 10494-1-AP; 1:5000), and β-actin (Cell Signaling Technology, USA; cat# 4967; 1:5000) at 4 °C overnight. After washing, membranes were incubated with HRP-conjugated goat anti-rabbit IgG (Boster, China; cat# BA1054; 1:3000) or HRP-conjugated goat anti-mouse IgG (Boster, China; cat# BA1050; 1:5000). Signals were detected using BeyoECL Plus kit (Beyotime, China) and imaged with a ChemiDoc XRS+ Imaging System (Bio-Rad, USA).

RNA extraction and qRT-PCR

Total RNA was extracted using NucleoZol reagent (Macherey-Nagel, Germany). cDNA was synthesized using a PrimeScript RT kit (Takara, Japan). qPCR was performed with SYBR Green (Beyotime, China) on a QuantStudio system (Thermo Fisher Scientific, USA) with primers (Table S3). GAPDH served as internal control. Relative mRNA expression was calculated using the 2^^−ΔΔCt method.

ChIP-qPCR

Cells were cross-linked with 1% formaldehyde, lysed in RIPA buffer, and sonicated using a Cole-Parmer ultrasonic processor (Cole-Parmer, USA). Chromatin was immunoprecipitated with an anti-HIF-1α antibody (Abcam, UK; cat# ab179483; 1:1) or IgG control (Zenbio, China) coupled to Protein G magnetic beads (MedChemExpress, China). DNA was purified and analyzed by qPCR with primers targeting the P4HA2 promoter (Table S4).

Cell proliferation assay

Cells were seeded in 96-well plates (2000 cells/well) 24 h post-transfection. Using the CCK-8 kit (Beyotime, China), cell viability was assessed every 24 h for 4 days. The absorbance at 450 nm was measured to plot the proliferation curve.

Colony formation assay

Cells were plated in 6-well plates (800–1000 cells/well) and cultured for 2 weeks. The resulting colonies were fixed, stained with crystal violet (Beyotime, China), and counted to determine the colony formation rate.

Wound healing assay

A sterile pipette tip was used to create a scratch on a confluent cell monolayer in a 12-well plate. After washing, cells were cultured in serum-free medium. Wound closure was monitored and imaged at 0, 12, 24, and 48 h to assess migration.

Cell migration assay

Cell migration was assessed using uncoated Transwell chambers (Corning, USA). Briefly, 5 × 10⁴ to 1 × 10⁵ cells in serum-free medium were seeded into the upper chamber, while the lower chamber was filled with medium supplemented with 10% FBS as a chemoattractant. After incubation for 24–72 h, the cells that had migrated to the lower surface of the membrane were fixed, stained with 0.1% crystal violet, and counted under microscope.

Cell invasion assay

Cell invasion was evaluated using Matrigel-coated Transwell chambers (Corning, USA). The assay procedure was similar to the migration assay, with the following modifications: the chambers were rehydrated with serum-free medium at 37 °C for 2 h prior to use, and a higher density of cells was seeded. After 24–72 h of incubation, the invaded cells on the lower membrane surface were stained with crystal violet and quantified.

Immunofluorescence (IF)

Cells were fixed with 4% PFA, permeabilized with 0.5% Triton X-100, blocked with BSA, and incubated with primary antibodies followed by Alexa Fluor-conjugated secondary antibodies (Cell Signaling Technology, USA) and rhodamine-phalloidin (UElandy, China). Nuclei were stained with DAPI (Beyotime, China). Images were captured using a laser scanning confocal microscope (Olympus, Japan).

Second harmonic generation (SHG) Imaging

Tissue sections were imaged using a two-photon confocal microscope (Olympus, Japan) with 890 nm excitation and 445 nm emission to visualize collagen fibers.

In vivo tumor xenograft study

Male BALB/c nude mice (4 weeks old; Fuzhou Jin’an Beida Trading Co., China) were randomly divided into two groups (n = 7 per group): the Vector group and the P4HA2 overexpression group. Mice in the Vector group were subcutaneously inoculated in the right flank with 2 × 10⁶ HN6 cells stably transfected with empty vector. Mice in the P4HA2 overexpression group were inoculated with the same number of HN6 cells stably overexpressing P4HA2.

Tumor size and mouse body weight were measured every day. Tumor volume was calculated using the formula: V = (length × width²) / 2. At the experimental endpoint (e.g., when the average tumor volume reached approximately 1000 mm³ or in accordance with ethical guidelines), all mice were euthanized. Tumors were carefully excised, weighed, and photographed. Subsequently, tumor tissues were divided into two parts: one part was fixed in 4% paraformaldehyde for paraffin embedding and subsequent IHC analysis, and the other part was snap-frozen in liquid nitrogen and stored at − 80 °C for Western blot analysis.

Hydroxyproline (HYP) quantification

Cells were hydrolyzed in 0.5 mL concentrated HCl at 95 °C for 6 h. After cooling, 0.5 mL Solution A and 0.25 mL Solution B were added (Elabscience, China), followed by dropwise addition of Solution B to adjust pH to 6.5–7.0. The volume was brought to 10 mL with distilled water, and 10 mg clarifying agent was added. After centrifugation at 1500 × g for 10 min, the supernatant was collected.

For the assay, 50 µL of standard or sample was mixed with 100 µL oxidizing working solution and incubated at 60 °C for 20 min. Then, 100 µL chromogenic working solution was added and incubated at 60 °C for another 20 min. Absorbance was measured at 558 nm, and HYP concentration was calculated from a standard curve.

Statistical analysis

Data are presented as mean ± standard deviation (mean ± SD). Comparisons between two groups were performed using Student’s t-test, while one-way analysis of variance (ANOVA) followed by an appropriate post hoc test was used for multi-group comparisons. Non-normally distributed data were analyzed using non-parametric tests (Mann-Whitney U test or Kruskal-Wallis test). All statistical analyses were performed using GraphPad Prism 8 software (GraphPad Software, USA), and image quantification was conducted using ImageJ software (National Institutes of Health, USA). All tests were two-sided, and a P-value < 0.05 was considered statistically significant. Significance levels are denoted as follows: ns (not significant), P ≥ 0.05; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Results

Overexpression of P4HA2 correlates with metastasis and shorter survival in HNSCC

As shown in Fig. 1A, analysis of data obtained from the TCGA database revealed that P4HA2 expression was significantly higher in HNSCC tissues compared to adjacent normal tissues. Survival analysis further demonstrated that patients with high P4HA2 expression had significantly shorter disease-specific survival (DSS) than those with low expression (Fig. 1B). To validate these findings, we performed transcriptome sequencing on five paired HNSCC and normal tissue samples, which confirmed significantly elevated mRNA levels of both P4HA2 and COL1A1 in tumor tissues (Fig. 1C). Western blot analysis of another five paired samples showed that protein expression levels of P4HA2 and COL1A1 were also significantly higher in cancerous tissues (Fig. 1D, E).

Fig. 1.

P4HA2 is Upregulated in HNSCC and Correlates with Poor Prognosis. (A) P4HA2 expression is significantly elevated in HNSCC tissues compared to adjacent normal tissues (n = 520, TCGA dataset). (B) Kaplan-Meier survival analysis shows that high P4HA2 expression (n = 234) is associated with shorter disease-specific survival (DSS) compared to the low expression group (n = 244) in patients with HNSCC. (C) Transcriptome sequencing reveals increased mRNA levels of P4HA2 and COL1A1 in tumor tissues (n = 5 pairs). (D, E) Western blot analysis confirms upregulation of P4HA2 and COL1A1 proteins in HNSCC tissues (n = 5 pairs). (F) Immunohistochemistry-based Kaplan-Meier analysis demonstrates that high P4HA2 expression predicts reduced overall survival in HNSCC patients (n = 108). (G) Representative immunohistochemical staining of P4HA2 in HNSCC clinical specimens, demonstrating negative (–), weak (+), moderate (++), and strong (+++) expression. Upper panels: low-power fields (scale bar, 500 μm). Lower panels: higher-magnification views of the boxed areas (scale bar, 100 μm). (*P < 0.05; **P < 0.01; ***P < 0.001)

We further examined the clinical significance of P4HA2 in 108 HNSCC specimens using immunohistochemistry. The positive rate of P4HA2 expression in HNSCC tissues was 66.67% (72/108), significantly higher than that in adjacent normal tissues (2/108). Representative images illustrating the spectrum of P4HA2 staining intensity are shown in Fig. 1G. To characterize the molecular context of our cohort, we also assessed p53 expression and HPV status. High p53 expression was observed in 86 cases (79.6%), while p16 positivity was detected in only 10 cases (9.3%), indicating that HPV-related HNSCC is uncommon in this cohort (Table 1, Figure S1A, B). Critically, Kaplan-Meier survival analysis indicated that high P4HA2 expression was significantly associated with shorter overall survival in HNSCC patients (Fig. 1F). Clinicopathological analysis revealed that P4HA2 expression was significantly correlated with tumor differentiation, lymph node metastasis, and perineural invasion, but not with patient gender, age, TNM stage, stromal invasion, or vascular emboli (Table 1). These clinical findings, combined with the established role of P4HA2 as a key enzyme in collagen synthesis and its association with poor outcomes in multiple malignancies, highlight the potential of P4HA2 as a prognostic biomarker and therapeutic target in HNSCC.

Table 1.

Correlation between P4HA2 expression and clinicopathological parameters in HNSCC

Clinical Feature	Total	P4HA2 Expression					z	P Value
		–	+		++	+++
Gender
Male	77	23		27	15	12	-0.980	0.327
Female	31	13		10	2	6
Age
≤ 60	76	25		28	11	12	-0.306	0.76
>60	32	11		9	6	6
Differentiation Degree
High	52	23		22	6	1	-3.903	<0.001***
Medium&Low	56	13		15	11	17
TNM Stage
Stage I - II	44	15		21	5	3	-1.681	0.093
Stage III - IV	64	21		16	12	15
Lymph Node Metastasis
No	71	27		29	9	6	-2.979	0.003**
Yes	37	9		8	8	12
Nerve Invasion
No	65	30		20	8	7	-3.512	<0.001***
Yes	43	6		17	9	11
Lymphatic Invasion
No	60	26		16	8	10	-1.668	0.095
Yes	48	10		21	9	8
Vascular Embolism
No	103	34		34	17	18	-1.078	0.281
Yes	5	2		3	0	0
p16 expression
Negative	98	29		35	16	18	-2.477	0.013
Positive	10	7		2	1	0
p53 expression
Low	22	11		5	4	2	-1.544	0.123
High	86	25		32	13	6

(**P < 0.01; ***P < 0.001)

Knockdown of P4HA2 inhibits proliferation, migration, and invasion of HNSCC cells

Combining both transient and stable knockdown approaches, the inhibitory effect of P4HA2 depletion on the proliferative capacity of HNSCC cells was consistently observed. Knockdown efficiency was confirmed at both transcriptional and protein levels at 48 h post-transfection (Fig. 2A, B). Transient siRNA-mediated knockdown of P4HA2 significantly suppressed cell proliferation, with reduced viability observed from 48 h onward (Fig. 2C). Consistently, stable P4HA2 knockdown using shRNA also attenuated clonogenicity and cell proliferation, as measured by CCK-8 growth curves, corroborating the transient knockdown results (Figure S3A). Colony formation ability was markedly impaired following P4HA2 reduction (Fig. 2D, Figure S2A, S3B, C). In wound healing assays, siRNA-transfected cells exhibited delayed migration (Fig. 2E, Figure S2B). Furthermore, Transwell assays demonstrated that P4HA2 knockdown strongly inhibited vertical migration and Matrigel invasion (Fig. 2F, G, Figure S2C).

Fig. 2.

Knockdown of P4HA2 Inhibits Malignant Phenotypes in HNSCC Cells., (A) P4HA2 mRNA levels are significantly reduced following siRNA-mediated knockdown. (B) Western blot analysis confirms decreased P4HA2 protein expression in knockdown cells. (C) CCK-8 assay demonstrates inhibited cell proliferation upon transient knockdown of P4HA2. (D) Colony formation assay reveals reduced clonogenic capacity in transient knockdown of P4HA2. (E) Wound healing assay shows impaired horizontal migration after transient knockdown of P4HA2. (F) Transwell migration and (G) invasion assays indicated that transient knockdown of P4HA2 suppressed both vertical migration and invasion capabilities. (*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001)

P4HA2 overexpression promotes proliferative and invasive phenotypes In Vitro and Tumor Growth In Vivo

P4HA2 was successfully overexpressed in HN6 and CAL27 cells (Fig. 3A). Overexpression of P4HA2 significantly promoted cell proliferation, as shown by increased viability and enhanced colony formation capacity (Fig. 3B, C, Figure S2D). Similarly, P4HA2 overexpression accelerated horizontal migration in wound healing assays (Fig. 3D, Figure S2E) and strengthened vertical migration and Matrigel invasion in Transwell assays (Fig. 3E, F, Figure S2F).

Fig. 3.

P4HA2 Overexpression Enhances Cell Proliferation, Migration, and Invasion in HNSCC Cells and Promotes Tumor Growth In Vivo (A) Western blot analysis confirms successful P4HA2 overexpression in HN6 and CAL27 cells. (B) CCK-8 assay shows enhanced proliferation in P4HA2-overexpressing cells. (C) Colony formation assay demonstrates increased clonogenic capacity upon P4HA2 overexpression. (D) Wound healing assay reveals accelerated horizontal migration in P4HA2-overexpressing groups. (E) Transwell migration and (F) invasion assays show enhanced vertical migration and invasion capabilities. (G) Representative images of nude mice and resected tumors from xenograft experiments. (H) Tumor growth curves (left) and final tumor weights (right) demonstrate that P4HA2 overexpression promotes tumor growth in vivo. (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001)

To validate these findings in vivo, a xenograft tumor model was established in nude mice. Tumors derived from P4HA2-overexpressing cells exhibited a significantly larger volume from day 18 onward (Fig. 3H, left) and a greater final tumor weight (Fig. 3H, right) compared to those from the control group. Representative images of the mice and resected tumors are shown in Fig. 3G. Collectively, these results demonstrate that P4HA2 functions as an oncogene by driving proliferation, migration, invasion, and tumor growth in HNSCC.

P4HA2 Regulates Collagen Synthesis and Fiber Architecture in HNSCC

Western blot analysis demonstrated that siRNA-mediated knockdown of P4HA2 significantly reduced COL1A1 protein expression in both HN6 and CAL27 cells compared to the NC group (Fig. 4A, Figure S4A). Conversely, P4HA2 overexpression markedly increased COL1A1 protein levels (Fig. 4B, Figure S4B). Immunofluorescence staining provided direct morphological evidence supporting these biochemical findings. Phalloidin staining (red) delineated cellular architecture, while collagen immunostaining (green) revealed substantially diminished fluorescence intensity and distribution area in P4HA2-knockdown cells relative to controls (Fig. 4C, Figure S4C). In contrast, P4HA2-overexpressing cells exhibited enhanced collagen deposition, as demonstrated by intensified green fluorescence signal and expanded distribution area (Fig. 4D, Figure S4D).

Fig. 4.

P4HA2 Regulates Collagen Synthesis and Fiber Architecture in HNSCC. (A) Western blot shows reduced COL1A1 protein levels following P4HA2 knockdown. (B) P4HA2 overexpression increases COL1A1 protein expression. (C) Immunofluorescence reveals diminished collagen deposition (green) in P4HA2-knockdown cells. Phalloidin (red) stains actin cytoskeleton. (D) Enhanced collagen deposition is observed in P4HA2-overexpressing cells. (E) HNSCC tissues were first classified into P4HA2‑high and P4HA2‑low groups by IHC. Left panels: low-power fields (scale bar, 500 μm). Right panels: higher-magnification views of the boxed areas (scale bar, 100 μm). SHG imaging of the same specimens reveals densely interwoven collagen fibers in P4HA2‑high tissues versus loose networks in P4HA2‑low tissues. (F) CT-fire analysis shows increased collagen fiber length and decreased width in P4HA2-high group. (G) No significant differences were observed in fiber straightness or angular distribution between the two groups. (H) Western blot analysis of xenograft tumors derived from HN6 cells confirms elevated P4HA2 and COL1A1 protein levels in the P4HA2‑overexpressing group compared with the control group. (ns P ≥ 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.)

Given that P4HA2 is an α‑ketoglutarate/Fe²⁺‑dependent dioxygenase that requires vitamin C as an essential cofactor for its enzymatic activity, we next examined whether exogenous vitamin C supplementation influences collagen production. As shown in Figure S5, vitamin C treatment alone significantly increased both COL1A1 protein levels and hydroxyproline content, confirming its critical role in collagen hydroxylation and maturation. P4HA2 overexpression also significantly increased both markers, consistent with the results in Fig. 4. No significant interaction was observed between the two factors, indicating that their effects are additive. These results confirm that while P4HA2 drives collagen production, optimal collagen maturation in vitro requires sufficient vitamin C availability.

Clinical validation using SHG imaging of six HNSCC specimens revealed distinct collagen organizational patterns. Tissues with high P4HA2 expression exhibited densely interwoven collagen fibers, whereas P4HA2-low tissues displayed loose, poorly organized collagen networks (Fig. 4E). Quantitative analysis using CT-fire software demonstrated that high P4HA2 expression was associated with significantly increased collagen fiber length and decreased width (Fig. 4F), although no significant differences were observed in straightness or angular distribution (Fig. 4G). Finally, Western blot analysis of xenograft tumors derived from HN6 cells confirmed that P4HA2 overexpression led to increased protein levels of both P4HA2 and COL1A1 in this model (Fig. 4H). Collectively, these results identify P4HA2 as a key regulator of collagen synthesis and architecture in HNSCC.

Hypoxia-Induced HIF-1α Directly Activates P4HA2 Transcription in HNSCC

To investigate the relationship between P4HA2 expression and hypoxic signaling, we first performed gene set enrichment analysis (GSEA) of TCGA data, which revealed significant enrichment of HIF-1 signaling pathway in P4HA2-high expressing tumors (Fig. 5A). Furthermore, P4HA2 mRNA expression showed positive correlations with hypoxia pathway genes and HIF-1α mRNA levels in the TCGA cohort (Fig. 5A).

Fig. 5.

Hypoxia-Induced HIF-1α Directly Activates P4HA2 Transcription in HNSCC. (A) GSEA analysis reveals significant enrichment of HIF‑1 signaling pathway in P4HA2‑high tumors (NES = 1.727, FDR < 0.001). Scatter plots show positive correlations of P4HA2 mRNA with hypoxia pathway genes (r = 0.218) and HIF‑1α mRNA (r = 0.357) in TCGA cohort. (B) Western blot demonstrates elevated HIF‑1α protein in HNSCC tissues compared to adjacent normal tissues. (C) Hypoxia upregulates HIF‑1α and P4HA2/COL1A1 protein expression in HN6 and CAL27 cells. (D) HIF‑1α knockdown under hypoxia reduces P4HA2 and COL1A1 expression. (E)Schematic of predicted HIF‑1α binding sites on the P4HA2 promoter. (F)Chromatin immunoprecipitation confirms hypoxia‑enhanced HIF‑1α binding to the P4HA2 promoter in HN6 and CAL27 cells. (G) IHC reveals co‑expression of P4HA2 and HIF‑1α in human HNSCC tissues. Left: representative images showing concurrent low or high expression of P4HA2 and HIF‑1α in adjacent sections of the same tumor region (scale bars, 100 μm). Right: bar plot showing the distribution of P4HA2 expression levels between HIF‑1α‑low and HIF‑1α‑high groups. Fisher’s exact test (OR = 18.13, P < 0.001). (H) Quantitative analysis confirms enhanced HIF‑1α expression in P4HA2‑overexpressing xenografts. (I) Representative immunohistochemical staining of HIF‑1α in xenograft tumors. Upper panels: low-power fields (scale bar, 500 μm). Lower panels: higher-magnification views of the boxed areas (scale bar, 100 μm). (nsP ≥ 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.)

Experimental validation demonstrated elevated HIF-1α protein levels in HNSCC tissues compared to adjacent normal tissues (Fig. 5B). Under hypoxic conditions, both HIF-1α, P4HA2 and COL1A1 protein expression were significantly upregulated in HN6 and CAL27 cells (Fig. 5C, Figure S6A). Importantly, HIF-1α knockdown under hypoxia substantially reduced P4HA2 and COL1A1 expression (Fig. 5D, Figure S6B), indicating HIF-1α’s essential role in regulating P4HA2.

Mechanistic investigation through chromatin immunoprecipitation revealed that HIF-1α directly binds to the P4HA2 promoter region. Using Cistrome Data Browser-predicted binding sites (Fig. 5E), we designed specific primers targeting the P4HA2 promoter. Hypoxic treatment enhanced HIF-1α binding by 15-fold in HN6 cells and 1.8-fold in CAL27 cells compared to normoxic conditions, with no significant enrichment in control regions (Fig. 5F).

Clinical validation in 108 HNSCC specimens showed that HIF-1α expression correlated with tumor differentiation status (Table 2). Representative immunohistochemistry images of sections from the same human HNSCC specimens revealed a consistent co‑expression pattern, with cases showing either high or low levels of both P4HA2 and HIF-1α (Fig. 5G, Table S5). Correlation analysis across all 108 HNSCC samples demonstrated a statistically significant positive association between HIF-1α and P4HA2 expression (Fisher’s exact test, OR = 18.13, P < 0.001), indicating that higher P4HA2 expression levels are significantly linked to HIF-1α‑high status (Fig. 5G). To functionally validate this clinical observation, we established xenograft models and found that tumors derived from P4HA2-overexpressing cells exhibited significantly enhanced HIF-1α protein stabilization compared to vector controls (Fig. 5H, I), supporting the association between P4HA2 expression and HIF-1α activation in the tumor microenvironment. Collectively, these findings from clinical specimens and in vivo models establish P4HA2 as a direct transcriptional target of HIF-1α and provide evidence linking hypoxic signaling to collagen remodeling in HNSCC progression.

Table 2.

Correlation between HIF-1α Expression and Clinicopathological Parameters in HNSCC

Clinical Feature	Total	HIF-1α Expression				χ²	P Value
		Low		High
Gender
Male	77	61		16		0.846	0.358
Female	31	22		9
Age
≤ 60	76	61		15		1.678	0.195
>60	32	22		10
Differentiation Degree
High	52	45		7		5.289	0.021*
Medium&Low	56	38		18
TNM Stage
Stage I - II	44	38		6		3.776	0.052
Stage III - IV	64	45		19
Lymph Node Metastasis
No	71	58		13		2.727	0.099
Yes	37	25		12
Nerve Invasion
No	65	53		12		2.016	0.156
Yes	43	30		13
Lymphatic Invasion
No	60	48		12		0.752	0.386
Yes	48	35		13
Vascular Embolism
No	103	78		25		1.579	0.209
Yes	5	5		0

(*P < 0.05)

Discussion

Collagen deposition and ECM remodeling are established hallmarks of the tumor microenvironment that drive cancer progression. In this study, we identified the collagen-modifying enzyme P4HA2 as a key contributor to this process in HNSCC. Analysis of TCGA data combined with clinicopathological validation revealed significant upregulation of P4HA2 in tumor tissues, where its elevated expression was significantly associated with aggressive pathological features including poor differentiation, lymph node metastasis, and perineural invasion. Clinically, high P4HA2 expression correlated markedly with shortened DSS, highlighting its potential as a robust prognostic biomarker and promising therapeutic target. Mechanistically, we elucidated a hypoxia-driven pathway wherein HIF-1α transcriptionally upregulates P4HA2 expression, leading to specific collagen architecture remodeling that ultimately promotes tumor proliferation, migration, and invasion. Collectively, our clinical and experimental findings demonstrate that P4HA2 plays a pivotal role in regulating HNSCC invasion and poor prognosis through ECM remodeling in the tumor microenvironment.

The pro-tumorigenic role of P4HA2 observed here aligns with previously reported findings in other malignancies [27] and more recently in HNSCC [28]. While these studies established an association between P4HA2 and HNSCC progression, the mechanistic link between P4HA2 and hypoxia-driven collagen remodeling remained unexplored. Here, we show that P4HA2 overexpression, often coordinated with its downstream product COL1A1, is associated with a dysregulated collagen metabolic program. Our data suggest that this ECM remodeling may not merely be a passive consequence but may actively contribute to malignancy. As an α‑ketoglutarate / Fe²⁺-dependent dioxygenase, P4HA2 requires vitamin C (ascorbate) as an essential cofactor for its enzymatic activity [29, 30]. Our supplementation experiments confirmed that exogenous ascorbate promotes collagen hydroxylation and maturation, underscoring the importance of considering cofactor availability when interpreting collagen‑related phenotypes [31]. A stiffened ECM provides structural support for tumors [32] and activates pro-survival and proliferative signaling cascades, such as the PI3K-AKT pathway, through integrin-mediated mechanotransduction [33, 34]. Importantly, collagen prolyl hydroxylation has been recognized as a broader regulatory mechanism that links metabolic cues to epigenetic remodeling and mesenchymal transition in cancer cells [35]. Our findings extend this concept by demonstrating that P4HA2-driven collagen remodeling in HNSCC is associated with tumor aggressiveness. Consistent with this, our functional assays showed that P4HA2 overexpression was associated with enhanced proliferative, migratory, and invasive capacities of HNSCC cells in vitro. This potential pro-tumorigenic role was further supported by in vivo experiments, in which P4HA2-overexpressing xenografts exhibited accelerated tumor growth in mice.

A key mechanistic insight from our work is the identification of HIF-1α as a direct transcriptional regulator of P4HA2 in HNSCC. The hypoxic core of solid tumors is a critical niche for progression, and our ChIP-qPCR data confirm that HIF-1α binds the P4HA2 promoter under low oxygen tension. Knockdown of HIF-1α suppressed P4HA2 and COL1A1 expression, establishing the “Hypoxia-HIF-1α-P4HA2” axis as a central adaptive response in HNSCC. This aligns with the established paradigm of hypoxia-inducible factor biology [36] and suggests a vicious cycle: hypoxia induces P4HA2, which remodels the ECM, potentially exacerbating hypoxia and further stabilizing HIF-1α activity [37, 38]. This feed-forward loop may be a critical driver of the metastatic cascade.

To further dissect how P4HA2 promotes tumor progression, we investigated its effect on collagen architecture. Using SHG microscopy and quantitative CT-Fire analysis, we discovered that P4HA2 overexpression generates a distinct ECM topology characterized by longer, thinner, and more densely packed collagen fibers while maintaining overall alignment. Our findings demonstrate that P4HA2-driven collagen remodeling reshapes the physical architecture of the ECM, a process that correlates with HNSCC progression. Notably, the resulting elongated and aligned collagen fibers are structurally reminiscent of tracks that could, in principle, serve as “highways” for directed cancer cell migration. We hypothesize that increased fiber length may provide extended guidance cues and adhesion sites [39, 40], while a reduction in fiber width could decrease physical resistance, thereby potentially allowing tumor cells to move more efficiently along these aligned structures [41]. These functional interpretations, while plausible based on prior literature, remain to be directly tested in the context of our model.

Beyond providing physical guidance for migration, the densely reorganized ECM could also hinder immune cell infiltration, as documented in other cancer types [42–44]. This offers a plausible ancillary mechanism of immune evasion consistent with P4HA2’s established pro-tumorigenic role. Moreover, P4HA2-induced ECM stiffening intensifies intra-tumoral hypoxia and nutrient deprivation, which sustains its own expression via HIF-1α activation. This establishes a self-reinforcing vicious cycle of “hypoxia-ECM stiffening-hypoxia” that progressively fuels tumor aggressiveness [45]. Intriguingly, a reciprocally regulatory mechanism has been reported in bladder cancer, where P4HA2 stabilizes HIF-1α protein, suggesting a potential positive feedback loop between P4HA2 and HIF-1α that may further amplify this vicious cycle [46]. Notably, the conserved straightness of collagen fibers amid alterations in other geometric parameters indicates that P4HA2 exerts precise, localized control over fiber architecture rather than causing global disorganization, consistent with its enzymatic function in stabilizing the collagen triple helix [47, 48].

In summary, our study delineates a HIF-1α-P4HA2-collagen remodeling axis as a critical driver of HNSCC progression. This pathway integrates hypoxic stimulus with ECM restructuring to generate a pro-tumorigenic microenvironment conducive to proliferation, migration, and invasion. While the current work offers substantial mechanistic insight, certain limitations remain. These include the need for validation in larger clinical cohorts and a more detailed dissection of the molecular mechanisms by which P4HA2 regulates downstream effectors such as COL1A1. Future research should therefore focus on the therapeutic targeting of the P4HA2-driven pathway, which may open new avenues for treating a cancer type still in urgent need of more effective strategies.

Conclusion

Our work establishes P4HA2 as a critical effector of hypoxia-driven collagen remodeling and tumor progression in HNSCC. We demonstrate that HIF‑1α directly transcriptionally upregulates P4HA2, which in turn promotes collagen deposition and extracellular matrix reorganization, thereby facilitating cancer cell proliferation, migration, invasion, and tumor growth. These findings provide mechanistic insight into hypoxic adaptation and nominate P4HA2 as a promising therapeutic target for disrupting the tumor-promoting microenvironment. Future studies evaluating P4HA2 inhibition in preclinical models hold significant potential for advancing treatment strategies against this aggressive malignancy.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Chaochao Zhang and Ruihuan Gan were responsible for study design, data collection, analysis, interpretation, and writing of the manuscript. Data collection was performed by Limei Wang, Qiuling Chen, and Yuxiang Yan. Dali Zheng and Youguang Lu contributed to study design. Ke Zheng contributed to study design, data interpretation, and writing of the manuscript. All authors were involved in writing the paper and had final approval of the submitted and published versions.

Funding

This work was supported by the Natural Science Foundation of Fujian Province (Grant No. 2023J01608) and the Scientific Research Project of National Key Clinical Specialty Construction (Grant No. 2022YBL-JB-04). The funding bodies had no role in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.

Data availability

All the data presented in this study and information supporting the results can be found in the article or in the supplementary information files and are available upon reasonable request from the corresponding authors.

Declarations

Ethical approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. This study involving human tissue samples was approved by the Institutional Ethics Committee of the First Affiliated Hospital of Fujian Medical University (Approval No. MRCTA ECFAH of FMU [2022]457). All animal procedures were conducted in accordance with the China Animal Welfare Legislation and were approved by the Ethics Committee of Fujian Medical University (Approval No. IACUC FJMU 2023-Y-0977).

Competing interests

The authors declare no competing interests.