CEDT

Introduction

Infantile Haemangioma (IH) affects 10-12% of infants and is the most common benign vascular tumour in childhood, characterised by rapid proliferation in the first year of life followed by gradual regression [1-3]. Risk factors include female sex (3:1 female to male ratio), low birth weight, prematurity, Caucasian ethnicity, maternal factors, and a family history of haemangiomas [4-7]. Although IH involutes spontaneously, 10% develop complications such as scarring, ulceration, tissue distortion and functional problems [4,7-11]. The pathogenesis of IH is not fully understood although it has been characterised to involve dysregulated angiogenesis, hypoxia-driven signalling, NOTCH and PI3K/ Akt/mTOR signalling pathways [12-14] . During the proliferating phase, endothelial cell differentiation and migration are regulated by pro-angiogenic factors such as VEGFR2, ANGPT2 and HIF-1α, with downstream MEK/ERK, PI3K/Akt/mTOR, PKC signalling further supporting growth [14-22].  This rapid formation of vascular network slows down in the plateau or quiescence phase. Involution is marked by adipogenesis, apoptosis, and lipid deposition, driven by PPARγ/CEBP and IGF-1 pathway, while suppressing endothelial survival and angiogenesis [12,23-25].  Current treatment approaches depend on location and size of the tumour with the risk of complications, and may include surgery, laser therapy, and β-blockers such as propranolol [1,8,9,26-29].  Propranolol has revolutionised IH management due to its ability to stop growth [21,30-34] . However, its long-term effects on vascular and metabolic development remain unclear, with poor response or relapse in females and severe lesions [7,35-39].

Mechanistic studies implicate regulation of SOX-18-mevalonate pathway, glycolysis and fatty acid oxidation [24,40-47].  The pronounced female predominance suggests hormonal modulation in IH development [48-52] . Metabolic profiling of haemangioma endothelial cells revealed elevated amino acid metabolism [54]. Given these hormonal and pathway alterations in IH, a comprehensive understanding of the proteomic landscape will improve our knowledge of its pathogenesis to fully characterise the underlying molecular landscape. Proteomic profiling, including mass-spectrometry-based approaches, will enable identification of proteins and associated post-translational modifications (PTMs), revealing alterations in differential signalling pathways. In this study, we investigate the proteomic changes across the developmental phases of IH in both sexes with healthy skin samples as controls for differential analysis. Significantly altered proteins were mapped and validated using western blot and PCR array. This represents the most extensive proteomic analysis of IH to date, providing insight to phase- and sex-based molecular differences, and associated pathways that may serve as promising therapeutic targets or biomarkers in IH.

Materials and Methods

Patient samples and study design

This case-control study involved proteomic profiling of IH patient tissue samples and healthy skin controls, collected at Centre for the Study & Treatment of Vascular Birthmarks (Lower Hutt, New Zealand) between 2015-2023. Snap-frozen samples were sourced from Gillies McIndoe Research Institute Biobank, New Zealand in accordance with the ethical standards, approved by the Wellington Ethics Committee (Ref.13/CEN/130). All procedures involving human subjects adhered to the principles of respect, confidentiality and ethical guidelines. Clinical and histological details are presented in Supplementary Table 1. IH samples were grouped as proliferating (0-10 months) (n = 12), quiescent or plateau (1218 months) (n = 8), involuting (18 months-4 years) (n = 8), and normal skin (8-11 years) (n = 8) with equal males and females. 

Protein extraction and digestion

Snap-frozen tissue samples were grounded in liquid nitrogen and lysed in 5% Sodium dodecyl sulphate (Sigma-Aldrich) buffer and protease inhibitor (Thermo Fisher). Lysates were homogenised, sonicated and quantified by BCA assay (Pierce BCA protein kit, Thermo Fisher). Protein reduction, alkylation, and on-column tryptic digestion were performed on S-Trap Micro spin column (Protifi) using established protocols [76,77].  Peptides were eluted, dried and reconstituted for analysis.

Quantification of protein and functional pathway analysis

Peptides were analysed by Liquid chromatography tandem mass spectrometry (LC-MS/MS) and searched against an in silico tryptic digest of Homo sapiens proteins from the UniProtKB/Swiss-Prot + TrEMBL sequence database (version August 2024). Proteins were detected with 95% probability (Protein False Discovery Rate = 0.1%) in Spectronaut (Biognosys), and label-free quantification (LFQ) performed using normalised peptide intensities. LFQ values were log2 transformed and filtered (≥ 70% valid values) in Perseus (Maxquant) [76] . Missing values were imputed from a normal distribution, and differential expression assessed via unpaired Student’s t-test (p ≤ 0.05; fold change ≥ 1.2). Ingenuity Pathway analysis (IPA) (Qiagen) identified enriched biological networks and predicted activation states (Fisher’s exact test, p ≤ 0.05).

RNA extraction and PCR array

Total RNA was extracted from proliferating IH (n = 8), involuting IH (n = 6) and normal skin (n = 6) samples using RNeasy kits (Qiagen). cDNA was synthesised using the RT2 First Strand kit (Qiagen) and profiled with the Human Lipoprotein Signalling and Cholesterol Metabolism RT2 Profiler PCR array (Qiagen). Ct values were normalised to housekeeping genes and analysed using ΔΔCt method.

Western blotting

Lysates from proliferating (n = 8) and involuting IH (n = 6) were separated on 8% or 4-12% Bis-Tris gels (Thermo Fisher), transferred to a nitrocellulose or polyvinylidene difluoride membrane (BioRad), and probed with primary antibodies: OXPHOS complexes (#ab110411, Mouse, 1:1000, Abcam), MAOA (#ab126751, Rabbit, 1:1000, Abcam), Na+K+ATPase (#23565, Rabbit, 1:1000, Cell Signalling Technologies), VDAC (#4661, Rabbit, 1:1000, Cell Signalling Technologies), FAS (#3180T, Rabbit, 1:1000, Cell Signalling Technologies), AMPK (#2532, Rabbit, 1:1000, Cell Signalling Technologies), followed by HRP-conjugated secondary antibodies. Bands were visualised using chemiluminescence and quantified using ImageLab/ ImageJ software. Uncropped blots are provided in (Supplementary Figures 1-2). 

Data analysis

Statistical analyses were performed with GraphPad Prism and Microsoft Excel. Two-group comparisons employed unpaired Student’s t-test or Mann-Whitney tests; multiple group comparisons used one-way ANOVA or Kruskal-Wallis test.

Statistical significance was defined as p ≤ 0.05.

Results

Protein identification

Mass-spectrometry-based proteomic analysis was performed to compare protein expression across the IH samples spanning the proliferating, plateau or quiescent and involuting phases, alongside normal skin controls. Spectral data analysed with Spectronaut identified 7977 proteins, followed by normalisation, log-transformation and imputation in Perseus. 5585 proteins were then analysed using IPA to examine the biological and molecular interactions (Supplementary Figure 3). Principal Component Analysis (PCA) revealed a clear distribution of IH and normal skin samples (Figure 1a). Venn diagram indicated substantial overlap of proteins across IH phases (Figure 1b).

Figure 1: Comparative analysis of the differential protein expression between IH and normal skin; (a) Principal Component Analysis shows clustering of IH samples by clinical phase - proliferating (red), plateau (green) and involuting phase (pink) compared to normal skin (blue); (b) Venn diagram depicting common and unique proteins identified across the three phases of IH. Volcano plot highlights the differential protein expression in proliferating phase; (c), plateau ;(d) and involuting phase ;(e) compared to normal skin. Proteins with a fold change ≥ 1.2 and p-value ≤ 0.05 were considered statistically significant (unpaired student’s t-test).

 Comparison of protein changes in Infantile Haemangioma and normal skin

Volcano plots showed the distribution of upregulated and downregulated proteins in each phase (Figure 1c-1e). Differential expression analysis used a threshold of fold change ≥ 1.2; p ≤ 0.05. Canonical Pathway Analysis (IPA) identified activated pathways (z scores from -5.831 to 8.310; p ≤ 0.05) (Figure 2). LXR/RXR activation, DHCR24 signalling pathway, xenobiotic metabolism (PXR and AHR), complement cascade and peroxisomal protein import were downregulated in proliferating and plateau phases but not involuting, suggesting early suppression of cholesterol biosynthesis, glutathione-mediated detoxification and activation of gene expression by SREBP. Glucose and triglyceride metabolism were increased in the plateau phase, whereas the apelin adipocyte signalling pathway, matrix metalloproteinase activation, gluconeogenesis and insulin receptor signalling were reduced in proliferating IH (Figure 2).  

Figure 2: Heatmap of top significantly expressed canonical pathways under Core Analysis in Ingenuity Pathway Analysis in the phases of IH – proliferating, plateau and involuting, compared to normal skin, p-value ≤ 0.05 (Fisher’s exact test).

Sex-specific pathway differences across IH development

Comparative analyses (fold change ≥ 1.2; p ≤ 0.05) examined molecular differences across sexes (Supplementary Table 2-5). LXR/RXR activation and the DHCR24 signalling pathway were downregulated in female proliferating versus involuting phase (Figure 3a). Key proteins, A1BG, AHSG, ALB, AMBP, APOB, APOC3, APOD, APOH were decreased (Supplementary Table 2), suggesting reduced lipoprotein and cholesterol homeostasis. Collagen biosynthesis and mitochondrial dysfunction were upregulated in the proliferating phase (Supplementary Table 2, Figure 3a), while plateau phase displayed activation of VEGF, RHOA, integrin, and actin cytoskeleton signalling (Figure 3b, Supplementary Table 3), consistent with persistent angiogenesis. Hence, the plateau phase is very similar to proliferating phase in females leading to a slow transition to the involuting phase. Oxidative phosphorylation (OXPHOS) complexes I (NDUFA4L2) and IV (COX4L2, COX7A2L) were selectively upregulated in proliferating female IH while MAOA and Na+K+ATPase decreased (Supplementary Table 2). 

LXR/RXR activation and DHCR24 signalling pathway were similarly downregulated in male proliferating and plateau phases versus involuting (Figure 3c-3d, Supplementary Table 4-5). DNA damage/telomere stress induced senescence were activated in male plateau IH demonstrating transition to involution (Figure 3d, Supplementary Table 5). Downstream biological processes indicated higher cell viability, migration, invasion and cellular homeostasis in female lesions compared to males (Figure 3e), suggesting aggressive growth and slow phase transitions in female IH.

Figure 3: Pathway analyses and functional annotation in IH, using IPA; Bar chart displays the top enriched canonical pathways identified through core analysis in proliferating phase in females; (a), plateau in females; (b), proliferating phase in males; (c), plateau in males; (d) compared to the involuting phase. Pathways are ranked based on -log(p-value) ≥ 1.3 (p-value ≤ 0.05); (e) Heatmap of comparison analysis of downstream biological processes in the female and male proliferating tumours, p-value ≤ 0.05 (Fisher’s exact test). Colour coding represents z scores: orange for activation (positive), white for no predicted change (0) and blue for inhibition (negative).

Validation by Western blotting

Proteins involved in mitochondrial dysfunction and lipoprotein signalling - OXPHOS complexes (I-V), MAOA, Na+K+ATPase, VDAC, FAS and AMPK were validated. OXPHOS complexes I and IV were upregulated in the female proliferating phase versus involuting (Figure 4a-4b, Supplementary Table 6); complex II and III reduced in female involuting IH, while complex V elevated in male involuting IH. MAOA expression rose from proliferation to involution (male, p = 0.004; female, p = 0.06) (Figure 4a4b). Na+K+ATPase expression was lower in females than males during both phases (proliferating, p = 0.1; involuting, p = 0.04). AMPK was significantly upregulated in the involuting phase in males compared to females (p = 0.01). FAS non-significantly increased in male involuting IH while VDAC was stable (Figure 4b, Supplementary Table 6).

Gene expression of cholesterol metabolism and Lipoprotein signalling

 LXR/RXR activation and DHCR24 signalling pathway are key regulators of cholesterol metabolism. 54 genes were differentially expressed in the Human Lipoprotein Signalling and Cholesterol metabolism PCR array in proliferating and involuting phases of IH versus normal skin (Supplementary Table 7), with 19 genes significantly altered (fold-change ≥ 1.5, p ≤ 0.05) between proliferating and involuting IH (Supplementary Table 8, Figure 4c).

They were further classified as, LDL associated proteins and receptors (ANKRA2, APOD, CEL, LRP10, LRP1B, STAB2), cholesterol homoeostasis (ABCG1, ANGPTL3, HDLBP, INSIG2, LCAT, PPARD, SREBF1) and cholesterol biosynthesis (ACAA2, CNBP, IDI2, NSHDL, PRKAA1, PRKAG2) (Figure 4c). Several sex-specific differences were observed in the proliferating and involuting phases as well, suggesting differential cholesterol regulation across IH development (Figure 4d-4e).

Figure 4: Validation of proteins using Western Blot and gene expression profiling in proliferating and involuting IH; (a) Western Blot analysis of FAS, Na+K+ATPase, AMPK, MAOA, OXPHOS Complexes I-V, VDAC across representative proliferating and involuting phase in female and male patients. β-actin was used as loading control; (b) Box and whisker plot representing Western Blot quantification of MAOA, OXPHOS Complexes I-V, Na+K+ATPase, FAS, VDAC, AMPK normalised to β-actin. Each box represents the interquartile range (Q1, 25th to Q3, 75th percentiles), the central horizontal line is the median value, and the whiskers represent the range. Statistical significance is shown by, *p-value ≤ 0.05, **p-value ≤ 0.01. Proliferating IH, n = 10 (5 male, 5 females), Involuting IH n = 8 (4 male, 4 females); (c) Bar graph shows the fold regulation of genes associated with LDL receptor, cholesterol biosynthesis, and homeostasis in the proliferating and involuting IH. Clinical details of the IH samples