Multi-omic analysis of human PHACTR1 signaling networks

Kathryn Wolhuter; Lijiang Ma; Nicole S Bryce; Osvaldo Contreras; Natalie Mellett; Ling Zhong; Chris Thekkedam; Siiri E Iismaa; Corey Giles; Richard P Harvey; Thomas Hennessy; Chris Fouracre; David Bradley; Peter J Meikle; Johan L M Björkegren; Jason C Kovacic

doi:10.1038/s42003-026-09542-w

. 2026 Jan 19;9:265. doi: 10.1038/s42003-026-09542-w

Multi-omic analysis of human PHACTR1 signaling networks

Kathryn Wolhuter ^1,², Lijiang Ma ³, Nicole S Bryce ^1,², Osvaldo Contreras ^1,², Natalie Mellett ⁴, Ling Zhong ⁵, Chris Thekkedam ¹, Siiri E Iismaa ^1,², Corey Giles ⁴, Richard P Harvey ^1,^2,⁶, Thomas Hennessy ⁷, Chris Fouracre ⁷, David Bradley ⁷, Peter J Meikle ^4,⁸, Johan L M Björkegren ^3,^9,^10,¹¹, Jason C Kovacic ^1,^2,^3,^✉

PMCID: PMC12913993 PMID: 41554990

Abstract

Genetic studies have linked PHACTR1 to a range of vascular diseases, underscoring its pivotal role in vascular biology. However, the full spectrum of PHACTR1-mediated signaling pathways remains largely unexplored. To bridge this gap, we employ a multi-omics pipeline combining pairwise differential expression analysis, multi-omics pathway integration, and feature-level correlation analyses across four distinct omics datasets to map the global signaling networks driven by PHACTR1. By integrating transcriptomic, proteomic, metabolic, and lipidomic profiles from human HT1080 cells with PHACTR1 overexpression or knockdown, and then validating key findings in primary human endothelial cells, here we show that PHACTR1 exerts broad control over fundamental cellular processes beyond cytoskeletal regulation. We demonstrate that PHACTR1 governs cell cycle progression, validating that increased expression alters key regulatory proteins. We also uncover a distinct function in iron metabolism, showing PHACTR1 regulates essential cellular iron-storage proteins and identify the PHACTR1 protein within the mitochondria where it directs morphology and bioenergetics through a signaling axis involving AKAP1 and Drp1. These mitochondrial changes align with observed shifts in lipid metabolism and correlations in human arterial tissue. These findings provide a systems-level blueprint of PHACTR1 function, revealing how this gene influences vascular health and offering potential targets for therapeutic intervention.

graphic file with name 42003_2026_9542_Figa_HTML.jpg

Subject terms: Cell signalling, Systems analysis

Integrated multi-omics analysis uncovers the role of the vascular disease-associated gene PHACTR1 in regulating mitochondrial dynamics, iron metabolism, and cell cycle.

Introduction

PHACTR1, which encodes the phosphatase and actin regulator 1 (PHACTR1) protein, plays a critical role in several vascular pathologies. Multiple genome-wide association studies (GWAS) have identified strong disease associations with the single nucleotide polymorphism (SNP) rs9349379, located within the PHACTR1 gene on chromosome 6p24.1¹. This SNP acts as an expression quantitative trait locus, modulating the expression of PHACTR1. Lower levels of PHACTR1, associated with the G/G genotype at rs9349379, correlate with an increased risk of coronary artery disease², coronary artery calcification³, and myocardial infarction⁴. In contrast, the A/A genotype, which correlates with higher PHACTR1 expression, is associated with migraine headaches⁵, fibromuscular dysplasia⁶, spontaneous coronary artery dissection⁷, cervical artery dissection⁸ and hypertension⁹. Notably, no other gene has been implicated in such a diverse range of vascular diseases, underscoring PHACTR1’s central role in vascular biology. These findings highlight PHACTR1 as a key genetic determinant of vascular diseases, reinforcing its pivotal role in human pathology.

Functionally, at a protein level, PHACTR1 is known to regulate cell motility and morphology through interactions with actin and protein phosphatase 1 (PP1)¹⁰. This regulatory activity has been linked to changes in cytoskeletal organization^11,12, matrix remodeling⁶, and angiogenesis¹³. Despite these insights, critical questions remain about how PHACTR1 exerts such a central role in cell function. We hypothesize that PHACTR1 influences additional aspects of cellular signaling that contribute to its wide-reaching impact on vascular disease. As such, the complete range of PHACTR1’s signaling responses is yet to be fully understood. To address this knowledge gap, here we take a discovery-based approach, utilizing integrative multi-omics to uncover the full extent of PHACTR1-mediated signaling in cells.

Integrative multi-omic technologies are revolutionizing our ability to explore the intricate molecular landscapes that govern biological systems. By concurrently analyzing multiple molecular layers, including transcriptomics, proteomics, metabolomics, and lipidomics, these approaches provide an unprecedented view of the complex interactions between genes, proteins, metabolites, and lipids¹⁴. This comprehensive analysis allows researchers to decode the dynamic interplay within cells, enabling the identification of key molecular targets and signaling pathways that drive biological processes and diseases^15,16. However, despite their potential, to date, the integration of multi-omics data is fraught with challenges that limit its broader application.

A primary challenge is the lack of reliable methods for integrating data across diverse ‘omics layers¹⁷. Many current approaches rely on simplistic techniques, such as concatenating raw data or using latent features, which often fail to capture the nuanced relationships among molecular components. These limitations become particularly pronounced when integrating more than two ‘omics datasets, often leading to outputs with limited biological relevance¹⁸. To overcome these barriers, we developed a multi-omic integration pipeline designed to identify critical biological pathways. This pipeline provides a robust framework for integrating diverse molecular data, focusing on pathway-level changes for a more biologically meaningful interpretation of multi-omics data, and ultimately facilitates the discovery of key regulatory mechanisms in complex diseases.

In this study, we generated high-coverage multi-omics datasets and utilized a streamlined integration pipeline to analyze the cellular and signaling changes resulting from PHACTR1 overexpression and knockdown, mirroring the A/A vs G/G effects of the rs9349379 SNP on PHACTR1 expression in humans. We hypothesized that these genetic manipulations would reveal distinct signaling pathways associated with PHACTR1, shedding light on its role in vascular disease.

By leveraging our multi-omic integration approach (Fig. 1A), we present new insights into PHACTR1’s regulatory mechanisms, paving the way for the development of targeted therapies.

Results

To examine the role of PHACTR1 in the flow of complex biological signals, we generated lentiviral constructs to overexpress (OE) and knockdown (KD) PHACTR1 in HT1080 fibrosarcoma cells (A/G genotype at rs9349379 SNP), generating stable cell lines expressing these constructs. These cells were chosen as our model system for several critical reasons that were paramount for the initial comprehensive multi-omics discovery phase of this study. Firstly, the relevance of PHACTR1 in specific vascular cell subtypes remains unclear, making a versatile and well-characterized cell line a suitable initial model. Secondly, HT1080 cells exhibit modest endogenous expression levels of PHACTR1, providing an ideal dynamic range for both overexpression and knockdown experiments. Most importantly, their robust growth characteristics and well-established amenability to lentiviral transduction were particularly beneficial for the efficient and consistent generation of stable cell lines. The creation of stable PHACTR1 overexpression and knockdown cell lines was essential to ensure consistent, long-term perturbation of PHACTR1 expression across the multiple biological replicates and the large quantities of cells required for four distinct omics datasets. This approach minimized the experimental variability that would be inherent with transient transfection methods.

Successful generation of PHACTR1 OE and KD lines was confirmed by qPCR, with a significant increase and decrease in PHACTR1 expression after transduction with their respective viral constructs compared to scrambled siRNA control (scrm ctrl) cells (Fig. 1B). As expected from previous studies¹⁹, phalloidin staining of these PHACTR1 OE and KD cell lines revealed alterations in their actin cytoskeleton (Fig. 1C). Characterization of the cell lines identified that 2-dimensional cell and nuclear area were significantly increased in PHACTR1 OE compared to scrm ctrl cells (Supplementary Fig. 1A, B), with these cells also having a marginal, but significant, decreased rate of migration (Supplementary Fig. 1C). PHACTR1 KD cells had no change in 2-dimensional cell area but significantly larger nuclei area compared to scrm ctrl cells with a reduced rate of migration (Supplementary Fig. 1A–C). There was no change in cell proliferation, scratch wound healing, adhesion or lamellipodia formation in either cell line compared to scrm ctrl cells (Supplementary Fig. 1D–I).

The PHACTR1 OE and KD cells, along with scrm ctrl cells, were then analyzed through our multi-omic pipeline (Fig. 1A). To ensure consistency of our experiments and results, we generated each replicate for all four ‘omics datasets in parallel using identical passage cells, grown in the same culture media, and harvested concurrently.

PHACTR1 levels control the expression of multiple downstream transcripts

To determine the effect of PHACTR1 expression on gene transcription, we performed bulk RNA-seq analysis on HT1080 cells, resulting in a comprehensive dataset encompassing 14,737 gene transcripts (Supplementary Data 1). Principal component analysis (PCA) showed distinct clustering of biological groups (Fig. 2A). Significant OE and KD of PHACTR1 were confirmed within the resulting transcriptomic dataset, identifying PHACTR1 as the most significantly up- and down-regulated gene in their respective cell lines when compared to scrm ctrl cells (Fig. 2B, D).

Fig. 2 — A PCA plot of *PHACTR1* KD, *PHACTR1* OE and scrm ctrl transcriptomes, n = 6. B Volcano plot of transcriptome in *PHACTR1* OE cells vs scrm ctrl. Significantly upregulated genes are shown in red and downregulated genes in blue. The full list of differentially expressed genes is found in Supplementary Data 2. C Top 10 significantly up- (red) and down-regulated (blue) gene pathways in *PHACTR1* OE cells from Ingenuity Pathway Analysis (IPA). The full list of significantly changed pathways is found in Supplementary Data 5. D Volcano plot of transcriptome in *PHACTR1* KD cells vs scrm ctrl. Significantly upregulated genes are shown in red and downregulated in blue. The full list of differentially expressed genes is found in Supplementary Data 3. E Top 10 significantly up- (red) and down-regulated (blue) gene pathways in *PHACTR1* KD cells from IPA. The full list of significantly changed pathways is found in Supplementary Data 8. F PCA plot of *PHACTR1* KD, *PHACTR1* OE and scrm ctrl proteomes, n = 6. G Volcano plot of the proteome in *PHACTR1* OE cells vs scrm ctrl. Proteins showing significantly increased abundance are shown in red and those showing decreased abundance in blue. The full list of proteins showing altered abundance is found in Supplementary Data 11. H Top 10 significantly increased (red) and decreased (blue) protein IPA pathways in *PHACTR1* OE cells. The full list of significantly changed pathways is found in Supplementary Data 12. I Volcano plot of proteome in *PHACTR1* KD cells vs scrm ctrl. Proteins showing significantly increased abundance are shown in red and those showing decreased abundance in blue. The full list of proteins showing altered abundance is found in Supplementary Data 13. J Top 10 significantly increased (red) and decreased (blue) protein IPA pathways in *PHACTR1* KD cells. The full list of significantly changed pathways is found in Supplementary Data 14.

Consistent with the discovery approach of this study, we aimed to capture the diversity of features altered by PHACTR1 expression and therefore used relatively relaxed stringency (FDR < 0.1) to identify differentially abundant features. This resulted in PHACTR1 OE cells having significantly increased the expression of 4672 gene transcripts and decreased expression of 4470 transcripts in comparison to scrm ctrl cells (Fig. 2B, and Supplementary Data 2). Similarly, PHACTR1 KD also significantly altered the transcriptome, albeit to a lesser extent, with 484 upregulated and 809 downregulated gene transcripts identified (Fig. 2D, and Supplementary Data 3).

Exploring options for feature mapping to biological pathways

A major challenge when mapping ‘omics features to biological pathways is selecting the most appropriate mapping tools. We used our transcriptomic dataset to test several common pathway mapping platforms, including Gene Ontology (GO) Cellular Component (GO:CC), Molecular Function (GO:MF), Biological Process (GO:BP), Ingenuity Pathway Analysis (IPA), the Kyoto Encyclopedia of Genes and Genomes (KEGG), Reactome, and WikiPathways (Supplementary Fig. 2, and Supplementary Data 4–9). Each of these databases has unique features suited to different types of research, with KEGG and Reactome being popular choices for pathway analysis, GO for functional annotation, IPA for deeper regulatory network insights, and WikiPathways for flexible, community-driven pathway data.

Although mapping the gene transcripts against GO yielded the greatest number of significantly altered terms (Supplementary Data 3 and 6), these were predominantly in GO:CC, which does not directly link to biological pathways, only cellular localization of the pathway components. Critically for our multi-omics integration goal, GO, including GO:MF, does not currently support the mapping or integrated analysis of metabolomic or lipidomic features alongside transcriptomic and proteomic data. This limitation makes GO unsuitable for achieving our primary objective of identifying conserved pathways across all four omics layers. GO analysis was therefore not used for further ‘omics mapping.

The remaining mapping tools are all capable of analyzing all our features of interest (genes, proteins, metabolites, and lipids). Among these, only IPA, KEGG, and Reactome offered comprehensive support across all four of our omics databases, making them suitable for true multi-omics integration. KEGG, Reactome and WikiPathways successfully identified significantly changed pathways in all comparisons, with the exception of upregulated pathways with PHACTR1 KD (Supplementary Fig. 2). However, pathway mapping by IPA returned pathways with a higher significance than KEGG, Reactome or WikiPathways and identified a greater number of up- and down-regulated signaling pathways in both PHACTR1 OE (Supplementary Fig. 2A, B) and KD (Fig. 2C, D) comparisons. This is perhaps unsurprising as IPA is an extensively curated database, the power of which is shown here by its outperforming other widely used tools in identifying a robust set of biologically relevant pathways for our specific experimental perturbations. From this point forward, pathway mapping across all ‘omics datasets was performed using only IPA (Fig. 2C, E).

The impact of PHACTR1 expression reaches across all ‘omics

We conducted untargeted proteomics analysis on fractionated protein digests from our HT1080 cell lines, identifying a total of 5666 proteins (Supplementary Data 10). PCA showed distinct clustering of biological groups (Fig. 2F). PHACTR1 OE significantly increased the abundance of 558 proteins and decreased the abundance of 797 (Fig. 2G, and Supplementary Data 11). These changes were associated with important alterations in biological signaling pathways, as shown by IPA (Fig. 2H, and Supplementary Data 12). Notably, pathways involved in gene transcription and cell cycle progression were significantly upregulated (Fig. 2H, in red), while pathways related to translation initiation and mitotic metaphase were downregulated (Fig. 2H, in blue). PHACTR1 KD resulted in an increased abundance of 338 proteins and a decreased abundance of 432 (Fig. 2I, and Supplementary Data 13). PHACTR1 KD was linked to IPA-identified pathway changes (Fig. 2J, and Supplementary Data 14), including increased nonsense-mediated decay (Fig. 2J, in red) and decreased Eukaryotic Initiation Factor 2 (EIF2) signaling (Fig. 2J, in blue).

To further explore the impact of PHACTR1 expression on cellular composition, we performed untargeted metabolomic analysis and targeted lipidomic analysis (Fig. 3). To enhance the detectability of metabolites, we employed two analytical liquid chromatography (LC) columns: hydrophilic interaction liquid chromatography (HILIC) method with higher affinity for hydrophobic molecules (Supplementary Data 15) and a reversed-phase (RP) method for detecting hydrophilic metabolites (Supplementary Data 18), which detected a total of 1752 unique features. PCA showed good clustering of PHACTR1 OE and scrm ctrl samples, with higher variance in the PHACTR1 KD samples from HILIC analysis (Fig. 3A). No clear clustering by experimental group was observed in the PCA from RP LC-MS (Fig. 3B). Recognizing the exploratory and unsupervised nature of PCA, we continued with supervised statistical methods to specifically test for significant differences in metabolites between groups.

In PHACTR1 OE cells, metabolomic analysis revealed 192 significantly altered features on HILIC (Fig. 3C, and Supplementary Data 16) and 9 on the RP method (Fig. 3D, and Supplementary Data 19). All 9 of the significantly changed features detected using the RP method were also detected using the HILIC method giving a total of 192 significantly changed features in PHACTR1 OE cells. Only features that could be spectrally validated at the MS2 level and had an associated KEGG or Human Metabolome Database (HMDB) ID were taken forward for IPA mapping, resulting in 118 mapped metabolites of the 192 total significantly changed features. Of these, 73 showed increased abundance and 45 decreased abundance when comparing PHACTR1 OE to scrm ctrl cells. IPA mapping of these metabolites revealed significant upregulation in pathways such as glycine and arginine degradation, and downregulation of tryptophan and choline catabolism (Fig. 3E, and Supplementary Data 21). In contrast, untargeted metabolomics of PHACTR1 KD cells detected only three features that showed reduced abundance: L-leucine, L-arginine, and 7-Oxoheptanoic acid (Fig. 3F, G, and Supplementary Data 17 and 21). Due to this limited number of significantly altered features, metabolite pathway analysis was not performed for PHACTR1 KD cells.

Targeted lipidomic analysis (Supplementary Data 22) showed distinct clustering of biological groups by PCA (Fig. 3H) and revealed changes across major lipid classes in both PHACTR1 OE and KD cells compared to scrm ctrl cells (Fig. 3I). Interestingly, lipid abundance generally trended in the same direction for both experimental conditions (e.g., free fatty acids increased in both PHACTR1 OE and KD cells), except for hydroxylated acylcarnitines, which increased in PHACTR1 OE cells and decreased in PHACTR1 KD cells. This was reflected in the lipid reaction network analysis that revealed notable changes in pathway activity, including the activation of acylcarnitine (AC) formation and suppression of free fatty acids (FA) and cholesterol esters (CE) in PHACTR1 OE cells (Fig. 3J) and increased phosphatidylethanolamine (PE) formation and suppression of Hex2Cer in PHACTR1 OE cells (Fig. 3M).

At the species level, of the 737 lipid species measured, 338 showed significant changes after PHACTR1 OE, with 117 species increased and 221 decreased (Fig. 3K, and Supplementary Data 23). IPA pathway analysis identified a subset of significantly upregulated and downregulated pathways following PHACTR1 OE (Fig. 3L, and Supplementary Data 24). In PHACTR1 KD cells, the abundance of 242 lipids was significantly altered compared to controls, with 27 species increased and 215 species reduced (Fig. 3N, and Supplementary Data 25). Consistent with the low number of increased lipids in PHACTR1 KD cells, IPA identified only downregulated pathways (Fig. 3O, and Supplementary Data 26). Whilst IPA analysis of lipidomics identified significantly changed pathways of interest, due to limitations in lipid species mapping tools (HMDB, KEGG or PubChem) only 25% of significantly changed lipids were mapped to IPA pathways.

Overall, consistent with its role in numerous diseases, these findings demonstrate that PHACTR1 gene expression levels exert profound and complex effects on downstream cell signaling that include levels of proteins, lipids and metabolites, as well as other transcripts. Moreover, numerous pathways of important biological relevance are affected by PHACTR1.

Multi-omic correlations reveal key nodes of interest for PHACTR1 signaling

Correlations between different ‘omics layers (e.g., genes and lipids) can reveal how changes at one molecular level cascade through the system to influence levels of other ‘omics. For example, a protein-coding gene that regulates a key enzyme could alter lipid metabolism, leading to changes in lipid composition. By performing multi-omic correlations, we aimed to visualize how PHACTR1 expression impacts the abundance of disparate molecular features. Multi-omic correlations have the added benefit of being agnostic to mapping IDs, allowing all detected metabolites and lipids to be included in these analyzes.

On the first pass, there was a high degree of correlation (R > 0.8) between each ‘omics dataset for both the PHACTR1 OE and KD cells when compared to scrm ctrl cells (Fig. 4). To increase stringency for subsequent analyzes, we limited each ‘omics dataset to features with >2-fold change from controls and an FDR < 0.01.

Gene transcripts are visualized in red, proteins in green, metabolites in purple and lipids in green. A Correlation plot of PCA vs correlation coefficients between ‘omics pairs for upregulated features in *PHACTR1* OE cells vs scrm ctrls. B Network analysis of significantly correlated features showing an increased abundance in *PHACTR1* OE cells. Data visualized as a circular plot in Supplementary Fig. 3. C Correlation plot of PCA vs correlation coefficients between ‘omics pairs for downregulated features in *PHACTR1* OE cells vs scrm ctrls. D Network analysis of significantly correlated features showing a decreased abundance in *PHACTR1* OE cells. Data visualized as a circular plot in Supplementary Fig. 4. E Correlation plot of PCA vs correlation coefficients between ‘omics pairs for upregulated features in *PHACTR1* KD cells vs scrm ctrls. F Network analysis of significantly correlated increased features from proteomic and lipidomic analysis of *PHACTR1* KD cells. G Correlation plot of PCA vs correlation coefficients between ‘omics pairs for downregulated features in *PHACTR1* KD cells vs scrm ctrls. H Network analysis of significantly correlated decreased features in *PHACTR1* KD cells. Data visualized as a circular plot in Supplementary Fig. 5. All data presents correlations with r > 0.95 between features with fold change>2 and FDR < 0.01 when compared to scrm ctrl cells.

Visualizing the positive correlations within upregulated features after PHACTR1 OE revealed a distinct subset of features that correlated across all four ‘omics datasets (Fig. 4A, and Supplementary Fig. 3). While most correlations were between gene and protein features, metabolites and lipids also contributed, highlighting that PHACTR1 expression has a significant role in regulating these features. Multi-omic correlation visualization identified the PHACTR1 protein as one of three highly correlated proteins within this network, along with Variable Charge Y-Linked (VCY) and Cluster of Differentiation 47 (CD47) (Fig. 4B). Correlations between downregulated features identified features of interest across all four ‘omics datasets (Fig. 4B, and Supplementary Fig. 4). Correlation visualization identified three key proteins: Ras Homolog Family Member C (RHOC), Ankyrin-2 (ANK2), and A0A94C0S0 (Sorting Nexin-25), along with a significant subset of lipids with a high degree of correlation (Fig. 4D), suggesting a role for PHACTR1 in the regulation of lipid-mediated signaling.

Additional gene-protein correlation analysis of PHACTR1 OE datasets (Supplementary Data 27–30) found strong positive correlation (>0.98) between upregulation of the genes involved in cell migration (MMP9, ITGA4, CXCL8) and transcription regulators (ZNF724, ZNF93, ZNF675) with the protein expression of PHACTR1 and glutathione S-transferase mu 3/GSTM3 (Supplementary Fig. 6A). IPA analysis of upregulated proteins that correlated with PHACTR1 identified pathways associated with antioxidant response by Nrf2 (NFE2L2) (Supplementary Fig. 6B, and Supplementary Data 35). Positive correlation of downregulated features showed strong correlation between genes involved in cytoskeletal organization (FLNC, ACTL8) and protein expression of myosin heavy chain 9 (Supplementary Fig. 6C). Negative gene-protein correlation was limited to correlations with correlation coefficient < −0.5 and found that several upregulated genes involved in mitochondrial bioenergetics (G6PD, BCKDK, SHMT2) and atherosclerosis (SHMT2, GPX1, G6PD, ID3) negatively correlated with the expression of the transmembrane receptor protein Notch1 (Supplementary Fig. 6D). In comparison, downregulated gene expression of AADAT, RPL36AP8 AND RPL5P3 negatively correlated with the centromere protein N/CENPN, a key mitosis regulator (Supplementary Fig. 6D).

4-way multi-omic correlations within the PHACTR1 KD dataset showed no significantly upregulated genes or metabolites and no significantly downregulated genes after implementing our stringent selection criteria (Fig. 4E). However, visualization of positive correlations for upregulated features after PHACTR1 KD identified 10 key proteins that were highly correlated with the abundance of 19 lipid species, predominantly PE (Fig. 4F). Positive correlations between proteomic, metabolomic, and lipidomic features downregulated in PHACTR1 KD cells identified two proteins: ANK2 and isoprenylcysteine carboxyl methyltransferase (ICMT), as well as a subset of lipids, mainly phosphatidylcholine (PC) species, that were highly correlated (Fig. 4G, H, and Supplementary Fig. 5).

To evaluate the correlation of genes across ‘omics datasets we performed gene-protein correlation analysis (Supplementary Fig. 7, and Supplementary Data 31–34). Positive correlation was found between the cytoskeletal-regulating gene ARHGAP30 and secretion regulating proteins including Arfaptin-2/ARFIP2 and Signal recognition particle 9 kDa protein/SRP9 (Supplementary Fig. 7A). Positive correlation of downregulated features included a strong correlation between the gene encoding coagulation factor III (F3) with several proteins including multivesicular regulator VPS36 and microtubule regulator Dynein Light Chain LC8-type 2/DYNLL2 (Supplementary Fig. 7B). IPA analysis of proteins that positively correlated with downregulated PHACTR1 expression found changes to Acyl-CoA hydrolysis and cellular response to mitochondrial stress (Supplementary Fig. 7C, and Supplementary Data 36). Negative correlations found upregulated SMOX correlated with decreased EEF2KMT (Eukaryotic Elongation Factor 2 Lysine Methyltransferase), a translation regulator whereas upregulated RAD51AP1 protein, which is involved in DNA repair, negatively correlated with the expression of several genes including the transcriptional regulator SOX9 (Supplementary Fig. 7D).

Integrative multi-omics identifies PHACTR1-mediated pathways

While feature-level correlation analyzes (Fig. 4, and Supplementary Figs. 6 and 7) are invaluable for identifying co-varying individual molecules and direct regulatory links, they do not inherently aggregate these findings into higher-order biological pathways across diverse omics types. Therefore, to identify overarching biological processes that were consistently impacted across all four distinct molecular layers, we employed a multi-omics IPA integration strategy. This approach allowed us to synthesize findings at the pathway level, providing a robust and comprehensive means to pinpoint convergent biological themes for downstream validation (Top 50 pathways displayed in Fig. 5, full dataset Supplementary Data 37–40). We ranked the overlapping IPA pathways, giving weight to pathways detected by a greater number of ‘omics types (i.e., transcript, protein, lipid, and/or metabolite).

Fig. 5 — A Top 50 overlapping significantly increased IPA pathways in *PHACTR1* OE cells detected by multi-omic analysis across all four ‘omics datasets. B Top 50 overlapping significantly decreased IPA pathways in *PHACTR1* OE cells detected by multi-omics analysis across all four ‘omics datasets. C Overlapping significantly increased IPA pathways in *PHACTR1* KD cells detected by multi-omic analysis across ‘omics datasets (overlapping pathways were only identified across transcriptomic and proteomic datasets). D Top 50 overlapping significantly decreased IPA pathways in *PHACTR1* KD cells detected by multi-omics analysis across ‘omics datasets (overlapping pathways were only identified across transcriptomic, proteomic and lipidomic datasets).

After PHACTR1 OE, the integration of IPA pathways identified one pathway upregulated in three ‘omics datasets (estrogen-dependent breast cancer signaling) and multiple pathways upregulated in two ‘omics datasets (Fig. 5A, and Supplementary Data 22). Additionally, eight pathways were downregulated across three ‘omics analyzes and several others in two ‘omics datasets (Fig. 5B, and Supplementary Data 23).

Since there were no significantly upregulated IPA pathways in the metabolomic or lipidomic datasets after PHACTR1 KD, pathway integration was performed using only the transcriptomic and proteomic datasets. We identified 32 upregulated pathways in PHACTR1 KD cells (Fig. 5C, and Supplementary Data 24). Downregulated IPA pathways were detected in all ‘omics analyzes except metabolomics. Integration identified several overlapping pathways downregulated in two ‘omics datasets (Fig. 5D, and Supplementary Data 25).

Across all analyzes, expected pathways were present (e.g., actin cytoskeleton signaling, Fig. 5C), validating that we can detect known PHACTR1-mediated signaling events using this methodology. More importantly, IPA integration identified several pathways not previously associated with PHACTR1 that are biologically relevant to the onset and progression of vascular disease. While numerous pathways could have been selected for validation, we focused on three that were unexpected and altered in both PHACTR1 OE and KD comparisons: cell cycle regulation, iron-mediated signaling, and mitochondrial biogenesis and bioenergetics.

PHACTR1 expression affects cell cycle progression and mitosis

Components of cell cycle regulation, including cell cycle checkpoints, mitotic cell cycle, and senescence pathways, were detected in all IPA integration comparisons (Fig. 5). To explore the possibility that PHACTR1 influences the cell cycle, we used OpenEMMU and flow cytometry to quantitatively evaluate the percentage of cells in each phase of the cell cycle, including G1 phase, DNA-replicating cells (S-phase), G2 phase, and mitosis²⁰. Consistent with the IPA integration findings, we found that PHACTR1 OE cells exhibit altered cell cycle progression compared to scrm ctrl cells, with a significant reduction in the proportion of G1 phase cells, accompanied by a marked increase in the proportion of cells in S-phase, as indicated by the percentage of 5-ethynyl-2′-deoxyuridine (EdU) positive cells (Fig. 6C, D). PHACTR1 OE cells also showed an increased proportion of G2/M phase cells (Fig. 6E). In contrast, PHACTR1 KD cells displayed more modest changes in cell cycle progression, with the only change being a reduction in the proportion of cells in the G2/M phases (Fig. 6E). Further experiments revealed that PHACTR1 OE cells have impaired mitotic entry, as evidenced by a reduced proportion of histone H3 p-S10+ cells (Fig. 6B, F). PHACTR1 KD cells did not show mitotic changes, suggesting an impairment specifically to the G2 phase. Overall, these results confirm the finding that PHACTR1 regulates the cycling activity of cells.

We next sought to evaluate the total protein levels of several cell cycle-specific proteins using immunoblotting (Fig. 6G–O). Among the Cyclin proteins tested, only Cyclin B1 (also known as CCNB1) was affected by PHACTR1 OE, with PHACTR1 OE cells showing reduced Cyclin B1 total levels (Fig. 6I). This result aligns with the role of Cyclin B1 in regulating the G2/M transition, mitotic entry, and chromosome condensation²¹. Additionally, total Cdt1 levels, a fundamental factor in DNA replication licensing, DNA firing and a marker of late G1 and early-S phase²², were also reduced in PHACTR1 OE cells (Fig. 6L), which aligns with the observed reduction in the proportion of cells in the G1 phase as determined by EdU uptake. Furthermore, immunoblot analysis for phosphorylated histone H3 p-S10 further confirmed our flow cytometry findings, demonstrating impaired mitosis in PHACTR1 OE cells (Fig. 6O). In contrast, PHACTR1 KD did not cause significant changes in total cell cycle protein levels (Fig. 6H–O). In agreement with IPA integration, these results further confirm that PHACTR1 functions as a regulator of the cell cycle.

To validate our finding that PHACTR1 regulates cell cycle progression and to ensure these effects were not specific to the HT1080 cell line or an artifact of stable lentiviral transduction, we next sought to confirm them in a more physiologically relevant primary cell model. We performed transient knockdown or overexpression of PHACTR1 in human umbilical vein endothelial cells (HUVECs, A/A genotype at rs9349379 SNP) for 48 h and evaluated the expression of key cell cycle proteins by immunoblotting (Fig. 7A).

Consistent with our primary findings, modulating PHACTR1 levels in HUVECs led to significant changes in cell cycle regulatory proteins (Fig. 7B). We observed a significant upregulation of Cyclin A2 in PHACTR1 OE cells (Fig. 7C). Furthermore, Cyclin E1 was reciprocally regulated, showing a significant increase in PHACTR1 KD and a decrease in PHACTR1 OE (Fig. 7E). This robustly implicates PHACTR1 in the control of the G1/S transition in endothelial cells. We also found that MCM2, a critical component of the DNA pre-replication complex, was significantly downregulated in PHACTR1 OE cells (Fig. 7G).

While the specific molecular markers affected in HUVECs differ from those observed in the stable HT1080 cell lines, e.g., Cyclin B1 (Figs. 6I and 7D) or Cdt1 (Figs. 6L and 7F), the results are consistent and point to the same biological process. For example, while Cdt1 was altered in HT1080 cells, the change in MCM2 in HUVECs likewise points to a clear role for PHACTR1 in regulating DNA replication licensing. These differences are potentially attributable to the distinct temporal nature of the experiments, where the acute, immediate effects of transient 48 h perturbation in HUVECs may differ from the chronic, adapted state of the stable HT1080 cell lines.

Ultimately, despite these differences in the specific protein players, these validation experiments in primary vascular cells strongly corroborate the broader discovery from our multi-omics analysis, that PHACTR1 is a fundamental regulator of the cell cycle.

Expression of major iron regulatory proteins is regulated by PHACTR1

Our multi-omics analysis identified significant changes in ferroptosis, an iron-dependent form of programmed cell death, in both PHACTR1 OE and KD cells compared to scrm ctrl (Fig. 5). To validate whether PHACTR1 modulates this pathway, we performed immunoblotting for key ferroptosis markers: 4F2 cell-surface antigen heavy chain (4f2hc)-SLC3A2, Kelch-like ECH-associated protein 1 (KEAP1), XCT2, nuclear receptor coactivator 4 (NCOA4), and glutathione peroxidase 4 (GPX4) (Fig. 6P). However, we detected no significant changes in the expression of these markers (Fig. 6Q–U), indicating that PHACTR1 OE and KD HT1080 cells were not actively undergoing ferroptosis.

Given the integral link between ferroptosis and cellular iron homeostasis, we next investigated the status of key iron-handling proteins that were also identified in our proteomic data. We observed differential expression of the iron storage protein ferritin heavy chain 1 (FTH1) and the iron-recycling enzyme heme oxygenase 1 (HMOX1) (Supplementary Data 11 and 13). Immunoblotting confirmed a significant reduction in both FTH1 and HMOX1 in PHACTR1 OE cells (Fig. 6P, V, W). This reduction aligns with a significant downregulation of the “Cytoprotection by HMOX1” pathway identified in our IPA analysis (Fig. 5B), confirming that PHACTR1 overexpression alters the cellular iron handling machinery. This may, in turn, sensitize cells to ferroptotic stimuli. Interestingly, although the “Cytoprotection by HMOX1” pathway was also downregulated in PHACTR1 KD cells (Fig. 5D), these changes were not accompanied by altered expression of FTH1 or HMOX1, indicating that further investigation is needed to clarify the specific mechanisms in the knockdown context.

To validate that PHACTR1’s role in regulating iron metabolism is a core biological function and not specific to the HT1080 model, we assessed key proteins in primary HUVECs following transient PHACTR1 perturbation (Fig. 7A, B).

In strong agreement with our findings in HT1080 cells, we observed a significant reduction in HMOX1 expression in PHACTR1 OE cells, and a similar, though non-significant, trend toward decreased FTH1 (Fig. 7L, M). This indicates that the fundamental impact of PHACTR1 on these core iron-handling proteins is conserved in primary vascular cells.

While the core iron storage and recycling proteins were similarly affected, we observed distinct changes in ferroptosis-regulatory proteins in HUVECs that were not seen in the stable HT1080 line. Specifically, GPX4 was significantly downregulated in PHACTR1 OE cells, and there was a significant increase in KEAP1 expression in PHACTR1 KD cells compared to PHACTR1 OE (Fig. 7J, K). These differences are potentially attributable to the acute (48 h) timeframe of transient transfection in HUVECs, which captures an immediate cellular response, versus the chronic, adapted state of the stable cell lines. Furthermore, the intrinsic differences in the metabolic wiring between primary endothelial cells and transformed fibrosarcoma cells may lead to distinct downstream compensatory responses.

Taken together, these results in primary HUVECs successfully validate the key discovery from our ‘omics analysis. They confirm that PHACTR1 plays a conserved role in regulating cellular iron metabolism, even if the precise downstream effectors involved in the ferroptotic response differ between cell types and perturbation models.

PHACTR1 expression impacts mitochondrial biogenesis

Another pathway of interest for validation was mitochondrial function. Multi-omics integration identified changes in mitochondrial biogenesis across transcriptomics, metabolomics, and lipidomics datasets (Fig. 5B, D), as well as changes in mitochondrial dysfunction in proteomics and transcriptomics (Fig. 5B).

To evaluate if PHACTR1 expression regulates mitochondrial morphology, we performed confocal microscopy to visualize cellular mitochondrial networks (Fig. 8A). Mitochondria in PHACTR1 KD cells were qualitatively observed to be smaller and more fragmented than the scrm ctrl mitochondria, and, in comparison, mitochondria from PHACTR1 OE cells were elongated. These data suggest potential PHACTR1-mediated changes in mitochondrial fission or fusion.

We next isolated mitochondria from PHACTR1 OE, KD, and scrm ctrl cells to probe the molecular mechanisms behind this change in mitochondrial morphology. To validate our isolations, we immunoblotted cellular fractions for GAPDH as a marker for the cytosol and a combination of mitochondrial respiration chain proteins as markers for mitochondria (Fig. 8B). Out of curiosity, we also blotted PHACTR1 protein expression. In addition to detecting PHACTR1 expression in the cytosol, unexpectedly, PHACTR1 was also localized to the mitochondrial fraction of PHACTR1 OE cells, which has not been previously reported (Fig. 8B).

We further immunoblotted the mitochondrial fractions for key markers of mitochondrial fission and fusion (Fig. 8C–G, and Supplementary Fig. 8). A significant increase in the fission regulator protein dynamin-related protein 1 (Drp1) phosphorylated at Ser637 was detected in PHACTR1 OE cells compared to PHACTR1 KD cells (Fig. 8C, F). This change in phosphorylation status is coupled with increased expression of mitochondrial fission factor (Mff, Supplementary Fig. 8A, E), a Drp1 binding protein that recruits Drp1 to the outer mitochondrial membrane²³. When Drp1 is phosphorylated at Ser637, its translocation from the cytosol to the mitochondria is inhibited. This reduces Drp1’s ability to mediate mitochondrial fission, leading to more elongated and fused mitochondria. Conversely, when Ser637 is dephosphorylated, Drp1 is recruited to the mitochondrial surface, where it promotes mitochondrial fission. Increased fission can result in more fragmented mitochondria; as such, this data is consistent with the observed changes by confocal microscopy.

Phosphorylation of Drp1 Ser637 at the mitochondrial membrane is regulated by protein kinase A (PKA)²⁴. Upregulated PKA-mediated signaling was detected by multi-omics IPA integration (Fig. 5A). While we were unable to immunoblot directly for PKA due to poor antibody specificity, we detected significant changes in the mitochondrial PKA scaffold A Kinase Anchoring Protein 1 (AKAP1), with protein expression significantly increased in PHACTR1 OE cells and decreased in PHACTR1 KD cells (Fig. 8C, G). These data indicate that PHACTR1 regulates mitochondrial biogenesis, in part due to alterations in PKA-mediated phospho-signaling, although there are likely other signaling pathways involved.

Extending a prior observation that levels of PHACTR1 were potentially linked to levels of AKAP1 in a gene regulatory network², the correlation between PHACTR1 and AKAP1 gene expression was further validated in the open-source GTEx dataset using RNAseq data from human arterial tissues. Linear regression analysis found a robust correlation between the two genes in the aorta (R = 0.64, Fig. 8H) as well as the coronary (R = 0.47, Supplementary Fig. 9A) and tibial arteries (R = 0.63, Supplementary Fig. 9B). These data confirm the biological relevance of the PHACTR1-AKAP1 relationship within the human vasculature.

Cellular bioenergetics are influenced by PHACTR1 expression

To evaluate if PHACTR1-mediated changes in mitochondrial biogenesis impact mitochondrial bioenergetics, we performed cell respirometry to measure cellular ATP production (Fig. 8I–K). The rate of total cellular ATP production was significantly increased in PHACTR1 KD cells (Fig. 8I). While there was no significant increase in mitochondrial ATP production in PHACTR1 KD cells compared to scrm ctrl cells (Fig. 8J), there was a significant elevation in the rate of ATP generated from glycolysis (Fig. 8K). Conversely, PHACTR1 OE cells showed no significant reduction in ATP generation from glycolysis compared to scrm ctrl cells (Fig. 8K), but mitochondrial ATP generation was significantly reduced (Fig. 8J). Although the 2-dimensional cell area was larger for PHACTR1 OE cells (Supplementary Fig. 1A), flow cytometry analysis of 3-dimensional cell volume showed no significant changes between cell lines (Supplementary Fig. 10A). Additionally, flow cytometry analysis found no changes in mitochondrial mass to explain the observed changes in mitochondrial bioenergetics (Supplementary Fig. 10B).

Alterations in mitochondrial metabolism were further supported by lipid profile changes, particularly in pathways associated with ATP production via mitochondrial β-oxidation (Supplementary Fig. 11). During β-oxidation, free fatty acids are converted into acyl-CoA, then to acylcarnitine, and back to acyl-CoA to ultimately enter the TCA cycle. Lipidomics data revealed that PHACTR1 KD cells had significantly reduced free fatty acids (Supplementary Fig. 11A), while PHACTR1 OE cells showed increased acylcarnitine accumulation (Supplementary Fig. 11B). This acylcarnitine buildup in PHACTR1 OE cells suggests impaired β-oxidation, potentially contributing to decreased mitochondrial ATP production. Conversely, the reduced free fatty acids in PHACTR1 KD cells, without significant acylcarnitine accumulation, may indicate lower fatty acid availability for β-oxidation, thus increasing reliance on glycolysis for ATP production. These data help validate the IPA analysis of proteins that positively correlate with decreased PHACTR1 expression, where the top changed pathway was “Acyl-CoA hydrolysis” (Supplementary Fig. 7C) and highlight the link between reduced PHACTR1 and altered cellular metabolism.

In sum, these experiments validate our multi-omics approach, demonstrating that PHACTR1 expression significantly influences mitochondrial structure and cellular bioenergetics.

Discussion

In this study, we present a comprehensive multi-omics analysis revealing the role of PHACTR1 in regulating cellular composition and function. By integrating transcriptomic, proteomic, metabolomic, and lipidomic data, we identified several new pathways associated with PHACTR1, highlighting its importance in vascular cell function. Our findings advance our understanding of PHACTR1 and reinforce the power of multi-omics integration in deciphering complex biological processes.

The success of our approach stems from the generation of high-quality ‘omics datasets spanning a multitude of biological entities. Our analysis encompassed 14,737 transcripts, 5666 proteins, 1752 metabolites, and 737 lipid species, providing comprehensive and highly diverse coverage of cellular components. Combined, this level of coverage surpasses single-omics studies, allowing for a more thorough examination of key cellular processes. For example, recent exemplary multi-omics studies in cardiovascular biology have typically focused on integrating two or three ‘omics layers^25–28, making our four-omics integration a significant advancement in the field.

While the field of multi-omics integration has expanded, the capacity to extract biologically significant pathways from diverse, high-throughput datasets remains a key challenge^29,30. What sets our approach apart is not the invention of new computational algorithms, but the successful demonstration of a robust and highly accessible pipeline that integrates four distinct omics datasets. Our strategy combined pathway-level and feature-level analyzes to build a comprehensive view of PHACTR1 function. The primary multi-omics IPA integration robustly identified overarching biological processes—such as cell cycle and iron handling—by identifying converging themes across the datasets, providing high confidence in selecting these broad themes for validation. Concurrently, our correlation analyzes offered a more granular perspective to pinpoint specific molecular targets. These corroborating strategies provide a more holistic view than single-omics approaches. Crucially, this methodology is built upon widely available tools and, as demonstrated by its success in guiding our experimental validation, is both straightforward to implement and powerful, positioning it as a valuable tool for deciphering the complex regulatory networks of other poorly understood genes without requiring extensive specialized bioinformatics expertise.

In the context of PHACTR1, our results provide major new insights into its role in cellular processes. Multi-omic pathway integration identified a role for PHACTR1 in several pathways in which this gene already has a well-characterized function, including actin regulation via Rho signaling³¹ or TGF-β³², as well as less established pathways, such as nitric oxide production¹³ and ERK/MAPK signaling³³. Identification of established PHACTR1-mediated pathways validated our ‘omics integration methodology, while also revealing previously unknown pathways that further elucidate PHACTR1’s biological significance. Importantly, we were able to validate PHACTR1’s regulatory role in three pathways not previously associated with this gene – cell cycle regulation, iron-mediated signaling, and mitochondrial biogenesis and bioenergetics. This demonstrated the power of our analysis to uncover new roles and functions of critical genes. These findings, particularly as validated in primary human vascular endothelial cells, have significant implications for understanding vascular disease mechanisms.

A striking finding from our multi-omics integration was the previously unrecognized role of PHACTR1 in cell cycle regulation. PHACTR1 likely influences the cell cycle, in particular progression from G2 to mitosis, via its regulatory effect on PP1. The involvement of PP1 in cell cycle control is well-established³⁴, however, PHACTR1 has not previously been implicated in the regulation of PP1 in this process. Localized activity of PP1 has been shown to regulate a key mitotic phosphatase relay required for coordinated mitotic progression³⁵. PP1 performs this role via the dephosphorylation of key target proteins including histone H3³⁶. As a PP1-interacting protein, elevated PHACTR1 may inappropriately anchor PP1 away from its targets needed to progress into and through mitosis, resulting in a decrease in histone H3 phosphorylation and mitosis in PHACTR1 OE cells, as observed in Fig. 6. Decreased protein expression of Cyclin B1 and Cdt1 in these cells suggests that PHACTR1 may regulate the cell cycle beyond targeting PP1-mediated dephosphorylation. Our results suggest that PHACTR1 regulation of PP1, and possibly other interactors, influence cell cycle progression. Furthermore, Cyclin E1, which was reciprocally regulated in PHACTR1 KD and OE primary endothelial cells, is essential to endothelial cell regeneration and restores vascular homeostasis post-inflammation³⁷. These data may partly account for the observation that A/A (increased PHACTR1) vs G/G (reduced PHACTR1) rs9349379 SNP status is associated with differing vascular pathologies^3,6,7,13.

Dysregulation of vascular smooth muscle and endothelial cell proliferation is a hallmark of various vascular pathologies, including atherosclerosis and restenosis³⁸. Specifically, abnormal cell cycle progression in endothelial cells contributes to vascular permeability and inflammation³⁷, while uncontrolled smooth muscle cell proliferation drives neointimal hyperplasia³⁸, both central to atherosclerotic plaque progression and post-intervention restenosis. Our validated findings of PHACTR1’s impact on key cell cycle regulators provide a molecular link between PHACTR1 and these fundamental processes, suggesting PHACTR1 as a potential modulator of vascular remodeling.

We also uncovered a link between PHACTR1 and iron signaling, with increased PHACTR1 expression resulting in decreased expression of key iron-binding proteins, FTH1 and HMOX1. This finding is intriguing in the context of vascular biology, as iron homeostasis plays a crucial role in endothelial function and the development of atherosclerosis³⁹. Dysregulation of iron can lead to oxidative stress and ferroptosis, a form of iron-dependent cell death, which is increasingly recognized as a contributor to vascular injury and atherosclerotic plaque instability. Our validation in HUVECs showing PHACTR1-dependent changes in GPX4 (a key ferroptosis regulator), KEAP1, HMOX1, and FTH1 directly implicates PHACTR1 in the pathways governing iron-mediated cellular defense and susceptibility to oxidative damage in vascular cells. HMOX1 has been tenuously linked to PP1 signaling via PP1-mediated regulation of Nek2 and Nuclear factor erythroid 2-related factor 2 (Nrf2)^40,41. However, the mechanism of their signaling and the role of PHACTR1 are unknown. To our knowledge, FTH1 has never previously been linked to PHACTR1 or PP1. As PP1 activity is dependent on the oxidation status of bound metal ions, including iron⁴², cellular availability of iron may directly impact PHACTR1-PP1-mediated signaling. Furthermore, PHACTR1 is associated with several diseases that exhibit dramatically different prevalences between males and females. A/A genotype at the rs9349379 SNP results in increased PHACTR1 and is associated with fibromuscular dysplasia and spontaneous coronary artery dissection, both of which overwhelmingly affect women⁴³. The possibility that the sex-biases of these diseases might be linked to PHACTR1-mediated differences in iron levels and iron handling between males and females is an intriguing possibility that warrants further consideration⁴⁴. Understanding how PHACTR1 modulates iron homeostasis could therefore offer new perspectives on sex-specific susceptibilities in vascular disease.

Perhaps most surprisingly, our study revealed a previously unknown connection between PHACTR1 and mitochondrial function. We identified PHACTR1 within the mitochondrial fraction for the first time and demonstrated PHACTR1-mediated changes in mitochondrial biogenesis and bioenergetics. These findings are particularly significant in light of the growing recognition of mitochondrial dysfunction as a key driver of vascular disease⁴⁵. Integrative IPA analysis identifying mitochondrial dysfunction as a significantly altered pathway across multiple ‘omics is supported by observations from single ‘omics datasets. This includes the detection of elevated acylcarnitines in PHACTR1 OE lipidomics and a concurrent reduction in free fatty acids in PHACTR1 KD cells. Changes in acylcarnitine signaling pathways are known to be linked to mitochondrial myopathy⁴⁶. It has also been suggested that decreased free fatty acids may contribute to coronary artery disease, linking our PHACTR1 KD cell model to this clinically relevant PHACTR1 phenotype⁴⁷.

Our data suggest that PHACTR1 regulates mitochondrial dynamics through PKA-AKAP1 signaling at Drp1 Ser637 via Mff on the outer mitochondrial membrane^23,48. Phosphorylation of Drp1 at Ser637 reduces Drp1 GTPase activity resulting in mitochondrial fission. This pathway has been implicated in endothelial cell function and angiogenesis⁴⁹, suggesting a potential mechanism by which PHACTR1 influences vascular health. The phosphorylation status of Drp1 Ser637 has also been implicated in the regulation of mitosis, potentially linking this finding to our observed change in the cell cycle²⁴. By confirming the correlation in expression between PHACTR1 and AKAP1 in human arterial tissues, we add weight to our hypothesis that PHACTR1 regulates mitochondrial dynamics in human vasculature via PKA-AKAP1 signaling. Further investigation into these pathways under disease conditions, such as coronary artery disease where mitochondrial leakage of reactive oxygen species is thought to drive disease progression⁴⁵, may reveal further insight into the pivotal role of PHACTR1 in pathology. Elucidating PHACTR1’s precise involvement in mitochondrial regulation could open avenues for understanding and potentially mitigating mitochondrial dysfunction in vascular diseases, offering new targets for therapeutic exploration.

While we did not detect significant global changes in the metabolome of PHACTR1 KD cells, it is important to consider the limitations of our untargeted discovery approach. This methodology, while comprehensive, may not always detect all metabolites with differential abundances, particularly if the changes are subtle, localized to specific metabolic branches, or if the metabolites are present at very low concentrations or are not well-represented in the detection library. This inherent sensitivity and coverage limitation is a potential factor contributing to the observed absence of significant changes in the PHACTR1 KD metabolome, even though functional changes in mitochondrial activity were observed, implying underlying metabolic shifts. Given these insights into mitochondrial function, future studies could employ more focused and sensitive isotope-assisted metabolic flux analysis to study components of glycolysis and the TCA cycle, deciphering precisely where in the ATP cycle PHACTR1 exerts its regulatory effects⁵⁰. This approach has been successfully used in vascular research to elucidate metabolic changes in endothelial cells under various conditions, and would provide a deeper, targeted understanding of PHACTR1’s metabolic influence⁵¹.

Our choice of IPA for multi-omics integration proved to be robust and informative. While IPA has been used for mapping cellular moieties in biomedical research^52,53, to our knowledge, this is the first time it has been used to integrate four ‘omics datasets in the context of vascular biology. By integrating ‘omics data using IPA mapping, we can quickly filter for relevant pathways and remove false positives. This is because each ‘omics dataset is not independent; together, they elucidate PHACTR1’s role in cell signaling. This integrative approach allowed us to identify numerous biologically relevant pathways directly impacted by PHACTR1 expression, revealing connections to cellular biology. Through the successful validation of three of these pathways, we underscore the utility of our multi-omics approach in revealing complex biological networks.

A limitation of any pathway mapping, including IPA, is that a feature must be cataloged in a major database (i.e., KEGG or Reactome) to create a “mappable” ID and place it within relevant signaling pathways⁵⁴. Any features without a respective ID are considered “unmappable” and are excluded from the analysis, resulting in incomplete pathway mapping. This issue is particularly prevalent for metabolites and lipids, as the field of research is expanding rapidly alongside our knowledge of these molecules⁵⁵.

To evaluate the significance of “unmappable” features, we determined multi-omic correlations between features in our four ‘omics datasets. This approach was particularly helpful when focusing on lipid signaling in PHACTR1 OE cells, which exhibited a predominant lipid signature with significant correlations to changes in the proteins RHOC, ANK2, and A0A94C0S0 (Sorting nexin-25), all of which have been linked to lipid signaling or are themselves lipid-modified^56–58. We also observed negative gene-protein correlations linking PHACTR1 expression to Notch1 protein, a known protective factor against atherosclerosis⁵⁹, to increased expression of genes associated with atherosclerosis. These combined observations strongly suggest that PHACTR1 may play a role in the regulation of lipid metabolism and vascular health. Given that the G/G genotype of the rs9349379 SNP is causally linked to atherosclerosis², our findings highlight that PHACTR1-dependent effects on lipids and Notch1 signaling warrant further investigation in the context of vascular disease.

In conclusion, we successfully integrated high-throughput transcriptomics, proteomics, lipidomics and metabolomics datasets from PHACTR1 OE and KD cells to define previously uncharacterized and unexpected aspects of the function of this critical gene. Our work not only advances the understanding of PHACTR1 but also showcases the power of multi-omics integration. By revealing connections between PHACTR1 and key cellular processes such as cell cycle regulation, iron homeostasis, and mitochondrial function, we have significantly expanded our understanding of how this gene may contribute to vascular homeostasis. This multi-omics study provides a powerful, hypothesis-generating roadmap for understanding PHACTR1 biology. While it offers a systems-level view of PHACTR1’s influence, the identified associations are inherently correlational and require further direct mechanistic validation. Techniques like co-immunoprecipitation for molecular interactions and super-resolution microscopy for subcellular localization will be crucial in future studies to establish the precise causal links and physical mechanisms suggested by our findings.

Although our multi-omics integration approach has provided valuable insights, it is important to acknowledge its limitations. The use of the HT1080 fibrosarcoma cell line was a necessary and strategic choice for the discovery phase of this project; however, we acknowledge the inherent limitations of this model system. As an immortalized cancer cell line, HT1080 cells have undergone significant genetic and epigenetic alterations, which may not fully recapitulate the biology of primary cells in their native tissue context. The findings from this study, therefore, represent a foundational step in understanding the molecular roles of PHACTR1, and future validation in in vivo systems is crucial. Whilst it would have been ideal to perform all analyzes in primary cells, their inherent variability, limited proliferative capacity, and challenges in obtaining the sheer volume of high-quality material from primary cells would have made the extensive multi-omics approach undertaken in this study practically unfeasible. The homogeneity and robust nature of the HT1080 stable cell lines provided the necessary consistency and scalability to generate high-quality, reproducible multi-omics data. This allowed for a broad and unbiased discovery of PHACTR1’s molecular network; indeed, key findings related to cell cycle and iron handling were subsequently validated in HUVECs with transient PHACTR1 knockdown or overexpression, thereby demonstrating the validity of our discovery-based strategy for generating key, translatable insights into PHACTR1 signaling. Additionally, discrepancies in dataset sizes and the challenges in mapping metabolites and lipids to known pathways highlight the need for continued development of bioinformatics tools and databases. As our knowledge of metabolite and lipid roles in cellular pathways expands, these limitations may be mitigated, allowing for even more comprehensive integrations in the future.

As multi-omic technologies and integration methods continue to evolve, we anticipate that this comprehensive approach to studying gene function will revolutionize our understanding of complex diseases and pave the way for new therapeutic strategies.

Methods

Cell culture

HT1080 fibrosarcoma cells (ECACC, Sigma-Aldrich, #85111505, lot#17A033, RRID:CVCL_0317) were maintained in Dulbecco’s Modified Eagle Medium (DMEM, Gibco, #11965-092) supplemented with 10% fetal bovine serum (FBS, Cytiva, #SH30071.03IH30-45) at 37 °C, 5% CO₂.

Human umbilical vein endothelial cells (HUVEC, ThermoFisher Scientific, #C0035C, lot#3027121) were maintained in Human Large Vessel Endothelial Cell Basal Medium (ThermoFisher Scientific, #M200500) supplemented with Large Vessel Endothelial Supplement (LVES, ThermoFisher Scientific, #A1460801) at 37 °C, 5% CO₂.

Routine in-house mycoplasma testing was conducted periodically using LookOut mycoplasma PCR detection kit (Sigma-Aldrich, #MP0035) to validate all experimental cells were mycoplasma negative.

rs9349379 SNP genotyping

Genomic DNA was extracted from cultured cells using Isolate II Genomic DNA Kit (Meridian Bioscience, #BIO-52067) as per manufacturer’s instructions. The genotype (A or G) of the rs9349379 SNP in human PHACTR1 was determined by qPCR using the TaqMan SNP Genotyping Assay (ThermoFisher, Taqman, #4351379; VIC = A,FAM = G) and 10 ng of genomic DNA (measured by Qubit) as per manufacturer’s instructions performed in a CFX384 Touch Real-Time PCR Detection System (BioRad) with the following cycling parameters: polymerase activation at 95 °C for 10 min, followed by 40 cycles of denaturation at 95 °C for 15 s and annealing/extension at 60 °C for 1 min.

Lentiviral transduction

The PHACTR1 overexpression (OE) gene construct (NM_030948.6) was cloned into the lentiviral expression vector (pLVX-EF1a-Ires-Puro, Takara Bio, #631988) and PHACTR1 siRNA knockdown (KD) construct (pSMART hEF1a with PHACTR1 siRNA, Horizon, #V3SH11240-225060755), along with the empty vector control and scrambled siRNA control (pSMART hEF1a with non-targeting control, Horizon, #DHA-VSC11) plasmids, were transfected into Lenti-X 293T cells (Takara, # Cat#632180) using the Lenti-X Packaging Single Shots system (Takara, #631275). Transfected cells were cultured in DMEM with 10% Tet System Approved FBS (Takara, # 631106) for 48 h. Cell culture media containing lentiviral particles for each plasmid were harvested and combined with 8 µg/mL polybrene before addition to HT1080 cells. Cells were incubated at 32 °C, 5% CO₂ for 6 h followed by 72 h at 37 °C, 5% CO₂. Transduced cells were selected by the addition of 0.4 μg/mL puromycin for 7 days. Stably transduced cell lines were maintained in complete DMEM with the addition of 0.2 μg/mL puromycin.

qPCR

RNA was isolated from cells using RNeasy kits (Qiagen, # 74004) with the addition of 0.1% β-mercaptoethanol to cell lysis buffer in combination with QIAshredders (Qiagen, #79656) to facilitate cell lysis. cDNA was synthesized and amplified from 1 µg total RNA using QuantiTect Reverse Transcription kit (QIAGEN, #205311). Gene expression was determined by qPCR using PHACTR1 primer: forward GCAGAGAAGAGCTGATAAAGCG, reverse AGCTCAGGGACTGCCCATTT and GAPDH primer: forward GTCTCCTCTGACTTCAACAGCG, reverse ACCACCCTGTTGCTGTAGCCAA with SsoAdvanced Universal SYBR Green Supermix (BioRad, #1725271) as per manufactures instruction. Samples were analyzed on a CFX384 Touch Real-Time PCR Detection System (BioRad) performed at 95 °C for 2 min, followed by 40 cycles of 95 °C for 5 s and 60 °C for 30 s. GAPDH was used as a control, and PHACTR1 expression was analyzed using the ΔΔCt method.

Confocal microscopy

Cells were fixed with 4% paraformaldehyde and permeabilized with 0.1% Triton X100. ATTO-488 labeled phalloidin ATTO-TEC GmbH (#AD 488-81) was used to visualize filamentous actin. Mitochondria were labeled with MitoTracker Orange CMTMRos (ThermoFisher, #M7510) and cell nuclei with Hoechst 33342 (ThermoFisher, #H1399). Images were taken on a Zeiss LSM800 confocal using a Plan-Apochromat 63x/1.40 Oil DIC M27 objective.

2-dimensional cell and nuclear area

Cells were plated at a density of 2.5 × 10⁴ cells/well onto uncoated glass coverslips and incubated overnight. Cells were fixed with 4% PFA for 20 min, washed twice with PBS, then concurrently permeabilized and stained with 1:100 10% Triton X-100, 1:1000 Phalloidin, 1:10,000 dapi and 1:5000 cell mask. Coverslips were washed twice with PBS, mounted onto glass slides, and imaged using a Nikon Ti2 microscope. The phalloidin and the DAPI/cell mask images were combined using image math in Image J. Cell and nuclear areas were measured using “analyze particles” function in Image J (v.1.54f).

Cell migration

HT1080 cell lines were plated at a density of 1.5 × 10⁴ cells/well into 6-well plates and incubated overnight. The following day, plates were transferred to a Nikon inverted live cell microscope and equilibrated for 1 h to minimize stage drift. Movies (6/well) were taken every 5 min for 4 h using the 20× objective. Movies were analyzed using the manual tracking module in Image J with the center of the nucleus taken as the tracking point.

Cell proliferation

Cells were plated at a density of 1.5 × 10³ cells/well into 96-well plates. At 24-h time point intervals, 8% PFA (final) was added to wells and incubated for 15 min at 37 °C. The media/PFA was removed, and cells were washed once with PBS. Cells were concurrently stained and permeabilized using 1:100 10% Triton-X 100 and 1:2000 Hoechst 33342 for 20 min. Wells were gently washed twice with PBS and 9 images per well covering the majority of the well area were taken using a Phenix microscope. Cell proliferation was assessed by automatically counting nuclei using Harmony software.

Scratch wound assay

Cells were plated at a density of 7.5 × 10⁴ cells/well into 4-well micro-inserts (ibidi, #80409) and incubated overnight. The following day, at time 0, the insert was removed to reveal the multi-shaped wound, and the media was replaced with fresh supplemented DMEM. Plates were imaged using a Nikon Ti2 microscope over 11 h at 2-h intervals starting 1 h post insert removal. For each well, a total of 12 movies were taken (3 per cross side) and the area of the wound over time was measured using Image J (v.1.54f).

Adhesion assay

Cell adhesion was measured using CytoSelect 48-Well Cell Adhesion Assay (Colorimetric Format, Cell Biolabs Inc, #CBA-070) as per the manufacturer’s instructions. In brief, 1 × 10⁶ cells/mL were plated into collagen-coated 48-well plates and left to adhere for 45 min. Plates were then aspirated, and wells were washed with PBS before the addition of Cell Stain Solution for 10 min. Wells were washed with deionized water and then air-dried. Once dry, the Extraction Solution was added and incubated for 10 min while shaking. The resulting cell extract was transferred to a 96-well plate, and cell adhesion was calculated indirectly as a read-out of OD 560 nm measured using a plate reader.

Quantification of lamellipodia

Cells were plated at a density of 2 × 10⁴ in 6-well plates. After 24 h, the cells were placed in an environmental chamber with a Nikon Ti2 microscope and incubated at 37 °C and 5% CO₂ in a humidified environment for 1 h prior and throughout the experiment. Images were taken every 6 s for 20 min using an Achromat LWD 40× Ph1 ADL objective. Kymographs were generated using ImageJ (v.1.54f).

Cell culture for ‘omics dataset generation

All HT1080 cell samples (n = 6/cell line) for multi-omic analysis (bulk RNAseq, untargeted proteomics, untargeted metabolomics and targeted lipidomics) were generated in parallel using the same passage cells, tissue culture reagents and culture conditions followed by concurrent harvesting. Cells were plated 24 h prior to harvest at 2.2 × 10⁵ cells/mL in DMEM with 10% FBS and 0.2 μg/mL puromycin for stable cell lines.

Bulk RNAseq

RNA was extracted from cells using RNeasy Mini Kit (QIAGEN, #74104), as per the manufacturer’s instructions. RNA libraries were prepared using RNA with RIN score >9.8 using the KAPA Total RNA Library Preparation kit (Roche, #08098093702). Barcoded libraries were subsequently analyzed using an Illumina NovaSeq 6000 (Kinghorn Centre for Clinical Genomics Sequencing Facility, Sydney, Australia) using NovaSeq 6000 S1 Reagent Kit v1.5 (300 cycles, Illumina, #20028317) with paired-end lengths of 150 base pairs with a depth of >40 million reads.

RNAseq data processing

Data quality control was performed using FastQC⁶⁰. Fastq files were mapped onto the human GENCODE 29 version using STAR 2.7.5⁶¹. The average reads/sample was 57M. No samples were excluded from analysis after quality control. Differential gene expression between PHACTR1 OE or KD vs scrm ctrl was analyzed using R package LIMMA-voom⁶². Raw counts were Log2 transformed and analyzed using the build-in function of empirical Bayes statistical analysis. Significantly changed features were identified with Benjamini–Hochberg FDR < 0.1 against scrm ctrl. All genes with FDR < 0.1 were used for downstream pathway analysis.

Proteomic sample preparation

Cells were washed twice with ice-cold PBS before lysis into 1% sodium deoxycholate (SDC) in Tris-HCl pH 7.5 with sonication (3 ×5 s bursts on ice). Cell lysates were reduced by addition of 10 mM tris(2-carboxyethyl)phosphine (TCEP) and incubation at 60 °C for 30 min with shaking. Samples were alkylated with 20 mM iodoacetamide (IOA) and incubated at room temperature for 1 h in the dark before the reaction was quenched with 2.5 mM dithiothreitol (DTT) and diluted to 0.1% SDC in 50 mM triethylammonium bicarbonate (TEAB) prior to proteolytic digest. MS-grade Trypsin (Thermo Scientific, #90057) was reconstituted in 20 mM acetic acid and added to cell samples at 1:100 (µg trypsin:µg total cellular protein) dilution before overnight incubation at 37 °C and then acidification with 0.5% formic acid (FA). Peptide digests were cleaned by solid phase extraction (SPE) using Strata-X 33 μm polymeric sorbent cartridges (Phenomenex, 30 mg/1 mL, #8B-S100-TAK), as per the manufacturer’s instructions, with final elutes using sequential addition 70% MS-grade acetonitrile (ACN) with 0.5% FA and 90% MS-grade ACN with 0.5% FA. Samples were lyophilized using an Eppendorf Concentrator at 37 °C until completely dry. Samples were fractionated using Pierce High pH Reversed-Phase Peptide Fractionation Kit (Thermo Scientific, #84868) into 8 fractions per sample and the resulting fractions were lyophilized at 37 °C until completely dry.

Untargeted proteomics

SPE-cleaned, fractionated peptide digests were resuspended in 0.2% trifluoroacetic acid (TFA) separated by nanoLC using an Ultimate nanoRSLC UPLC and autosampler system (Dionex, Amsterdam, Netherlands). Samples (2.5 µl) were concentrated and desalted onto a micro C18 precolumn (300 µm × 5 mm, Dionex) with H₂O:ACN (98:2, 0.2% TFA) at 15 µL/min. After a 4 min wash, the pre-column was switched (Valco 10-port UPLC valve, Valco, Houston, TX) into line with a fritless nano column (75 µm × ~ 15 cm) containing C18AQ media (1.9 µ, 120 Å Dr Maisch, Ammerbuch-Entringen Germany). Peptides were eluted using a linear gradient of H₂O:ACN (98:2, 0.1% FA) to H₂O:ACN (64:36, 0.1% FA) at 200 nL/min over 30 min. High voltage 2000 V was applied to low volume Titanium union (Valco) and the tip positioned ~0.5 cm from the heated capillary (T = 275 °C) of a Orbitrap Fusion Lumos (Thermo Electron, Bremen, Germany) mass spectrometer. Positive ions were generated by electrospray and the Fusion Lumos operated in data-dependent acquisition mode (DDA).

A survey scan m/z 350–1750 was acquired in the orbitrap (resolution = 120,000 at m/z 200, with an accumulation target value of 400,000 ions) and lockmass enabled (m/z 445.12003). Data-dependent tandem MS analysis was performed using a top-speed approach (cycle time of 2 s). MS2 spectra were fragmented by HCD (NCE = 30) activation mode and the ion-trap was selected as the mass analyzer. The intensity threshold for fragmentation was set to 25,000. A dynamic exclusion of 20 s was applied with a mass tolerance of 10 ppm.

Proteomic data processing

Mass spectrometry raw data files were searched using MaxQuant (v2.1.3.0) against the Uniprot human canonical proteome (downloaded April 2023). N-terminus acetylation and M oxidation were set as dynamic variables and carbamidomethylated C as a fixed modification. The minimum number of razor peptides for protein identification was 1 with 0.01 protein FDR. MaxQuant Protein Group files were used for downstream analysis performed by LFQ-Analyst using Perseus-type imputation, where missing values were replaced by random numbers drawn from a normal distribution of 1.8 standard deviation down shift and with a width of 0.3 of each sample⁶³. Label free quant (LFQ) ion intensity data was Log2 transformed, and statistical analysis was two-sample Student’s t-test built in LFQ-Analyst tool. Significantly changed features were identified with Benjamini–Hochberg FDR < 0.1 against scrm ctrl. All proteins with FDR < 0.1 were used for downstream pathway analysis.

Metabolomic sample preparation

Aqueous metabolites were extracted using Agilent’s biphasic MeOH:CHCl₃:H₂O (2:2:1.8) extraction protocol (Agilent Technologies application note 5989-7407). In brief, cells were washed once with ice-cold PBS before quenching with 300 µl 2:1 ice-cold MeOH:H₂O on dry ice. Plates were transferred to wet ice to scrape cells. Collected cell pellets were moved to pre-cooled 2 mL Eppendorf and plates were washed with 300 µl ice-cold MeOH:H₂O. The wash was collected and added to cell pellets. Samples and a blank extract containing 600 µl MeOH:H₂O, were kept on wet ice throughout extraction by the addition of 240 µl ice-cold CHCl₃ spiked with a mixture of internal standards containing L-valine-13C5,15N (Sigma-Aldrich, #600148) and unlabeled caffeine (Sigma-Aldrich, #1000949613). Samples were vortexed, then 240 µl ice-cold H₂O was added with vortexing followed by 240 µl ice-cold CHCl₃. Samples were vortexed 3 × 10 s then centrifuged at 5000 × g for 40 min at 4 °C. The aqueous phase was transferred to a fresh Eppendorf tube and extracts were lyophilized at room temperature and stored at −80 °C. Prior to analysis, samples were reconstituted in 90 µl MeOH:ACN:H₂O (1:1:1.6). Samples were vortexed and sonicated in a water bath for 20 min at room temperature. Samples were then centrifuged (21,000 × g, 10 min, 4 °C) and transferred into sample vials with glass inserts for analysis.

Untargeted metabolomics

In brief, extracted cellular metabolites were analyzed in one batch along with technical quality control (TQC) samples spaced 1 in every 10. Samples were analyzed by MS1 followed by 5 iterative QC (iQC) samples at the end of the worklist analyzed by MS2. Samples (5 µl) were separated on a 1290 Infinity II Bio High-Speed Pump and Infinity II Bio Multisampler HPLC system (Agilent). Samples were run in duplicate on two HPLC columns: (1) InfinityLab Poroshell 120 HILIC-Z (2.1 × 100 mm, 2.7 µm, Agilent) with UHPLC Guard Poroshell 120 HILIC-Z (2.1 × 5 mm, 2.7 µm, Agilent) and (2) reversed-phase ZORBAX Eclipse XDB-C18 (2.1 × 100 mm, 1.8 µm, Agilent) with UHPLC Guard ZORBAX (2.1 × 5 mm, 1.8 µm, Agilent). All samples were run with the thermostat set at 35 °C. The solvent system consisted of Solvent (A) H₂O with 10 mM ammonium formate and 0.1% FA and solvent (B) H₂O:ACN (10:90) containing 10 mM ammonium formate and 0.1% FA with metabolite elution using a stepped linear gradient at 0.25 mL/min.

The gradients were as follows. For HILIC-Z: starting at 98% B for 3 min followed by a decrease to 70% B over 8 min, to 70% B over 1 min, to 5% B over 4 min. The solvent was held at 5% B for 2 min (total 18 min). Equilibration was as follows: the solvent was increased from 5% B to 98% B over 1 min and held for 1 min within the run time. This was followed by a 4 min post run time (total 24 min sample cycle/sample). For XBD: starting at 0% B for 2 min followed by an increase of solvent to 50% B over 15 min then 95% B over 2 min. The solvent was held at 95% B for 5 min (total 22 min). Equilibration was as follows: the solvent was decreased from 95% B to 0% B over 0.5 min. This was followed by a 4 min post run time (total 26.5 min sample cycle/sample).

Mass spectrometry analysis was performed on an Agilent 6546 QTOF mass spectrometer in positive ion mode (ESI+). The following mass spectrometer conditions were used; gas temperature 200 °C, gas flow rate 10 L/min, nebulizer 35 psi, Sheath gas temperature 300 °C, capillary voltage 3500 V and sheath gas flow 12 L/min. MS1 data was acquired for m/z range 100–1200 at a scan rate for 3 spectra/s. Iterative MS2 acquisitions were performed using: MS1 m/s range 100–1200, MS1 scan rate 5 spectra/s, MS2 m/z range 50–1600, MS2 scan rate 10 spectra/s, isolate width 1.3 amu, accumulation target value 25,000 ions, active exclusion after 2 spectra for 0.2 min with a mass tolerance of 20 ppm. MS1 TQCs were monitored for changes in peak area, width, and retention time to determine the performance of the LC-MS analysis.

Metabolomic data processing

Feature picking and identification from MS1 spectral data was performed using MassHunter Explorer 1 (Agilent). Feature identification settings: Ion height filter (counts): 600, Ion Species: +H, +Na, +NH4, RT correction against TQC, Max time shift between samples: 0.5 min + 0.5%, RT tolerance: 0.3 min, Mass tolerance: 20 ppm. The total ion intensity of features were Log2 transformed and filtered (80% frequency, ≤25% coefficient of variation in at least one group). Features were identified against MassHunter METLIN Metabolite accurate mass Personal Compound Database and Library with mass tolerance of 5 ppm. Statistical analysis was performed in R using two-sample Student’s t-test. Significantly changed features were identified with Benjamini–Hochberg FDR < 0.1 against scrm ctrl.

For MS2 validation of significantly changed features, iTQC spectral data was analyzed in MassHunter Qualitative Analysis (Agilent) using Find by Molecular Feature algorithm against MassHunter METLIN accurate mass Personal Compound Database and Library. Spectral library search settings: min m/z: 30, min match score: 50, precursor tolerance 20 ± 0.05 ppm, MS/MS peak height (counts) ≥10, limit to the largest 100, isotope spacing tolerance 0.0025 m.z/7 ppm, isotope model: common organic molecules. Additional Molecular Feature database search settings: Use peaks with height ≥600, Modifiers: +H, +Na, +NH₄, Mass match tolerance: 2.00 mDa, Extract separate MS/MS spectrum per collision energy: True. MS/MS spectral data for all five iTQC were exported as a single combined MGF. Significant features found in MassHunter Explorer were then manually validated against the combined MGF using SIRIUS 5.8.6 by Molecular Formular Identification, CSI:FingerID Fingerprint Prediction and Structural Database Searching against HMDB and KEGG⁶⁴. Only metabolites with FDR < 0.1 and MS2-validated human metabolites with HMDB and/or KEGG IDs were used for pathway analysis.

Lipidomic sample preparation

HT1080 cells (2 × 10 cm plate per experiment sample) were washed twice with 2 mL of ice-cold PBS. PBS was removed by aspiration and cells were collected by scraping into pre-cooled Eppendorf tubes before centrifugation at 500 × g for 2 min at 4 °C. Supernatants were discarded, and the resulting cell pellet was resuspended in 200 µL PBS before cell lysis by sonication (3 ×5 s on ice). Protein concentration was measured using a standard bicinchoninic acid (BCA) assay and samples were diluted to 5 mg/mL prior to lipid extraction. Cellular lipids were extracted from 10 µL of cellular lysate by mixing with 100 µL of C₄H₉OH:MeOH (1:1) with 10 mM ammonium formate containing a mixture of internal standards⁶⁵. Samples were vortexed and sonicated in a water bath for 1 h at room temperature. Samples were then centrifuged (14,000 × g, 10 min, 20 °C) and transferred into sample vials with glass inserts for analysis. NIST (National Institute of Standards and Technology) human plasma standard reference material 1950 was extracted to provide alignment of lipidomic results with existing literature.

Targeted lipidomics

Targeted lipidomics to measure 737 lipid species across 43 classes⁶⁵. In brief, extracted cellular lipids were analyzed in one batch along with quality control samples spaced throughout. These included pooled plasma QCs (PQC) (1 in every 10 samples), technical QCs (TQC) (1 in every 10 samples) and NIST1950 (1 in every 10 samples). Samples (1 µL) were separated on an Agilent 1290 series HPLC system and a ZORBAX eclipse plus C18 column (2.1 × 100 mm 1.8 μm, Agilent) with the thermostat set at 60 °C. The solvent system consisted of Solvent (A) H₂O:ACN:IPA (50:30:20) with 10 mM ammonium formate and solvent (B) H₂O:ACN:IPA (1:9:90) containing 10 mM ammonium formate with lipid elution using a stepped linear gradient at 0.4 mL/min with a 15 min cycle time per sample.

The gradient was as follows: starting at 10% B and increasing to 45% B over 2.7 min, then to 53% over 0.1 min, to 65% over 6.2 min, to 89% over 0.1 min, to 92% over 1.9 min and finally to 100% over 0.1 min. The solvent was then held at 100% B for 0.8 min (total 11.9 min). Equilibration was as follows: solvent was decreased from 100% B to 10% B over 0.1 min and held for an additional 0.9 min. Flow rate was then switched to 0.6 mL/min for 1 min before returning to 0.4 mL/min over 0.1 min. Solvent B was held at 10% B for a further 0.9 min at 0.4 mL/min for a total cycle time of 15 min.

Mass spectrometry analysis was performed on an Agilent 6490 QQQ mass spectrometer in positive ion mode (ESI+) with dynamic scheduled multiple reaction monitoring (MRM) using published MRM transitions for each lipid class, subclass, and individual species⁶⁵. In brief, the following mass spectrometer conditions were used; gas temperature, 150 °C, gas flow rate 17 L/min, nebulizer 20 psi, Sheath gas temperature 200 °C, capillary voltage 3500 V and sheath gas flow 10 L/min. Isolation widths for Q1 and Q3 were set to “unit” resolution (0.7 amu). TQCs were monitored for changes in peak area, width, and retention time to determine the performance of the LC-MS/MS analysis.

Lipidomic data processing

Quantification of lipid species was determined by comparison to the relevant internal standards⁶⁵. In brief, peaks were integrated using Masshunter (Agilent), and peak areas were reported. Areas were put through a QC pipeline followed by background subtractions (mean concentration of each lipid species present in blank samples). Concentrations (pmol lipid/mg of cell homogenate) were calculated post-QC. Lipid species data (pmol lipid/mg protein) was Log2 transformed and statistical analysis was performed in R using two sample Student’s t-test. Significantly changed features were identified with Benjamini–Hochberg FDR < 0.1 against scrm ctrl. Only lipids with FDR < 0.1 and HMDB and/or PubChem IDs were used for pathway analysis.

Function and pathway analysis

Functional assessment of differentially expressed genes was analyzed with Gene Ontology (GO) against a reference background of “whole genome in human” using a Bonferroni corrected P < 0.05 for statistical significance⁶⁶. KEGG, Wiki and Reactome pathways for differentially expressed genes were analyzed using KEGG pathway, WikiPathways, and Reactome pathway databases, respectively^67–69. Lipid reaction network analysis of lipidomics data was performed using LipidSig 2.0⁷⁰. Differentially expressed genes, proteins, lipids, and metabolites were separately analyzed using Ingenuity Pathway Analysis (IPA, Ingenuity System) against a reference background of “whole genome in animal and human” to identify changes in biological pathways for each ‘omics dataset.

Multi-omic correlation analysis

Raw counts or measurements from upregulated or downregulated transcriptomics, proteomics, lipidomics, and metabolomics data in PHACTR1 overexpression or knockdown versus scramble experiments were normalized and analyzed using mixOmics R package for integration and multivariate correlation network⁷¹. For conditions that had two datasets reach significance threshold, such as in PHACTR1 KD upregulated group, (regularized) Canonical Correlation Analysis (rCCA) method was used. For groups that had three (PHACTR1 KD downregulated group) or four datasets (PHACTR1 OE upregulated and downregulated groups), N-Integration Discriminant Analysis with Data Integration Analysis for Biomarker discovery using Latent variable approaches for Omics (DIABLO) was used for the integration of multiple datasets in a supervised analysis⁷². To optimize the number of molecules presented in the networks, the cut-off value for correlation was 0.95 for integrating three or more datasets, or 0.7 for integrating two datasets. Results from correlation network analysis were imported to for visualization⁷³.

Gene-protein correlations

For the differentially expressed genes and proteins (FDR < 0.1), expression matrices were normalized, and correlation was performed between the two ‘omics dataset for the knockdown versus scramble and overexpression versus scramble conditions in both up- and down-regulated groups. The top 25 negatively correlated gene-protein pairs in each group of experimental conditions with cut-off value for correlation below −0.5 were selected to make negative correlation plots. The analysis was performed, and graphs were made using R.

Integrative pathway analysis

IPA pathways from up- or down-regulated transcriptome, proteome, lipidome, and metabolome data in PHACTR1 KD or OE groups were merged into multi-omics IPA pathways for each condition and plotted using ggplot2 in R.

Flow cytometry and cell cycle discrimination

5-Ethynyl-2 deoxyuridine (EdU) flow cytometry analysis was performed using OpenEMMU^20,74. Briefly, 23 h after media change, EdU (10 μM final concentration) was added to the culture medium and incubated for 1 h. For negative staining controls, we included cells that had not been exposed to EdU. Once each treatment was finished, cells were washed with PBS and dissociated in 1 mL (6-well plate) of pre-warmed TrypLE (ThermoFisher Scientific, #12604013). After TrypLE incubation, cold PBS was added, cells were fully dissociated by pipetting, and samples collected. Samples were centrifuged at 500 × g for 5 min. Then, the supernatant was discarded, and the pellet of cells resuspended with 0.4 mL of cold PBS. Cells were fixed by adding 0.5 mL of 4% PFA into ~0.5 mL of cell suspension. Cells were incubated for 10 min at room temperature, and then ~2% PFA was washed three times with PBS. When cells were ready to work with, 0.1% BSA in PBS was added, and cells pelleted at 500 × g for 5 min. Pelleted cells were flicked and 400 µL of a saponin-based permeabilization and wash reagent (0.2% saponin containing 4% FBS, 1% BSA and 0.02% Sodium Azide in PBS) was added and incubated for 15 min. Then the cells were centrifuged (500 × g for 5 min).

EdU detection was performed using an in-house developed click EdU reaction cocktail (OpenEMMU) made of 200 nM AZDyeTM 488 Azide (Click Chemistry Tools, #1275), 800 μM Copper (II) sulfate, and 5 mM Ascorbic acid in PBS. In brief, 400 µL per sample of the click reaction cocktail was added to the pellet, and the cells resuspended and incubated for 30 min at room temperature, protected from light. After click EdU reaction cocktail incubation, the cells were washed three times with saponin-based permeabilization and wash reagent and pelleted at 500 × g for 5 min, leaving 50 µL of pellet per tube which was resuspended by flicking. Then, 100 µL of conjugated antibody prepared in perm/wash buffer (Alexa Fluor 594 anti-histone H3 Phospho (Ser10), clone 11D8, Mouse IgG2b, κ, 1:250 dilution, BioLegend) was added and incubated overnight at room temperature. After antibody incubation, 800 µL of 0.1% BSA in PBS was added to each tube, and samples were spun down at 500 × g for 6 min. The supernatants were removed and 50 µL pellets resuspended by flicking and incubated with 300 µL of Hoechst 33342 (10 μg/mL final concentration, Sigma-Aldrich, #B2261) for 10 min at room temperature in saponin-based permeabilization and wash reagent.

Samples were analyzed by flow cytometry for DNA content and EdU labelled cells using a BD LSR Fortessa Laser Cell Analyzer (BD Biosciences, Erembodegem, Belgium) equipped with five excitation lasers (UV 355 nm, Violet 405 nm, Blue 488 nm, Yellow/Green 561 nm, and Red 633 nm). FSC-H versus FSC-A, FSC-H versus FSC-W and SSC-H versus SSC-W cytograms were used to discriminate and gate out doublets/cell aggregates during analysis. EdU-AZDyeTM 488 Azide and pH3S10-Alexa Fluor 594 fluorescence were detected with logarithmic amplification using the B530 (530/30) and YF610 (610/20) detectors, respectively, whereas Hoechst fluorescence was detected with linear amplification using the V450 (V450/50) detector.

Flow cytometry measurements were run at a mid-flow rate, and the core stream allowed to stabilize for 5 s prior to acquisition. Data were collected using FacsDIVA 8 software. For optimal Hoechst signal detection and cell cycle progression analyzes, an event concentration of <800 events/s was used, and 20,000 events were captured. All flow cytometry data were analyzed using FlowJo Portal (v.10.8.1, Becton Dickinson & Company) using Mac OS X operating system.

Cell size measurements and mitochondrial mass quantification protocol

Live cell size measurements were performed by flow cytometry, using single-cell cytograms to identify and exclude doublets or cell aggregates. Gating was based on the forward scatter area (FSC-A) linear values, expressed as arbitrary units (AU), to ensure accurate size determination using a BD LSR Fortessa Laser Cell Analyzer.

To quantify mitochondrial mass, MitoSpy™ Green FM (BioLegend) was used following the manufacturer’s protocol for live-cell labeling using a BD LSR Fortessa Laser Cell Analyzer. In brief, cells were dissociated, resuspended, and incubated with 200 nM MitoSpy™ Green FM in complete medium without serum at 37 °C for 10 min. After incubation, the cells were washed thoroughly with 1× PBS to remove excess dye. MitoSpy™ Green FM fluorescence intensity was measured using flow cytometry with the B-530A detector, ensuring that fluorescence signals were acquired from live cells. The data were analyzed to quantify mitochondrial mass based on fluorescence intensity values.

Gel electrophoresis and transfer

Cells were plated 24 h prior to harvest at 2.2 × 10⁵ cells/mL in DMEM with 10% FBS and 0.2 μg/mL puromycin in 10 cm plates. Cells were washed once in ice-cold PBS and lysed into RIPA buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 5 mM EDTA, 1% sodium deoxycholate, 1% Triton X-100, 0.1% sodium dodecyl sulfate [SDS]) with the addition of 1X Halt Protease and Phosphatase Inhibitor Cocktail (PI, ThermoFisher Scientific, #78440). Plates were incubated on ice for 10 min before being scrapped into pre-cooled Eppendorf tubes. Samples were vortexed and centrifuged at 21,000 × g for 10 min at 4 °C. The resulting supernatant was transferred to a fresh Eppendorf tube and BCA was performed to determine protein concentration. Protein lysates were made up at 0.75 µg/µL in sample buffer (50 mM Tris-HCl pH 6.8, 2% SDS, 0.1% bromophenol blue, 10% glycerol) with the addition of 5% β-mercaptoethanol. Samples were heated to 95 °C for 10 min before centrifugation at 21,000 × g for 10 min. Samples (25 µg protein) were loaded onto 4–15% Mini-PROTEAN TGX Stain-Free Protein Gels (Bio-Rad, #4568086) and resolved by gel electrophoresis. Gels were transferred to PVDF membranes (Bio-Rad, #1704156) using Trans-Blot Turbo Transfer System.

Immunoblotting

PVDF membranes were blocked in EveryBlot Blocking Buffer (Bio-Rad, #12010020) for 30 min followed by overnight incubation with primary antibodies diluted 1:1000 in EveryBlot Blocking Buffer, unless specified. Primary antibodies used were GAPDH (Sigma-Aldrich, #G8795, diluted 1:10,000), AKAP1 (Sigma-Aldrich, #SAB4503247, diluted 1:500), PHACTR1 (Abcam, #ab229120), Total OXPHOS Blue Native WB Antibody Cocktail (Abcam, #ab110412), HMOX1 (Santa Cruz Biotechnology, #sc-7695) and the following all from Cell Signaling: CDT1 (#8064), Thymidine Kinase 1 (#28755), Phospho-Histone H3 (Ser10, #53348), Cyclin A2 (#91500), Cyclin B1 (#12231), Cyclin E1 (#20808), Phospho-cdc2 (Tyr15, #4539), Tom20 (#42406), OPA1 (#80471), DRP1 (#8570), Phospho-DRP1 (Ser616, #4494), Phospho-DRP1 (Ser637, #6319), MFF (#84580), Mitofusin-1 (#14739), Mitofusin-2 (#11925), GPX4 (#52455), NCOA4 (#66849), KEAP1 (#8047), FTH1 (#4393) and 4F2hc/SLC3A2 (#47213). Membranes were washed for 3 × 5 min with 0.1% Tween 20 (TBS-Tween) followed by a 2 h incubation at room temperature with the appropriate secondary antibody: anti-mouse IgG HRP-linked (Cell Signaling, #7076) or anti-rabbit IgG HRP-linked (Cell Signaling, #7074) diluted 1:1000 in EveryBlot Blocking Buffer. Membranes were washed 3 × 5 min with TBS-Tween and were detected using Clarity Western ECL Substrate (Bio-Rad, #1705061) and imaged on a ChemiDoc MP Imaging System (Bio-Rad). Quantification of protein bands was performed using Image Lab (v.6.0, Bio-Rad).

Transient transfection of HUVEC

PHACTR1 was knocked down or overexpressed in HUVECs using transient chemical transfection. HUVECs (P3-5) were plated in 6-well plates 24 h prior to transfection. For PHACTR1 overexpression, 2.5 µg PHACTR1 OE construct in expression vector (as used above for lentiviral generation) was transfected into cells using Lipofectamine LTX with plus reagent (Invitrogen, #15338-100). PHACTR1 was knocked down by siRNA transfection using 15 pmol Silencer Select PHACTR1 siRNA (ThermoFisher Scientific, #s48095) with Lipofectamine RNAiMAX Transfection Reagent (ThermoFisher Scientific, #13778150). Control cells were transfected with Silencer Select Negative Control No. 1 siRNA (ThermoFisher Scientific, #4390843). All transfections were performed in a final volume of 2 mL Opti-MEM (Gibco, #11058021) per well. HUVECs were incubated for 4 h with Lipofectamine complex before the media was changed to complete Human Large Vessel Endothelial Cell Basal Medium. Cells were harvested in RIPA buffer supplemented with 1X PI 48 h post transfection and analyzed by gel electrophoresis and immunoblotting, as described above.

Mitochondrial isolation

Mitochondria were isolated from HT1080 cells lines using a Mitochondria Isolation Kit for Cultured Cells (Thermo Scientific, #89874) as per the manufacturer’s instructions. 2 × 10⁷ cells per condition, grown over 2 × 150 mm plates, were trypsinized and harvested by centrifugation at 880 × g for 2 min. Cell pellets were resuspended in 800 µL Reagent A supplemented with protease inhibitors and vortexed before being incubated on ice for 2 min before being transferred to a pre-cooled Dounce homogenizer. Cells were lysed using 60 strokes of the homogenizer before being transferred to a fresh Eppendorf tube containing 800 µL Reagent C supplemented with protease inhibitors. The Dounce was washed with 200 µL Reagent A, the wash was added to the cell lysate and samples were mixed by inversion. Samples were centrifuged at 700 × g for 10 min at 4 °C following which the supernatant was transferred to a fresh tube and centrifuged at 12,000 × g for 5 min at 4 °C. The resulting pellet, containing isolated mitochondria, was washed with 500 µL Reagent C and centrifuged again at 12,000 × g for 5 min at 4 °C. Pellets were resuspended in 100 µL sample buffer with the addition of 5% β-mercaptoethanol and analyzed by gel electrophoresis and immunoblotting, as described above.

Linear regression of GTEx gene expression

PHACTR1 and AKAP1 gene expression values in human aorta, coronary artery and tibial artery tissues were obtained from the Genotype-Tissue Expression (GTEx) v8 (gtexportal.org) bulk RNAseq transcript reads per kilobases per million (RPKM) files. RPKM divides total reads by 1,000,000 (“per million”) scaling factor (RPM), which normalizes for sequencing depth and further divides the RPM values by the length of the gene, in kilobases. Correlations between PHACTR1 and AKAP1 normalized gene expression RPKM values in different arterial tissues were analyzed using R.

Cell respirometry

Cellular adenosine triphosphate (ATP) production from glycolysis and mitochondria respiration was measured using the ATP Seahorse XF Real-Time ATP Rate Assay Kit (Agilent, #103592-100) on a Seahorse XF24 Analyzer (Agilent). In brief, HT1080 cell lines were plates at 3.5 × 10⁴ cells/well in XF24 cell culture microplates (Agilent, # 102340-100). Cells were left to adhere for 90 min at room temperature followed by an overnight incubation at 37 °C, 5% CO₂. 1 h prior to analysis, cells were washed with Seahorse XF DMEM medium (Agilent, #103575-100) supplemented with 10 mM of XF glucose, 1 mM of XF pyruvate and 2 mM of XF glutamine. The media was replaced with 500 µL supplemented Seahorse XF DMEM medium and plates were incubated at 37°C while the Seahorse XF24 Analyzer underwent calibration. The media on cells were replaced with fresh supplemented Seahorse XF DMEM medium and a pre-soaked XFe24 sensor cartridge (Agilent, #102340-100) was placed onto the cell culture plate. ATP production was measured over 72 min with the addition of 1.5 µM Oligomycin at 24 min and 0.5 µM Rotenone-Antimycin A at 48 min. At the end of the assay, media was removed, and cells were lysed in 1% SDS. A standard BCA was performed on cell lysates and Seahorse data was normalized to mg protein. Seahorse raw data was analyzed using Agilent Seahorse Analytics portal (seahorseanalytics.agilent.com).

Statistics and reproducibility

Statistical analysis and results presentation of ‘omic data were conducted using R v4.3.3. Statistical analysis and results presentation of non-omic data were conducted using GraphPad Prism v10.4.0, with the exception of the linear regression of gene expression analysis which was conducted using R v4.3.3.

Sample size for all ‘omic dataset generation was n = 6 per group. One lipidomics replicate from scrm ctrl group was excluded as an outlier, identified by Z-score extremeness & PCA. The sample size for pathway validation by immunoblotting was n = 3–6 per group in HT1080 cell lines and n = 4 per group in HUVECs. FACS sample size was n = 3 per group. Linear regression of gene expression was conducted with n = 432 samples from human aorta, n = 241 from coronary artery, and n = 664 from tibial artery from the GTEx dataset. Seahorse bioenergetics samples size was n = 11–12 per group. All replicates are biological replicates.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Supplementary information

Transparent Peer Review file^{(822.7KB, pdf)}

Supplementary Information^{(6MB, pdf)}

42003_2026_9542_MOESM3_ESM.pdf^{(38.4KB, pdf)}

Description of Additional Supplementary Files

Supplementary Data 1-40^{(230MB, xlsx)}

Supplementary Data 41^{(211.5KB, xlsx)}

Reporting Summary^{(2.5MB, pdf)}

Acknowledgements

J.C.K. acknowledges research support from New South Wales health grant RG194194, the Leducq Foundation (25CVD02), the Bourne Foundation, the Angles Family Foundation, Snow Medical, SCAD Research Inc. and Agilent. K.W. acknowledges research support from a Victor Chang Cardiac Research Institute Mid-Career Research Grant. RPH was supported by grants from the National Health and Medical Research Council (NHMRC) of Australia (2000615; 2008743). OC acknowledges support from the Australian Research Council (DP210102134), Medical Research Futures Fund Stem Cell Therapy Mission (2024/MRF2032746), The Victor Chang Cardiac Research Institute Outstanding EMCR Grant, and Miltenyi Research Grant. We acknowledge the Victor Chang Cardiac Research Institute’s Innovation Centre, funded by the NSW Government, and funding from the Freedman Foundation for the Metabolomics Facility. We especially thank Dr Muhammad Ali and Esther Kristianto from the Metabolomics Facility of the Innovation Centre for technical advice with mass spectrometry analysis. The Genotype-Tissue Expression Project was supported by the Common Fund of the Office of the Director of the National Institutes of Health, and by NCI, NHGRI, NHLBI, NIDA, NIMH, and NINDS. We thank the teams at ArrayExpress, the Mass Spectrometry Interactive Virtual Environment (MassIVE), and the Metabolomics Workbench for curating and providing public access to the data. The Metabolomics Workbench is supported by Metabolomics Workbench/National Metabolomics Data Repository (U2C-DK119886), Common Fund Data Ecosystem (3OT2OD030544) and Metabolomics Consortium Coordinating Center (1U2C-DK119889). BioRender was used to create Fig. 1A: Wolhuter, K. (2025) BioRender.com/e53c630 and the Graphical Abstract: Wolhuter, K. (2026) BioRender.com/2tkfj30.

Author contributions

Conceptualization, J.C.K. and K.W.; Methodology, K.W., L.M., N.S.B., O.C., N.M., L.Z., C.T., S.E.I., C.G., R.P.H., P.J.M., J.L.M.B. and J.C.K.; Software, L.M., J.L.M.B. and J.C.K; Formal Analysis, L.M., K.W., N.S.B, O.C., N.M. and L.Z.; Investigation, K.W., L.M., N.S.B., O.C., N.M., L.Z. and C.T.; Writing – Original Draft, K.W. and J.C.K., Writing – Review & Editing, all authors; Visualization, K.W. and L.M.; Supervision, T.H., C.F. and D.B.; Funding Acquisition, K.W. and J.C.K.

Peer review

Peer review information

Communications Biology thanks Seyed Amir Malekpour and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary Handling Editors: Jian-Da Lin and David Favero. A peer review file is available.

Data availability

The ‘omics datasets generated and analyzed during the current study are publicly available. The transcriptomics data have been deposited in the ArrayExpress database (https://www.ebi.ac.uk/biostudies/arrayexpress) under accession number E-MTAB-15583⁷⁵. The proteomics data have been deposited in the MassIVE database under Project ID MSV000099266 (10.25345/C5XD0RB0M)⁷⁶. The metabolomics and lipidomics data have been deposited in Metabolomics Workbench/National Metabolomics Data Repository under Project ID PR002646 with Study IDs ST004197 and ST004202 (10.21228/M8MV84)⁷⁷. Source data for all graphs are provided in Supplementary Data 41. The FACS gating strategy is provided in Supplementary Fig. 12. Uncropped immunoblots are provided in Supplementary Figs. 13–18. All other data are available from the corresponding author upon reasonable request.

Competing interests

The authors declare the following competing interests: D.B., T.H., and C.F. are employees and shareholders of Agilent Technologies. J.C.K. is the recipient of an Agilent Thought Leader Award, and part of the award for this prize was used to support the research in this study. All other authors declare no competing interests.

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

The online version contains supplementary material available at 10.1038/s42003-026-09542-w.

References

1.Beaudoin, M. et al. Myocardial infarction-associated SNP at 6p24 interferes with MEF2 binding and associates with PHACTR1 expression levels in human coronary arteries. Arterioscler. Thromb. Vasc. Biol.35, 1472–1479 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Hao, K. et al. Integrative prioritization of causal genes for coronary artery disease. Circ. Genom. Precis. Med.15, e003365 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
3.O’Donnell, C. J. et al. Genome-wide association study for coronary artery calcification with follow-up in myocardial infarction. Circulation124, 2855–2864 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Myocardial Infarction Genetics, C. et al. Genome-wide association of early-onset myocardial infarction with single nucleotide polymorphisms and copy number variants. Nat. Genet.41, 334–341 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Georges, A. et al. Genetic investigation of fibromuscular dysplasia identifies risk loci and shared genetics with common cardiovascular diseases. Nat. Commun.12, 6031 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Kiando, S. R. et al. PHACTR1 is a genetic susceptibility locus for fibromuscular dysplasia supporting its complex genetic pattern of inheritance. PLoS Genet.12, e1006367 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Adlam, D. et al. Association of the PHACTR1/EDN1 genetic locus with spontaneous coronary artery dissection. J. Am. Coll. Cardiol.73, 58–66 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Debette, S. et al. Common variation in PHACTR1 is associated with susceptibility to cervical artery dissection. Nat. Genet.47, 78–83 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Fujimaki, T. et al. Association of a transcription factor 21 gene polymorphism with hypertension. Biomed. Rep.3, 118–122 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Fedoryshchak, R. O. et al. Molecular basis for substrate specificity of the Phactr1/PP1 phosphatase holoenzyme. eLife9, e61509 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Fils-Aimé, N. et al. MicroRNA-584 and the protein phosphatase and actin regulator 1 (PHACTR1), a new signaling route through which transforming growth factor-β Mediates the migration and actin dynamics of breast cancer cells. J. Biol. Chem.288, 11807–11823 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Kelloniemi, A. et al. The early-onset myocardial infarction associated PHACTR1 gene regulates skeletal and cardiac alpha-actin gene expression. PLoS ONE10, e0130502 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Ma, X. et al. PHACTR1, a coronary artery disease risk gene, mediates endothelial dysfunction. Front. Immunol.13, 958677 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Dai, X. & Shen, L. Advances and trends in omics technology development. Front. Med.9, 911861 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Johnson, K. W. et al. Enabling precision cardiology through multiscale biology and systems medicine. JACC Basic Transl. Sci.2, 311–327 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Mohammadi-Shemirani, P., Sood, T. & Pare, G. From ‘omics to multi-omics technologies: the discovery of novel causal mediators. Curr. Atheroscler. Rep.25, 55–65 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Santiago-Rodriguez, T. M. & Hollister, E. B. Multi ‘omic data integration: a review of concepts, considerations, and approaches. Semin. Perinatol.45, 151456 (2021). [DOI] [PubMed] [Google Scholar]
18.Tarazona, S., Arzalluz-Luque, A. & Conesa, A. Undisclosed, unmet and neglected challenges in multi-omics studies. Nat. Comput. Sci.1, 395–402 (2021). [DOI] [PubMed] [Google Scholar]
19.Wiezlak, M. et al. G-actin regulates the shuttling and PP1 binding of the RPEL protein Phactr1 to control actomyosin assembly. J. Cell Sci.125, 5860–5872 (2012). [DOI] [PubMed] [Google Scholar]
20.Contreras, O. et al. OpenEMMU: a versatile, open-source EdU multiplexing methodology for studying DNA replication and cell cycle dynamics. iScience28, 113380 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Gavet, O. & Pines, J. Activation of cyclin B1-Cdk1 synchronizes events in the nucleus and the cytoplasm at mitosis. J. Cell Biol.189, 247–259 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Rona, G. & Pagano, M. CDT1, a licensing factor that limits rereplication. Mol. Cell83, 1–3 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Liu, R. & Chan, D. C. The mitochondrial fission receptor Mff selectively recruits oligomerized Drp1. Mol. Biol. Cell26, 4466–4477 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Ko, H. J. et al. The phosphorylation status of Drp1-Ser637 by PKA in mitochondrial fission modulates mitophagy via PINK1/Parkin to exert multipolar spindles assembly during mitosis. Biomolecules1110.3390/biom11030424 (2021). [DOI] [PMC free article] [PubMed]
25.Lau, E. et al. Integrated omics dissection of proteome dynamics during cardiac remodeling. Nat. Commun.9, 120 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Garmany, R. et al. Multi-omic architecture of obstructive hypertrophic cardiomyopathy. Circ. Genom. Precis. Med.16, e003756 (2023). [DOI] [PubMed] [Google Scholar]
27.Wang, W. et al. Metabolic characterization of hypertrophic cardiomyopathy in human heart. Nat. Cardiovasc. Res.1, 445–461 (2022). [DOI] [PubMed] [Google Scholar]
28.Hill, M. C. et al. Integrated multi-omic characterization of congenital heart disease. Nature608, 181–191 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Subramanian, I., Verma, S., Kumar, S., Jere, A. & Anamika, K. Multi-omics data integration, interpretation, and its application. Bioinform. Biol. Insights14, 1177932219899051 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Chen, C. et al. Applications of multi-omics analysis in human diseases. MedComm4, e315 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Mouilleron, S., Wiezlak, M., O’Reilly, N., Treisman, R. & McDonald, N. Q. Structures of the Phactr1 RPEL domain and RPEL motif complexes with G-actin reveal the molecular basis for actin binding cooperativity. Structure20, 1960–1970 (2012). [DOI] [PubMed] [Google Scholar]
32.Fils-Aime, N. et al. MicroRNA-584 and the protein phosphatase and actin regulator 1 (PHACTR1), a new signaling route through which transforming growth factor-beta Mediates the migration and actin dynamics of breast cancer cells. J. Biol. Chem.288, 11807–11823 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Ihara, D. et al. The BDNF-ERK/MAPK axis reduces phosphatase and actin regulator1, 2 and 3 (PHACTR1, 2 and 3) mRNA expressions in cortical neurons. Drug Discov. Ther.18, 255–259 (2024). [DOI] [PubMed] [Google Scholar]
34.Dohadwala, M. et al. Phosphorylation and inactivation of protein phosphatase 1 by cyclin-dependent kinases. Proc. Natl. Acad. Sci. USA91, 6408–6412 (1994). [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Grallert, A. et al. A PP1-PP2A phosphatase relay controls mitotic progression. Nature517, 94–98 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Chadee, D. N. et al. Increased Ser-10 phosphorylation of histone H3 in mitogen-stimulated and oncogene-transformed mouse fibroblasts. J. Biol. Chem.274, 24914–24920 (1999). [DOI] [PubMed] [Google Scholar]
37.Liu, M. et al. Sox17 is required for endothelial regeneration following inflammation-induced vascular injury. Nat. Commun.10, 2126 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Wolhuter, K., Kong, S. M. Y., Stanley, C. P. & Kovacic, J. C. The role of oxidants in percutaneous coronary intervention-induced endothelial dysfunction: can we harness redox signaling to improve clinical outcomes? Antioxid. Redox Signal38, 1022–1040 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Vinchi, F. et al. Atherosclerosis is aggravated by iron overload and ameliorated by dietary and pharmacological iron restriction. Eur. Heart J.41, 2681–2695 (2020). [DOI] [PubMed] [Google Scholar]
40.Wu, J. et al. NEK2 affects the ferroptosis sensitivity of gastric cancer cells by regulating the expression of HMOX1 through Keap1/Nrf2. Mol. Cell Biochem.10.1007/s11010-024-04960-y (2024). [DOI] [PMC free article] [PubMed]
41.Pi, J. et al. Molecular mechanism of human Nrf2 activation and degradation: role of sequential phosphorylation by protein kinase CK2. Free Radic. Biol. Med.42, 1797–1806 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]
42.Salvi, F. et al. Effects of stably incorporated iron on protein phosphatase-1 structure and activity. FEBS Lett.592, 4028–4038 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Kim, E. S. H., Saw, J., Kadian-Dodov, D., Wood, M. & Ganesh, S. K. FMD and SCAD: sex-biased arterial diseases with clinical and genetic pleiotropy. Circ. Res.128, 1958–1972 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Ryan, B. J., Charkoudian, N. & McClung, J. P. Consider iron status when making sex comparisons in human physiology. J. Appl. Physiol.132, 699–702 (2022). [DOI] [PubMed] [Google Scholar]
45.Yu, E., Mercer, J. & Bennett, M. Mitochondria in vascular disease. Cardiovasc. Res.95, 173–182 (2012). [DOI] [PubMed] [Google Scholar]
46.Vissing, C. R., Duno, M., Wibrand, F., Christensen, M. & Vissing, J. Hydroxylated long-chain acylcarnitines are biomarkers of mitochondrial myopathy. J. Clin. Endocrinol. Metab.104, 5968–5976 (2019). [DOI] [PubMed] [Google Scholar]
47.Pan, Y. et al. Decreased free fatty acid levels associated with adverse clinical outcomes in coronary artery disease patients with type 2 diabetes: findings from the PRACTICE study. Eur. J. Prev. Cardiol.30, 730–739 (2023). [DOI] [PubMed] [Google Scholar]
48.Chang, C. R. & Blackstone, C. Cyclic AMP-dependent protein kinase phosphorylation of Drp1 regulates its GTPase activity and mitochondrial morphology. J. Biol. Chem.282, 21583–21587 (2007). [DOI] [PubMed] [Google Scholar]
49.Shenouda, S. M. et al. Altered mitochondrial dynamics contributes to endothelial dysfunction in diabetes mellitus. Circulation124, 444–453 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
50.Moiz, B., Li, A., Padmanabhan, S., Sriram, G. & Clyne, A. M. Isotope-assisted metabolic flux analysis: a powerful technique to gain new insights into the human metabolome in health and disease. Metabolites1210.3390/metabo12111066 (2022). [DOI] [PMC free article] [PubMed]
51.Moiz, B. et al. (13)C Metabolic flux analysis indicates endothelial cells attenuate metabolic perturbations by modulating TCA activity. Metabolites1110.3390/metabo11040226 (2021). [DOI] [PMC free article] [PubMed]
52.Alcazar, O. et al. Parallel multi-omics in high-risk subjects for the identification of integrated biomarker signatures of type 1 diabetes. Biomolecules1110.3390/biom11030383 (2021). [DOI] [PMC free article] [PubMed]
53.Yadikar, H. et al. Deciphering early and progressive molecular signatures in Alzheimer’s disease through integrated longitudinal proteomic and pathway analysis in a rodent model. Int. J. Mol. Sci.2510.3390/ijms25126469 (2024). [DOI] [PMC free article] [PubMed]
54.Karp, P. D., Midford, P. E., Caspi, R. & Khodursky, A. Pathway size matters: the influence of pathway granularity on over-representation (enrichment analysis) statistics. BMC Genom.22, 191 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Tsouka, S. & Masoodi, M. Metabolic pathway analysis: advantages and pitfalls for the functional interpretation of metabolomics and lipidomics data. Biomolecules1310.3390/biom13020244 (2023). [DOI] [PMC free article] [PubMed]
56.Adamson, P., Marshall, C. J., Hall, A. & Tilbrook, P. A. Post-translational modifications of p21rho proteins. J. Biol. Chem.267, 20033–20038 (1992). [PubMed] [Google Scholar]
57.Lorenzo, D. N. & Bennett, V. Cell-autonomous adiposity through increased cell surface GLUT4 due to ankyrin-B deficiency. Proc. Natl. Acad. Sci. USA114, 12743–12748 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
58.Lauzier, A. et al. Snazarus and its human ortholog SNX25 modulate autophagic flux. J. Cell Sci.13510.1242/jcs.258733 (2022). [DOI] [PubMed]
59.Briot, A. et al. Endothelial NOTCH1 is suppressed by circulating lipids and antagonizes inflammation during atherosclerosis. J. Exp. Med.212, 2147–2163 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
60.Andrews, S. FastQC: A Quality Control Tool for high-throughput Sequence Data. http://www.bioinformatics.babraham.ac.uk/projects/fastqc/ (2023).
61.Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics29, 15–21 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
62.Ritchie, M. E. et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res.43, e47 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
63.Shah, A. D., Goode, R. J. A., Huang, C., Powell, D. R. & Schittenhelm, R. B. LFQ-analyst: an easy-to-use interactive web platform to analyze and visualize label-free proteomics data preprocessed with MaxQuant. J. Proteome Res.19, 204–211 (2020). [DOI] [PubMed] [Google Scholar]
64.Duhrkop, K. et al. SIRIUS 4: a rapid tool for turning tandem mass spectra into metabolite structure information. Nat. Methods16, 299–302 (2019). [DOI] [PubMed] [Google Scholar]
65.Huynh, K. et al. High-throughput plasma lipidomics: detailed mapping of the associations with cardiometabolic risk factors. Cell Chem. Biol.26, 71–84 e74 (2019). [DOI] [PubMed] [Google Scholar]
66.Gene Ontology, C. et al. The gene ontology knowledgebase in 2023. Genetics22410.1093/genetics/iyad031 (2023).
67.Kanehisa, M. & Goto, S. KEGG: kyoto encyclopedia of genes and genomes. Nucleic Acids Res.28, 27–30 (2000). [DOI] [PMC free article] [PubMed] [Google Scholar]
68.Agrawal, A. et al. WikiPathways 2024: next generation pathway database. Nucleic Acids Res.52, D679–D689 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
69.Milacic, M. et al. The reactome pathway knowledgebase 2024. Nucleic Acids Res.52, D672–D678 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
70.Liu, C. H. et al. LipidSig 2.0: integrating lipid characteristic insights into advanced lipidomics data analysis. Nucleic Acids Res.52, W390–W397 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
71.Rohart, F., Gautier, B., Singh, A. & Le Cao, K. A. mixOmics: An R package for ‘omics feature selection and multiple data integration. PLoS Comput. Biol.13, e1005752 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
72.Singh, A. et al. DIABLO: an integrative approach for identifying key molecular drivers from multi-omics assays. Bioinformatics35, 3055–3062 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
73.Shannon, P. et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res.13, 2498–2504 (2003). [DOI] [PMC free article] [PubMed] [Google Scholar]
74.Contreras, O. & Harvey, R. P. Single-cell transcriptome dynamics of the autotaxin-lysophosphatidic acid axis during muscle regeneration reveal proliferative effects in mesenchymal fibro-adipogenic progenitors. Front. Cell Dev. Biol.11, 1017660 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
75.Wolhuter, K. & Kovacic, J. C. RNA-seq of PHACTR1 Knockdown and Overexpression in HT1080 Stable Cell Lines [Data set] ArrayExpresswww.ebi.ac.uk/biostudies/arrayexpress/studies/E-MTAB-15583 (2025).
76.Wolhuter, K. & Kovacic, J. C. Proteomic Analysis of PHACTR1 Overexpressing and Knockdown HT1080 Stable Cell Lines [Data set] MassIVE10.25345/C5XD0RB0M (2025).
77.Wolhuter, K. & Kovacic, J. C. Metabolomic Analysis of PHACTR1 Overexpressing and Knockdown HT1080 Stable Cell Lines [Data set] Metabolomics Workbench10.21228/M8MV84 (2025).

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Transparent Peer Review file^{(822.7KB, pdf)}

Supplementary Information^{(6MB, pdf)}

42003_2026_9542_MOESM3_ESM.pdf^{(38.4KB, pdf)}

Description of Additional Supplementary Files

Supplementary Data 1-40^{(230MB, xlsx)}

Supplementary Data 41^{(211.5KB, xlsx)}

Reporting Summary^{(2.5MB, pdf)}

Data Availability Statement

[CR1] 1.Beaudoin, M. et al. Myocardial infarction-associated SNP at 6p24 interferes with MEF2 binding and associates with PHACTR1 expression levels in human coronary arteries. Arterioscler. Thromb. Vasc. Biol.35, 1472–1479 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR2] 2.Hao, K. et al. Integrative prioritization of causal genes for coronary artery disease. Circ. Genom. Precis. Med.15, e003365 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR3] 3.O’Donnell, C. J. et al. Genome-wide association study for coronary artery calcification with follow-up in myocardial infarction. Circulation124, 2855–2864 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR4] 4.Myocardial Infarction Genetics, C. et al. Genome-wide association of early-onset myocardial infarction with single nucleotide polymorphisms and copy number variants. Nat. Genet.41, 334–341 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR5] 5.Georges, A. et al. Genetic investigation of fibromuscular dysplasia identifies risk loci and shared genetics with common cardiovascular diseases. Nat. Commun.12, 6031 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR6] 6.Kiando, S. R. et al. PHACTR1 is a genetic susceptibility locus for fibromuscular dysplasia supporting its complex genetic pattern of inheritance. PLoS Genet.12, e1006367 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR7] 7.Adlam, D. et al. Association of the PHACTR1/EDN1 genetic locus with spontaneous coronary artery dissection. J. Am. Coll. Cardiol.73, 58–66 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR8] 8.Debette, S. et al. Common variation in PHACTR1 is associated with susceptibility to cervical artery dissection. Nat. Genet.47, 78–83 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR9] 9.Fujimaki, T. et al. Association of a transcription factor 21 gene polymorphism with hypertension. Biomed. Rep.3, 118–122 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR10] 10.Fedoryshchak, R. O. et al. Molecular basis for substrate specificity of the Phactr1/PP1 phosphatase holoenzyme. eLife9, e61509 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR11] 11.Fils-Aimé, N. et al. MicroRNA-584 and the protein phosphatase and actin regulator 1 (PHACTR1), a new signaling route through which transforming growth factor-β Mediates the migration and actin dynamics of breast cancer cells. J. Biol. Chem.288, 11807–11823 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR12] 12.Kelloniemi, A. et al. The early-onset myocardial infarction associated PHACTR1 gene regulates skeletal and cardiac alpha-actin gene expression. PLoS ONE10, e0130502 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR13] 13.Ma, X. et al. PHACTR1, a coronary artery disease risk gene, mediates endothelial dysfunction. Front. Immunol.13, 958677 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR14] 14.Dai, X. & Shen, L. Advances and trends in omics technology development. Front. Med.9, 911861 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR15] 15.Johnson, K. W. et al. Enabling precision cardiology through multiscale biology and systems medicine. JACC Basic Transl. Sci.2, 311–327 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR16] 16.Mohammadi-Shemirani, P., Sood, T. & Pare, G. From ‘omics to multi-omics technologies: the discovery of novel causal mediators. Curr. Atheroscler. Rep.25, 55–65 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR17] 17.Santiago-Rodriguez, T. M. & Hollister, E. B. Multi ‘omic data integration: a review of concepts, considerations, and approaches. Semin. Perinatol.45, 151456 (2021). [DOI] [PubMed] [Google Scholar]

[CR18] 18.Tarazona, S., Arzalluz-Luque, A. & Conesa, A. Undisclosed, unmet and neglected challenges in multi-omics studies. Nat. Comput. Sci.1, 395–402 (2021). [DOI] [PubMed] [Google Scholar]

[CR19] 19.Wiezlak, M. et al. G-actin regulates the shuttling and PP1 binding of the RPEL protein Phactr1 to control actomyosin assembly. J. Cell Sci.125, 5860–5872 (2012). [DOI] [PubMed] [Google Scholar]

[CR20] 20.Contreras, O. et al. OpenEMMU: a versatile, open-source EdU multiplexing methodology for studying DNA replication and cell cycle dynamics. iScience28, 113380 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR21] 21.Gavet, O. & Pines, J. Activation of cyclin B1-Cdk1 synchronizes events in the nucleus and the cytoplasm at mitosis. J. Cell Biol.189, 247–259 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR22] 22.Rona, G. & Pagano, M. CDT1, a licensing factor that limits rereplication. Mol. Cell83, 1–3 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR23] 23.Liu, R. & Chan, D. C. The mitochondrial fission receptor Mff selectively recruits oligomerized Drp1. Mol. Biol. Cell26, 4466–4477 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR24] 24.Ko, H. J. et al. The phosphorylation status of Drp1-Ser637 by PKA in mitochondrial fission modulates mitophagy via PINK1/Parkin to exert multipolar spindles assembly during mitosis. Biomolecules1110.3390/biom11030424 (2021). [DOI] [PMC free article] [PubMed]

[CR25] 25.Lau, E. et al. Integrated omics dissection of proteome dynamics during cardiac remodeling. Nat. Commun.9, 120 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR26] 26.Garmany, R. et al. Multi-omic architecture of obstructive hypertrophic cardiomyopathy. Circ. Genom. Precis. Med.16, e003756 (2023). [DOI] [PubMed] [Google Scholar]

[CR27] 27.Wang, W. et al. Metabolic characterization of hypertrophic cardiomyopathy in human heart. Nat. Cardiovasc. Res.1, 445–461 (2022). [DOI] [PubMed] [Google Scholar]

[CR28] 28.Hill, M. C. et al. Integrated multi-omic characterization of congenital heart disease. Nature608, 181–191 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR29] 29.Subramanian, I., Verma, S., Kumar, S., Jere, A. & Anamika, K. Multi-omics data integration, interpretation, and its application. Bioinform. Biol. Insights14, 1177932219899051 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR30] 30.Chen, C. et al. Applications of multi-omics analysis in human diseases. MedComm4, e315 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR31] 31.Mouilleron, S., Wiezlak, M., O’Reilly, N., Treisman, R. & McDonald, N. Q. Structures of the Phactr1 RPEL domain and RPEL motif complexes with G-actin reveal the molecular basis for actin binding cooperativity. Structure20, 1960–1970 (2012). [DOI] [PubMed] [Google Scholar]

[CR32] 32.Fils-Aime, N. et al. MicroRNA-584 and the protein phosphatase and actin regulator 1 (PHACTR1), a new signaling route through which transforming growth factor-beta Mediates the migration and actin dynamics of breast cancer cells. J. Biol. Chem.288, 11807–11823 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR33] 33.Ihara, D. et al. The BDNF-ERK/MAPK axis reduces phosphatase and actin regulator1, 2 and 3 (PHACTR1, 2 and 3) mRNA expressions in cortical neurons. Drug Discov. Ther.18, 255–259 (2024). [DOI] [PubMed] [Google Scholar]

[CR34] 34.Dohadwala, M. et al. Phosphorylation and inactivation of protein phosphatase 1 by cyclin-dependent kinases. Proc. Natl. Acad. Sci. USA91, 6408–6412 (1994). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR35] 35.Grallert, A. et al. A PP1-PP2A phosphatase relay controls mitotic progression. Nature517, 94–98 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR36] 36.Chadee, D. N. et al. Increased Ser-10 phosphorylation of histone H3 in mitogen-stimulated and oncogene-transformed mouse fibroblasts. J. Biol. Chem.274, 24914–24920 (1999). [DOI] [PubMed] [Google Scholar]

[CR37] 37.Liu, M. et al. Sox17 is required for endothelial regeneration following inflammation-induced vascular injury. Nat. Commun.10, 2126 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR38] 38.Wolhuter, K., Kong, S. M. Y., Stanley, C. P. & Kovacic, J. C. The role of oxidants in percutaneous coronary intervention-induced endothelial dysfunction: can we harness redox signaling to improve clinical outcomes? Antioxid. Redox Signal38, 1022–1040 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR39] 39.Vinchi, F. et al. Atherosclerosis is aggravated by iron overload and ameliorated by dietary and pharmacological iron restriction. Eur. Heart J.41, 2681–2695 (2020). [DOI] [PubMed] [Google Scholar]

[CR40] 40.Wu, J. et al. NEK2 affects the ferroptosis sensitivity of gastric cancer cells by regulating the expression of HMOX1 through Keap1/Nrf2. Mol. Cell Biochem.10.1007/s11010-024-04960-y (2024). [DOI] [PMC free article] [PubMed]

[CR41] 41.Pi, J. et al. Molecular mechanism of human Nrf2 activation and degradation: role of sequential phosphorylation by protein kinase CK2. Free Radic. Biol. Med.42, 1797–1806 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR42] 42.Salvi, F. et al. Effects of stably incorporated iron on protein phosphatase-1 structure and activity. FEBS Lett.592, 4028–4038 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR43] 43.Kim, E. S. H., Saw, J., Kadian-Dodov, D., Wood, M. & Ganesh, S. K. FMD and SCAD: sex-biased arterial diseases with clinical and genetic pleiotropy. Circ. Res.128, 1958–1972 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR44] 44.Ryan, B. J., Charkoudian, N. & McClung, J. P. Consider iron status when making sex comparisons in human physiology. J. Appl. Physiol.132, 699–702 (2022). [DOI] [PubMed] [Google Scholar]

[CR45] 45.Yu, E., Mercer, J. & Bennett, M. Mitochondria in vascular disease. Cardiovasc. Res.95, 173–182 (2012). [DOI] [PubMed] [Google Scholar]

[CR46] 46.Vissing, C. R., Duno, M., Wibrand, F., Christensen, M. & Vissing, J. Hydroxylated long-chain acylcarnitines are biomarkers of mitochondrial myopathy. J. Clin. Endocrinol. Metab.104, 5968–5976 (2019). [DOI] [PubMed] [Google Scholar]

[CR47] 47.Pan, Y. et al. Decreased free fatty acid levels associated with adverse clinical outcomes in coronary artery disease patients with type 2 diabetes: findings from the PRACTICE study. Eur. J. Prev. Cardiol.30, 730–739 (2023). [DOI] [PubMed] [Google Scholar]

[CR48] 48.Chang, C. R. & Blackstone, C. Cyclic AMP-dependent protein kinase phosphorylation of Drp1 regulates its GTPase activity and mitochondrial morphology. J. Biol. Chem.282, 21583–21587 (2007). [DOI] [PubMed] [Google Scholar]

[CR49] 49.Shenouda, S. M. et al. Altered mitochondrial dynamics contributes to endothelial dysfunction in diabetes mellitus. Circulation124, 444–453 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR50] 50.Moiz, B., Li, A., Padmanabhan, S., Sriram, G. & Clyne, A. M. Isotope-assisted metabolic flux analysis: a powerful technique to gain new insights into the human metabolome in health and disease. Metabolites1210.3390/metabo12111066 (2022). [DOI] [PMC free article] [PubMed]

[CR51] 51.Moiz, B. et al. (13)C Metabolic flux analysis indicates endothelial cells attenuate metabolic perturbations by modulating TCA activity. Metabolites1110.3390/metabo11040226 (2021). [DOI] [PMC free article] [PubMed]

[CR52] 52.Alcazar, O. et al. Parallel multi-omics in high-risk subjects for the identification of integrated biomarker signatures of type 1 diabetes. Biomolecules1110.3390/biom11030383 (2021). [DOI] [PMC free article] [PubMed]

[CR53] 53.Yadikar, H. et al. Deciphering early and progressive molecular signatures in Alzheimer’s disease through integrated longitudinal proteomic and pathway analysis in a rodent model. Int. J. Mol. Sci.2510.3390/ijms25126469 (2024). [DOI] [PMC free article] [PubMed]

[CR54] 54.Karp, P. D., Midford, P. E., Caspi, R. & Khodursky, A. Pathway size matters: the influence of pathway granularity on over-representation (enrichment analysis) statistics. BMC Genom.22, 191 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR55] 55.Tsouka, S. & Masoodi, M. Metabolic pathway analysis: advantages and pitfalls for the functional interpretation of metabolomics and lipidomics data. Biomolecules1310.3390/biom13020244 (2023). [DOI] [PMC free article] [PubMed]

[CR56] 56.Adamson, P., Marshall, C. J., Hall, A. & Tilbrook, P. A. Post-translational modifications of p21rho proteins. J. Biol. Chem.267, 20033–20038 (1992). [PubMed] [Google Scholar]

[CR57] 57.Lorenzo, D. N. & Bennett, V. Cell-autonomous adiposity through increased cell surface GLUT4 due to ankyrin-B deficiency. Proc. Natl. Acad. Sci. USA114, 12743–12748 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR58] 58.Lauzier, A. et al. Snazarus and its human ortholog SNX25 modulate autophagic flux. J. Cell Sci.13510.1242/jcs.258733 (2022). [DOI] [PubMed]

[CR59] 59.Briot, A. et al. Endothelial NOTCH1 is suppressed by circulating lipids and antagonizes inflammation during atherosclerosis. J. Exp. Med.212, 2147–2163 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR60] 60.Andrews, S. FastQC: A Quality Control Tool for high-throughput Sequence Data. http://www.bioinformatics.babraham.ac.uk/projects/fastqc/ (2023).

[CR61] 61.Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics29, 15–21 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR62] 62.Ritchie, M. E. et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res.43, e47 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR63] 63.Shah, A. D., Goode, R. J. A., Huang, C., Powell, D. R. & Schittenhelm, R. B. LFQ-analyst: an easy-to-use interactive web platform to analyze and visualize label-free proteomics data preprocessed with MaxQuant. J. Proteome Res.19, 204–211 (2020). [DOI] [PubMed] [Google Scholar]

[CR64] 64.Duhrkop, K. et al. SIRIUS 4: a rapid tool for turning tandem mass spectra into metabolite structure information. Nat. Methods16, 299–302 (2019). [DOI] [PubMed] [Google Scholar]

[CR65] 65.Huynh, K. et al. High-throughput plasma lipidomics: detailed mapping of the associations with cardiometabolic risk factors. Cell Chem. Biol.26, 71–84 e74 (2019). [DOI] [PubMed] [Google Scholar]

[CR66] 66.Gene Ontology, C. et al. The gene ontology knowledgebase in 2023. Genetics22410.1093/genetics/iyad031 (2023).

[CR67] 67.Kanehisa, M. & Goto, S. KEGG: kyoto encyclopedia of genes and genomes. Nucleic Acids Res.28, 27–30 (2000). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR68] 68.Agrawal, A. et al. WikiPathways 2024: next generation pathway database. Nucleic Acids Res.52, D679–D689 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR69] 69.Milacic, M. et al. The reactome pathway knowledgebase 2024. Nucleic Acids Res.52, D672–D678 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR70] 70.Liu, C. H. et al. LipidSig 2.0: integrating lipid characteristic insights into advanced lipidomics data analysis. Nucleic Acids Res.52, W390–W397 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR71] 71.Rohart, F., Gautier, B., Singh, A. & Le Cao, K. A. mixOmics: An R package for ‘omics feature selection and multiple data integration. PLoS Comput. Biol.13, e1005752 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR72] 72.Singh, A. et al. DIABLO: an integrative approach for identifying key molecular drivers from multi-omics assays. Bioinformatics35, 3055–3062 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR73] 73.Shannon, P. et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res.13, 2498–2504 (2003). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR74] 74.Contreras, O. & Harvey, R. P. Single-cell transcriptome dynamics of the autotaxin-lysophosphatidic acid axis during muscle regeneration reveal proliferative effects in mesenchymal fibro-adipogenic progenitors. Front. Cell Dev. Biol.11, 1017660 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR75] 75.Wolhuter, K. & Kovacic, J. C. RNA-seq of PHACTR1 Knockdown and Overexpression in HT1080 Stable Cell Lines [Data set] ArrayExpresswww.ebi.ac.uk/biostudies/arrayexpress/studies/E-MTAB-15583 (2025).

[CR76] 76.Wolhuter, K. & Kovacic, J. C. Proteomic Analysis of PHACTR1 Overexpressing and Knockdown HT1080 Stable Cell Lines [Data set] MassIVE10.25345/C5XD0RB0M (2025).

[CR77] 77.Wolhuter, K. & Kovacic, J. C. Metabolomic Analysis of PHACTR1 Overexpressing and Knockdown HT1080 Stable Cell Lines [Data set] Metabolomics Workbench10.21228/M8MV84 (2025).

PERMALINK

Multi-omic analysis of human PHACTR1 signaling networks

Kathryn Wolhuter

Lijiang Ma

Nicole S Bryce

Osvaldo Contreras

Natalie Mellett

Ling Zhong

Chris Thekkedam

Siiri E Iismaa

Corey Giles

Richard P Harvey

Thomas Hennessy

Chris Fouracre

David Bradley

Peter J Meikle

Johan L M Björkegren

Jason C Kovacic

Abstract

Introduction

Fig. 1. Multi-omics pipeline to uncover the PHACTR1 signaling network.

Results

PHACTR1 levels control the expression of multiple downstream transcripts

Fig. 2. PHACTR1 expression impacts the cellular transcriptome and proteome.

Exploring options for feature mapping to biological pathways

The impact of PHACTR1 expression reaches across all ‘omics

Fig. 3. Changes in metabolite and lipid abundance and associated signaling pathways regulated by PHACTR1 expression.

Multi-omic correlations reveal key nodes of interest for PHACTR1 signaling

Fig. 4. Multi-omic correlations identify PHACTR1-mediated links across datasets.

Integrative multi-omics identifies PHACTR1-mediated pathways

Fig. 5. Multi-omic Ingenuity Pathway Analysis (IPA) integration reveals the biological functions of PHACTR1.

PHACTR1 expression affects cell cycle progression and mitosis

Fig. 6. Multi-omic integration validation: cell cycle phase processes and iron handling proteins under direct regulation by PHACTR1.

Fig. 7. PHACTR1-dependent changes to cell cycle and iron handling in primary endothelial cells.

Expression of major iron regulatory proteins is regulated by PHACTR1

PHACTR1 expression impacts mitochondrial biogenesis

Fig. 8. Expression of PHACTR1 impacts mitochondrial signaling and bioenergetics.

Cellular bioenergetics are influenced by PHACTR1 expression

Discussion

Methods

Cell culture

rs9349379 SNP genotyping

Lentiviral transduction

qPCR

Confocal microscopy

2-dimensional cell and nuclear area

Cell migration

Cell proliferation

Scratch wound assay

Adhesion assay

Quantification of lamellipodia

Cell culture for ‘omics dataset generation

Bulk RNAseq

RNAseq data processing

Proteomic sample preparation

Untargeted proteomics

Proteomic data processing

Metabolomic sample preparation

Untargeted metabolomics

Metabolomic data processing

Lipidomic sample preparation

Targeted lipidomics

Lipidomic data processing

Function and pathway analysis

Multi-omic correlation analysis

Gene-protein correlations

Integrative pathway analysis

Flow cytometry and cell cycle discrimination

Cell size measurements and mitochondrial mass quantification protocol

Gel electrophoresis and transfer

Immunoblotting

Transient transfection of HUVEC

Mitochondrial isolation

Linear regression of GTEx gene expression

Cell respirometry

Statistics and reproducibility

Reporting summary

Supplementary information

Acknowledgements

Author contributions