Robotic handling preserves iPSC-derived vascular smooth muscle cell differentiation Authors: Monique Bax, Jeya Ramalingam, Valentin Romanov, Muhammad A Alsherbiny, Keerat Junday, Jordan Thorpe, Emily Hurley, Melanie Lim, Ingrid Tarr, Stephanie Hesselson, Eleni Giannoulatou, Renjing Liu, Adam Hill, Siiri Iismaa, Robert M Graham Published: 2026 ABSTRACT AbstractTo tackle the global cardiovascular disease epidemic, vascular cell culture requires scalable, efficient, and reproducible methods for differentiating cells. However, no studies have validated whether automation, essential for scale-up, preserves differentiation quality of cells. Here, we developed a differentiation protocol for induced pluripotent stem cell (iPSC)-derived vascular smooth muscle cells (iVSMCs) that is applicable to both manual, and automated robotic culture. iPSC lines generated from male and female donors across various ages were used to optimize the generation of iVSMCs. iVSMCs had consistent morphology and protein expression, which were largely similar to those of primary VSMCs, including expression of myosin heavy chain-11 (MYH11), α-smooth muscle actin (α-SMA), transgelin (TAGLN) and calponin (CNN1). Functionally, these iVSMCs responded to the vasoconstrictor carbachol. Coupling this protocol with an automated Hamilton liquid-handling robotics system allowed the generation of iVSMCs in large quantities, that were morphologically similar to manually differentiated iVSMCs. Comparative proteomic analysis confirmed that protein expression was substantially the same between automated and manual differentiation methods. Automation markedly reduced manual labor and facilitated increased production without sacrificing cell quality. This study demonstrates the feasibility of automating iVSMC differentiation, marking a significant step towards scalable VSMC manufacture for three-dimensional applications, and for the ever-growing demands of organoid and tissue engineering applications.Keywords: Cardiovascular disease, Vascular smooth muscle cells, Tissue engineering, Proteomics, Liquid, Handling robotics, High-throughput, Drug screening Introduction Cardiovascular diseases are projected to cost $1.1 trillion by 2035 in the US alone, with blood vessel-associated conditions including high blood pressure and coronary heart disease accounting for, or contributing to, the overwhelming majority of cases [1]. Vascular smooth muscle cells (VSMCs) are a central component of blood vessels, responsible for wall contractility, regulating blood flow and blood pressure [2], and are accordingly a central focus of efforts to improve vascular health. As the prevalence of these vascular-linked diseases continues to increase globally, so too does the demand for advanced VSMC research, such as drug screens and bioengineered products for regenerative medicine applications [3], which require large volumes of VSMCs. VSMCs present unique challenges for in vitro modeling. Unlike cardiac muscle cells (cardiomyocytes) or skeletal muscle cells, VSMCs are not terminally differentiated [4]. This plasticity allows them to switch between contractile and synthetic states to facilitate normal vascular growth and repair, but also leaves these cells vulnerable to the pathogenesis of a myriad of vascular disorders, including atherosclerosis and arterial dissections [[5], [6], [7]]. Synthetic VSMCs are characterized by active cell proliferation, migration, and extracellular matrix (ECM) production [8,9]. Fluid dynamics is a key driver for phenotype switching in VSMCs, which sense altered pressure, and turbulence [10,11]. To best model this dynamic behavior both in the context of development, as well as in health and disease, induced pluripotent stem cell (iPSC)-derived VSMCs (iVSMCs) are increasingly employed, capturing the diverse genetic backgrounds of individuals in a human-derived system [3,12]. Current iPSC differentiation methods for VSMC generation have provided invaluable insights into vascular biology and disease mechanisms, examining both the effects of the embryological backgrounds of the cells and phenotypic variations [13,14]. However, challenges remain for the field, including a lack of information regarding the effects of scalability and automation, a high cost of implementation, particularly labor costs, as well as reagent costs, and limited reporting on the robustness of protocols across multiple lines [12]. Scalability requires a highly efficient differentiation protocol to generate a consistent cell type, with low batch-to-batch variability, and low phenotypic variability. Automating liquid handling represents an attractive option to address standardization in tissue culturing through minimization of human variation, and errors in handling; to reduce institutional liability by limiting exposure of staff to compounds that are potentially toxic; to reduce labor costs, and to maximize staff productivity [15]. However, to our knowledge, the impact of automated liquid handling on differentiation has not yet been examined for any cell type. For VSMCs, which alter their cellular identity in response to fluid dynamics [16,17], determining whether robotic liquid-handling impacts differentiation and maintenance of iVSMCs is as an important step in assessing its appropriateness for the automated production of cells in large scale. Recognizing these critical gaps, we report a novel differentiation protocol that addresses the urgent need for robust, functionally relevant iVSMCs. This protocol is designed to be easy to use, with low-variability and low-cost, as well as being weekend-free, which significantly enhances the reproducibility and accessibility of iVSMCs generation. Using multiple iPSC lines generated from male and female donors of varying ages, we show that iVSMCs can be generated both by manual handling (traditional tissue culture), as well as by automated means. The resultant cells expressed hallmark VSMC proteins, including myosin heavy chain-11 (MYH11), α-smooth muscle actin (α-SMA also known as ACTA2), transgelin (TAGLN also known as SM22) and calponin (CNN1). The iVSMCs produced also contracted in response to the cholinergic agonist, carbachol and, importantly, proteomic analysis of cells produced either manually or by the automated protocol showed no major phenotypic differences in terms of the protein pathways expressed. As proof-of-principle for the use of the cells in future applications, we demonstrate that they can form three-dimensional vascular rings. The current protocol developed to automate iVSMC differentiation, paves the way for their scalable generation, serving to improve disease modeling, drug discovery, and tissue engineering to combat vascular disease. Methods iPSC reprogramming from fibroblasts and peripheral blood mononuclear cells All assays were conducted in compliance with the St. Vincent's Hospital Human Research Ethics Committee (HREC/16/SVH/338, protocol number SVH 16/245). Cells were cultured at 37°C, under normoxic conditions (5% CO2). All reagents were purchased from ThermoFisher Scientific unless specified. Peripheral blood mononuclear cells (PBMCs) were reprogrammed as described previously [18,19]. Briefly, PBMCs were obtained from local pathology laboratories, following separation by Ficoll centrifugation, the cells were cultured in StemSpan erythroid progenitor medium (Stem Cell Technologies; SCT) according to the manufacturer's guidelines. PBMCs at a density of 1 × 105 were transduced using Cytotunes 2.0 Sendai Virus. After an incubation period of 24 h, the cells were centrifuged, a full media change was performed, and cells were seeded in a single well of a 12-well tissue culture plate. After four days of culture, the cells were transferred to Matrigel (Corning) coated plates. The culture medium was supplemented with ReproTeSR (SCT), maintaining a 50:50 media ratio, and daily medium changes were conducted. Fibroblast samples were obtained from a skin punch biopsy, and reprogrammed as previously described [20]. The iPSC colonies that emerged were manually selected based on their characteristic iPSC morphology starting from day 15. Clonal characterization Clonal characterization was carried out as previously described [18] unless otherwise stated. Briefly, iPSCs were first screened for morphology indicative of iPSCs, and Sendai Virus presence via PCR. Karyotyping was carried out by Monash Health Services, Victoria (Lines 4 and 5), or by Sullivan Nicolaides Pathology, Queensland. For all karyograms, a minimum of 15 cells were viewed at a resolution of 400 bands per haploid set by a cytogeneticist. All cell lines were karyotypically normal prior to initiating experiments, including Line 4 (via g-banding performed in 2019). A separate later culture of Line 4 showed trisomy 12 (via g-banding in 2025) in 1 of 15 metaphases following ten passages of the original stock. The karyotypic phenotype of this line could therefore not be established retrospectively, however their differentiation capacity and proteomic profiles were consistent with the other four lines. Expression of pluripotency markers was determined immunocytochemically and by mRNA expression, using the Pluripotency Scorecard assay. Capacity for trilineage differentiation of the selected representative clones was similarly determined by protein and mRNA expression. iPSCs were maintained using ReLeSR passaging, seeding on Matrigel-coated plates in mTeSR Plus media. Differentiation to synthetic and contractile phenotype VSMCs iPSCs were single cell dissociated with Accutase (SCT) and plated at a density of 21 - 56 x 103 cells/cm2 onto Matrigel with Y-27632 and supplemented with mTeSR Plus media overnight. A full media change using a chemically defined media (WALL media; Dulbecco's modified eagle medium/nutrient mixture F-12, GlutaMAX, MEM non-essential amino acids, penicillin-streptomycin, insulin-transferrin-selenium) supplemented with bone morphogenetic protein 4 (BMP4) (50 ng/mL; SCT) and fibroblast growth factor 2 (FGF2) (20 ng/mL; SCT), was performed on day 1, and a partial media change was performed on day 3 for mesodermal progenitor generation. A partial media change was also performed on days 6, 8, and 10, using WALL media supplemented with vascular endothelial growth factor (VEGF) (50 ng/mL) and Repsox (25 μM). Robotic liquid handling differentiation Simultaneous differentiation via robotic liquid handling and manual handling was conducted using the Hamilton MicroLab STAR (Hamilton Company, Bonaduz, Switzerland) liquid-handling robot (Supplementary Fig. 1a). All consumables were from Hamilton, unless stated. Custom deck layout included tips, conical tubes, reservoirs, plate holder plinths, and lid holder plinths. The automated workflow was developed to perform media exchange using a predefined custom deck layout containing 300 μL and 1000 μL conductive pipette tips, 60 mL reagent reservoirs, culture plates, and 50 mL conical tubes (Supplementary Fig. 1b). The workstation was programmed to execute the required aspiration, dispensing, and plate-handling steps under aseptic conditions. Each plate was transferred to the barcode reader, and the measured barcode was verified against the requested plate identity before processing. The plate was then transferred to the tilt module and the lid was removed. The plate was tilted by 5° during aspiration of the spent culture medium. Identical to manual media changes, 800 μL of spent medium was aspirated from each well and replaced with 1 mL of fresh medium using two 1000 μL tips simultaneously. Tips were changed between cell lines. Plates were returned to level orientations and new media was replaced into two wells at a time. The plate lids were replaced and the plate was returned to its original incubator stacker and level, and the incubator inventory was updated. All scripts and configurations used to run the Hamilton system are publicly available via Zenodo (https://doi.org/10.5281/zenodo.17355820). Immunocytochemistry and imaging Adherent cells were processed as described [21], including a 4% paraformaldehyde fixation (10 min, room temperature; RT), permeabilization (12 min, 0.1% Triton X/PBS) and 10% goat serum block for 1h at room temperature, and then incubated overnight with antibody (Supplementary Table 2).Vascular rings were fixed for 20 min and then permeabilized for 20 min using 0.1% Triton X/PBS. Samples were washed with Tween 0.1% in PBS to remove residual primary antibody and incubated with a corresponding secondary antibody (1h RT), and DAPI stained. Fluorescence microscopy was performed as indicated using an EVOS cell imaging system (ThermoFisher), or a Nikon Eclipse Ti2-E inverted microscope, or Celigo (Revvity) imaging cytometry. Characterization of differentiation efficiency was quantified by unbiased imaging using an automated confocal microscopy (Opera Phenix, Revvity). Confluence and live cell tracking was performed using an in-incubator imager (Incucyte). Images were processed (overlay, scale bar addition, crop) with Image J. Mass spectrometry sample preparation Cells were harvested in RIPA buffer supplemented with protease inhibitor (cOmplete protease inhibitor; Roche) following aspiration of media and subsequent washes with PBS. Lysate was stored at −80°C. Cells were then lysed by thawing and repeated trituration through a 26.5 G needle, prior to centrifugation (20,000 x g for 30 min at 4°C) and harvesting of the supernatant fraction. Protein levels were determined using a bicinchoninic acid (BCA) assay in accordance with the manufacturer's guidelines (Pierce Rapid Gold BCA Protein Assay Kit #A53226, ThermoScientific, USA). Protein samples (100 μg in 100 μL of conditioned media) were prepared using the EasyPep Mini MS Sample Prep Kit (#A40006), following the instructions provided by the manufacturer. This process included a pre-digestion treatment with PNGaseF, a deglycosylating enzyme, before overnight trypsin digestion (ThermoFisher Scientific, USA). The resulting clean peptide samples were then dried by vacuum centrifuge and subsequently reconstituted in 100 μL of 0.1% formic acid in water, in preparation for LC-MS analysis. Label-free bottom-up quantification via micro-high-performance liquid chromatography coupled with quadrupole time-of-flight mass spectrometry (micro-HPLC-qTOF-MS) LCMS data was acquired according to a previous study with some modifications in the liquid chromatography [22]. Briefly, a micro liquid chromatography system, Waters M-Class, coupled to SCIEX 6600 Triple-TOF mass spectrometer operating in positive electron spray ion mode (ESI +), was used to analyse the tryptic peptides within 33 min of acquisition. 4 μg of the peptide digest was injected onto nanoEase M/Z HSS T3 1.8 μm 300μmx150mm (Waters, 186009249) column coupled to a Zorbax 300SB-C18, (5um, 5 × 0.3 mm) guard column (Agilent, USA). The column temperature was maintained at 40°C, and 98% water (2% acetonitrile) and acetonitrile containing 0.1% Formic acid were used as mobile phase A and B, respectively, at 5 μL/min flow rate with initial loading and later column cleaning and equilibration at 7 μL/min. Briefly, the gradient started at a 7 μL/min flow rate at 2-10% B for 1.66 min and inclined to 10-25% B at 1.67-21.67 min at 5 μL/min flow rate, then 40-95% B at 23.33-24.67 min and kept for 2 min, then equilibrated for 9 min with 2% B at 7 μL/min flow rate. The TripleTOF 6600 system was equipped with a DuoSpray source and 25 μm internal diameter electrode and controlled by Analyst 1.8.1 software and Analyst Device Driver for LC control. Ion source and MS parameters were set as follows: GS1 25; GS2 15; curtain gas 20; ion source temperature 150; ionSpray voltage floating of 5500; lower m/z limit 350; upper m/z limit 1250; window overlap (Da) 1.0; collision energy spread was set at 5. The Sequential windowed acquisition of all theoretical fragment ion spectra (SWATH™) acquisition method involved an MS1 scan ranging from m/z 350 to m/z 1250 with an accumulation time of 50 ms, followed by 40 MS2 scans, each with a 35 ms accumulation time The MS2 scans utilized a variable precursor isolation width covering the m/z range of 400 to 1250 (Supplementary Table 4). MS2 spectra were collected in the range of m/z 100 to 2000 in high-resolution mode, with ∼1.5 s total cycle time. PepCalMix calibrant (SCIEX, P/N 5045759, 10 fmol/μL solution that was prepared by 1:100 dilution of the stock solution in dilution buffer consisting of 5% acetic acid and 2% acetonitrile) was injected every 12 samples for mass calibration along with the pooled quality control sample to evaluate the whole batch.Six pooled QC samples were injected to generate DIA-only library using gas phase fractionation workflow [23] with the following m/z mass ranges: 400–500, 500–600, 600–700, 700–800, 800–900, 900–1000, and 1000–1250. The precursor isolation window was set to m/z 5, with collision energy spread of 5 eV except for 700-990 and 990-1000 was 8 and 10 eV, respectively (Supplementary Table 5). The cycle time was 2.14 s, consisting of high- and low-energy scans with 40 ms MS2 accumulation time, and data were acquired in ‘high-resolution’ mode. Proteomic data processing and availability Data were processed using Spectronaut v19.5 where DirectDIA + analysis using BGS Factory workflow was implemented using canonical human FASTA database imported from Uniprot (25th September 2024; 20,420 entries). Briefly, Pulsar search was done using the following parameters trypsin/P and LysC/P cleavage rules, specific digest type with 7-52 peptide length and max 2 missed cleavages. Carbamidomethyl (C) was set as a fixed modification and Acetyl (protein N-term), and oxidation (M) were set as variable modification with max variable modification set to 5 and identification FDR of 1% for PSM, peptide and protein group were applied for directDIA + workflow with Automatic LFQ protein quantification method using the areas at MS2 level and automatic normalization strategy. Automatic protein inference workflow using IDPicker algorithm was used for protein inference. The raw and processed data have been deposited to the ProteomeXchange Consortium via the PRoteomics IDEntifications (PRIDE) repository [24] with the dataset identifier PXD070184 and 10.6019/PXD070184. Proteomic data analysis and network analysis Relative protein intensity was statistically analyzed via Metaboanalyst [25]. Data was filtered for a minimum requirement of 60% expression per feature, log transformed, and Pareto scaled. Pearson's correlation of samples, hierarchical clustering heatmaps of normalized data (Standardization: Auto-scale features; Distance measure: Euclidean, Clustering method: Ward) and volcano plots were all generated with the platform. Fold change and multiple comparison adjusted P-value were as per in-text description. To analyse pathway-level interactions, Cytoscape software (version 3.8X) was implemented, including network clustering using Markov Cluster (MCL). Calcium imaging To measure calcium transients, iVSMCs were treated with Cal 520 a.m. dye (21130, AAT Bioquest) at a concentration of 5 μM for 30 min at 37°C. Dye was aspirated and cells were returned to iVSMC media (phenol red-free DMEM/F12) and allowed to equilibrate for 15 min. Cells were imaged in a live cell chamber (37°C and 5% CO2) using a Nikon Eclipse Ti2-E inverted microscope, utilizing a Nikon Plan Fluor 10× objective (NA, 0.3), and images were captured using an Andor Zyla sCMOS (Oxford Instruments). Cells were imaged for 10 min at baseline, followed by a vehicle control media change and carbachol treatment (2 mM ab141354, Abcam). Data acquisition was performed with a temporal resolution of 20 ms. Data analysis was conducted using a custom MATLAB script, as previously described [26] to establish traces. Sum of signaling area was analyzed using a custom Python pipeline generated using Google Colab, with assistance of ChatGPT 4o. Pipeline development incorporated scikit-image, OpenCV, NumPy, SciPy, pandas, and matplotlib. Grayscale video frames were extracted after skipping the initial 10% of the sequence to avoid transient surges. A baseline image was generated by averaging multiple sampled frames. Regions of interest (ROIs) were identified using Otsu thresholding, Gaussian filtering, and distance-transformed watershed segmentation (scikit-image). Noise was excluded by applying a minimum area threshold. ROI boundaries were visualized using the ‘find contours’ function and overlaid on grayscale images with randomly assigned colors. Mean intensity values were extracted for each ROI across the time course (OpenCV, NumPy), normalized to baseline (F/F0), and exported alongside the summed intensity signal for all ROIs per frame. Script is available via Zenodo (https://doi.org/10.5281/zenodo.21493764). Cell size quantification iVSMC contractions were imaged using a Nikon Eclipse Ti2-E Inverted Microscope (10X), capturing images every 10s for 10 min via Andor Zyla sCMOS. Contractility was measured using Image J using a standard measure analysis at T0 and T10, randomly selecting cells. To measure cell size of HAoSMC, a standard measuring analysis using Image J analysis was also performed. For these media comparisons, images of iVSMCs were taken using an EVOS brightfield microscope (ThermoFisher Scientific, USA) and blinded images were adjusted using Brightness and Contrast, and Threshold settings to outline cells. n = 20 cells were catalogued as Regions of Interest and measured per image. Cell yield quantification Cells were imaged at 4X and 10X magnification using an EVOS brightfield microscope (ThermoFisher Scientific, USA) at day 14 to assess cell density. Density of cells per square area was calculated using a custom python script run through Terminal generated with 5.6 Sol to analyse the 10X magnification, and verified against the 4X, using a conservative approach. Images were read as 8-bit grayscale arrays and scaled to the interval 0–1 for response calculation. Dark local cell cores were enhanced using the difference between Gaussian-smoothed images with physical standard deviations of 16.0 μm and 3.2 μm. To suppress detections in smooth cell-free areas, a local grayscale standard deviation map was calculated using a 40 μm Gaussian neighbourhood; pixels with local standard deviation greater than 6 grayscale units were classified as attached-cell regions. Candidate centres were defined as local maxima of the response above its 80th percentile. Non-maximum suppression used a magnification-adjusted window corresponding to 22 μm, with the rounded window size forced to the next odd integer and a minimum of 3 pixels. A border equal to the larger of 12 pixels or the suppression-window size was excluded. A candidate contributed one object to the count only when it was above the response threshold, was a local maximum and lay within the attached-cell mask. Cell masks were visually assessed to ensure accuracy. Density per square centimetre was calculated using a 0.3087 pixels per μm for 4X images, and 0.782 pixels per μm ratio. An average of each differentiation (n = 3) per cell line (n = 5) was compared between manual culture and robotic culture. Statistical comparisons were performed following normality testing, using a one-way ANOVA with Tukey's multiple comparisons test, and visualized via bar graph with dot points representing each sample in Prism (v11); P > 0.05 was considered non-significant. Descriptive statistical analyses (value minimum, maximum, range, mean, standard deviation, standard error of mean), and two-way ANOVA with Tukey's post hoc analyses were also performed in Prism (v11). Total cellular yield was interpolated by calculating the cells per square centimetre, and multiplying to account for total well size (9.6 cm2 total area), as density was consistent throughout the well. Script is available via Zenodo (https://doi.org/10.5281/zenodo.21493764) Protein yield quantification Protein yields, determined via BCA assay, were analyzed to assess consistency of protein yield between differentiations. An average of each differentiation (n = 5) per cell line (n = 5) was compared between manual culture and robotic culture. Statistical comparisons were performed following normality testing, using a one-way ANOVA with Tukey's multiple comparisons test, and visualized via bar graph with dot points representing each sample in Prism (v11); P > 0.05 was considered non-significant. Descriptive statistical analyses (value minimum, maximum, range, mean, standard deviation, standard error of mean), and two-way ANOVA with Tukey's post hoc analyses were also performed in Prism (v11). Fabrication of iVSMC vascular rings Master molds (negatives) for fabricating PDMS pillars to support ring formation were 3D printed using an Asiga Max X35 (365 nm) stereolithography printer using Asiga DentaMODEL resin. Print parameters were set to 25 μm layer thickness, 0.27 s exposure time, 32°C build temperature, and 30% light intensity. After printing, molds were first washed gently in 99% isopropyl alcohol for at least 1 h on an orbital shaker to remove uncured resin, then UV cured for 3 × 15 min, and finally baked overnight at 65°C. PDMS was prepared by mixing Sylgard 184 (10:1 base:curing) and degassing in a vacuum chamber for 1 h. Once ready, each well inside a 6 well plate (Corning) was cleaned with 100% ethanol and 2 mL of PDMS was pipetted into each well to form a thin layer of PDMS. This layer was then cured for 45 min at 65°C. This layer helps to ensure that molds detach without ripping off the surface of the well plate. Next, 3D printed molds were placed in an oxygen plasma chamber and exposed to plasma for 30 s, followed by pipetting of PDMS into each mold and degassing for 2 min to remove trapped air bubbles. The molds were then inverted and placed on top of the cured PDMS layers within each well of the 6-well plate. The assembled system was cured overnight at 65°C. After cooling, the 3D printed molds were carefully removed using pliers leaving behind the PDMS positives. Tissues were generated by seeding ∼1 x 107 iVSMCs (1 well of a 6 well plate) on day 10 of the differentiation in 100 μL of Matrigel:laminin (1:1 ratio) into the ring wells and left to gel for 30 min prior to gently filling the well with media. Media changes were carried out as normal. Statistical analysis Numerical data are presented as mean ± S.E.M., unless otherwise stated. Sample sizes and number of independent experiments are as stated and specified in figure legends. Graphical representations and statistical analyses were conducted using R and GraphPad Prism version 10. To assess the significance of differences between a control and experimental group, normality was tested to determine if parametric or non-parametric testing was appropriate. Two-tailed Student's t-test were employed when comparing two groups, unless stated. For comparisons among multiple groups, one-way ANOVA was utilized with post hoc testing as described for each. Results Generation of novel healthy female and male donor-derived iPSC lines to validate differentiation protocol robustness To evaluate the robustness of our differentiation protocol, we generated a cohort of induced pluripotent stem cell (iPSC) lines from both female (n = 3) and male (n = 2) donors across a broad age range. One iPSC line (Line 4; SCVI15) was kindly provided by the Stanford Cardiovascular Institute Biobank [27], while all other lines were successfully reprogrammed from fibroblasts or peripheral blood mononuclear cells (PBMCs) using Sendai virus as previously described [18,19,28]. All lines were subjected to a characterization pipeline (Fig. 1a). The cell donors of this cohort included females and males ranging from 16 to 71 years of age at time of collection (Fig. 1b). The pluripotent status of the reprogrammed lines was confirmed via immunostaining by the pluripotency markers OCT4 and SSEA4 (Fig. 1c). The capability for trilineage differentiation was also demonstrated through immunocytochemistry; the potential of these iPSC lines to differentiate into cells expressing hallmark proteins, including PAX6, TAGLN and SOX17, which are indicative of successful differentiation to ectoderm, mesoderm, and endoderm, respectively (Fig. 1d-f). Gene expression profiles were analyzed using the hPSC Scorecard to further validate that these iPSC lines were capable of self-renewal and trilineage differentiation (Fig. 1g). Development and characterization of iPSC lines from healthy donors of diverse age and sex a) Schematic of male and female donors of various ages and with differing health statuses, and b) details of iPSC lines that were generated from fibroblasts or PBMCs from male and female donors of various ages and with differing health statuses using Sendai virus reprogramming. iPSC: induced pluripotent stem cell, HAoSMC: Human Aortic Smooth Muscle Cells, F: Female, M: Male, NR: Not Reported, MI: Myocardial Infarct, WHO Rec: World Health Organization recommended levels of exercise (c) Representative immunofluorescence images show pluripotency marker expression (OCT4, red; SSEA4, green; nuclear stain, blue). Immunocytochemistry following trilineage differentiation demonstrates capacity for three germ layer differentiation: d) PAX6 (green; DAPI nuclear stain, blue) expression indicates ectoderm differentiation, e) GATA4 (green; nuclear stain, blue) expression indicates mesoderm differentiation, and f) SOX17 expression indicates endoderm differentiation. g) Representative hPSC Scorecard analysis depicting the self-renewal capabilities and differentiation potential that were further validated by RNA microarray analysis of the lineage-specific differentiation potential of the iPSC lines. Robust iVSMC differentiation Having validated our iPSC lines, we sought to develop a weekend-free, automatable protocol that was relatively low cost, and low variability (WALL) that could generate differentiated cells at scale. This design philosophy addresses the practical barriers that limit VSMC production in both academic and industrial settings. A range of variables including media composition, dissociation methods, cell seeding density, and maturation time were trialed that ultimately resulted in a robust, monolayer-based protocol that can be generated under xenogenic-free conditions (Fig. 2). Representative vascular smooth muscle cell differentiation of iPSC lines a) Schematic of the iVSMCs differentiation protocol describing full media change (FMC) and partial media changes (PMC), with suggested timelines to follow for weekend-free culturing, and with representative images of cells at day 0, 3, 8, 14; b) Phenol red-free DMEM/F12 media reduced the cellular yield of mesodermal progenitors; c) Expression of the early marker GATA4 in progenitor cells at day 6 decreased as cells began to mature into VSMC-like cells on day 10. d) Cells seeded over a range of densities yielded more linear, elongated cells at higher density e) Cell seeding densities were adjusted for plate size to account for the differences in tissue culture plates. f) Quantification of key VSMC contractile-state proteins (myosin heavy chain 11, MYH11; transgelin, TAGLN; calponin 1, CNN1; α−smooth muscle actin, αSMA) expressed in iVSMCs (lines 1-5) compared to primary adult human aortic smooth muscle cells (P), as determined from proteomic analysis (raw intensity; n = 3-5 differentiations per line). WALL media To ensure clinical relevance, a chemically defined, xenogenic-free WALL medium was developed comprising DMEM/F12, insulin-transferrin-selenium, non-essential amino acids, and GlutaMAX (Supplementary Table 1). To assess the effect of WALL media on protein expression of iVSMC compared to common commercially available media, proteomic analysis was performed on primary VSMC (human aorta-derived smooth muscle cells; HAoSMC) cultured in either WALL media or smooth muscle growth media 2 (SmGM-2, Lonza), for 8 days (Supplementary Fig. 3). Cells cultured in either medium exhibited similar morphology (Supplementary Fig. 3a). WALL media-cultured cells proliferated slower (Supplementary Fig. 3b), and this more closely resembled the turnover rates of quiescent contractile VSMCs in vivo, that under normal physiological conditions turnover less than once per year [29]. Qualitative analysis of the proteomes of HAoSMCs cultured in WALL media compared to LONZA media indicated near identical protein expression, where 97.6% of the 3443 proteins detected were in common between the two conditions (Supplementary Fig. 3c). While hierarchical analysis indicated sample clustering by media type (Supplementary Fig. 3d), the overall proteomes of the WALL media-cultured cells were similar to those in commercial media, as indicated by similar Pearson's correlation values between Lonza samples (0.870 – 0.926) and WALL samples (0.823 – 0.904), indicating consistent profiles between passages (Supplementary Fig. 3e). Student's t-test identified differentially expressed proteins between the two groups (Supplementary Fig. 3f), with differences primarily due to expression of proteins driven by the growth factors, IGF and PDGF, which are not utilized in the WALL medium differentiation (Supplementary Fig. 3g). Initiation format To harness the ability of iVSMCs and their progenitors to synthesize biologically relevant ECMs in situ, and to ensure the protocol would lend itself to automation, we chose to use a monolayer iVSMC progenitor approach, rather than an embryoid body formation approach. Supplementation of the chemically defined media with BMP4 and FGF2 induced morphological changes indicative of mesodermal progenitor differentiation. Cells grew in colonies, prior to reaching confluence, and were GATA4 positive on day 6, with reduced expression by day 10 as they matured (Fig. 2c). Cell seeding density was the largest contributor to cellular variability. Testing of different cell densities (5 x 103 cells/cm2, 12.5 x 103 cells/cm2 or 25 x 103 cells/cm2 in 24 well plates) indicated that cells grown at higher seeding densities had a more elongated and aligned morphology when stained for VSMC markers, including TAGLN, suggesting increased maturation in these cells (Fig. 2d). Plate well-size also impacted differentiation, with smaller wells, such as those in 96-well plates, requiring a higher seeding density. This is likely because fluidic forces to which cells are exposed, including well surface area to volume ratios, base curvatures, and media handling differences, vary between tissue culture-plates of different sizes. Different seeding densities were thus established for different plate sizes, ranging from 21 to 56 x 103 cells/cm2 (Fig. 2e). Mid-differentiation passaging was avoided wherever possible, as repeated enzymatic digestion of the endogenously generated ECM markedly altered cell morphology and encouraged the development of the synthetic phenotype. If replating was required, this was performed on day 8 when the cells maintained more cellular plasticity. Confluence also varied markedly with substrate stiffness (glass compared to plastic), which has been previously reported, and is consistent with the ability of VSMCs to sense and respond to stiffness with shifts in proliferation [10]. To generate a robust culture of contractile iVSMCs, we employed Repsox, an inhibitor of TGF-b1 receptors [30]. Though many iVSMC protocols commonly use PDGF-BB and TGF-β1, we, similar to others [31,32], observed that the use of these proteins increased the presence of a synthetic phenotype. Use of VEGF alone generated a more diverse culture with both contractile and synthetic cells, which may be of interest in disease modeling, while the combination of VEGF and Repsox produced a robust contractile culture. Proteomic analysis of these cells confirmed robust expression of hallmark VSMC proteins including MYH11, α-SMA, TAGLN and CNN1 (Fig. 2f). Protein expression consistent with robust differentiation of iVSMCs To examine batch-to-batch variation, all five iPSC lines were differentiated concurrently, repeating the differentiations three to five times to examine protein expression via immunocytochemistry and proteomics. Immunostaining indicated the differentiated iVSMCs were positive for hallmark VSMC markers, and quantification of positive cells (minimum 12,000 cells analyzed per line for each marker, n = 5 lines) indicated robust expression of MYH11 (97 ± 0.82 SD % positive cells) Fig. 3a), TAGLN (97.59 ± 2.2 SD % positive cells; Fig. 3b) and CNN1(93 ± 1.3 SD % positive cells; Fig. 3c), also validating the proteomic identification of these markers (in Fig. 2f). Representative Immunocytochemical Expression of VSMC Marker Proteins a) myosin heavy chain 11 (MYH11; red), b) transgelin (TAGLN; dark magenta) and c) calponin 1 (CNN1; dark magenta), with nuclear stain DAPI (blue). Protein quantification was performed using ImageJ software applied to a minimum of 12,000 cells per line, per marker in n = 3-5 differentiations per line. Quantification was analyzed by one-way ANOVA. No significant difference was detected between lines. WALL differentiated iVSMCs are contractile To assess iVSMC function, intracellular calcium dynamics and contractile responses to carbachol (2 mM) were evaluated (Fig. 4). Time-lapse microscopy using phase contrast imaging of the iVSMCs (Fig. 4a), revealed significant carbachol-induced contraction of cells over a 10 min period, as compared to baseline and to vehicle controls (Fig. 4b–Supplementary Video 1). Carbachol-treated cells exhibited a 16 - 27% decrease in cell area, consistent with the normal range of VSMC contraction under physiological conditions. Characteristic calcium transient alterations were observed within the first minute of carbachol treatment (Fig. 4c). An example of an individual iVSMC shifting its contractile response from single calcium peaks to rolling oscillations is shown in Fig. 4d. Individual traces demonstrated a fast initial calcium spike, followed by a plateau phase lasting several seconds, often taking >20 s to return to baseline. Quantification of the sum of Cal-520AM dye fluorescence area revealed a ∼7-fold increase with carbachol treatment (Fig. 4e). Together these findings indicate that iVSMCs generated with the WALL protocol were responsive to vasoactive signals, a key feature of smooth muscle physiology. Functional characterization of iVSMCs a) Representative phase contrast images of iVSMC pre- and post-carbachol treatment (2 mM, 10 min). b) Blinded single cell quantification of cell area prior to media change (no treatment), in response to vehicle control, and in response to carbachol treatment. Carbachol induced reductions in cell surface area by 16-27% (n = 5 lines, represented by colors, 10 cells/line; *p < 0.01 determined via one-way ANOVA with Tukey's post hoc test) indicating their contractile competency. c) Fluorescence imaging of Cal 520 a.m. dye-treated iVSMCs following vehicle (control) or carbachol treatment(2 mM, 1 min; top panels) and Region of Interest detection analysis via custom Python script of these areas (bottom panels). d) iVSMC calcium transients quantified via MATLAB. e) Sum of the area of fluorescence detected via Cal 520 a.m. imaging using a Nikon Eclipse Ti2-E Inverted Microscope. iVSMC proteome unaltered by robotic media changes We next investigated the effects of using a large-scale liquid handling robotics system, Hamilton MicroLab STAR liquid handling robot (Fig. 5a), for carrying out the differentiation procedure and performed comparisons with concurrently cultured, manually-differentiated cells. Gross morphologies were not altered by robotic liquid dispensing as observed by brightfield imaging (Fig. 5b). Overall cell expansion assessment by cellular yield quantification revealed no significant difference in cell density as assessed by one-way ANOVA of the 10 samples (Supplementary Fig. 6a; Supplementary Table 4). Two-way ANOVA demonstrated a slight decrease in cell number in manual culture of line 2 (n = 3), being modest, but significantly lower than lines 4 and 5, however no difference was detected in the robotically handled culture. The total cell count at experimental endpoint was similar between the two methods; the average yield between five lines was 984,171 ± 7% cells/well for manual handling and 971,118 ± 2% cells/well for robotic handling. Similarly, total protein concentration following cell harvest did not differ (P > 0.05) between manual and robotic cultures, as assessed by both one-way, and two-way ANOVA analyses of the replicates (n = 5 samples per line; n = 5 lines) (Supplementary Fig. 6b; Supplementary Table 4). Together, these findings indicate that robotic liquid handling did not alter overall cell growth or protein recovery. The iVSMC Proteome is Unaltered by Automated Robotic Liquid Compared to Manual Handling. a) iVSMCs differentiation was tested for feasibility of automation using a Hamilton MicroLab STAR Liquid handler coupled to an incubator system. b) Representative brightfield imaging using Incucyte shows similar morphology of cells cultured manually and robotically. c) Qualitative Venn diagram comparisons of the mean proteomes of five iVSMC lines (mean of n = 5 lines differentiated concurrently, with n = 5 consecutive differentiation replicates) either manually, or via robotic handling. d) With the exception of one outlier (line 3 sample), principal component analysis clustered the average of samples by cell line (indicated by number) rather than by culture technique (manual (M) in red, robotic (R) in blue). e) Hierarchical analysis of the iVSMC proteomes indicated clustering of samples based on the cell lines (indicated by number) rather than the culture technique – Robotic (R) or manual static culture (M). f) Volcano plot representation of t-test scores indicates that few proteins are differentially expressed. g) Expression of the VSMC marker proteins, MYH11, TAGLN, CNN1 and α-SMA was not different between cells cultured using robotic vs manual techniques. To determine if robotic differentiation induced global changes, proteomic analyses were performed. iVSMCs from the five iPSC lines were differentiated manually or robotically in parallel across five sequential differentiations to ensure reproducibility and robustness. The quantified proteomes consisted of 6,003 detected proteins. Following filtering for robust expression across proteins 5,974 proteins were robustly detected. Classic VSMC-associated proteins were detected, including α-SMA, which was the tenth most abundant protein (average of n = 5 lines, 2 culture conditions), as well as MYH11, TAGLN, and CNN1, which were also confirmed by immunocytochemistry (Fig. 3) Variability of marker concentration between differentiations was relatively low between cell lines (Supplementary Fig. 4). Protein detection was highly consistent between the robotically cultured and manually cultured cells, with only 1.00% of proteins undetected between these culturing techniques (Fig. 5c). Principal component analysis and hierarchical clustering indicated that sample similarity, with partitioning driven by cell line identity rather than culture technique (Fig. 5d). Hierarchical clustering similarly indicated that sample partitioning was determined by cell line, not culturing technique (Fig. 5e). Comparison of the robotic and manual culturing of iVSMC Student's t-test revealed no significant differences between robotic and manual iVSMC cultures using standard thresholds (fold change ≥2, FDR ≤0.1). Even with relaxed thresholds (fold change ≥1.5, p ≤ 0.05), only 18 proteins (0.29%) were differentially expressed between the two conditions (Fig. 5f; Supplementary Fig. 5). STRING pathway analysis revealed no pathway level enrichment between these 18 proteins (protein-protein interaction enrichment P-value = 0.89). No significant differences were detected in the hallmark VSMC proteins, MYH11, α-SMA, TAGLN and CNN1 (Fig. 5g). iVSMCs capable of 3D microtissue formation To investigate the application of these cells in a microtissue engineering setting, we utilized iVSMCs differentiated via automation to generate vascular rings. Using custom-designed polydimethylsiloxane (PDMS) ring-molds (Fig. 6a–d), we generated vascular ring tissues by combining Matrigel (∼10 mg/mL) and laminin (1 mg/mL) with iVSMCs (∼1 x 107 cells/mL) dissociated on day 8 of the protocol, at a ratio of 1:1:1. Interestingly, resuspension in Matrigel alone was not sufficient to maintain the integrity of the vascular ring structures. Vascular rings were cultured for 4 days, at which time they held enough structural integrity to be removed from the pillar using tweezers. Immunocytochemical analysis indicated cells within the ring were positive for TAGLN (Fig. 6e–f), suggesting cells could be appropriate for future 3D applications, though further characterization is required. Differentiated iVSMCs were capable of forming 3D microtissues. a) Bespoke vascular ring templates were designed via Fusion360 to generate vascular rings in a 6-well format to accommodate the same liquid handling workflow. b) The design with a tapered center to facilitate removal of ring for future applications. c) Polydimethylsiloxane molds were cast with polylactic acid. d) iVSMCs formed microtissues of sufficient integrity to be handled with tweezers for removal for downstream assays on day 4. e) Representative ring, approximately the size of a single well in a 96-well plate and f) overlay of immunocytochemical analysis of a ring showing positive expression of TAGLN (red), as well as DAPI (blue) staining, confirmed using a Celigo imaging cytometer connected to a Hamilton Robotics system. Discussion The design of the WALL protocol was guided by several key considerations aimed at maximizing scalability, reproducibility, and clinical relevance. These included the selection of a chemically defined, xenogenic-free medium, optimization of cell seeding and culture format, and the choice of small molecule modulators to promote a contractile VSMC phenotype. Throughout the iPSC-derived modeling field, there is an increased demand for chemically-defined media due to the variable nature of animal-derived products, such as fetal bovine serum (FBS). The use of xenogenic-free reagents overcomes the possible development of xenogenic immune responses, and reduces animal to human transmittable diseases (xenozoonosis) [33]. Decreasing reliance on animal-based products in laboratory settings has also been encouraged following the COVID-19 pandemic, to further reduce risk of future zoonotic epidemics [34]. In vascular models, this practice is particularly important as FBS can alter VSMC extracellular matrix deposition and TGF-β signaling [35]. Luo and colleagues have previously demonstrated a xenogenic-free iVSMC differentiation protocol has no significant impact on the capacity of iPSCs to differentiate, moving the field closer towards clinical application of engineered VSMC cells [36]. Accordingly, here we have similarly employed a xenofree ‘WALL’ media comprising of DMEM/F12, insulin-transferrin-selenium, non-essential amino acids and GlutaMAX in lieu of undefined components (Supplementary Table 1). Interestingly, we observed a marked reduction in cell viability and proliferation towards mesenchymal progenitors using phenol red-free DMEM/F12, thus standard DMEM/F12 was utilized, although phenol red-free medium could be used in week 2 of the protocol (Fig. 2b, Supplementary Fig. 2). This effect of phenol red-free DMEM/F12 may be associated with reduced estrogen signaling as the phenol-based dye is a known estrogen mimic which is known to influence proliferation in VSMCs [37,38]. Partial media changes were performed following an initial full media change to WALL media. While this is common practice in iPSC differentiation protocols in other fields [21,39], it is not commonly stipulated in iVSMC protocols. Partial media changes capitalize on the growth factors and extracellular matrix (ECM) generated endogenously by the differentiating cells, and reduce effects of large fluctuations in osmolality and pH [40]. A monolayer differentiation approach was selected over embryoid body formation to harness the ability of differentiating cells to generate biologically relevant ECM in situ, and to ensure compatibility with automated liquid handling. The observation that cell seeding density was the largest contributor to cellular variability, and that plate well-size influenced optimal density, likely reflects the differential fluidic forces, surface area to volume ratios, and base curvatures inherent to different tissue culture formats. Mid-differentiation passaging was minimized as repeated enzymatic disruption of endogenously generated ECM promoted a synthetic phenotype, consistent with the known mechanosensitivity of VSMCs to their surrounding matrix. Confluence also varied with substrate stiffness, in agreement with prior reports that VSMCs sense and respond to mechanical properties with shifts in proliferation [10]. The use of Repsox, a TGF-β1 receptor inhibitor [36], rather than the more commonly employed PDGF-BB and TGF-β1, was critical in promoting a robust contractile phenotype, as we, consistent with others [37,38], observed that these growth factors encouraged a more synthetic VSMC phenotype. VEGF alone generated a mixed contractile-synthetic population potentially useful for disease modeling, while the combination of VEGF and Repsox most reliably yielded contractile iVSMCs. A key strength of this study was the inclusion of iPSC lines from donors of both sexes and across a broad age range, enabling evaluation of protocol robustness in a biologically relevant cohort. To ensure our differentiation protocol was robust and broadly applicable, we prioritized diversity in donor age and sex when generating this cohort of iPSC lines. The inclusion of both female and male lines in studies is particularly important, as vascular cells are known to exhibit sex-specific differences from birth, regardless of hormones [41,42]. Moreover, because some epigenetic signatures of aging can be retained following reprogramming [43]. Together, diversity across these two parameters enabled the evaluation of differentiation efficiency in a biologically relevant cohort. All lines in this initial cohort were derived from individuals of Caucasian genetic diversity (except line SCVI15 which was deidentified), reflective of the European ancestry majority in Australia [44]. While this enabled certain baseline comparisons, we acknowledge the lack of ethnic diversity as a limitation. In future studies we aim to broaden genetic ancestry representation to better reflect the global population and to ensure inclusivity. A central finding of this study was the demonstration that robotic liquid handling did not alter the proteome of iVSMCs relative to manual differentiation (Table 1). PCA of the cell line averages for each condition did not reveal partitioning based on culturing techniques, with cell line identity driving tight clustering in four of the five lines. This suggests that the cell line signature is retained – an important feature for future applications in disease modelling. Hallmark VSMC proteins MYH11, α-SMA, TAGLN, and CNN1 were unaffected. This is particularly noteworthy given that VSMCs are mechanosensitive cells whose phenotype is readily influenced by fluid dynamics and physical handling cues [16,17]. These findings address a critical gap in the field, as the impact of automated liquid handling on iPSC differentiation outcomes had not previously been assessed at the proteome level for any cell type, to our knowledge. The equivalence of robotic and manual differentiation demonstrated here establishes a precedent for the application of automation to iPSC-derived cell manufacturing, supporting the feasibility of standardized, large-scale production [15]. Table 1Summary of comparisons of manual and automated robotic iVSMC differentiation.Table 1Manual techniqueAutomated robotic techniqueComparisonMorphologyElongated VSMC-like morphologyElongated VSMC-like morphologyNo visible difference; Morphologically similar to manual culturesAverage cellular yield984,171 cells/well-c971,118 cells/wellNo significant difference (P > 0.05); One-way ANOVAAverage Protein yield3,606.2 μg/mL2,872.8 μg/mLNo significant difference (P > 0.05); One-way ANOVAVSMC marker expressionMYH11, α-SMA, TAGLN and CNN1 detectedMYH11, α-SMA, TAGLN and CNN1 detectedNo significant difference (P > 0.05); One-way ANOVAGlobal proteome5,936 proteins quantified5,952 proteins quantifiedNo significant proteins at the standard threshold; 0.29% differed at a relaxed threshold Summary of comparisons of manual and automated robotic iVSMC differentiation. As proof-of-principle for downstream applications, iVSMCs were used to generate 3D microtissues. Cells collected from one well of a six well plate were a sufficient quantity to generate one 3D construct with our custom design. The ability of these cells to form structurally intact microtissues of sufficient integrity for manual handling by day 4 suggests that the generated cells may be suitable for microtissue engineering applications in the future, such as the development of vascular grafts [45]. Interestingly, resuspension in Matrigel alone was not sufficient to maintain ring integrity, suggesting that laminin supplementation provides essential structural or signaling cues. The implementation of chemically defined, xenogenic-free conditions further aligns this approach with good manufacturing practice (GMP) standards required for translational applications [46]. The WALL protocol, combined with robotic automation, is well-positioned to meet the growing demand for scalable iVSMC production for high-throughput drug screening, and vascular organ-on-a-chip models; With further validation of the microtissues generated with these cells it also has potential future applications in microtissue engineering, and ultimately regenerative medicine [3]. Conclusion In this study, we showcase a WALL protocol, a weekend-free, automatable, low-cost, low-variation protocol, to differentiate iPSCs into VSMCs. We demonstrated for the first time that differentiation via robotic liquid handling does not significantly alter iVSMC differentiation at the functional, protein-expression, or pathway-interaction level, compared with manual processing. Applied across a diverse cohort of iPSC lines from male and female donors of different ages, the protocol consistently generated high-quality iVSMCs, supporting its robustness and reproducibility. The implementation of a chemically-defined differentiation medium also marks a critical step toward a fully xenogenic-free workflow, increasing compatibility with GMP standards for future translational purposes. Together, these findings establish the WALL protocol as a scalable and standardized platform for the large-scale generation of iVSMCs. Its compatibility with automation could facilitate high-throughput drug screening, vascular organ-on-a-chip studies, and with further validation of 3D microtissues, microtissue engineering, and, ultimately, regenerative medicine applications. CRediT authorship contribution statement Monique Bax: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Jeya Ramalingam: Writing – review & editing, Software, Resources, Project administration, Methodology. Valentin Romanov: Writing – review & editing, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation. Muhammad A. Alsherbiny: Writing – review & editing, Software, Resources, Methodology, Investigation, Data curation. Keerat Junday: Writing – review & editing, Investigation. Jordan Thorpe: Writing – review & editing, Project administration, Methodology, Investigation. Emily Hurley: Writing – review & editing, Visualization, Project administration, Methodology, Data curation. Melanie Lim: Writing – review & editing, Project administration, Investigation. Ingrid Tarr: Writing – review & editing, Formal analysis. Stephanie Hesselson: Writing – review & editing, Project administration. Eleni Giannoulatou: Writing – review & editing, Supervision, Funding acquisition. Renjing Liu: Writing – review & editing, Supervision, Conceptualization. Adam Hill: Writing – review & editing, Supervision, Methodology. Siiri Iismaa: Writing – review & editing, Supervision, Resources, Project administration, Funding acquisition. Robert M. Graham: Writing – review & editing, Supervision, Resources, Funding acquisition. Declaration of generative AI and AI-assisted technologies in the manuscript preparation process During the preparation of this work, the authors used Google Search, Google Scholar, ChatGPT (4o, v5, v5.2 Thinking, 5.6 Sol-High), and Claude (v1.1.1520) were used for literature discovery. Microsoft Word, and ChatGPT were utilized for spelling and grammar checks, and language editing. R Studio, Google Colab, and ChatGPT were used to assist in coding for analysis as described. The authors reviewed and edited the output as needed and take full responsibility for the content of the published article. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements The authors wish to acknowledge the generous contributions of our cell donors, the ongoing support of the SCADaddle Inc. group, and the valuable input of the VCCRI Arteriopathy-SCAD Cohort (VASC) team. The authors wish to also acknowledge Dr Stephanie Hesselson for coordinating sample collection and donor data curation. We wish to thank Prof Joseph Wu, who kindly provided one of the iPSC lines used in this work. We also wish to thank Dr Osvaldo Contreras and Prof Richard Harvey for use of their tissue culture facilities. This work was partially supported by funding from the National Health & Medical Research Council of Australia (APP1161200), NSW Health (Senior Clinician Cardiovascular Grant RG193092, and Senior Scientist Cardiovascular Grant RG04492) and a L3 National Health and Medical Research Council, Australia Investigator Grant (ID: 2020203) awarded to RMG. Animal component-free work was supported by a Medical Advances Without Animals Fellowship and an AMP Tomorrow Maker grant awarded to MB. Portions of this research were conducted using the Victor Chang Cardiac Research Institute Innovation Centre, established with funding from the New South Wales Government. JT and VR were supported by Office of Health and Medical Research Non-animal Technologies Network; NSW Cardiovascular Capacity building scheme collaborative research grant that supported JR, and MRFF Stem cell mission (2024443) also supported JT. Footnotes Supplementary data to this article can be found online at https://doi.org/10.1016/j.mtbio.2026.103638. Contributor Information Monique Bax, Email: mcb202@cam.ac.uk. Robert M. Graham, Email: b.graham@victorchang.edu.au. Supplementary data The following are the supplementary data to this article: Multimedia component 1 Multimedia component 2 Multimedia component 3 Multimedia component 4 Multimedia component 5 Multimedia component 6 Video 1 Data availability We have shared the link to our data References Associated Data Supplementary Materials Multimedia component 1 Multimedia component 2 Multimedia component 3 Multimedia component 4 Multimedia component 5 Multimedia component 6 Video 1 Data Availability Statement We have shared the link to our data