Cardiovascular Organoids With Adjustable Endothelial Composition via SOX17‐Engineered hPSCs Authors: Po‐Yu Liang, Gyuhyung Jin, Nathan R Petrucci, Xiaojun Lance Lian, Xiaoping Bao Published: 2026 ABSTRACT ABSTRACTOrganoids are considered a novel modeling platform for studying human biology and advancing health research. With the ability to demonstrate complex 3D structure and multicellular interactions, organoids have advanced studies in all major organs as a reliable model. In this study, we generated an advanced cardiovascular organoid by using a genome‐edited human pluripotent stem cell line with inducible SOX17 expression, enabling controlled endothelial specification, adjustable cell‐type composition, and human heart‐like morphology. Our organoids recapitulated the cardiotoxic phenotypes of FDA‐approved chemotherapeutic doxorubicin, manifesting as decreased cell viability and diminished contractile activity. Cryoinjury‐induced myocardial infarction in our organoids led to reduced beating, viability, and α‐actinin expression, along with increased fibroblast formation, which were mitigated by Captopril. Lastly, isoproterenol treatment increased peak Ca2+ transient amplitude and shortened APD50 in our organoids, consistent with previously reported β‐adrenergic responses. In summary, we established a protocol for generating in vitro 3D cardiovascular organoids with controllable cellular composition and heart‐like structures, providing a robust and easy‐to‐produce platform for future studies of human heart disease.Keywords: cardiomyocyte, cardiovascular disease, genome editing, human pluripotent stem cell, organoid, SOX17 Introduction Cardiovascular diseases cause a great challenge to global health care system as the leading cause of death around the world (Murphy 2024; Rossi et al. 2021; Sallam et al. 2022). Establishing a reliable and reproducible in vitro cardiovascular model has been crucial for cardiovascular study (Dardano et al. 2024; Eisen et al. 2019; Halloin et al. 2019; Hofbauer et al. 2021; Knight et al. 2021; Volmert et al. 2023; Yi et al. 2021). In the past few years, many platforms have been established to create 3D cardiac models, such as 3D printing organoids (Kupfer et al. 2020; Radhakrishnan et al. 2025; Ye et al. 2024), optogenetics‐ or biomaterials‐mediated patterning (Hellwarth et al. 2021; Hoang et al. 2021; Repina et al. 2020; Repina et al. 2023), heart‐on‐a‐chip system (Butler and Reyes 2024; Kieda et al. 2024; Landau et al. 2024; Mozneb et al. 2024), and embryoid body differentiation of cardiovascular organoids (Ergir et al. 2022; Schmidt et al. 2023; Song et al. 2024). However, recapitulating two key features of cardiogenesis when differentiating in vitro 3D cardiac models remains a significant challenge: (1) Specific cardiovascular cell type composition; (2) Intricate intercellular interaction (Farah et al. 2024; Giacomelli et al. 2017; Giacomelli et al. 2020; Qu et al. 2022; Voges et al. 2023; Yang et al. 2022). Previously, we reported a protocol for 3D vascularized cardiac organoids (vCO) with a chamber‐like structure (Liang et al. 2022). While the model shares similar features to the human heart, the ratio between cardiomyocytes (CMs) and endothelial cells (ECs) could not be precisely modulated. To solve the problem, we investigated gene‐mediated vascularization in cardiovascular organoids. SOX17 has been identified as a critical regulator of cardiogenesis, promoting endothelial differentiation in both human pluripotent stem cells (hPSCs) and mouse embryos (Liu et al. 2014; Ream et al. 2024; Saba et al. 2019). To make reproducible heart‐like cardiac organoids (hCOs), we utilized CRISPR/Cas9 technology to construct a doxycycline‐inducible SOX17 hPSC line and demonstrated that induced overexpression of SOX17 at cardiac progenitor stage leads to reproducible EC generation within our 3D cardiac organoids (COs). By combining the SOX17‐edited and unedited hPSCs, we developed a one‐step protocol for generating endothelialized COs in a reproducible manner. Our hCOs displayed organized morphology and diverse cell types resembling those of the human heart (Protze et al. 2019), providing a valuable platform for future cardiovascular research. Material and Methods Key Resources Table: ReagentSourceIdentifierAntibodiesSOX17‐APC (1:50)R&D SystemsCat#IC1924ACD31‐FITC (1:100)Miltenyi BiotecCat#130‐110‐668NKX2.5InvitrogenCat# PA5‐85768VE‐cadherin (1:50)Santa CruzCat#sc9989cTnTAbcamCat#ab91605WT1ThermoFisher ScientificCat#PA5‐18548MYL2ProteintechCat#10906‐1‐APMYL7InvitrogenCat# MA5‐26390VIMR&D SystemsCat#Af2105Propidium iodide (PI)BD PharmingenCat#556463ACTNInvitrogenCat#ab68167Fluo‐4, AMInvitrogenCat#F14201Chemicals, peptides, and proteinsMatrigelCorningCat#3554277iMatrix‐511Iwai North America IncCat#N‐892021mTeSR™ PlusSTEMCELL TechnologiesCat#100‐0276RPMI 1640GibcoCat#11875093B‐27™ Supplement (50X), serum freeGibcoCat#17504044CellTiter‐Glo 2.0 AssayPromegaCat#G9241MethoCult H4434STEMCELL TechnologiesCat#04434EDTAThermo Fisher ScientificCat#15575020Y‐27632Cayman ChemicalCat#10005583CHIR99021Cayman ChemicalCat#13122Ascorbic acid (L‐ascorbic acid 2‐phosphate)Sigma‐AldrichCat#A8960Bovine serum albuminSigma‐AldrichCat#A1923Human VEGFPEPROTECHCat#100‐2016% FormaldehydeThermo Fisher ScientificCat#28906Triton X‐100Thermo Fisher ScientificCat#A16046. AEBlotting‐grad blocker (dry milk)Bio‐Rad Laboratories, IncCat#1706404Hoechst33342InvitrogenCat#AM‐105Dulbecco's phosphate‐buffered salineSigma‐AldrichCat#D8537Doxorubicin hydrochlorideSigma‐AldrichCat#25316‐40‐9Isoproterenol hydrochlorideThermo Fisher ScientificCat#51‐30‐9CaptoprilThermo Fisher ScientificCat#62571‐86‐2Experiment models: Cell linesHuman pluripotent stem cell: H9WiCellRRID:CVCL_9773Software and algorithmsImageJSchindelin et al. https://imagej.net/ij/ OtherSterile biosafety cabinetsThermo Fisher ScientificCat#1323TSHumidified tissue culture incubatorThermo Fisher ScientificCat#51026282Leica DMi8 microscopeLeica MicrosystemsCat# 12‐550‐15Zeiss LSM 900 confocal microscopeZEISS Methods Maintenance of Human Pluripotent Stem Cells H9 cell lines were purchased from WiCell and cultured in mTeSR Plus medium on Matrigel‐coated plates at 37°C in humidified atmosphere with 5% CO2. Cell passaging was conducted using 0.1 mM EDTA when cells reached 80% ~ 90% confluency and replated in mTeSR plus medium with 5 μM Y27632 (Lian et al. 2014). Hanging Drop Preparation for Embryoid Body Formation For embryoid body (EB) formation, 20 μL droplets were formed by pipetting on the inner side of a 96‐well U‐bottom culture plate lid. Droplets contained 100,000 hPSCs/mL and were composed of mTeSR Plus supplemented with 5 μM Y‐27632 (Cayman Chemical) and iMatrx‐511 (1:100). 100 μL Phosphate‐Buffered Saline (PBS) was added to each well before returning the lid to culture plate. After 24 h, EBs were transferred with a pipette to a new 96‐well U‐bottom culture plate (single organoid per well) and cultured in mTeSR Plus with 5 μM Y‐27632 for 1 day. The resulting EBs were then ready for further differentiation (Banerjee and Bhonde 2006; Wang and Yang 2008). Cardiovascular Organoid Differentiation Rested EBs were treated with 6 μM CHIR99021 (Cayman Chemical) in RPMI 1640 medium for 48 h at day 0. On day 3, medium was replaced with RBL medium (0.1056 g L‐ascorbic acid 2‐phosphate and 0.25 g Bovine Serum Albumin (BSA) in 500 ml RPMI 1640 medium) containing 2 μM Wnt‐C59. On day 4, the medium was replaced with RBL medium. For cardiomyocyte organoid differentiation, RPMI 1640 medium supplemented with 1× B27 (RPMI + B27) was added on day 6. Starting on day 7, the medium was changed daily to RPMI + B27 supplemented with 3 μM doxycycline (DOX) until day 10. After day 10, organoids were maintained in RPMI + B27, with the medium replaced every 3 days. For cardiovascular organoid differentiation, RPMI + B27 with 3 μM CHIR99021 was added on Day 6. Starting on Day 7, the medium was changed daily to RPMI + B27 supplemented with 2 μM DOX until Day 10. After day 10, organoids were maintained in RPMI + B27, with the medium replaced every 3 days. Immunostaining and Imaging Differentiated organoids were fixed in PBS −/− containing 4% formaldehyde for 90 min at 4°C. Following fixation, the solution was aspirated, and organoids were washed twice with PBS. Organoids were incubated overnight at 4°C with the appropriate primary antibodies (1:50) in PBS containing 0.1% Triton X and 5% BSA supplemented with 5% nonfat dry milk. After washing organoids with PBS twice, organoids were stained with secondary antibody (1:50) in PBS containing 0.1% Triton X and 5% BSA at 4°C for 2 h. Hoechst 33342 (2′‐(4‐ethoxyphenyl)−5‐(4‐methylpiperazin‐1‐yl)−2,5′‐bibenzimidazole) (16.2 μM) was added to the solution at 1:1000 ratio during the final 30 min of incubation. After 2 h, staining medium was aspirated, and organoids were washed with PBS twice. Organoids were imaged using a Leica DMi8 microscope or a Zeiss LSM 900 confocal microscope. The images were processed using ImageJ. Doxorubicin (DXR) Toxicity Test DXR stock solution (5 mM) was prepared in DMSO. The toxicity solution was prepared by diluting DXR in RPMI + B27 at final dilution ratios of 1:1000 or 1:5000. Organoids were exposed to the toxicity solution for 3 days, replacing the solution daily. Following the 3‐day treatment, organoids were maintained in RPMI + B27. Cell Viability Assay CellTiter‐Glo 2.0 Assay was conducted by following Promega Corporation CellTiter‐Glo 2.0 Assay protocol. 50 μL of RPMI + B27 and 50 μL CellTiter‐Glo reagent was added to organoids. Organoids were incubated on a shaking incubator for 2 min followed by 10 min under room temperature. Luminescence was recorded using plate reader (Molecular Devices; San Jose, CA, SpectraMax iD3). Myocardial Infarction (MI) Model A MI model was established by inducing cryoinjury in cardiovascular organoids. Cryoinjury was induced using a sterile needle chilled in liquid nitrogen for 1 min. After removing the RPMI + B27 with insulin medium from the cardiovascular organoids, the frozen needle was applied directly to the organoids for 5 s. Fresh RPMI + B27 with insulin medium, with or without captopril (CAP), was then added for recovery culture. Cryosection and Immunofluorescence Organoids were rinsed in PBS twice and fixed in 4% paraformaldehyde solution overnight at 4°C. After washing with PBS three times, organoids were incubated in 30% sucrose at 4°C overnight. Organoids were subsequently embedded in optimal cutting temperature (O.C.T) compound and frozen at −80°C until cryosectioning. Frozen organoids were sliced in 50 μM sections and collected on glass slides. The cryosection was washed three times with PBS and incubated overnight at 4°C with the appropriate primary antibodies (1:50) in PBS containing 0.1% Triton X and 5% BSA supplemented with 5% nonfat dry milk. After washing organoids with PBS twice, organoids were stained with secondary antibody (1:50) in PBS containing 0.1% Triton X and 5% BSA at 4°C for 2 h. Hoechst 33342 (2′‐(4‐ethoxyphenyl)−5‐(4‐methylpiperazin‐1‐yl)−2,5′‐bibenzimidazole) (16.2 μM) was added to the solution at 1:1000 ratio during the final 30 min of incubation. After 2 h, staining medium was aspirated, and organoids were washed with PBS twice. Organoids were imaged using a Leica DMi8 microscope or a Zeiss LSM 900 confocal microscope. Isoproterenol (ISO) Model Analysis ISO stock solution (100 mM) was prepared in PBS. Working ISO medium was generated by diluting the stock 1:1000 in RPMI + B27. Organoids were treated with ISO‐containing medium for 3 days, with a daily medium change. Following treatment, organoids were maintained in RPMI + B27. Membrane Calcium Voltage Analysis Calcium voltage staining was conducted by following Invitrogen standard Fluo‐A AM protocol. The labeling solution was prepared by dissolving 50 μg of Fluo‐4 AM in 50 μL of DMSO. The staining solution was prepared by diluting the labeling solution 1:200 in RPMI + B27 with insulin. Beating organoids were incubated with the staining solution for 40–60 min at 37°C in the dark. Following incubation, organoids were gently washed twice with PBS and transferred to fresh RPMI + B27 with insulin medium for imaging. Calcium transients were recorded within 30 min after staining using a microscope equipped with a FITC channel. After recording, the intensity of FITC was quantified by ImageJ. The amplitude is measured by dividing the maximum intensity to the lowest intensity in one beat. A representative recording of calcium activity is shown in Video S1. Viability Analysis Organoids were incubated with propidium iodide (1:100 in PBS) for 30 min at room temperature. During the final 10 min of incubation, Hoechst 33342 was added at a 1:1000 dilution. After staining, the solution was aspirated, and organoids were washed twice with PBS. Organoids were then imaged by fluorescence microscopy, and cell numbers were quantified using ImageJ. Results Construction of AAVS1 XLone‐SOX17 hPSCs To control endothelial differentiation in our organoid system (Liu et al. 2024), we generated an H9 cell line with an AAVS1‐targeted, doxycycline (DOX)‐inducible SOX17 overexpression construct (SOX17‐H9) to enable inducible SOX17 expression during differentiation. Successful AAVS1 insertion in H9 cells was confirmed by genotyping following drug selection (Figure 1A). One of the homozygous clones C3 was expanded and used for subsequent experiments. DOX was tested at concentrations ranging from 1 to 7 μM in our inducible SOX17 H9 cell line to minimize DOX‐associated toxicity while ensuring robust SOX17 expression (Ahler et al. 2013). While 1 μM DOX failed to induce SOX17 expression, treatment with 3 μM DOX activated SOX17 in the SOX17 H9 cell line, as confirmed by SOX17 immunostaining (Figure 1B, Supporting Figure 1). To minimize DOX‐associated toxicity during differentiation, we used 3 μM DOX in all subsequent experiments. We applied CM differentiation protocol for hPSCs (Lian et al. 2015; Mehta et al. 2014) to our SOX17‐H9 EBs with the addition of DOX during cardiac progenitor stage between Day 7 to Day 10 to initiate EC differentiation (Figure 1C). NKX2.5 expression on day 6 confirmed the differentiation of EBs into cardiac progenitor cells (Supporting Figure 2). SOX17 signal gradually emerged from day 8 and eventually formed a clear vascular structure around day 12 (Figure 1C) while no spontaneous beating occurred. The vascular structure located at the center of our organoids likely reflects their self‐organizing properties. Based on the observed vascular structures, we hypothesized that SOX17 was expressed in ECs, which was confirmed by co‐staining with VE‐cadherin and SOX17 in the resulting organoids (Figure 1D). Co‐expression of SOX17 and VE‐cadherin indicates that SOX17 can serve as a marker of endothelial cells, consistent with previous studies (Corada et al. 2013; Kim et al. 2023). To test the potential of orchestrating the ratio of ECs in COs, we cultured EBs with different H9 to SOX17‐H9 ratios (1:0; 2:1; 0:1) (Supporting Figure 3). After differentiation, H9 COs mainly showed cTnT signal with little VE‐cadherin expression while SOX17‐H9 COs demonstrated great vascular VE‐cadherin signal with no cTnT expression (Figure 1E). By combining both cell lines, we successfully created COs with both endothelial networks and beating CMs. The endothelial signal appeared to correlate with the initial SOX17‐H9 ratio (Figure 1E), suggesting that the number of endothelial cells in COs may be tunable. To generate organoids exhibiting both spontaneous beating and endothelial differentiation, we used a 2:1 mixture of H9 and SOX17‐H9 cells for all subsequent experiments. Endothelialized structure in hPSC‐derived cardiac organoids by SOX17 overexpression. (A) Schematic of the AAVS1 safe‑harbor knock‑in strategy and PCR genotyping of edited clones XLone‐SOX17 H9. Five clones were homozygous and two were heterozygous for the knock‑in. (B) Verification of SOX17 overexpression under doxycycline of AAVS1 safe‑harbor knock‑in XLone‐SOX17 H9. Scale bar = 200 μM. (C) A protocol of differentiating embryoid bodies into cardiac organoids and representative brightfield images showing the emergence of SOX17+ endothelial networks between Day 7 and Day 15. Red = SOX17. Scale bar = 200 μM. (D) Representative immunofluorescence images demonstrating endothelialized SOX17+ structures and co‐expression of VE‐cadherin and SOX17 within cardiac organoids. Green = VE‐cadherin, red = SOX17, blue = Hoechst. Scale bar = 100 μM. (E) Representative immunofluorescence images and quantification of VE‐cadherin, cTnT, and Hoechst of organoids with different H9 to SOX17 cell line ratios (1:0, 2:1, 0:1). Green = VE‐cadherin, yellow = cTnT, blue = Hoechst. Scale bar = 200 μM. n = 3 independent organoids per condition. Data presented as normalized to Hoechst. Values = mean ± standard deviation (SD), one‐way ANOVA. Generation of Multicellular Cardiac Organoids To generate cardiac organoids with three major cardiovascular cell types (ECs, CMs, and epicardial cells), we activated Wnt signaling pathway at cardiac progenitor stage to promote epicardial differentiation as previously reported (Bao et al. 2016; Bao et al. 2017; Branco et al. 2022b; Fernandes et al. 2023). On day 6, we added 3 μM CHIR to our organoids with RPMI + B27. After 1 day of Wnt activation, the medium was replaced with RPMI + B27 containing DOX and maintained for 3 days (Figure 2A). COs under new protocol preserved spontaneous beating and vascular SOX17 expression. The presence of each major cardiovascular cell type was verified by immunostaining (Figure 2A). To assess the morphology of our cardiovascular organoids, we co‐stained COs with major cardiovascular markers: VE‐cadherin for ECs, WT1 for epicardial cells, and cTnT for CMs (Figure 2B). Our COs showed consistent, lineage‐driven self‐organization, with internal lumen‐like cavities forming in 63.6% of samples (7/11) (Supporting Figure 4A). Immunofluorescence and cryosection analyses revealed clear spatial compartmentalization: ECs (SOX17+/VE‐cadherin+) were predominantly localized to the interior, lining the lumen‐like voids, whereas CMs and epicardial cells occupied the peripheral layers (Figure 2B, Supporting Figure 4B‐C). Co‐differentiation of cardiovascular cell types has been shown to promote CM maturation (Mikryukov et al. 2021; Saba et al. 2019). Although H9 COs expressed higher levels of the immature atrial‐associated marker MYL7, COs derived from mixed H9 and SOX17‐H9 cells showed reduced MYL7 expression accompanied by increased MYL2 expression (Supporting Figure 5). This shift in myosin light chain isoforms suggests a trend toward CM specification and maturation, potentially driven by cardiovascular intercellular interactions. Altogether, our cardiac organoids exhibit robust self‐organization, forming internal lumen‐like structures and a layered architecture, while displaying molecular features consistent with early CM subtype specification. One‐step differentiation of cardiovascular organoids with major cardiac lineages and heart‐like morphology. (A) A protocol of differentiating cardiovascular organoids and representative immunofluorescence images of three major cardiovascular cell types within organoids. Green = VE‐cadherin, yellow = cTnT, purple = WT1, blue = Hoechst. Scale bar = 200 μM. (B) Representative confocal images of cardiovascular organoids demonstrating chamber formation, self‐organization, and the presence of major cardiovascular cell types. Green = VE‐cadherin, yellow = cTnT, purple = WT1, blue = Hoechst. Scale bar = 200 μM. Chemotherapy Induced Cardiotoxicity Modeling Doxorubicin (DXR) is a widely used chemotherapy drug for cancer, which may cause irreversible heart damage (Burridge et al. 2016; Sheibani et al. 2022). To assess the extent to which our model recapitulates aspects of human heart biology, hCOs were treated with DXR under different concentration for 3 days. We measured the contractility and viability of hCOs after treatment to quantify the effect of DXR (Figure 3A). While the beating rate of each group remained in a certain range, DXR treatment significantly decreased the number of organoids which demonstrated spontaneous visible beating (Figure 3B). hCOs exposed to a higher DXR concentration (5 μM) eventually stop beating, whereas those treated with a lower dose (1 μM) retain measurable contractility (Figure 3C). We then analyzed the beating behavior of hCOs by measuring calcium transients. Although calcium transients were still detectable in hCOs treated with 5 μM DXR, their beating was markedly disrupted and failed to produce stable, consistent contractions (Figure 3D). Both calcium transient intensity and APD50 were significantly reduced under 5 μM DXR treatment (Figure 3E, Supporting Video 1,2), mirroring the cardiotoxic effects of DXR reported in human hearts (J. Yang et al. 2024). We then measured the viability of hCOs after 6 days of DXR treatment using Promega CellTiter‐Glo 2.0 kit. Treatment with 5 μM DXR resulted in more than 80% cell death in hCOs, although the organoids retained their spherical morphology (Figure 3F). These results suggest that hCOs may serve as a useful platform for drug screening. Advanced cardiovascular organoids simulate doxorubicin‐induced cardiotoxicity. (A) Schematic diagram of simulating and analyzing doxorubicin‐induced cardiotoxicity with cardiovascular organoids. (B) Time‐ and dose‐ dependent effects of doxorubicin on beating frequency of cardiovascular organoids. n = 6–8 independent organoids per condition. Values = mean ± SD. (C) Quantification of the percentage of visibly beating organoids under brightfield microscopy after treatment with different concentrations of doxorubicin. n = 8–13 independent organoids in each concentration. (D) Representative image of time‐lapse recordings of calcium transient activity in cardiovascular organoids following treatment with control (Ctrl), 1 μM, or 5 μM doxorubicin. (E) Quantification of amplitude (F/F0) and APD50 of calcium transient in cardiovascular organoids under different concentration (Ctrl, 1 µM, 5 μM) of doxorubicin treatment. n = 11–14 independent organoids per condition. Values = mean ± SD, Welch's t‐test. (F) Quantification of cardiovascular organoid viability following treatment with different concentrations (Ctrl, 1 μM, 5 μM) of doxorubicin. n = 11–14 independent organoids per condition. Values = mean ± SD, ordinary one‐way ANOVA. In Vitro Modeling of Cardiac Injury To evaluate the utility of hCOs for disease modeling, we induced cryoinjury to simulate myocardial infarction (MI) (Kar et al. 2024; Richards et al. 2020). We induced cryoinjury on day 15 post‐differentiation (Figure 4A). We monitored the beating rate and ratio following cryoinjury and subsequently assessed calcium transients and viability to evaluate the extent of damage to hCOs (Figure 4A). Captopril (CAP), an FDA‐approved medication for myocardial infarction, was administered after cryoinjury to assess its ability to mitigate functional impairment. Beating activity was rapidly impaired following cryoinjury. In the absence of CAP, hCOs showed a marked decline in both beating rate and beating area, with contractions ceasing entirely by day 6 post‐injury (Figure 4B‐C). In contrast, CAP‐treated hCOs retained limited but irregular beating activity and demonstrated higher viability (Figure 4D,E). We then stained hCOs with a common muscle cell marker, sarcomeric alpha actinin, to assess CM integrity. hCOs treated with cryoinjury and CAP demonstrated higher sarcomeric alpha actinin intensity compared to hCOs treated with cryoinjury without CAP (Figure 4F). To validate if fibroblasts were differentiated from epicardial cells due to cryoinjury (Ciucci et al. 2023), hCOs were stained with VIMENTIN (VIM), a fibroblast marker, after treatment. hCOs subjected to cryoinjury alone exhibited elevated fibroblast signals, as evidenced by enhanced VIM expression, potentially indicating increased fibrosis. CAP treatment mitigated fibroblast formation under the same conditions (Xia et al. 2025) (Figure 4F). Taken together, these findings indicate that our hCOs may serve as a promising platform for modeling cardiovascular diseases. Advanced cardiovascular organoids model myocardial infarction by cryoinjury and captopril. (A) Schematic diagram of modeling and analyzing myocardial infarction with advanced cardiovascular organoids by cryoinjury. (B) Time‐dependent effects of cryoinjury and captopril on beating frequency of cardiovascular organoids. n = 6–8 independent organoids per condition. Values = mean ± SD. (C) Quantification of percentage of organoids visibly beating under brightfield microscopy under each condition (Ctrl, Cryo, Cryo + Cap). n = 8–13 independent organoids in each condition. (D) Representative images of time‐lapse recording of calcium transient activity in cardiovascular organoids after cryoinjury or cryoinjury plus captopril treatment. (E) Quantification of cardiovascular organoid viability under different conditions (Ctrl, Cryo, Cryo + Cap). n = 8–10 independent organoids per condition. Values = mean ± SD, two‐way ANOVA. (F) Quantification of sacromeric alpha actinin and VIMENTIN in organoids under each condition (Ctrl, Cryo, Cryo + Cap). n = 4–6 independent organoids per condition. Data presented as normalized to Hoechst. Values = mean ± SD, one‐way ANOVA. Isoproterenol (ISO) is a drug commonly used to treat bradycardia condition (Ebihara et al. 2017). We aimed to investigate if our hCOs can recapitulate human hearts’ response toward ISO. hCOs were treated with ISO for 3 days (Figure 5A). Following ISO treatment, hCOs exhibited an increased beating amplitude (Video S3) while maintaining a similar beating frequency compared to pre‐treatment levels (Figure 5B,C). To evaluate the effect of ISO on contractile behavior, calcium transient assays were performed on hCOs. ISO treatment significantly increased the peak amplitude (from 1.14 to 1.67) and shortened APD50 (Figure 5D,E). These findings were consistent with previous reports (Gardner et al. 1990) and further supported the utility of our hCOs as a platform for evaluating the effects of drugs on human hearts. Advanced cardiovascular organoids as a platform for isoproterenol function assay. (A) Schematic diagram of mimicking isoproterenol's effect on human heart with advanced cardiovascular organoids by cryoinjury. (B) Time‐ and dose‐ dependent effects of isoproterenol on beating frequency of cardiovascular organoids. n = 6–8 independent organoids per condition. Values = mean ± SD. (C) Quantification of percentage of visibly beating organoids under brightfield microscopy with or without isoproterenol. n = 6–8 independent organoids in each concentration. (D and E) Representative images of time‐lapse recording of calcium transient activity in cardiovascular organoids after isoproterenol treatment (D) and the quantification of amplitude (F/F0) and APD50 (E) are shown. n = 10–14 independent organoids per condition. Values = mean ± SD, Welch's t‐test. Discussion While ETV2 is an established master regulator that initiates stem cell differentiation toward endothelial lineages and is transiently expressed in early mesoderm (Wang et al. 2020), SOX17 acts later in development to specify arterial identity and stabilize the endothelial phenotype (Grath and Dai 2024; Robinson et al. 2014). To better mimic the temporal hierarchy of embryonic development, we therefore employed SOX17 to drive endothelial cell specification from cardiac progenitors within the organoids. Expression of SOX17 during cardiac progenitor stage stimulated endothelial differentiation in our cardiac organoids, thereby facilitating controlled adjustment of cardiovascular cell type ratios to better mimic the human heart. By combining varying proportions of unedited and genome‐edited cell lines, we successfully generated organoids with tunable levels of vascularized ECs. By co‐differentiating multiple cardiovascular cell types, the intricate paracrine interplay between different cells may be retained, as partially demonstrated by improved CM maturation. Our COs exhibited robust self‐organization and developed 3D morphological features, including internal lumen‐like cavities and a layered architecture reminiscent of the human heart. This structural organization may reflect, in part, the stabilizing role of SOX17, which functions during development to refine endothelial identity and promote arterial fate. The platform showed strong potential for cardiovascular disease modeling, as demonstrated by human heart‐like physiological responses to isoproterenol and doxorubicin (Tatekoshi et al. 2024). Overall, our study indicates that SOX17 overexpression at the cardiac progenitor stage efficiently promotes endothelial specification, providing an adaptable strategy for generating organoids with a tunable endothelial cell composition. This approach may be broadly useful for complex disease modeling and therapeutic development in cardiovascular research (Ng et al. 2021; Wauchop et al. 2023). Although our organoids show considerable potential as a platform for cardiovascular research, several limitations remain. Our cardiovascular organoids comprised the three major cardiac cell types—ECs, CMs, and epicardial cells—but also contained minor, undesired cell populations, likely due to the inherent heterogeneity of stem cell differentiation (Cheng et al. 2021; Misfeldt et al. 2009; Moriwaki et al. 2023; Pinto et al. 2016; Silva et al. 2021). To develop a model that more closely resembles the human heart, these minor cell types cannot be overlooked. Technically, the dense 3D extracellular matrix and low total cell numbers per organoid also make it difficult to generate a high‐quality single‐cell suspension for accurate quantification of cellular composition by flow cytometry or other assays. Another limitation of this study was the lack of detailed elucidation of intercellular interactions (Lee et al. 2020; Nakajima et al. 2019; Odiete et al. 2012; Palencia‐Desai et al. 2015; Quijada et al. 2021). Early communication between cardiovascular cell types plays a crucial role in cardiogenesis. Although our protocol preserved some degree of crosstalk among these cells, the specific mechanisms by which they influence one another warrant further investigation. Addressing these limitations provides opportunities and direction for future research. Author Contributions Po‐Yu Liang, Xiaojun Lance Lian, and Xiaoping Bao designed research. Po‐Yu Liang, Gyuhyung Jin, and Nathan R. Petrucci, performed research. Po‐Yu Liang, Xiaojun Lance Lian, and Xiaoping Bao wrote the paper. Conflicts of Interest The authors declare no conflicts of interest. Supporting information Supporting Video 1 Supporting Video 2 Supporting Video 3 Supporting Video 4 Supporting File 1 Acknowledgments We thank members of the Bao and Lian groups for their discussion of this manuscript. We also thank the NSF CBET‐2143064 (to X.B.), NSF CBET‐2425704 (to X.B.), NSF CBET‐1943696 (to X.L.L.), NIH R37CA265926 (to X.B.), NIH R21EB026035 (to X.L.L.), and NIH R56DK133147 (to X.L.L.) for their funding support. Liang, P.‐Y. , Jin G., Petrucci N. R., Lian X. L., and Bao X.. 2026. “Cardiovascular Organoids With Adjustable Endothelial Composition via SOX17‐Engineered hPSCs.” Biotechnology and Bioengineering 123: 2655–2667. 10.1002/bit.70275. Po‐Yu Liang and Gyuhyung Jin are co‐first authors. Contributor Information Xiaojun Lance Lian, Email: lian@psu.edu. Xiaoping Bao, Email: bao61@purdue.edu. Data Availability Statement The data that support the findings of this study are available from the corresponding author upon reasonable request. References Associated Data Supplementary Materials Supporting Video 1 Supporting Video 2 Supporting Video 3 Supporting Video 4 Supporting File 1 Data Availability Statement The data that support the findings of this study are available from the corresponding author upon reasonable request.