Deepak Srivastava, M.D. | Gladstone Institutes and University of California San Francisco

Introduction

Heart disease remains the number one non-infectious cause of death worldwide in adults and newborns. Survivors with heart disease are often left with damaged hearts due to the inability of the mammalian heart to regenerate itself. As a result, globally over 25 million people suffer from heart failure where the cardiovascular system is unable to support the body’s normal activities. Genetic causes of valve disease and congenital malformations, present in one percent of all live births, have been discovered, but the cause for a majority is still unknown and no disease-modifying therapies are available clinically.

A deep understanding of gene networks involved in cardiac cell fate and morphogenesis during development has enabled relatively precise control of cell fate decisions in vitro. The ability to control programming and reprogramming of human cells is now possible and is used to not only generate human induced pluripotent stem (iPS) cells, but also to differentiate pluripotent cells into desired cell types for transplantation, to understand disease mechanisms, and to discover new therapeutics. Here we will provide some examples of how control of fate decisions involving cellular reprogramming is providing hope for cardiac regeneration and new therapeutic development.

Cellular Reprogramming for Regenerative Medicine

In recent years, redeployment of key networks by discrete combinations of transcription factors (TFs) and translational regulators has enabled reprogramming of somatic cells to induced pluripotent stem (iPS) cells, demonstrating the power of a limited number of factors to dictate cell-fate decisions (1, 2). More recently, somatic cells have been reprogrammed directly toward multiple cell types in vitro using combinations of transcription factors and/or microRNAs with varying degrees of efficiency, including the cardiomyocyte, neuronal, and hepatocyte fates, among other cell types (3). Because the heart has limited regenerative potential and has a vast pool of fibroblasts, the ability to reprogram endogenous cardiac fibroblasts into cardiomyocyte-like cells has emerged as a promising approach to restore function to damaged hearts.

Our lab initially pioneered the conversion of cardiac fibroblasts toward a cardiomyocyte-like cell with a combination of Gata4, Mef2c and Tbx5 (GMT) (4, 5). Many groups have reproduced and advanced the technology with various combinations of genes and microRNAs and chemicals (6). We termed these cells “induced cardiomyocytes” (iCMs), as they developed sarcomeric structures and calcium transients typical of cardiomyocytes. Although there was significant heterogeneity to the iCMs, the more fully reprogrammed iCMs had action potentials that were most similar to adult ventricular myocytes, consistent with lineage tracing evidence that reprogrammed cells did not go through a cardiac progenitor stage during their transition (4). However, most of the cells transdifferentiated in vitro were only partially reprogrammed.

Because the objective with this strategy was to harness the endogenous cardiac fibroblasts for regeneration without needing to use cell-based therapy, we delivered GMT in vivo retrovirally after ischemic injury in mice and successfully converted resident non-myocytes to cardiomyocyte-like cells. Genetic lineage-tracing studies in mice were performed to demonstrate that dividing non-myocytes infected by retroviruses could be converted into iCMs (5) progressively over a period of 4 weeks. iCMs in this setting developed sarcomeres, with ~50% of reprogrammed cells developing contractile activity when isolated in single-cell suspension, compared to less than 0.1% when reprogrammed in vitro. Importantly, we found evidence for electrical coupling of the in vivo reprogrammed iCMs with endogenous cardiomyocytes and other iCMs, and reprogrammed cells were most similar to ventricular cardiomyocytes electrically and transcriptomically. In vivo delivery of GMT intramyocardially decreased scar size and attenuated cardiac dysfunction after coronary ligation, as assessed by MRI and echocardiography. As expected, cardiomyocytes within the scar area of GMT-treated mice represented newly born iCMs as determined by lineage tracing experiments (Fig. 1).

To advance cardiac reprogramming technology, we and others described overlapping but distinct combinations of factors that could reprogram human fibroblasts into a more cardiomyocyte-like state. This included a combination of MEF2C, its co-activator, MYOCARDIN, and TBX5 (7, 8). This technology is now being developed toward a clinical trial by Tenaya Therapeutics, where reprogramming factors have been packaged into a single AAV vector with initial efficacy observed in the pig model. Future clinical trials in patients with heart failure will determine if this regenerative approach can result in improved cardiac function and avoidance of heart transplant or death.

Cellular Reprogramming for Disease Modeling and Drug Discovery

Human mutations with large effect size, as observed for monogenic diseases, hold the greatest promise for successful disease modeling using human iPS-derived cells, with the hope that underlying mechanisms will be relevant to more common forms of disease. However, use of iPS-derived models for understanding more complex disease will also be important, albeit more challenging. Here, we will consider examples of both cases as they relate to heart disease.

iPS-Derived Therapeutic Development for Calcific Aortic Valve Disease

Calcific aortic valve disease (CAVD) is the third leading cause of adult heart disease and is responsible for over 100,000 valve replacements annually in the United States alone. The disease progresses in an age-dependent fashion. Bicuspid aortic valve (BAV), a congenital malformation which occurs in 1-2% of the population and involves the formation of two rather than the normal three valve leaflets, is a major risk factor for early valve calcification, although the mechanism for the calcification had been unknown. Our group previously reported two families with heterozygous nonsense mutations in the membrane-bound transcription factor, NOTCH1 (N1), which led to BAV and severe aortic valve calcification in adults (9). Valve thickening also occurred, and pathology ranged from neonatal to adult onset. Further studies have identified N1 mutations in additional familial cases of BAV and CAVD, as well as in approximately 4% of sporadic CAVD cases, underscoring the importance of N1 in this disease (10, 11). In mice, endothelial cell (EC)-specific deletion of the N1 ligand JAGGED1 results in aortic valve calcification and malformed valves (12), consistent with a critical role for the ECs lining the valve in the malformation. Despite the recognition of N1 mutations as a cause of CAVD, as well as SMAD6 (13), there are currently no medical treatments available for CAVD patients.

Recent studies from our group suggest the pathology involves reprogramming of valve endothelial cells into osteoblast-like cells with activation of gene networks that promote calcification. Using human iPSC-derived ECs, we showed that heterozygous nonsense mutations in N1 disrupt the epigenetic architecture, resulting in de-repression of latent pro-osteogenic and -inflammatory gene networks (Fig. 2) (14). Hemodynamic shear stress activated anti-osteogenic and anti-inflammatory networks in N1^+/+, but not N1^+/– iPSC-derived ECs (14). N1 haploinsufficiency altered H3K27ac at N1-bound enhancers determined by chromatin immunoprecipitation followed by sequencing (ChIP-seq), dysregulating downstream transcription of over 1000 genes. The gene pathways that were perturbed were implicated in osteogenesis and inflammation. Computational analyses of the disrupted N1-dependent gene network by integrating datasets revealed regulatory nodes, particularly the transcription factors, SOX7, TCF4 (mediating Wnt signaling) and SMAD1 (mediating Bmp signaling), that were upregulated in the mutant setting (Fig. 2). Remarkably, knockdown of just SOX7 and TCF4 restored the gene network dysregulated by N1 haploinsufficiency toward the wild-type (WT) state (14). Importantly, primary ECs grown from explanted valves from patients with CAVD show a similar dysregulation in gene expression (15), supporting the clinical relevance of these findings.

We introduced the mouse N1^– allele into mice lacking the telomerase RNA component (TERC), resulting in shortened telomeres over successive generations of breeding. TERC^–/– mice are relatively normal for up to 5 generations of breeding, each with progressively shorter telomeres. Although N1^+/– mice have normal echocardiographic findings at baseline, crossing this strain to mice lacking TERC over successive generations (mTR^−/−generation 1-3, referred to as mTR^G1-mTR³) demonstrated that in the setting of shortened telomeres, N1^+/– mice develop age-dependent AV thickening, calcification, and stenosis as well as pulmonary valve (PV) stenosis, mimicking the range of human disease caused by N1 haploinsufficiency (16). 40-50% of N1^+/– mTR^G2 mice developed significant histologic evidence of aortic valve calcification within 2 months of age, and ~30% had severe enough aortic or pulmonary valve stenosis to detect echocardiographically, represented by acceleration of blood flow across the valve. Most strikingly, immunohistochemistry of aortic valve sections revealed the presence of Runx2-positive cells in the valve. Runx2 is a “master transcriptional regulator” of the osteoblast fate, consistent with the conclusion from the human iPSC study that the underlying pathology in CAVD is a cellular reprogramming event of a valve cell into an osteoblast-like state. Thus, this mouse model both effectively recapitulated many aspects of the human disease state and supported conclusions of the human iPSC model. It also provided an in vivo model in which to test potential therapies.

Given the mechanistic insights we developed through our gene network analyses, we sought to identify small molecules that could correct the gene expression dysregulated by N1 haploinsufficiency. Small molecules are traditionally screened for their effects on one to several outputs at most, from which their predicted efficacy on the disease as a whole is extrapolated. However, determining the gene regulatory networks driving human disease allows the design of therapies targeting the underlying disease mechanism rather than primarily symptomatic management. In principle, mapping the architecture of the dysregulated network could enable screening for molecules that correct a gene network’s core regulatory elements rather than peripheral downstream effectors that will likely have only limited influence on the disease process.

Accordingly, we designed a targeted RNA-seq strategy assaying expression of over 100 genes that were either predicted central regulatory nodes or peripheral genes positioned within varied regions of the N1-dependent network in human iPSC-derived ECs determined by whole transcriptome RNA-seq (14, 15). We evaluated these genes in isogenic wild-type (WT) or N1^+/– iPS-derived ECs exposed to either DMSO or each of 1595 small molecules. To screen small molecules for this effect, we used machine learning approaches to classify the network gene expression by targeted RNA-seq as WT or N1^+/– based on isogenic ECs of each genotype exposed to vehicle. This strategy resulted in a total of 7 hits that were ultimately validated. The molecule, XCT790, had the largest corrective impacts on the network.

We performed a pre-clinical trial with the drug candidates by treating 4-week-old N1^+/– mTR^G2 mice with daily intraperitoneal injection of each compound for 30 days. Consistent with the gene network shift in human iPSCs, XCT790 treatment was the most effective in vivo and was sufficient to prevent aortic valve stenosis in vivo by echocardiography and showed a trend of reducing pulmonary valve stenosis by echocardiography (Fig. 3). Compared to control solvent, XCT790 also significantly reduced the thickness of treated aortic and pulmonary valves, and calcification of aortic valves.

We tested whether the effect of XCT790 could generalize to primary AV ECs from multiple patients with sporadic CAVD. We performed RNA-seq in primary human AV ECs cultured from explanted normal tricuspid AVs (nTAVs, n=5), calcified tricuspid AVs (cTAVs, n=9), and calcified bicuspid AVs (cBAVs, n=12) treated with XCT790 or DMSO. Overall, there was a significant overlap in genes dysregulated in N1-haploinsufficient ECs with those dysregulated in the same direction in cTAV ECs and cBAV ECs. XCT790 was effective in broadly correcting the dysregulated genes back to the normal state in both primary cTAV and cBAV ECs, including the key nodes regulators, SOX7, TCF4 and SMAD1 (15) (Fig. 4).

XCT790 is annotated to be a highly specific compound that targets the orphan nuclear receptor ERRa (estrogen-related receptor a), with little cross-reactivity with the estrogen receptor (ER). Thus, by inhibiting ERRa, XCT790 may function to block the aberrantly activated pro-osteogenic signaling in N1-haploinsufficient cells to prevent valve stenosis and calcification. Further development of XCT790 or a chemical derivative is underway for clinical advancement of this potential medical therapy which was based on human genetic studies, iPS-based interrogation of mechanism, gene network correcting drug discovery, and in vivo drug efficacy.

Discovery of novel gene candidates for congenital heart disease by intersecting iPS-based proteomics and genetics

While monogenic causes of heart disease have been informative, these are relatively rare. Genetic analyses of over 3,000 proband-parent trios in the Pediatric Cardiac Genomics Consortium (PCGC) revealed that de novo monogenic aberrations were found to collectively contribute to ~10% of congenital heart disease (CHD) cases, while rare inherited and copy number variants have been identified in ~1% and 25% of cases, respectively (17). Additionally, polygenic and oligogenic inheritance models, where multiple genetic variants with epistatic relationships are implicated, have been proposed as mechanistic explanations for certain complex phenotypes. A recent study from our group highlighted the involvement of genetic modifiers in human cardiac disease (18), but the net contribution of oligogenic inheritance remains to be determined. A barrier to a complete understanding of CHD’s etiology is its immense genetic heterogeneity. Estimates based on de novo mutations alone indicate that more than 390 genes may contribute to CHD pathogenesis (19). Despite the growing catalogue of human genome variants, the cause of over 50% of CHD cases remains unknown (17).

Cardiac malformations have been linked to variants in tissue-enriched cardiac transcription factors that are expressed more widely. Such transcription factors typically form complexes with other tissue-enriched and ubiquitous proteins to orchestrate specific developmental gene programs. Missense variants in transcription factors can disrupt specific interactions with other proteins, affecting their transcriptional cooperativity and causing disease (20). In CHD specifically, an excess of protein-altering de novo variants from the Pediatric Cardiac Genomic Consortium’s cohort were found in ubiquitously expressed chromatin regulators that partner with cardiac transcription factors to regulate the expression of key developmental genes (21). This led us to hypothesize that protein-protein interactors of transcription factors associated with CHD may be enriched in disease-associated proteins, even if these proteins are not tissue-specific.

GATA4 and TBX5 are two essential transcription factors and among the first identified monogenic etiologies of familial CHD. Subsequent studies demonstrated that TBX5 and GATA4 cooperatively interact on DNA throughout the genome to regulate heart development (20, 22). Disruption of the physical interaction between these proteins or with other specific co-factors by missense variants can impair transcriptional cooperativity and lineage specification, and ultimately cause cardiac malformations (20, 23). Therefore, the unbiased identification of human GATA4 and TBX5 (GT) protein interactors during cardiogenesis could highlight disease mechanisms and aid in predicting the impact of protein-coding variants in CHD.

To identify the GATA4 and TBX5 protein interactome (GT-PPI) in the relevant human cardiac cells, we used human induced pluripotent stem cell-derived cardiac progenitors (CPs) and identified antibodies against each endogenous factor that were effective for affinity purification and mass spectrometry (AP-MS). Using CRISPR Cas9-gRNA ribonucleoproteins, we generated clonal TBX5 or GATA4 homozygous knockout hiPSC lines as negative controls. This approach yielded 272 proteins in total, which comprised several of the previously reported GATA4 and TBX5 interactors as well as novel interactors (24). Mutations in several of these interactors have been previously associated with human or mouse cardiac malformations, highlighting the potential of this approach for disease-gene discovery. Use of human iPS-derived cardiac progenitors was essential for this approach, as the same type of AP-MS resulted in a set of interactors that were not enriched for variants in CHD probands.

To determine whether the GT-interactors identified in human CPs might help predict genetic risk factors for CHD, we assessed their intersection with de novo variants (DNVs) and very rare (minor allele frequency (MAF) <10^-5) inherited loss-of-function variants found in over 3,000 CHD probands from the PCGC. We used a permutation-based statistical test to analyze the frequency of variants in GT-interacting proteins among the CHD probands compared to the control group. The analysis indicated that protein-altering DNVs were significantly more likely to be found within GT interactors in the CHD cohort relative to the control cohort by nearly 6-fold for GATA4 interactors, and 4-fold for TBX5 interacting proteins (Fig. 5). By contrast, very rare inherited loss-of-function variants occurred in GT-PPI proteins with the same frequency in control and CHD groups.

We developed a prioritization score for the DNVs that incorporated gene, residue and proband information that appeared to effectively rank the potential impact of each missense DNV. To experimentally test the importance of missense mutations predicted to play a role in CHD, we focused on a high-scoring GATA4-interacting protein, GLYR1, which is a chromatin reader involved in chromatin modification and regulation of gene expression through nucleosome demethylation (25). The GLYR1 missense CHD DNV we detected involved the substitution of a highly conserved proline with a leucine at amino acid (aa) 496. This amino acid change disrupted GLYR1 interaction with GATA4 (24).

Analyses of DNA occupancy in iPS-CPs found a statistically significant overlap between GLYR1 and GATA4-bound gene bodies, identifying a defined subset of GATA4 and GLYR1-bound genes, mostly upregulated in CPs vs hiPS cells and with greater enrichment in heart development GO terms compared to GLYR1-only and GATA4-only occupied gene bodies.

Evaluation of a human iPS cell line containing the GLYR1 proline to leucine (P496L) missense mutation found in the patient suggested that the P496L variant affects GLYR1 DNA occupancy and transcriptional regulation of a discrete set of target genes co-bound by GATA4, several of which have been involved in human cardiac malformations and cardiomyopathies. Overall, these data demonstrated a detrimental impact of the GLYR1 P496L variant in CM differentiation, associated with altered GLYR1 genomic occupancy and gene regulation at a discrete set of loci co-bound by GATA4.

In order to assess the biological importance of the GLYR1 P496L variant in vivo, we generated a mouse line harboring a P495L single nucleotide variant in GLYR1 (Glyr1^P495L/+), homologous to human P496L, using CRISPR-Cas mediated genome editing. Over half of the mutant mice died at birth, many with ventricular septal defects (VSDs). Thus, this model provided evidence for the biological importance of GLYR1 in cardiac development, and demonstrates a deleterious effect of the P495L variant in vivo.

To assess whether there was a GATA4-GLYR1 genetic interaction in mice, we crossed Glyr1^P495L/+ mice to GATA4-mutant mice. Whole mount, histology and echocardiography analysis showed that compound Glyr1^P495L/+:Gata4^+/- hearts had complete penetrance of cardiac septal defects, including about 80% represented by atrio-ventricular septal defects (AVSDs). These data provide in vivo evidence for the biological relevance of the GLYR1 P496L variant and its interaction with GATA4 in human disease. Furthermore, the results experimentally validate the approach of identifying the interactomes of disease-causing proteins and evaluating genetic variants in the interactors.

Conclusions

In summary, the use of cellular reprogramming technologies has facilitated novel methods of cardiac regeneration by directly reprogramming resident cardiac fibroblasts into new cardiomyocyte-like cells, led to the recognition of aberrant cellular reprogramming as the basis for aortic valve disease, and revealed novel genes contributing to CHD. Each of these findings establish the foundation for novel therapeutic approaches to human heart disease.

References

1. Takahashi K, Yamanaka S. Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors. Cell. 2006;126(4):663-76. PubMed PMID: 16904174.
2. Yu J, Vodyanik MA, Smuga-Otto K, Antosiewicz-Bourget J, Frane JL, Tian S, Nie J, Jonsdottir GA, Ruotti V, Stewart R, Slukvin, II, Thomson JA. Induced pluripotent stem cell lines derived from human somatic cells. Science (New York, NY. 2007;318(5858):1917-20. PubMed PMID: 18029452.
3. Srivastava D, DeWitt N. In vivo cellular reprogramming: The next generation. Cell. 2016;166(6):1386-96. doi: 10.1016/j.cell.2016.08.055. PubMed PMID: 27610565; PMCID: PMC6234007.
4. Ieda M, Fu J, Delgado-Olguin P, Vedantham V, Hayashi Y, Bruneau BG, Srivastava D. Direct reprogramming of fibroblasts into functional cardiomyocytes by defined factors. Cell. 2010;142:375-86. PubMed PMID: 20691899; PMCID: PMC2919844.
5. Qian L, Huang Y, Spencer CI, Foley A, Vedantham V, Liu L, Conway SJ, Fu J, Srivastava D. In vivo reprogramming of murine cardiac fibroblasts into induced cardiomyocytes. Nature. 2012;485:593-8; PMCID: PMC3369107.
6. Xie Y, Liu J, Qian L. Direct cardiac reprogramming comes of age: Recent advance and remaining challenges. Semin Cell Dev Biol. 2022;122:37-43. Epub 2021/07/27. doi: 10.1016/j.semcdb.2021.07.010. PubMed PMID: 34304993; PMCID: PMC8782931.
7. Fu JD, Stone NR, Liu L, Spencer CI, Qian L, Hayashi Y, Delgado-Olguin P, Ding S, Bruneau BG, Srivastava D. Direct reprogramming of human fibroblasts toward a cardiomyocyte-like state. Stem Cell Reports. 2013;1(3):235-47. Epub 2013/12/10. doi: 10.1016/j.stemcr.2013.07.005. PubMed PMID: 24319660; PMCID: PMC3849259.
8. Mohamed TM, Stone NR, Berry EC, Radzinsky E, Huang Y, Pratt K, Ang YS, Yu P, Wang H, Tang S, Magnitsky S, Ding S, Ivey KN, Srivastava D. Chemical enhancement of in vitro and in vivo direct cardiac reprogramming. Circulation. 2017;135:978-95. doi: 10.1161/CIRCULATIONAHA.116.024692. PubMed PMID: 27834668; PMCID: PMC5340593.
9. Garg V, Muth AN, Ransom JF, Schluterman MK, Barnes R, King IN, Grossfeld PD, Srivastava D. Mutations in NOTCH1 cause aortic valve disease. Nature. 2005;437(7056):270-4. PubMed PMID: 16025100.
10. Foffa I, Ait Ali L, Panesi P, Mariani M, Festa P, Botto N, Vecoli C, Andreassi MG. Sequencing of NOTCH1, GATA5, TGFBR1 and TGFBR2 genes in familial cases of bicuspid aortic valve. BMC Med Genet. 2013;14:44. doi: 10.1186/1471-2350-14-44. PubMed PMID: 23578328; PMCID: PMC3637327.
11. Mohamed SA, Aherrahrou Z, Liptau H, Erasmi AW, Hagemann C, Wrobel S, Borzym K, Schunkert H, Sievers HH, Erdmann J. Novel missense mutations (p.T596M and p.P1797H) in NOTCH1 in patients with bicuspid aortic valve. Biochem Biophys Res Commun. 2006;345(4):1460-5. PubMed PMID: 16729972.
12. Hofmann JJ, Briot A, Enciso J, Zovein AC, Ren S, Zhang ZW, Radtke F, Simons M, Wang Y, Iruela-Arispe ML. Endothelial deletion of murine Jag1 leads to valve calcification and congenital heart defects associated with Alagille syndrome. Development. 2012;139(23):4449-60. doi: 10.1242/dev.084871. PubMed PMID: 23095891; PMCID: PMC3509736.
13. Gillis E, Kumar AA, Luyckx I, Preuss C, Cannaerts E, van de Beek G, Wieschendorf B, Alaerts M, Bolar N, Vandeweyer G, Meester J, Wunnemann F, Gould RA, Zhurayev R, Zerbino D, Mohamed SA, Mital S, Mertens L, Bjorck HM, Franco-Cereceda A, McCallion AS, Van Laer L, Verhagen JMA, van de Laar I, Wessels MW, Messas E, Goudot G, Nemcikova M, Krebsova A, Kempers M, Salemink S, Duijnhouwer T, Jeunemaitre X, Albuisson J, Eriksson P, Andelfinger G, Dietz HC, Verstraeten A, Loeys BL, Mibava Leducq C. Candidate gene resequencing in a large bicuspid aortic valve-associated thoracic aortic aneurysm cohort: SMAD6 as an important contributor. Front Physiol. 2017;8:400. Epub 2017/07/01. doi: 10.3389/fphys.2017.00400. PubMed PMID: 28659821; PMCID: PMC5469151.
14. Theodoris CV, Li M, White MP, Liu L, He D, Pollard KS, Bruneau BG, Srivastava D. Human disease modeling reveals integrated transcriptional and epigenetic mechanisms of NOTCH1 haploinsufficiency. Cell. 2015;160:1072-86. PubMed PMID: 25768904; PMCID: PMC4359747.
15. Theodoris CV, Zhou P, Liu L, Zhang Y, Nishino T, Huang Y, Kostina A, Ranade SS, Gifford CA, Uspenskiy V, Malashicheva A, Ding S, Srivastava D. Network-based screen in iPSC-derived cells reveals therapeutic candidate for heart valve disease. Science (New York, NY. 2021;371(6530). Epub 2020/12/12. doi: 10.1126/science.abd0724. PubMed PMID: 33303684; PMCID: PMC7880903.
16. Theodoris CV, Mourkioti F, Huang Y, Ranade SS, Liu L, Blau HM, Srivastava D. Long telomeres protect against age-dependent cardiac disease caused by NOTCH1 haploinsufficiency. J Clin Invest. 2017;127(5):1683-8. doi: 10.1172/JCI90338. PubMed PMID: 28346225; PMCID: PMC5409071.
17. Zaidi S, Brueckner M. Genetics and genomics of congenital heart disease. Circ Res. 2017;120(6):923-40. Epub 2017/03/18. doi: 10.1161/CIRCRESAHA.116.309140. PubMed PMID: 28302740; PMCID: PMC5557504.
18. Gifford CA, Ranade SS, Samarakoon R, Salunga HT, de Soysa TY, Huang Y, Zhou P, Elfenbein A, Wyman SK, Bui YK, Cordes Metzler KR, Ursell P, Ivey KN, Srivastava D. Oligogenic inheritance of a human heart disease involving a genetic modifier. Science (New York, NY. 2019;364(6443):865-70. Epub 2019/05/31. doi: 10.1126/science.aat5056. PubMed PMID: 31147515; PMCID: PMC6557373.
19. Homsy J, Zaidi S, Shen Y, Ware JS, Samocha KE, Karczewski KJ, DePalma SR, McKean D, Wakimoto H, Gorham J, Jin SC, Deanfield J, Giardini A, Porter GA, Jr., Kim R, Bilguvar K, Lopez-Giraldez F, Tikhonova I, Mane S, Romano-Adesman A, Qi H, Vardarajan B, Ma L, Daly M, Roberts AE, Russell MW, Mital S, Newburger JW, Gaynor JW, Breitbart RE, Iossifov I, Ronemus M, Sanders SJ, Kaltman JR, Seidman JG, Brueckner M, Gelb BD, Goldmuntz E, Lifton RP, Seidman CE, Chung WK. De novo mutations in congenital heart disease with neurodevelopmental and other congenital anomalies. Science (New York, NY. 2015;350(6265):1262-6. Epub 2016/01/20. doi: 10.1126/science.aac9396. PubMed PMID: 26785492; PMCID: PMC4890146.
20. Ang YS, Rivas RN, Ribeiro AJ, Srivas R, Rivera J, Stone NR, Pratt K, Mohamed TM, Fu JD, Spencer CI, Tippens ND, Li M, Narasimha A, Radzinsky E, Moon-Grady AJ, Yu H, Pruitt BL, Snyder MP, Srivastava D. Disease model of GATA4 mutation reveals transcription factor cooperativity in human cardiogenesis. Cell. 2016;167(7):1734-49. doi: 10.1016/j.cell.2016.11.033. PubMed PMID: 27984724; PMCID: PMC5180611.
21. Zaidi S, Choi M, Wakimoto H, Ma L, Jiang J, Overton JD, Romano-Adesman A, Bjornson RD, Breitbart RE, Brown KK, Carriero NJ, Cheung YH, Deanfield J, DePalma S, Fakhro KA, Glessner J, Hakonarson H, Italia MJ, Kaltman JR, Kaski J, Kim R, Kline JK, Lee T, Leipzig J, Lopez A, Mane SM, Mitchell LE, Newburger JW, Parfenov M, Pe'er I, Porter G, Roberts AE, Sachidanandam R, Sanders SJ, Seiden HS, State MW, Subramanian S, Tikhonova IR, Wang W, Warburton D, White PS, Williams IA, Zhao H, Seidman JG, Brueckner M, Chung WK, Gelb BD, Goldmuntz E, Seidman CE, Lifton RP. De novo mutations in histone-modifying genes in congenital heart disease. Nature. 2013;498(7453):220-3. doi: 10.1038/nature12141. PubMed PMID: 23665959; PMCID: PMC3706629.
22. Luna-Zurita L, Stirnimann CU, Glatt S, Kaynak BL, Thomas S, Baudin F, Samee MA, He D, Small EM, Mileikovsky M, Nagy A, Holloway AK, Pollard KS, Muller CW, Bruneau BG. Complex Interdependence Regulates Heterotypic Transcription Factor Distribution and Coordinates Cardiogenesis. Cell. 2016;164(5):999-1014. doi: 10.1016/j.cell.2016.01.004. PubMed PMID: 26875865; PMCID: PMC4769693.
23. Garg V, Kathiriya IS, Barnes R, Schluterman MK, King IN, Butler CA, Rothrock CR, Eapen RS, Hirayama-Yamada K, Joo K, Matsuoka R, Cohen JC, Srivastava D. GATA4 mutations cause human congenital heart defects and reveal an interaction with TBX5. Nature. 2003;424(6947):443-7. PubMed PMID: 12845333.
24.  Gonzalez-Teran B, Pittman M, Felix F, Thomas R, Richmond-Buccola D, Huttenhain R, Choudhary K, Moroni E, Costa MW, Huang Y, Padmanabhan A, Alexanian M, Lee CY, Maven BEJ, Samse-Knapp K, Morton SU, McGregor M, Gifford CA, Seidman JG, Seidman CE, Gelb BD, Colombo G, Conklin BR, Black BL, Bruneau BG, Krogan NJ, Pollard KS, Srivastava D. Transcription factor protein interactomes reveal genetic determinants in heart disease. Cell. 2022;185(5):794-814 e30. Epub 2022/02/20. doi: 10.1016/j.cell.2022.01.021. PubMed PMID: 35182466.
25. Yu S, Li J, Ji G, Long Ng Z, Siew J, Ning Lo W, Ye Y, Yuan Chew Y, Chau Long Y, Zhang W, Guccione E, Han Loh Y, Jiang Z-H, Yang H, Wu Q. Npac is a co-factor of histone H3K36me3 and regulates transcriptional elongation in mouse embryonic stem cells. Genomics, Proteomics & Bioinformatics. 2021. doi: https://doi.org/10.1016/j.gpb.2020.08.004.