Lateral thinking in syndromic congenital cardiovascular disease

Comparative observations in different model organisms guide our understanding of how the human body develops, functions and manifests congenital anomalies. Numerous rare congenital diseases present with complex syndromic phenotypes that affect seemingly disconnected organs or cell types. Studying such disease phenotypes in model organisms can (1) reveal new, clinically meaningful insights about a potentially shared embryonic origin or molecular mechanism, and (2) inform about possible additional phenotypes in affected patients. However, decoding the underlying developmental and molecular causes of complex syndromic birth defects remains challenging when we lack a unifying concept to connect the observed phenotypes.

Heart formation provides a powerful example for the close developmental and genetic relationship of embryonic lineages with seemingly distant locations in the final body plan (Bulatovic et al., 2016; Diogo et al., 2015; Meilhac and Buckingham, 2018; Stainier et al., 1993). The heart develops from the lateral plate mesoderm (LPM) that initially emerges at the edge of the forming vertebrate embryo (Meilhac and Buckingham, 2018; Prummel et al., 2020; Saint-Jeannet et al., 1992; Stainier et al., 1995) (Fig. 1). Curiously, the emerging heart field is part of a wider progenitor cell population called the cardiopharyngeal field (CPF) that includes progenitor cells forming various muscle groups of the neck and the head (Figs 1 and 2) (Chan et al., 2016; Diogo et al., 2015; Gopalakrishnan et al., 2015; Lescroart et al., 2015, 2021; Stolfi et al., 2010). In addition to the heart, the LPM also gives rise to several smooth muscle lineages, ventral dermis, mesothelia and connective tissue of the limbs, including long bones (Lane and Smith, 1999; Selleck and Stern, 1991; Warga and Nüsslein-Volhard, 1999) (Fig. 1). The kidneys, traditionally described as intermediate mesoderm in birds and mammals, share clonal and gene expression relationships with other LPM-derived progenitors across vertebrates (Fig. 1) (Davidson et al., 2019; Mansour et al., 2022; Mattonet et al., 2022; Takasato and Little, 2015; Takasato et al., 2014). This complex developmental and genetic architecture renders the LPM sensitive to defects in its patterning and cell fate specification, with broad consequences for several of its descendant organs and tissues. Case in point, heart-hand syndromes manifest as cardiovascular and forelimb anomalies (Camarata et al., 2010; Yin et al., 2018, 2020). Although these phenotypes present in seemingly distant organs, the progenitors for both the heart and forelimb skeleton share a close developmental connection as they emerge from adjacent cell territories within the LPM, which also have a considerable number of co-expressed genes (Figs 1 and 2) (Mori and Bruneau, 2004; Silverman and Hurst, 1968; Tamari and Goodman, 1974; Yin et al., 2020).

Overlaying complex syndromic disease phenotypes with developmental information provides a conceptual framework to discover underlying mechanisms and possible comorbidities relevant to long-term patient health. Here, we discuss the LPM as a developmental concept to connect seemingly disparate phenotypes in human birth defects affecting the heart and beyond with known and unknown causative genes. We further present as-of-yet orphan syndromes missing causative gene associations that manifest as potential LPM diseases due to their affecting two or more LPM-derived tissues. We lastly outline how clinical, computational and developmental biology laboratories can productively collaborate towards revealing the causes of complex congenital syndromes affecting the heart and comorbidities.

LPM formation and patterning is regulated by deeply conserved principles across chordates (Box 1) (Prummel et al., 2020). Its emergence is closely tied to dorsal-ventral patterning of the vertebrate embryo, which is driven by secreted Bmp and Nodal signaling molecules that control downstream Smad transcription factors (Graff et al., 1996; Liu et al., 1996; Müller et al., 1999; Savage et al., 1996; Schier and Talbot, 2005; Sekelsky et al., 1995; Shi et al., 1997; Zhang et al., 1996). While Bmp regulates the dorsal-ventral arrangement of the body axis, Nodal activity additionally supports ectoderm versus mesendoderm patterning (Conlon et al., 1991; Davidson et al., 2015; Gritsman et al., 1999; Hayes et al., 2021; Kattman et al., 2011; Martyn et al., 2019; Schier and Talbot, 2005; Zhou et al., 1993). During gastrulation, the LPM emerges as ventral and lateral mesoderm territory (Garcia-Martinez and Schoenwolf, 1993; Gurdon, 1995; Lane and Smith, 1999; Mead et al., 1996; Selleck and Stern, 1991). In zebrafish, Smad proteins cooperate with Eomesa, Foxh1 and Mixl1 to induce the LPM, yet whether this combination is necessary and sufficient across vertebrates remains to be determined (Bjornson et al., 2005; Prummel et al., 2019; Slagle et al., 2011). Nonetheless, the basic concept of ventral LPM-inducing activity seems to be conserved across chordates and sets up the emerging LPM as a dedicated progenitor field for its various downstream cell fates (Figs 1 and 2, Box 1) (Prummel et al., 2019, 2020).

A hallmark of LPM architecture is the shared expression of transcription factor genes across seemingly disparate lineage progenitors; whether this shared expression reflects a clonal origin of co-expressing progenitors or co-opted gene regulation for similar regulatory purposes remains a major open question. In the emerging heart field, Tbx5 synergizes with Nkx2.5 (Nkx2-5), Gata4, Hand1/2 and other cardiac transcription factors to regulate heart tube precursors that develop from the LPM-embedded first and second heart fields (Greulich et al., 2011; Hiroi et al., 2001; Pi-Roig et al., 2014; Steimle and Moskowitz, 2017). The first heart field contains early-differentiating progenitors that form the primitive heart tube; descendant cells of the second heart field differentiate subsequently at both poles of the emerging heart tube (Dyer and Kirby, 2009; Laura et al., 2020; Rochais et al., 2009) (Figs 1 and 2). In the LPM immediately posterior to the emerging heart field, forelimb/pectoral fin bud initiation requires Tbx5 expression in the fin/limb-specifying field to induce Fgf10, which then activates Fgf8 in the overlying apical ectodermal ridge (Cunningham and Duester, 2015; Garrity et al., 2002; Minguillon et al., 2009, 2012; Moreau et al., 2019; Parrie et al., 2013; Rodriguez-Esteban et al., 1999; Takeuchi et al., 1999) (Fig. 2). Hand2 establishes the anterior-posterior forelimb axis through activating Shh and antagonizing Gli3 (Charité et al., 2000; Galli et al., 2010; Osterwalder et al., 2014). In the hindlimb-specifying field, Pitx1 activates Tbx4, which in turn induces Fgf10, achieving an analogous feedback loop with Fgf8 in the apical ectoderm ridge. Although forelimbs and hindlimbs display unique morphologies, Tbx4 can function in place of Tbx5 when expressed in the forelimb-specifying region (Don et al., 2016; Minguillon et al., 2005, 2012; Rodriguez-Esteban et al., 1999; Takeuchi et al., 1999) (Fig. 2). Despite their physical proximity and the fact that both the heart and the forelimb fields express Tbx5 and Hand2, it remains unclear whether these two lineages share a common progenitor or whether they arise from disparate progenitor fields (Ahn et al., 2002; Charité et al., 2000; Fernandez-Teran et al., 2000; Garrity et al., 2002; Rodriguez-Esteban et al., 1999; Steimle and Moskowitz, 2017; Yelon et al., 2000).

Expanding the developmental context of heart formation within the LPM, emerging heart progenitors are part of the CPF that expresses Tbx1 and Nkx2.5/7 and contributes to craniofacial muscles beyond the heart (Comai et al., 2019; Diogo et al., 2015; Lescroart et al., 2010; Meilhac et al., 2004; Nomaru et al., 2021; Xu et al., 2004; Zhou et al., 2008). Whether the entire CPF is one lineage field or a composite of the LPM and other anterior mesodermal lineages remains to be clarified (Figs 1 and 2). Nonetheless, the activities of Mesp1, Nkx2.5 and Tbx1/10, as well as their related gene-regulatory networks, indicate a homology between tunicate trunk ventral cells and the vertebrate CPF (Stolfi et al., 2010; Tolkin and Christiaen, 2012; Wang et al., 2013) (see also Box 1).

During development, other LPM lineages intersect with cardiac precursors in origin, gene expression or both. The endocardium emerges from anterior Scl/Tal1- and Etv2-expressing progenitors that have a distinct morphology and migration path compared to that of other cranial endothelial cells (Bussmann et al., 2007; DeRuiter et al., 1992; Harris and Black, 2010; Ivanovitch et al., 2017; Misfeldt et al., 2009; Proulx et al., 2010) (Fig. 1). In the trunk, Scl/Tal1-expressing progenitors develop into trunk endothelium and first erythrocyte progenitors, which form the initial circulatory system components (De La Garza et al., 2019; Garcia-Martinez and Schoenwolf, 1992; Mead et al., 1996; Pardanaud et al., 1996; Stainier et al., 1996) (Figs 1 and 3). In parallel, Pax2- and Pax8-positive LPM progenitors form the kidney, as initially discovered by the contribution of Pax2/8 perturbation to kidney disease (Bouchard et al., 2002; Eccles et al., 2004; Keller et al., 1994; McMahon, 2016; Poleev et al., 1992; Torres et al., 1995) (Fig. 1). Recent work in zebrafish has documented that pax2a-expressing LPM progenitors contribute to endothelial and erythroid fates besides forming kidney lineages (Mattonet et al., 2022). The kidney precursors initially reside laterally to endothelial progenitors at least in the teleost LPM, contrary to in mammals, in which they reside between somites and endothelial precursors as ‘intermediate mesoderm’ (McMahon, 2016; Mullins et al., 1996; Perens et al., 2016; Takasato and Little, 2015; Terashima et al., 2014). These data suggest that kidney progenitors can be regarded as an LPM lineage. Lastly, the lateralmost LPM forms the Hand2-expressing mesothelium progenitors that surround internal organs and the body cavities, as well as different smooth muscle types (Ariza et al., 2016; Gays et al., 2017; Prummel et al., 2022; Shin et al., 2009). Work in chicken has documented interactions and stepwise separation between smooth muscles, blood and endothelial progenitors as driven by Hand2, Wnt, Bmp and Notch (Shin et al., 2009). Further underlining non-autonomous cross-talk between the stripes, loss of hand2 in zebrafish perturbs the balance between endothelial progenitors and kidney formation (Perens et al., 2016, 2021). Subsequent blood flow has a major impact on tissue remodeling and even the emergence of key cell fates, as, after the onset of circulation, mechanical forces influence the budding of Runx1-expressing definitive hematopoietic stem cell progenitors from the ventral wall of the dorsal aorta (Boisset et al., 2010; Downing et al., 1993; Kissa and Herbomel, 2010; Lancino et al., 2018; Müller et al., 1994; North et al., 2009; Traver et al., 2003).

Taken together, the LPM harbors progenitor populations for multiple organs, providing a model to link perturbations in shared precursors to causation of congenital defects affecting two or more LPM-derived organs (Gays et al., 2017; Nishimoto and Logan, 2016; Prummel et al., 2019, 2020, 2022; Tanaka, 2016). Yet, the LPM is not of single clonal origin, and major organ precursors segregate as early as during gastrulation. Consequently, alternative mechanisms contributing to comorbidities could involve perturbed organ progenitor positioning, signaling interactions and shared gene expression across LPM domains. For this Review, we consider LPM-affecting diseases as those that affect two or more organs or cell types of LPM origin and/or that share LPM-centered gene expression profiles. The developmental and evolutionary peculiarities of the LPM underline how mutations in genes affecting common progenitors or shared regulatory mechanisms could trigger syndromic defects with complex comorbidities in seemingly unrelated organ systems.

Joint heart and upper limb anomalies occur in several human congenital syndromes (Fig. 2). Affected patients share common clinical presentations that vary in their phenotypic expressivity, described under the umbrella term heart-hand or upper limb-cardiovascular syndromes. Although the genetic bases for several of these diseases have been uncovered (Table 1,Box 2), many remain without an assigned genetic cause (Table 2). These syndromes often present with additional respiratory, endocrine, metabolic, hematopoietic or neurologic symptoms for which clinical urgency can dominate over the heart and hand phenotypes, including symptoms that arise as complications from the primary cardiac defects (Metra et al., 2011; Tamari and Goodman, 1974) (Tables 1 and 2).

Holt-Oram syndrome is the prototypic heart-hand syndrome. It manifests as autosomal-dominant upper (anterior) limb and shoulder girdle abnormalities with a spectrum of congenital heart defects [Online Mendelian Inheritance in Man (OMIM) #142900] (Hurst et al., 1991; Temtamy, 1985). The most common presentation of Holt-Oram syndrome is triphalangeal thumbs and cardiac atrial septal defects, although the severities of limb and heart defects vary widely. In 1997, two separate groups confirmed that mutations in TBX5 were causative for Holt-Oram syndrome, resulting in predicted instances of haploinsufficiency (Basson et al., 1997; Yi et al., 1997). Ever since linking TBX5 mutations to the syndrome, a growing number of variant TBX5 alleles have been uncovered in affected patients. Loss-of-function mutations or TBX5-spanning large deletions cause the most severe phenotypes (Basson et al., 1997; Gruenauer-Kloevekorn and Froster, 2003) (Borozdin et al., 2006). Missense variants may cause distinct phenotypes depending on the position of the mutation (Yang et al., 2000). The human syndromic phenotypes are recapitulated in recessive Tbx5 loss-of-function mouse models (Bruneau et al., 2001; Rallis et al., 2003), and zebrafish tbx5a mutants and morphants have co-occurring heart and pectoral fin anomalies (Fig. 2, Table 3) (Ahn et al., 2002; Garrity et al., 2002). Select putative gain-of-function TBX5 alleles cause atypical Holt-Oram syndrome with mild skeletal defects and paroxysmal atrial fibrillation (Postma et al., 2008). A 48 kb duplication encompassing TBX5 exons 2 through 9 is associated with mild limb defects and non-septal cardiac defects (Patel et al., 2012).

Loss of the Tbx5 paralog Tbx4 in mice (Naiche and Papaioannou, 2003; Rodriguez-Esteban et al., 1999), and tbx4 in zebrafish (Don et al., 2016), results in hindlimb or pelvic fin absence (Fig. 2, Table 3), although mutant mice do not survive embryonic stages due to fatal lung failure (Arora et al., 2012) (Table 3). This combination of developmental phenotypes links to Tbx4 expression in the hindlimb bud, as well as in the future lung mesothelium that is likely to be of LPM origin. Curiously, dominant TBX4 perturbations in humans largely cause cardiopulmonary defects, including pulmonary arterial hypertension and patent ductus arteriosus, in which the pulmonary circuit fails to remodel into a functional unit before birth (D'Alto and Mahadevan, 2012; Haarman et al., 2019; Haarman et al., 2020; Karolak et al., 2023; Welch and Chung, 2020; Yarboro et al., 2022). Patients with heterozygous TBX4 variants have mild limb anomalies affecting the patella or joints (Haarman et al., 2019, 2020; Karolak et al., 2023). Notably, the human phenotypes affecting hindlimbs are distinct from those arising in mouse loss-of-function models, which seem to stem from possible species-specific partial compensation of Tbx4 by Pitx1 and from Isl1/Ldb function in murine hindlimb development (Duboc et al., 2021; Itou et al., 2012; Narkis et al., 2012) (Fig. 2).

Mutations in HAND1 and HAND2 have been associated with cardiac anomalies such as tetralogy of Fallot, hypoplastic left ventricle and others (Lu et al., 2016; Reamon-Buettner et al., 2008, 2009). These basic helix-loop-helix (bHLH) transcription factors have widespread expression in the early LPM and in the subsequent cardiac, pharyngeal arch, limb and mesothelial progenitors (Charité et al., 2000; Gibert et al., 2006; Prummel et al., 2022). Hand1 and Hand2 have unique as well as redundant functions in cardiac and limb patterning in mice, controlling epithelium formation, extracellular matrix composition, cell migration and the expression of downstream transcription factor genes (Barnes and Firulli, 2009; Barnes et al., 2010; Conway et al., 2010; Firulli et al., 1998, 2010). Curiously, zebrafish only retained hand2 in their genome, and hand2 mutants develop cardiac and branchial arch defects, hypoplastic pectoral fins and reduced mesothelium that results in herniation (Garavito-Aguilar et al., 2010; Hu et al., 2011; Perens et al., 2016; Prummel et al., 2022; Yelon and Stainier, 2005; Yelon et al., 2000). Compared to genes encoding other cardiac transcription factors, HAND1/2 seem to be less frequently mutated in congenital disease, possibly due to their partial redundancy and potential to compensate for one another upon perturbation (Barnes et al., 2011; Charité et al., 2000; McFadden et al., 2005). Notably, partial trisomy of distal chromosome 4q (4q+; no OMIM reference), which contains the HAND2 locus, is associated with cardiac, digit, kidney and craniofacial defects, indicating broader LPM perturbation during development (Battaglia et al., 2005; Lundin et al., 2002; Lurie, 2005). Elegant studies have linked mouse Hand2 dosage to developmental phenotypes that recapitulate major aspects of human 4q+ syndrome (Table 3) (Tamura et al., 2013). Together, these shared heart and limb transcription factors provide clear examples for LPM-based comorbidities.

Heart-hand syndromes feature obvious LPM-linked comorbidities, yet additional heart-affecting syndromes with complex comorbidities seem also to be influenced by shared LPM progenitors and/or shared genetic programs (Figs 1 and 2, Tables 1 and 2, Box 2). DiGeorge syndrome (OMIM #188400) is caused by autosomal-dominant 22q11.2 deletions (Wilson et al., 1993). Varying widely between patients, the most consistent symptoms of DiGeorge syndrome are hypocalcemia from parathyroid hypoplasia, thymus hypoplasia and T-cell deficiencies, submucosal cleft palate and cardiac outflow tract defects. Patients with 22q11.2 deletion can also present with variable craniofacial defects, encompassing conditions known as Shprintzen (velocardiofacial) syndrome or Takao syndrome, associated with dysphagia and a broad range of heart defects (Karpinski et al., 2022; Motahari et al., 2019; Shprintzen, 1994; Yagi et al., 2003; Yamagishi et al., 2016). Nonetheless, 22q11.2 harbors several genes and as-of-yet incompletely charted regulatory elements (Fernández et al., 2005; Saitta et al., 2004), and no single gene has been linked to the full phenotype spectrum. Consequently, DiGeorge syndrome could comprise multiple syndromes caused by individual genes in the 22q11.2 locus (Burn, 1999; Wilson et al., 1993).

The major phenotypes of DiGeorge syndrome can potentially be explained by TBX1 haploinsufficiency, with other symptoms attributed to the cumulative effects from loss of other 22q11.2 genes (Jerome and Papaioannou, 2001; Lindsay et al., 2001; Paylor et al., 2006; Yagi et al., 2003). Notably, patients with point mutations in TBX1 display characteristic phenotypes of 22q11.2 deletion patients (Yagi et al., 2003), although they do not develop the full spectrum of DiGeorge syndrome. Mice hemizygous for the highly conserved 22q11.2 region display mild cardiac outflow tract defects, as phenocopied in heterozygous Tbx1 mutant mice (Lindsay et al., 1999, 2001). Homozygous Tbx1 mutant mice (Jerome and Papaioannou, 2001; Merscher et al., 2001) and tbx1 mutant zebrafish (van gogh, vgo) (Piotrowski et al., 2003) recapitulate the major structural defects observed in patients with severe DiGeorge syndrome, including heart, thymus and craniofacial defects (Table 3), but do not recapitulate all aspects of human DiGeorge syndrome. Notably, the expression pattern of Tbx1 is not single lineage or LPM specific, but rather a mix of different anterior/rostral cell types, including the LPM, neural crest and additional head mesoderm, that awaits concise definition. Consequently, untangling individual contributions of Tbx1 function in individual cell types, as well as in cell-cell crosstalk, in head and heart structures remains an ongoing challenge in the field.

In addition, patients with atypical 22q11.2 microdeletions that leave TBX1 unaffected still show craniofacial and cardiac defects, as well as velopharyngeal insufficiency (Shi and Wang, 2018; Verhagen et al., 2012). One possibility is that these patients lack cis-acting regulatory elements required for modulating TBX1 expression; alternatively, deletion of CRKL, another 22q11.2 gene that has also been linked to craniofacial and heart anomalies in mouse models (Guris et al., 2001, 2006; Motahari et al., 2019; Verhagen et al., 2012), triggers the craniofacial and cardiac defects of DiGeorge syndrome. Further, modifying alleles outside the 22q11.2 microdeletion may synergize with TBX1 to create unique, atypical phenotypic manifestations (Aggarwal and Morrow, 2008). Although DiGeorge syndrome is complex and incompletely understood, it provides a powerful example for the complex interplay of cell lineages associated with heart development across the LPM and beyond.

Several predominantly hematopoietic disorders feature seemingly unrelated structural comorbidities (Tables 1 and 2, Box 2). A classic example is thrombocytopenia-absent radius (TAR) syndrome (OMIM #274000), a rare congenital disease affecting less than 1:100,000 live births. In its most selective manifestation, TAR syndrome patients feature a severely reduced number of platelet-forming megakaryocytes (<30×10⁹ platelets/l of blood), which correlates with hypo-megakaryocytic thrombocytopenia, and absent radius bones (Bonsi et al., 2009; Greenhalgh, 2002; Hall et al., 1969; Klopocki et al., 2007; Petit and Boussion, 2022; Thompson et al., 2001). The blood coagulation defect, which may be transient, is seen in all cases, and, in over 90% of patients, manifests as symptomatic within the first 4 months of life (Hall et al., 1969). Curiously, the coagulation defect resolves with increasing age, but it needs to be determined whether this is due to the migration of hematopoietic progenitors from the fetal liver to the fetal bone marrow. The non-hematopoietic syndromic phenotypes have variable expressivity, ranging from mild joint issues to absent radii, while several TAR syndrome patients also present with variable heart and kidney anomalies, intolerance to cow's milk presenting as persistent diarrhoea and failure to thrive, and can experience exacerbation of thrombocytopenia (Whitfield and Barr, 1976) later in life. More rarely, patients present with mild craniofacial anomalies. Genetically, TAR syndrome associates with a heterozygous 1q21.1 microdeletion removing at least ten genes, including the gene encoding the exon junction complex factor RBM8A/Y14 that ubiquitously contributes to the regulation of mRNA splicing, nonsense-mediated decay and mRNA localization (Gehring et al., 2005; Klopocki et al., 2007; Palacios et al., 2004). Hypomorphic perturbation of RBM8A is assumed to cause TAR syndrome, as the 1q21.1 microdeletion, combined with putative hypomorphic RBM8A/Y14 alleles, has been found in most sequenced patients (Albers et al., 2012, 2013). However, the mechanistic link between hypomorphic RBM8A and the peculiar phenotype combination of TAR syndrome remains to be elucidated. Notably, TAR syndrome affects predominantly LPM-derived cell lineages in the blood, forelimb skeleton, heart and kidneys. Considering TAR syndrome an LPM-associated pathology might provide a framework to investigate its mechanistic underpinnings (Kocere et al., 2023 preprint) (Figs 1 and 3).

Another ultra-rare syndrome, autosomal-dominant radioulnar synostosis with amegakaryocytic thrombocytopenia (RUSAT1; OMIM #605432), features severe coagulation defects with restricted pronation-supination of the forearm (Thompson and Nguyen, 2000; Thompson et al., 2001). Two RUSAT1-affected families have been reported to carry mutations in HOXA11, a Hox gene involved in regional bone, muscle and tendon patterning (Hawkins et al., 2021; Swinehart et al., 2013); how HOXA11 perturbation causes the hematopoietic defects remains unknown. By contrast, other RUSAT1 cases have been associated with mutations in the MECOM locus that spans the MDS1 and EVI1 genes (Niihori et al., 2015); the EVI1 transcription factor broadly controls cell cycle genes and is dynamically expressed in various tissues, including the forming limbs (Bard-Chapeau et al., 2012; Celá et al., 2013; Li et al., 2022; Nucifora, 1997). Notably, Evi1 is an upstream regulator of hematopoietic transcription factors, including the megakaryocyte-erythrocyte regulator Gfi1 (Ayoub et al., 2018; Chiba et al., 2020; Konantz et al., 2016).

In addition to TAR and RUSAT1, other rare hematopoietic disorders with co-occurring limb anomalies have been reported. Individuals with Fanconi anemia (OMIM #609054) can present with digit anomalies, such as unilateral radial anomalies that include duplicated thumbs (Bhandari et al., 2022; Tischkowitz and Hodgson, 2003). Other examples include WT limb-blood syndrome (OMIM #194350) (Gonzalez et al., 1977), Diamond-Blackfan anemia 3 (OMIM #610629) (Gazda et al., 2006; Landowski et al., 2013) and congenital hemidysplasia with ichthyosiform erythroderma and limb defects (OMIM #308050) (Happle et al., 1996). Notably, TAR, Fanconi anemia and Diamond-Blackfan anemia are seemingly caused by mutations in ubiquitous regulatory factors, suggesting an early developmental susceptibility to their activity levels that remains to be elucidated. Nonetheless, whether the root of these syndromic phenotypes is an earlier developmental LPM phenotype remains unclear, yet provides a working hypothesis to study their embryonic origins.

In the examples above, we outlined the rationale for considering a shared LPM origin for syndromic congenital heart and cardiovascular diseases as part of complex syndromes with comorbidities affecting LPM-associated structures. A conceptual framework to connect syndromic phenotypes to a joint developmental basis extends to linking structural birth defects with fetal exposure to environmental factors such as teratogenic compounds (Box 3). Ultimately, combining syndromes with similar phenotypes via shared molecular pathways or developmental processes may shed light on orphan diseases with as-of-yet unknown causes and, at times, no clear all-encompassing diagnoses. Orphan diseases (Table 2, Box 2) remain understudied, as they are exceedingly rare, at times affecting single families or, alternatively, may have complex polygenic origins (Chen and Altman, 2017).

Holt-Oram syndrome and Duane-radial ray syndrome, which we discuss in Box 3, present similar phenotypes due to genetic perturbations in a shared molecular pathway (Harvey and Logan, 2006): Tbx5 acts directly upstream of Sall4 during heart and appendage development in mice (Koshiba-Takeuchi et al., 2006) and in zebrafish (Harvey and Logan, 2006). Other heart-hand syndromes may result from developmental perturbation of the same or interacting pathways and the progenitor cell populations they control. Heart-hand syndrome type 3 (also known as Spanish type; OMIM #140450) features the characteristic upper limb deformities and congenital heart defects. Affecting three individuals of a single Spanish family, no causative gene has been assigned to the condition to date (Ruiz de la Fuente and Prieto, 1980). Additional orphan diseases that primarily present as congenital heart and limb defects include patent ductus arteriosus and bicuspid aortic valve with hand anomalies (OMIM #604381) (Gelb et al., 1999) and long-thumb brachydactyly syndrome (OMIM #112430) (Gelb et al., 1999; Hollister and Hollister, 1981) (Table 2). Despite the overlapping phenotypes, a shared progenitor cell or genetic origin has not been established in these orphan diseases. Therefore, investigating potential perturbations of TBX5, SALL4 and other cardiac/forelimb transcription factors in the LPM-derived heart and limb mesoderm warrants further consideration (Fan et al., 2009; Harvey and Logan, 2006; Koshiba-Takeuchi et al., 2006).

Numerous orphan diseases outside the heart-hand syndromes feature symptoms that affect LPM-derived organs or tissues. Examples include the previously mentioned WT limb-blood syndrome, with patients presenting endothelial, limb and craniofacial defects (Gonzalez et al., 1977), and thoracoabdominal syndrome (OMIM #313850), which affects craniofacial and neck structures, the heart, blood and kidneys (Carmi et al., 1990; Gonzalez et al., 1977). These clearly complex phenotypes could result from early perturbations in LPM patterning yet could also result from more pleiotropic developmental issues. Considering the joint LPM origin of various affected organs as the developmental stencil to narrow down possible connecting causes could advance the investigation of underlying mutations for these orphan diseases. Although orphan diseases are exceedingly rare, collectively, affected patients make up a large percentage of genetic diseases and deserve diagnostic and therapeutic investigation (Orphanet).

As outlined above, a LPM disease-associated gene may be expressed in multiple organ lineages. This is based on lineage connections or co-option of gene regulatory mechanisms during organ development in disparate cell types (McQueen and Rebeiz, 2020; Monteiro, 2012; Olson, 2006; Shubin et al., 2009). Throughout this Review, we have highlighted how model organisms have fundamentally contributed to our understanding of developmental birth defects, gene ontology and biological mechanisms (Fig. 3, Table 3). For example, the notion that head and neck muscles, together with the heart, are affected in DiGeorge syndrome indicates that they emerge from the Tbx1-expressing CPF (Diogo et al., 2015; Greulich et al., 2011; Lescroart et al., 2015; Swedlund and Lescroart, 2019). Lineage tracing experiments have established firm evidence for shared clonal origins of heart and head muscle lineages (Diogo et al., 2015; Lescroart et al., 2010, 2015; Mahadevan et al., 2004; Meilhac et al., 2004). Consequently, mutations in undifferentiated early CPF progenitors, such as mutations in TBX1, can potentially influence all downstream lineages. However, how Tbx1 is regulated in the CPF and whether its downstream expression in different muscle lineages uses shared or divergent mechanisms in each lineage remains to be clarified.

TBX5, as the causative gene in Holt-Oram syndrome, provides another potent example for joint expression in different LPM descendants. The recessive zebrafish mutant heartstrings carries a mutant tbx5a allele (Garrity et al., 2002). heartstrings mutants, as well as tbx5a morphants, display heart and pectoral fin defects, recapitulating the Tbx5 mutant mouse and human Holt-Oram phenotypes (Ahn et al., 2002; Garrity et al., 2002) (Table 3). Studies in zebrafish tbx5a mutants and morphants have provided new insights into how Tbx5 contributes to left-right laterality, cardiac looping and myocardial conductivity (Boyle Anderson and Ho, 2018; Mao et al., 2015; Mao et al., 2021; Mosimann et al., 2015; Parrie et al., 2013). These insights provide basic concepts of the cardiac manifestations that could arise in patients carrying TBX5 gene variants. Mice heterozygous for a Tbx5 deletion (Tbx5^del/+) recapitulate the hallmarks of the human syndrome with congenital forelimb and heart defects, whereas mice with homozygous deletions (Tbx5^del/del) are embryonic lethal (Bruneau et al., 2001; Mori and Bruneau, 2004; Zhou et al., 2005). Mechanistic work revealed that Tbx5 and Nk2.5 synergistically interact to promote the expression of the gap junction protein Cx40 (Gja5), reduction of which is assumed to contribute to the cardiac anomalies in Holt-Oram syndrome (Bruneau et al., 2001; Zhou et al., 2005). In the limbs, absence of Tbx5 abrogates the Fgf10-Fgf8 signaling loop, which is critical for limb bud outgrowth (Agarwal et al., 2003; Arora et al., 2012; Minguillon et al., 2012). However, as we discussed earlier, whether Tbx5 activity in both cardiac and forelimb progenitors reflects a joint clonal LPM origin of these organ precursors, or whether Tbx5 expression and function are co-opted by forelimb precursors in a separate gene-regulatory event, remains unresolved. The mouse Tbx5 locus harbors several regulatory elements, including a dedicated forelimb enhancer (Cunningham et al., 2018; Minguillon et al., 2012) and multiple cardiac enhancers (Richter et al., 2020; Smemo et al., 2012); more regulatory elements (and their upstream control) await discovery (Bickley and Logan, 2014; Bruneau et al., 1999). Similarly, zebrafish hand2 and mouse Hand2 mutants have provided key insights into the developmental mechanisms in heart, limbs and additional LPM organs influenced by this deeply conserved transcription factor (Barnes et al., 2011; Firulli et al., 1998; George et al., 2022; Hashimoto et al., 2019; Keegan et al., 2005; Laurent et al., 2017; McFadden et al., 2005; Osterwalder et al., 2014; Perens et al., 2016; Prummel et al., 2022; Vincentz et al., 2021; Yamagishi et al., 2000; Yelon and Stainier, 2005). The upstream regulation of Hand2 in mouse and zebrafish, beyond the pharyngeal arches (Charité et al., 2001; Iklé et al., 2012), remains vastly uncharacterized, as regulatory elements for broad early LPM expression or for heart and limb/pectoral fin activity have not been identified to date.

Multiple lineages may also develop in an interconnected manner, whereby perturbation of one may have consequences for another. Beyond Tbx5, the heart and the pectoral fin/forelimb also provide an example for inter-organ signaling interactions during development. In the absence of retinoic acid signaling in the prospective pectoral fin/forelimb field, the heart field expands in cell number (Cunningham and Duester, 2015; Duong et al., 2021; Gibert et al., 2006; Zhao et al., 2009). This interaction might reflect the evolutionary relationship between heart and forelimb, with models proposing that delimiting the heart field was a prerequisite to paired appendage evolution (Tanaka, 2013, 2016). Beyond the developmental impact of perturbing retinoic acid signaling and its upstream regulation, avoiding embryo or fetal exposure to retinoic acid analogs remains paramount (Box 3) (Cunningham and Duester, 2015; Duester, 2008; Lammer et al., 2010; Mondal et al., 2017; Niederreither et al., 2001). In the forming trunk, the kidney, hematopoietic and endothelial progenitors form in close proximity within the posterior LPM and are influenced by signaling interactions that include the lateralmost mesothelial progenitors (Fig. 1) (Chapman, 2019; Cheng et al., 2015; Dressler, 2009; Fleming et al., 2013; Perens et al., 2016; Shin et al., 2009; Wingert and Davidson, 2008). Consequently, perturbations of these developmental interactions could lead to multi-organ defects that manifest as complex syndromic phenotypes.

While work across animal models has discovered causes and consequences of congenital heart anomalies and their comorbidities, animal models also emphasize inter-species differences in genetic susceptibility to structural heart disorders. In humans, the majority of genetic congenital heart disease patients carry heterozygous mutations (Richards and Garg, 2010; Andersen et al., 2014; Nees and Chung, 2020; Pierpont et al., 2018; Williams et al., 2019; Yasuhara and Garg, 2021). By contrast, mouse and zebrafish models of these disorders frequently require bi-allelic loss of function of the orthologous gene(s) to reveal a phenotype, as documented for mutations in Tbx5, Hand2, Sall4, Gata family members and others (Bruneau et al., 2001; Garrity et al., 2002; Granados-Riveron et al., 2012; Harvey and Logan, 2006; Yelon et al., 2000) (Fig. 3, Table 3). Beyond complete loss of function, numerous mutations identified in patients result in amino acid substitutions or splicing and regulatory alterations with more subtle impact on gene function. Therefore, recapitulating the exact patient-derived mutation may be necessary to appropriately model a particular syndrome or patient cohort (Albadri et al., 2017; Kimura et al., 2014; Prykhozhij et al., 2018). Rather than reducing the validity of animal models, these nuances reveal the critical roles of gene dosage or of compensatory genetic mechanisms that await further elucidation (Rossi et al., 2015; Sztal and Stainier, 2020).

Taken together, the LPM provides a conceptual framework to connect seemingly disperse disease phenotypes, either by common clonal origin, close lineage connection during development or shared regulatory gene expression among progenitors. Considering developmental and molecular connections between cells and organs has vast potential in directing diagnostics for possible and plausible comorbidities, avoiding potential incomplete diagnoses, and facilitating the design of complex treatment strategies.

Next-generation sequencing has advanced ground-breaking insights into mutations and gene variants associated with developmental disorders and invigorated association studies for disease susceptibility loci (Austin-Tse et al., 2022; Lightbody et al., 2019; Morash et al., 2018; Yang and Kim, 2018). Several publicly available databases curate information on disease genetics, such as the Monarch Initiative and OMIM (McMurry et al., 2016; Shefchek et al., 2020; Motulsky, 1993) (Fig. 3). Although they are exceedingly powerful in aggregating diverse data from clinical and functional studies, any developmental or mechanistic connections for LPM-associated diseases, such as joint LPM origin of affected organs, remains to be determined by the database user. Further, drawing functional or causative conclusions from individual mutations or genetic variation is heavily influenced by insights gained in model organism studies (Alghamdi et al., 2022), in particular, assessing the phenotypic impact of individual gene perturbations on structural birth defects. As nowhere near all orthologs and paralogs of human genes have been functionally tested in developmental model organisms (Table 3, Box 2), the community still faces a gap in gene ontology knowledge that restricts the sequence-based prediction of disease traits. CRISPR-mediated mutagenesis for basic functional testing of candidate genes is possible by microinjection in zebrafish and Xenopus: these CRISPR-induced somatic mutants, so-called crispants (Burger et al., 2016), can recapitulate major aspects of the candidate gene's function. Nonetheless, although crispants provide a powerful discovery tool for first functional insights before mouse knockout studies, they require stringent controls and germline mutant validation in species that enable this workflow (Burger et al., 2016; Naert and Vleminckx, 2018; Shah et al., 2015; Trubiroha et al., 2018; Wu et al., 2018) (Fig. 3).

The broad lack of interfaces linking patient genetic data to functional in vivo follow-up, such as missing cross-species integration with animal models in individual databases, remains a key gap. The Model organism Aggregated Resources for Rare Variant ExpLoration (MARRVEL) website curates available model organism resources for human disease-relevant genes, providing a powerful resource for overcoming this gap (Wang et al., 2017a). Additional information, such as genomic coordinates, or interconnection with databases mapping non-coding features of the genome, could significantly expand patient-derived sequence data. Further, the varied versions of the human genome assembly used across individual databases renders comparisons challenging. Although individual genes of interest can be cross-referenced manually, the systematic generation of candidate gene lists to validate in model organisms remains labor intensive and not broadly accessible. Augmenting databases towards driving experimental validation of mutations, variants and causal gene discovery overall would provide fertile grounds for collaborations between clinicians, bioinformaticians and developmental biologists to accelerate research on rare LPM disorders, as we started to define throughout this Review, and beyond.

Ultimately, a major hurdle to diagnose and fully care for syndromic congenital anomalies remains the sheer complexity of human phenotypes such as heart-hand disorders. Patients with a specific organ-affecting disease often require highly specialized care. Performing a full diagnostic phenotype assessment of potentially dozens of symptoms in multiple organs is frequently not possible for reasons involving time, finances, technology or lacking consent. Nonetheless, providing insights into the underlying biology of linked comorbidities of syndromic congenital disease has the potential to improve long-term care and consequently the quality of life of affected patients. Predicting or anticipating care-relevant congenital phenotypes that can co-occur based on embryonic connections gleaned from developmental biology is only the starting point towards this goal.

This article is part of a collection ‘Moving Heart Failure to Heart Success: Mechanisms, Regeneration & Therapy’, which was launched in a dedicated Special Issue guest edited by Jeroen Bakkers, Milena Bellin and Ravi Karra. See related articles in this collection at https://journals.biologists.com/collection/8169/Moving-Heart-Failure-to-Heart-Success.

We thank the Mosimann laboratory and the members of the TBX4Life initiative for constructive input on the manuscript.