Zhao H et al. (2026), HER2 deficiency causes a developmental disorder...

XB-ART-61828

J Clin Invest 2026 Apr 30; doi: 10.1172/JCI199043.

Show Gene links Show Anatomy links

HER2 deficiency causes a developmental disorder with growth retardation and craniofacial malformations.

Zhao H, Wang P, Jiao Y, Huang H, Yu M, He Q, Pan C, Guo S, Huang W, Jia Y, Kong Q, Peng H, Han Y, Hou Y, Ren Z, Tao Y, Huang F, Jiang H, Sun S, Dong Y, Lin J, Yin C, Zhu X, Chen F, Ding Y.

???displayArticle.abstract???
The human epidermal growth factor receptor 2 (HER2) is a major therapeutic target in cancer. While the oncogenic effects of HER2 hyperactivation are well-characterized, the biological consequences of its deficiency remain poorly defined. Here, through exome sequencing analyses of a cohort of 720 families affected by isolated or syndromic orofacial clefts, we unexpectedly identified five distinct rare germline HER2 variants in five unrelated families with growth deficits, orofacial clefts, and other craniofacial, skeletal, and auditory anomalies. In Xenopus embryos, these variants failed to recapitulate the developmental effects of wild-type HER2. In cultured cells, they disrupted HER2 protein stability, membrane localization, or site-specific phosphorylation, resulting in diminished ERK signaling. Strikingly, knock-in mice expressing a patient-derived HER2 variant and mice maternally exposed to Tucatinib, a recently approved anti-HER2 drug, both replicated patient phenotypes: retarded growth and diverse craniofacial abnormalities, including ocular dysgenesis, short jaws, and cleft palate. Collectively, our findings define a developmental disorder that we designate GRACE syndrome (Growth Retardation and Craniofacial Malformations Caused by HER2 Deficiency), establish HER2's essential role in human growth and craniofacial morphogenesis, and reveal that HER2-targeted therapies during pregnancy can induce craniofacial defects and lifelong growth impairment in fetuses. 5.

???displayArticle.pubmedLink??? 42060361
???displayArticle.pmcLink??? PMC13262720
???displayArticle.link??? J Clin Invest

???attribute.lit??? ???displayArticles.show???

	Figure 1. Heterozygous missense HER2 variants cause growth impairment and craniofacial malformations.(A) Pedigrees of 5 families. Heterozygous HER2 variants and genotypes of available family members are indicated. Squares denote male family members, circles female family members, open symbols unaffected, and black symbols affected. mut, mutant. (B) Images of affected family members. Arrowheads mark cleft lip (top row), and arrows denote cleft palate (bottom row). For patient F3-II:2, lip and palate images were obtained at the ages of 1 year and 5 years, respectively. (C) Genomic and protein locations of HER2 variants (left) and the topological feature of HER2 in the plasma membrane (right). Variants are represented with stars. HER2 is comprised of an extracellular domain (ECD) with 4 subdomains (I, II, III, and IV), a single transmembrane domain (TM), and a cytoplasmic tyrosine kinase domain (TKD) followed by a C-terminal tail. (D) Protein sequence alignment of HER2 across various vertebrate species reveals the conservation of amino acids affected by the variants. Ser1150 is less conserved, while all other residues are highly conserved. (E) Structural modeling of the p.A87T variant. Ala87 is located on the C-terminal side of the parallel β-sheet in the extracellular domain I. It is buried within a hydrophobic cavity formed by Leu33, Leu85, Leu117, Thr63, and His88. Substituting alanine with threonine introduces a larger side chain containing a polar hydroxyl group, which may cause steric clashes (indicated by red disks) and disrupt the local hydrophobic environment. (F) Structural modeling of the p.R970W variant. Arg970 is positioned on the N-terminal side of Helix-I in the tyrosine kinase domain. Its positively charged side chain forms a salt bridge with Asp838 (indicated by yellow dashed lines). Substitution with tryptophan, which lacks a charged side chain, is predicted to disrupt this electrostatic interaction, potentially destabilizing the kinase domain.
	Figure 2. Patient variants impair HER2/ERK signaling.(A) HER2 variants are less effective than WT HER2 in inducing the formation of supernumerary tails. Arrowheads indicate extra tails. Lateral view, anterior to right. Scale bar: 1 mm. Xenopus images were reproduced from Development (35). (B) WT, but not mutant HER2 induces ectopic xbra expression in the ectoderm. Arrowheads indicate ectopic xbra signals in the ectoderm. Lateral view, animal pole to top. Scale bar: 0.5 mm. Xenopus images were reproduced from Development (35). (C) HER2 variants are less efficient than WT HER2 in rescuing impaired twist1 expression in her2-depleted embryos. twist1 expression was detected in the 4 streams of CNCCs migrating towards the pharyngeal arches (marked by white dashed circles): the mandibular (no. 1), hyoid (no. 2), anterior branchial (no. 3), and posterior branchial (no. 4) streams. Lateral view, anterior to left. Scale bar: 0.5 mm. (D) Patient variants reduce HER2 and ERK activities. HER2 knockout HEK293T cells were transfected, stimulated with EGF, and harvested for immunoblot analysis. (E) p.A87T, p.G603S, and p.R970W variants disrupt HER2 membrane localization. Hela cells were transfected and processed for immunofluorescent analysis. Scale bar: 10 μm. (F) p.A87T HER2 shows shortened half life. HER2 knockout HEK293T cells were transfected, treated with cycloheximide (CHX), and harvested for immunoblot analysis. The quantified HER2 levels normalized to β-ACTIN are at the bottom. (G) Bafilomycin A1 (BFA1) partially normalizes p.A87T HER2 expression level. HER2 knockout HEK293T cells were transfected, treated with BFA1, and harvested for immunoblot analysis. (H) p.S1151W variant abolishes HER2 phosphorylation at Ser1151. HER2 knockout HEK293T cells were transfected, stimulated with EGF, and harvested for immunoprecipitation (IP) and immunoblot analysis. (I) p.T1242M HER2 displays diminished threonine phosphorylation. The experiment was performed as in panel H, except that the p.T1242M variant was used. (J) Summary of the pathogenic mechanisms of HER2 variants.
	Figure 3. HER2 is expressed in growth plate proliferating chondrocytes and craniofacial tissues in mice.(A) Schematic diagram of mouse growth plate. RZ, resting zone; PZ, proliferating zone; HZ, hypertrophic zone. (B and C) Immunostaining of HER2 (red) in the P21 mouse proximal tibial growth plate shows its predominant expression in proliferating chondrocytes. The boxed region in panel B is shown at higher magnification in panel C. (D–F) Schematic diagrams of coronal sections of mouse craniofacial region at indicated stages (36, 37). (G–I) Overview of HER2 immunostaining (red) in coronal sections of mouse heads at E13.5, E15.0, and E15.5. Aside from epidermis and oral epithelium, HER2 is also expressed in the eyes and cranial neural crest–derived (CNC-derived) mesenchyme. (J–L) Magnified views of the maxillary region from panels G–I. HER2 shows strong expression in the differentiated osteoblasts and osteogenic front (green dashed circles in panels K and L) and the tooth germs (white dashed lines in panels J–L). In addition, HER2 expression is detected in midline epithelial seam formed by the fusing palatal shelves at E15.0 and diminishes as the seam underwent degradation at E15.5 (white arrowheads in panels K and L). (M–O) Magnified views of the mandibular region from panels G–I. Strong HER2 expression is detected in differentiated osteoblasts and osteogenic front (green dashed circles in panels M–O), but not in Meckel’s cartilage and perichondrium (yellow dashed circles in panels M–O). HER2 expression is also observed in the mandibular tooth germs (white dashed lines in panels M–O) and in the tongue muscle (white arrows in panels N and O). T, tongue; PS, palatal shelf; SP, secondary palate; MES, midline epithelial seam; CNC, cranial neural crest. Scale bar: 200 μm in all panels.
	Figure 4. Her2 p.A87T knock-in mice exhibit growth deficits and craniofacial anomalies.(A) Her2A87T/A87T mice display reduced body length and weight at P21. Representative images (left; scale bar: 1 cm) and quantification (right; n = 4 per genotype; P < 0.01) are shown. (B) Penetrance of developmental defects in E18.5 Her2A87T/A87T embryos, including reduced crown-rump length (CRL), ocular defects, maxillary (Mx) and mandibular (Md) hypoplasia, and cleft palate. MUT, mutant. (C and D) Her2A87T/A87T embryos (7.56%, 9 of 119) display reduced crown-rump length. Representative stereomicroscope images (C) and microcomputed tomography images (D) of E18.5 embryos (scale bar: 2 mm) and quantification (n = 4 per genotype; P < 0.001) are shown. (E) Her2A87T/A87T embryos exhibit maxillary and mandibular hypoplasia (12.6%, 15 of 119) and ocular defects (anophthalmia, 5.04%, 6 of 119) at E18.5. Dashed lines with double arrowheads indicate maxillary (top) and mandibular (bottom) lengths. Arrows indicate missing eyes. Each column corresponds to the same embryo. R, right view; L, left view. Scale bar: 2 mm. (F) Microcomputed tomography images of E18.5 craniofacial skeleton confirming reduced maxillary and mandibular lengths and narrower skull width in Her2A87T/A87T embryos. Scale bar: 2 mm. Quantification is shown (n = 4 per genotype; P < 0.05, **P < 0.01). (G) Her2A87T/A87T embryos (4.2%, 5 of 119) exhibit cleft palate at E18.5. Boxed regions are magnified in the bottom. Arrows indicate complete (middle) or incomplete (right) cleft palate. Scale bar: 1 mm. (H) Representative scanning electron micrographs confirming the cleft palate in Her2A87T/A87T embryos shown in panel G. Scale bar: 1 mm. (I and J) Immunofluorescence (I) and immunoblot (J) showing reduced HER2 and phospho-ERK (pERK) signals, but unchanged ERK level, in the palate shelf (PS), mandible (Md), and tongue (T) of E13.5 Her2A87T/A87T embryos. Boxed regions in panel I are magnified on the right. Scale bar: 200 μm. MUT, mutant. Unpaired t tests (A, D, and F).
	Figure 5. Maternal exposure to Tucatinib results in growth deficits and craniofacial anomalies in mice.(A) Schematic of Tucatinib administration. (B) Immunoblot showing reduced HER2 and pERK levels in E15.5 palate shelves following Tucatinib treatment. (C) Mice exposed to Tucatinib exhibit reduced body length and weight at P21. Representative images (left; scale bar: 1 cm) and quantification (right; n = 4 per group; *P < 0.01) are shown. (D) Penetrance of developmental defects in E18.5 embryos exposed to Tucatinib, including reduced crown-rump length (CRL), ocular defects, maxillary (Mx) and mandibular (Md) hypoplasia, and cleft palate. P < 0.05, P < 0.01, P < 0.001. (E and F) Tucatinib-treated embryos (35.14%, 26 of 74) display reduced crown-rump length. Representative stereomicroscope images (E) and microcomputed tomography images (F) of E18.5 mouse embryos (scale bar: 2 mm) treated with vehicle or Tucatinib and quantification (n = 4 per group; P < 0.001) are shown. (G) Tucatinib-treated embryos exhibit ocular defects (microphthalmia, 12.16%, 9 of 74; anophthalmia, 9.46%, 7 of 74) and maxillary and mandibular hypoplasia (18.92%, 14 of 74) at E18.5. Dashed lines with double arrowheads indicate maxillary and mandible lengths. Black arrow, microphthalmic eye; white arrows, missing eyes. Insets show the contralateral eye. Scale bar: 2 mm. (H) Microcomputed tomography images of E18.5 craniofacial skeletons confirming reduced maxillary and mandibular lengths and narrower skull width following Tucatinib exposure. Scale bar: 2 mm. Quantification (right; n = 4 per group; *P < 0.001) is shown. (I) Tucatinib treatment induces cleft palate (9.5%, 7 of 74) in E18.5 embryos. Scanning electron micrographs show complete (middle) or incomplete (right) cleft palate. Yellow arrowheads indicate cleft; dashed lines mark unfused palate shelves. Scale bar: 2 mm. Unpaired t tests (C, F, and H) and χ2 tests (D).