Dissecting the role of vitamin B₁₂ metabolism in craniofacial development through analysis of clinical phenotypes and model organism discoveries

Abstract

Vitamin B₁₂, otherwise known as cobalamin, is an essential water-soluble vitamin that is obtained from animal derived dietary sources. Mutations in the genes that encode proteins responsible for cobalamin uptake, transport, or processing cause inborn errors of cobalamin metabolism, a group of disorders characterized by accumulation of homocysteine and methylmalonic acid, neurodevelopmental defects, ocular dysfunction, anemia, and failure to thrive. Mild to moderate craniofacial phenotypes have been observed but these phenotypes are not completely penetrant and have not been consistently recognized in the literature. However, in the most recent decade, animal models of cblX and cblC, two cobalamin disorder complementation groups, have documented craniofacial phenotypes. These data indicate a function for cobalamin in facial development. In this review, we performed a literature review of all cobalamin complementation groups to identify which groups, and which human variants, are associated with dysmorphic features, microcephaly, or marfanoid phenotypes. We identified dysmorphic facial features in cblC, cblX, cblG, cblF, and cblJ, which are caused by mutations in MMACHC, HCFC1, MTR, LMBRD1, and ABCD4, respectively. Other complementation groups were associated primarily with microcephaly. Animal models (zebrafish and mouse) of cblC and cblX support these clinical phenotypes and have demonstrated neural crest cell deficits that include reduced expression of prdm1a, sox10, and sox9, key molecular markers of neural crest development. Characterization of a zebrafish mmachc germline mutant also suggests atypical chondrocyte development. Collectively, these data demonstrate an essential role for cobalamin in facial development and warrant future mechanistic inquiries that dissect the cellular and molecular mechanisms underlying human facial phenotypes in cobalamin disorders.

1. Structure and absorption of cobalamin

Cobalamin (Cbl; vitamin B₁₂) is a water-soluble micronutrient, structurally composed of a corrin ring that supports a cobalt metal ion, a base (5,6-dimethyl benzimidazole; DMB), a sugar, and an aminopropanol group (Fig. 1) (Hodgkin et al., 1956; Rickes et al., 1948; Smith, 1948). The Cbl corrin ring is covalently bonded to the core cobalt, which can form 4–6 bonds depending on oxidation state (Gailus et al., 2010). The bonds include 4 inner nitrogen bonds, one bond with DMB, and a final bond with an R group ligand. The R group or β-upper axial ligand is variable. The R group defines the derivative of Cbl and can include adenosyl (AdoCbl), methyl (MeCbl), hydroxy (OHCbl), and cyano (CNCbl) groups (Barker et al., 1960, 1960, 1960; Hannibal et al., 2008; Randaccio et al., 2010; Stabler, 2020). The central cobalt atom exists in either cob(I)alamin (Cbl⁺) (Schrauzer et al., 1968), cob(II)alamin (Cbl⁺²) (Stich et al., 2005), or cob(III)alamin (Cbl⁺³) oxidation states (Watkins and Rosenblatt, 2022). AdoCbl, MeCbl, CNCbl, and OHCbl are Cbl⁺³ forms and adopt a configuration where the nitrogen base is co-ordinated with the cobalt in the lower axial position (Froese and Gravel, 2010).

Fig. 1.

Schematic diagram of the cobalamin structure. The various R upper β-axial ligands are shown (R Ligands Box). Cobalamin (Cbl; vitamin B₁₂) is structurally composed of a corrin ring that supports a cobalt metal ion, a base (5,6-dimethyl benzimidazole; DMB), a sugar, and an aminopropanol group. The most common R-upper axial ligands include methylcobalamin, hydroxycobalamin, cyanocobalamin, and adenosylcobalamin.

Cbl is synthesized by microorganisms (i.e., bacteria and archaea) within the digestive tract of animals (Andres and Dali-Youcef, 2020; Kósa et al., 2022; Stabler, 2020) and for humans to acquire high enough concentrations of Cbl it must be obtained from the human diet (i.e., fish, meat, and dairy) (Andres and Dali-Youcef, 2020; Kósa et al., 2022). Cbl is bound to proteins in the food that humans ingest and must be processed into an active form. Errors in the processing and absorption of Cbl lead to disease and deficiency. The processing of Cbl begins in the mouth because saliva helps to release Cbl from the proteins it is naturally bound to. Once Cbl is released from the proteins in food, it binds to a protein called haptocorrin (R-protein), which is secreted by the salivary glands. The process continues in the stomach, where gastric acid is released, and any remaining Cbl associated with protein from food will be disassociated. All the free Cbl binds to R-protein creating the Cbl-R complex. Cbl-R is protected from the acidic environment of the stomach. Cbl is dissociated from R-protein in the duodenum of the intestine, enabling Cbl to bind to intrinsic factor (IF), a transport protein secreted by parietal cells in the stomach. Cbl-IF binds to the receptor cubilin that is in the ileum at the apical membrane of intestinal epithelial cells. Cubilin interacts with two major proteins, Amnionless and Lrp2 (low density lipoprotein receptor-related protein 2) (Mathew et al., 2024). The Cubilin complex is required for receptor mediated endocytosis of Cbl. Endocytosis causes a decrease in pH and the acidic condition separates IF from Cbl (Kósa et al., 2022). The free Cbl binds to transcobalamin II (TCbl) or holocobalamin and can be transported to all the tissues in the body through the circulatory system. The TCbl complex binds to the transcobalamin receptor (TCblR) and is internalized by endocytosis where it is shuttled across the lysosomal membrane. Inside the lysosome, Cbl departs from the carrier protein and exits into the cytosol (Fig. 2). This process is mediated through the interaction of two lysosomal membrane proteins, ABCD4 and LMBRD1, and a cytosolic processing enzyme, MMACHC (Hannibal and Jacobsen, 2022). Cytosolic Cbl will undergo changes in oxidation states through enzymatic reactions mediated by the MMACHC protein (Hannibal and Jacobsen, 2022). MMACHC is a unique enzyme catalyzing the removal of the β-axial ligand “R” with a reduction of the cobalt center. Processed Cbl will bind with two enzymes, methionine synthase (MS) and mitochondrial methylmalonyl-CoA mutase (MUT). The interaction is mediated by adaptor methylmalonic aciduria type D and homocystinuria (MMADHC). MS and MUT activity depend upon the presence of MeCbl and AdoCbl, respectively.

Fig. 2.

An abbreviated schematic of the cobalamin processing pathway. Cobalamin (Cbl) is transported into the lysosome by receptor mediated endocytosis. It is released from the lysosome where it is modified through the enzymatic function of MMACHC. Multiple proteins facilitate the release of cobalamin from the lysosome, which include ABCD4 and LMBRD1. Mutations in the genes that encode specific proteins (red lettering) and cause specific complementation groups are shown (indicated in black). In the cytosol, cobalamin binds to the MMACHC enzyme/carrier protein. Mutation of MMACHC causes cblC. Cobalamin undergoes various oxidative states and is transferred to the MTR enzyme (mutation causes cblG). Methylcobalamin (MeCbl) is required for the activity of MTR and conversion of homocysteine to methionine. In an alternative pathway cobalamin is transferred into the mitochondria where adenosylcobalamin (AdoCbl) is utilized by the MUT enzyme to convert methylmalonic acid into succinyl co-A, which feeds into the Krebs cycle. Additional defects in mitochondrial processing are indicated. Additional complementation groups occur due to mutations in the upstream regulators of MMACHC expression, indicated in the nucleus. THF refers to tetrahydrofolate and oxidation states are indicated with superscripts on the Cbl molecule.

1.1. Cellular requirements for Cbl and development

Cbl influences key biochemical functions such as DNA synthesis, DNA methylation, methionine synthesis, fatty acid production, folate, and mitochondrial metabolism (Anwar et al., 2024; Wiedemann et al., 2022). Cbl is a major cofactor in the cell and MeCbl is required for the formation of methionine, which has multiple functions in the cell. Methionine is required for protein synthesis, but it is also a precursor for molecules such as glutathione and the synthesis of polyamine spermine, which reduce reactive oxygen species and affect cell division, respectively (Cellarier et al., 2003). Methionine is also a primary source of methyl groups in the cell, which affects DNA methylation. DNA methylation is a critical mediator of gene expression and during development, dynamic changes in DNA methylation can help to establish tissue specific gene expression (Moore et al., 2013). Hypomethylation has been associated with increased risk for neural tube defects (Chen et al., 2010). Neural tube closure and neural induction are tightly integrated with formation of the neural crest, a transient cell population that produces cartilage and bone of the head, neck, and face. The role of Cbl in DNA methylation, via methionine deficiency could have a critical impact on the regulation of cell fate and differentiation because DNA methylation patterns change over the course of cell differentiation (Kim and Costello, 2017). Cbl can also have a protective effect on the cell, through the remediation of reactive oxygen species (ROS). This is because Cbl can exist in multiple oxidation states, some of which (Cbl⁺²) are capable of removing superoxide anions (Offringa et al., 2021). Insufficient Cbl can lead to increased oxidative stress (Brito et al., 2017) which can cause DNA damage, lipid peroxidation, amino acid oxidation, and reduced cell survival (Burton and Jauniaux, 2011). ROS can also have a positive effect on cells, which include activation of signaling cascades, consequently a balance in the equilibrium of ROS is required. Homeostasis of ROS equilibrium is also critically important in the balance between pluripotency and differentiation. Low levels of oxidative stress help to maintain pluripotency, whereas a moderate increase in oxidative stress promotes differentiation (Lee et al., 2018). Thus, changes in the ROS balance could impact neural crest cell differentiation. Vitamin B₁₂ also helps to generate energy for the body and assists in the breakdown of proteins, phospholipids, and neurotransmitters (Anwar et al., 2024). Based on the cellular functions of Cbl, it is not surprising that Cbl is required for fetal development where it helps to regulate gastrulation, nervous system development, and the formation of hemoglobin (Buers et al., 2016).

2. Inborn errors of cobalamin metabolism

Inborn errors of metabolism are a collection of genetic diseases with heterogeneous phenotypes and a spectrum of severity. They are generally caused by an inherited mutation of genes that encode for enzymes known to regulate key metabolic pathways (Ferreira and Van Karnebeek, 2019). Individually, they are considered rare, but collectively, they occur in 1 out of every 2500 births (Ferreira and Van Karnebeek, 2019). Those that affect Cbl metabolism are referred to as inborn errors of cobalamin metabolism. One of the hallmark features of Cbl deficiency is the accumulation of toxic byproducts due to deficiencies in the availability of MeCbl and AdoCbl. The heterogeneity of the clinical symptoms is dependent on the age of onset. There are two onset forms described in patients: early and late. Early-onset patients are typically diagnosed within the first year of life and most clinical cases are early onset. Early onset Cbl disorders are characterized by megaloblastic anemia, failure to thrive, brain malformations, developmental delay, and craniofacial dysmorphic features (Agarwal et al., 2021; Andres and Dali-Youcef, 2020; Cerone et al., 1999; Sloan et al., 2021). Additional phenotypes exist and multiple combinations of individual phenotypes have been described across patients.

2.1. Biochemical disruptions associated with inborn errors of Cbl

A defining feature of Cbl deficiency is the buildup of metabolites, homocysteine (HC), and methylmalonic acid (MMA) in the blood or urine. Without the processed forms of Cbl, these byproducts will rise to toxic levels. MMA is produced when methylmalonyl CoA (MCA), the substrate for the MUT enzyme, is not readily converted to succinyl-CoA because of reduced access to AdoCbl in the mitochondria. Methylmalonyl CoA is a common byproduct of cholesterol decomposition, odd-chain fatty acids, and branched-chain amino acids. Within the context of inborn errors of Cbl metabolism, the levels of MMA increase in blood and urine because Cbl is not processed correctly. At the level of the cytoplasm, MS and MeCbl are necessary for the synthesis of methionine from HC. In this reaction, the methyl group from MeCbl is transferred to HC synthesizing methionine. Free Cbl feeds into the folate cycle where it is remethylated using 5-methyltetrahydrofolate (5-methyl-THF) as a methyl donor which produces tetrahydrofolate (THF) as a byproduct (Fig. 2). Atypical MS activity, reduces conversion of HC to methionine, causing the accumulation of methyl-tetrahydrofolate, otherwise known as folate trapping (Green and Miller, 2022). This is significant because tetrahydrofolate is required to produce 5,10 methylenetetrahydrofolate, which is required for the conversion of deoxyuridine to thymidine. Inadequate thymidine synthesis results in defective DNA synthesis affecting rapidly dividing cells in the bone marrow and elsewhere. Neural crest cells are one example of a rapidly dividing cell population. Methionine is also a key methyl donor for the synthesis of S-adenosylmethionine (SAM). SAM is a source of methyl groups for the synthesis of creatine, phospholipids, neurotransmitters, DNA, RNA, and protein methylation. After donating the methyl group, S-adenosylhomocysteine (SAH) is formed, which is cleaved to HC and adenosine by SAH-hydrolase. The generated HC is either remethylated to form methionine or condensed with serine to form cystathionine by a vitamin B₆-dependent enzyme, cystathionine beta-synthase.

2.2. Complementation groups

Cbl disorders can result from mutations in more than one of the genes required for Cbl processing. Cbl syndromes are inherited autosomal recessive disorders divided into genetic complementation groups based on the pathogenic mutations that cause each subgroup (Table 1). Table 1 provides Online Mendelian Inheritance in Man numbers for each group (MIM#). The genes mutated in Cbl disorders include ABCD4, LMBRD1, MMAA, MMAB, MMACHC, MMADHC, MTR, MTRR, THAP11, or ZNF143. Mutations in HCFC1 are unique because of X-linked inheritance (Yu et al., 2013). Fig. 2 specifies the location/function of each protein in Cbl processing and the complementation group that results from mutation of that specific gene. For example, cblC is caused by mutations in the MMACHC gene, cblD by mutations in the MMADHC gene, cblF by mutations in the LMBRD1 gene, and cblJ is caused by mutations in the ABCD4 gene. There are also complementation groups caused by pathogenic mutations that affect upstream genetic regulators of the MMACHC gene. These include mutations in HCFC1 (cblX disorder) (Yu et al., 2013), THAP11 (cblX-like) (Quintana et al., 2017), and ZNF143 (cblX-like) (Pupavac et al., 2016). Assignment of the complementation class has traditionally been achieved by somatic cell complementation analysis but more recently, causative genes associated with each complementation group can be identified through sequencing approaches. Next-generation sequencing has been used for patients with unknown genetic mutations. However, if no causal mutations are identified or the pathogenicity of novel sequence variants is unclear, biochemical tests of cultured patient fibroblasts can establish a diagnosis of abnormal metabolism, but the causative gene may remain undiagnosed. Importantly, the results from patient fibroblasts are not always indicated in open-source databases that report variants of unknown significance.

Table 1.

Inborn errors of cobalamin metabolism.

Disorder	MIM #	Gene	Chromosome Location
cblA	251100	MMAA	4q31.21
cblB	251110	MMAB	12q24.11
cblC	277400	MMACHC	1p34.1
cblD	277410	MMADHC	2q23.2
cblE	236270	MTRR	5p15.31
cblF	277380	LMBRD1	6q13
cblG	250940	MTR	1q43
cblJ	614857	ABCD4	14q24.3
cblX	309541	HCFC1	Xq28

3. Complementation groups associated with craniofacial phenotypes

Craniofacial anomalies are one of the most common human birth defects, with significant functional, aesthetic, and social consequences (Trainor and Richtsmeier, 2015). The head and face are vulnerable to prenatal genetic and environmental insults. The most common congenital orofacial malformations in the United States are isolated cleft palate (CP) and cleft lip with or without cleft palate (CL/P). Their estimated prevalence is 16.86 cases per 10,000 live births (Stanbouly et al., 2024). Twin studies have revealed a 40–60% concordance rate among monozygotic twins, indicating a significant genetic component in the etiology of CL/P (Stanbouly et al., 2024). Mild to moderate craniofacial dysmorphia has been reported in Cbl metabolic disorders. However, these phenotypes are not prominent and are typically mild, such that the prevalence of facial phenotypes has been underestimated across multiple complementation groups. Maternal vitamin B₁₂ deficiency has been associated with an increased risk for CL/P (Munger et al., 2021), though these phenotypes are dramatically different than the mild to moderate facial phenotypes observed in Cbl disorders. There is no explanation for the discrepancy between Cbl disorder phenotypes and maternal deficiency associated with facial dysmorphia. Interestingly, most of the Cbl complementation groups are associated with microcephaly as a phenotype, however, this phenotype is unique from “dysmorphic features or facial features” in the literature and mechanistically. Since microcephaly is an independent phenotype, we categorized microcephaly, dysmorphic/facial features, and Marfan features as unique phenotypes in our analysis of the human variants in each of the following subsections. This approach proved to be useful (Table 2) because it allowed for the distinct identification of pathogenic variants exclusively associated with craniofacial dysmorphia across all Cbl groups.

Table 2.

Variants associated with inborn errors of cobalamin metabolism, cblA-J.

Gene	Disease	Phenotype	cDNA Variant	Protein Variant	Citation
MMAA	cblA	Microcephaly	c.433C > T	p.Arg145X	Merinero et al. (2008). Journal of inherited metabolic disease, 31(1), 55–66
MMAB	cblB	Microcephaly	c.454G > T	p.Glu152X	Keeratichamroen et al. (2007). Biochemical genetics, 45(5–6), 421–430
MMACHC	cblC	Microcephaly	c.331C > T/c.271dupA	p.Arg111X/p.Arg91LysfsX14a	Almannai et al. (2017) Molecular Genetics and Metabolism, 122(1–2) 60–66
MMACHC	cblC	Dysmorphic features	c.440G > C	p.Gly147Ala	Valentino et al. (2021) Brain Sciences, 11(7) 936
MMACHC	cblC	Marfanoid features, microcephaly	c.1A > G/c.271dupA	p.Met1?/p.Arg91LysfsX14a	Heil et al. (2007) Journal of Inherited Metabolic Diseases, 30(5) 811
MMACHC	cblC	Microcephaly	c.271dupA/c.90G > A	pArg91LysfsX14/p.Trp30Xa	Richard et al. (2009) Hum Mutat Nov; 30 (11):1558–66
MMACHC	cblC	Microcephaly, Marfanoid features	c.394C > T	p.Arg132X	Kaur et al. (2021) Amino acids, 53(2), 253–264
MMACHC	cblC	Microcephaly	c.316G > A	p.Glu106Lys	Kaur et al. (2021) Amino acids, 53(2), 253–264
MMACHC	cblC	Microcephaly, Marfanoid features	c.347T > C	p.Leu116Pro	Kaur et al. (2021) Amino acids, 53(2), 253–264
MMACHC	cblC	Microcephaly	c.567dupT/c.609G > A	p.Ile190Tyrfs13X/p.Trp203Xa	Weisfeld-Adams et al. (2013) Molecular Genetics and Metabolism, 110(3), 241–247
MMACHC	cblC	Microcephaly	c.271dupA	p.Arg91LysfsX14	Weisfeld-Adams et al. (2013) Molecular Genetics and Metabolism, 110(3), 241–247
MMACHC	cblC	Microcephaly	c.658_660delAAG	p.Lys220del	Liu et al. (2023) Frontiers in Nutrition, May 12:10:1124387
MMACHC	cblC	Dysmorphic features	c.609G > A/c.643T > C	p.Trp203X/p.Tyr215Hisa	Liu et al. (2022b) Clinica Chimica Acta, 533, 31–39
MMACHC	cblC	Microcephaly	c.394C > T	p.Arg132X$	Waheed et al. (2021a) Egyptian Journal of Human Genetics 75
MMACHC	cblC	Microcephaly	c.331C > T/c.271dupA	p.Arg111X/p.Arg91LysfsX14a	Brooks et al. (2016) Ophthalmology Mar; 123(3): 571–582
MMACHC	cblC	Microcephaly	c.457C > T/c.271dupA	p.Arg153X/p.Arg91LysfsX14a	Brooks et al. (2016) Ophthalmology Mar; 123(3): 571–582
MMACHC	cblC	Microcephaly	c.600G > A/c.271dupA	p.Tyr200X/p.Arg91LysfsX14a	Brooks et al. (2016) Ophthalmology Mar; 123(3): 571–582
MMACHC	cblC	Microcephalyb	c.271dupA/c.565C > A	p.Arg91Lysfs14/p.Arg189Sera	Nogueira et al. (2008) Mol Genet Metab. Apr; 93(4):475–80
MMACHC	cblC	Microcephalyb	c.666C > A c.666C > A/c.394C > T	p.Tyr222X p.Tyr222X/p.Arg132Xa	Nogueira et al. (2008) Mol Genet Metab. Apr; 93(4):475–80
MMACHC	cblC	Microcephalyb	c.3G > A/c.271dupA	p.Met1Ile/p.Arg91Lysfs14a	Nogueira et al. (2008) Mol Genet Metab. Apr; 93(4):475–80
MMACHC	cblC	Microcephalyb	c.3G > A/c.481C > T	p.Met1Ile/p.Arg161Xa	Nogueira et al. (2008) Mol Genet Metab. Apr; 93(4):475–80
MMACHC	cblC	Microcephalyb	c.271dupA/c.544T > C	p.Arg91LysfsX14/p.Cys182Arga	Nogueira et al. (2008) Mol Genet Metab. Apr; 93(4):475–80
MMACHC	cblC	Microcephalyb	c.271dupA/c.394C > T	p.Arg91LysfsX14/p.Arg132Xa	Nogueira et al. (2008) Mol Genet Metab. Apr; 93(4):475–80
MMACHC	cblC	Microcephalyb	c.394C > T/c.666C > A	p.Arg132X/p.Tyr222Xa	Nogueira et al. (2008) Mol Genet Metab. Apr; 93(4):475–80
MMACHC	cblC	Microcephalyb	c.331C > T	p.Arg111X	Nogueira et al. (2008) Mol Genet Metab. Apr; 93(4):475–80
MMACHC	cblC	Microcephalyb	c.271dupA/c.468_469delCT	p.Arg91LysfsX14/p. Trp157ValfsX24a	Nogueira et al. (2008) Mol Genet Metab. Apr; 93(4):475–80
MMACHC	cblC	Microcephalyb	c.331C > T/c.457C > T	p.Arg111X/Arg153Xa	Nogueira et al. (2008) Mol Genet Metab. Apr; 93(4):475–80
MMACHC	cblC	Microcephalyb	c.394C > T/c.468_469delCT	p.Arg132X/p.Trp157ValfsX24a	Nogueira et al. (2008) Mol Genet Metab. Apr; 93(4):475–80
MMACHC	cblC	Facial Features	c.482G > A/c.567dupT	p.Arg161Gln/p.Ile190Tyrfs13X a	Wang et al. (2015) Int J Clin Exp Pathol. 8(8):9337–9341
MMACHC	cblC	Dysmorphic features	Unknown	Unknown	D’Alessandro et al. (2010) Minerva Stomatologica, 59(3)
MMACHC	cblC	Dysmorphic features	Unknown	Unknown	Rosenblatt et al. (1997) Journal of Inherited Metabolic Disease 20(4), 528–538
MMACHC	cblC	Dysmorphic features and microcephaly	Unknown	Unknown	Andersson et al. (1999) Genetics in Medicine, 1(4), 146–150
MMACHC	cblC	Dysmorphic features	Unknown	Unknown	Mamlok et al. (1986) Neuropediatrics, 17, 94–99
MMACHC	cblC	Dysmorphic features	Unknown	Unknown	Fischer et al. (2014) Journal of Inherited Metabolic Disease, 37(5), 831–840
MMACHC	cblC	Dysmorphic features	Unknown	Unknown	Cerone et al. (1999) Journal of Inherited Metabolic Disease, 22(3), 247–250
MMADHC	cblD	Marfanoid Features	c.748C > T	p.Arg250X	Coelho et al. (2008) The New England journal of medicine, 358(14), 1454–1464.
MMADHC	cblD	Microcephaly (parental description only)	c.746A > G	p.Tyr249Cys	Atkinson et al. (2014) JIMD reports, 17, 77–81
MTRR	cblE	Microcephaly	c.903 + 469T > C	p.? (intronic variant)	Zavadáková et al. (2005). Human mutation, 25(3), 239–247
MTRR	cblE	Microcephaly	c.903 + 469T > C/c.15574_1557+3del7	p.?/p.Val488_Lys519dela	Zavadáková et al. (2005). Human mutation, 25(3), 239–247
LMBRD1	cblF	Dysmorphic features, cleft palate	c.1056delG	p.Leu352fsX18	Constantinou et al. (2015) Molecular Syndromology, 6(5), 254–258
LMBRD1	cblF	Dysmorphic features	c.515_516del	p.Thr172ArgfsX10	Braz et al. (2022) Clinical and experimental dermatology, 47(4), 812–815
LMBRD1	cblF	Microcephaly and dental features	c.1156C > T	p.Arg366X	Altawil et al. (2020) JAAD Case Reports, 6 (9), 882–885
LMBRD1	cblF	Dental features	c.1056delG	p.Leu352fsX18	Rutsch et al. (2009) Nature Genetics, 41, 234–239
LMBRD1	cblF	Dental features	c.1056delG/?	p.Leu352fsX18/?a	Rutsch et al. (2009) Nature Genetics, 41, 234–239
LMBRD1	cblF	Dysmorphic features	c.70-4298_246 + 2311del6785/c.1056delG	p.Ala24_Lys82del/p.Leu352fsX18a	Oladipo et al. (2011) Pediatrics, 128: e1636-e1640
LMBRD1	cblF	Dysmorphic features	Unknown	Unknown	Shih et al. (1989) American Journal of Human Genetics, 33(4), 555–563
MTR	cblG	Dysmorphic features	c.2699del	p.Asn900IlefsX2	Huemer et al. (2015) Journal of inherited metabolic disease, 38(5), 957–967
MTR	cblG	Microcephaly	c.IVS-166A > G/c.2112delTC	p.?/NM_000254.3 (NP_000245.2) p.Leu705Glnfsb4 (predicted only)	Wilson et al. (1998) American journal of human genetics, 63(2), 409–414
MTR	cblG	Dysmorphic features, microcephaly	c.3518C > T	p.Pro1173Leu	Komulainen-Ebrahim et al. (2017) Neuropediatrics, 48(6), 467–472
ABCD4	cblJ	Dysmorphic features	c.542+1G > T/c.1456G > T	Intronic variant resulting in p.Asp143_Ser181del and p.Gly443_Ser485del/p.Gly486Cysa	Coelho et al. (2012) Nature Genetics, 44(10), 1152–1155

^aIndicates a compound heterozygous inheritance model of indicated variants.

^bIndicates that the phenotype was found in a cohort of individuals where the variant was identified. The phenotype is therefore only indirectly correlated with the variant indicated.

3.1. cblA and cblB

cblA and cblB cause isolated accumulation of MMA due to defects in the processing of AdoCbl and result from mutations in the MMAA and MMAB genes, respectively (Table 1; Fig. 2). At the time of this review, the ClinVar database (Landrum et al., 2014) indicated there are 80 pathogenic variants of the MMAA gene and 64 pathogenic variants of the MMAB gene, excluding changes in copy number. Analysis of literature associated with these variants did not identify any variants exclusively associated with dysmorphic or facial features but did reveal one variant of MMAA (c.433C > T) and a single variant of MMAB (c.454G > T) associated with microcephaly (Table 2) (Keeratichamroen et al., 2007; Merinero et al., 2008). We also analyzed ClinVar for pathogenic missense variants in the MUT gene, which also results in isolated accumulation of MMA (Moreno-Garcia et al., 2012). Out of the 96 putative missense pathogenic variants in ClinVar for the MUT gene, we did not find any associated with craniofacial dysmorphia. Based on these data, we predict that the presence of facial dysmorphia with isolated accumulation of MMA is rare.

3.2. cblC

cblC is caused by mutations in the MMACHC gene, located on chromosome 1p34.1 (Table 1). According to ClinVar there are 119 pathogenic variants in cblC excluding copy number variants. Pathogenic variants affect the synthesis of AdoCbl and MeCbl (Lerner-Ellis et al., 2006) and cause combined MMA and HC accumulation (Antony et al., 2023). Among the complementation groups, cblC syndrome is the most common (Lerner-Ellis et al., 2009). The most frequently reported MMACHC mutations include c.271dupA, c.394C > T, c.331C > T, and the c.609G > A in Chinese populations (Wang et al., 2019). The c.271dupA mutation produces p.Arg91LysfsX14 and this particular variant is associated with microcephaly in homozygous or compound heterozygous inheritance models (Table 2). We identified 19 MMACHC pathogenic variants (Table 2) associated with craniofacial phenotypes or microcephaly. Many of the variants have a compound heterozygous inheritance and therefore, the causal relationship of individual variants in each phenotype was difficult to decipher. Two compound heterozygous variant combinations; the p.Trp203X/p.Tyr215His (c.609G > A/c.643T > C) and the p.Arg161Gln/p.Ile190fsX (c.482G > A/c.567dupT) were associated with dysmorphic features in the literature (Liu et al., 2022a; Wang et al., 2015). These variants are in the C-terminal/TonB domain of MMACHC. We found one variant, the p. Gly147Ala (c.440G > C) found in the Cbl binding domain also associated with dysmorphic features (Table 2) (Valentino et al., 2021). Variants exclusively associated with microcephaly include p.Arg91LysfsX14 identified in compound heterozygous inheritance with either the p.Arg111X, p.Met1? p.Trp30X, p.Arg153X, p.Tyr200X, p.Arg189Ser, or p.Met1Ile (Table 2 details the genomic location) variants (Almannai et al., 2017; Brooks et al., 2016; Heil et al., 2007; Kaur et al., 2021; Liu et al., 2023; Nogueira et al., 2008; Richard et al., 2009; Waheed et al., 2021a; Weisfeld-Adams et al., 2013). Weisfeld-Adams et al. also observed microcephaly in homozygous carriers of the p.Arg91LysfsX14 (c.271dupA) variant (Weisfeld-Adams et al., 2013). Some variants were identified in cohorts of microcephaly but without clear direct association genotype: phenotype correlation. These are indicated with a footnote in Table 2. Most of the variants associated with microcephaly lead to premature stop codons before or within the Cbl binding domain. Patient variants associated exclusively with dysmorphic features were primarily located within the C-terminal/TonB domain.

Facial phenotypes in the dysmorphic category were generally characterized as mild to moderate. Cerone et al. observed minor facial abnormalities characterized by a long face, a high forehead, wide floppy low-set ears, and a flat philtrum (Cerone et al., 1999). In 1997, Rosenblatt et al. documented craniofacial malformations in two female patients with cblC syndrome. They presented with an elongated face, prominent forehead, oversized and downwardly positioned ears, and a flattened groove between the nose and upper lip (Rosenblatt et al., 1997). D’Alessandro et al. described the facial features in an 11-year-old patient diagnosed with early onset cblC as a high forehead, epicanthal folds, broad nasal bridge, a long and flat philtrum, and a low set of floppy large ears. His description included the presence of an overbite, competent lips, and an inability to express correct motor gestures of the face (D’Alessandro et al., 2010). Dysmorphic features reported by Isenberg et al. included a hatchet-shaped head, arachnodactyly, and a high arched palate (Isenberg et al., 1986). Many observed dysmorphic features were described in patients without any indication of the specific variant underlying causality (Andersson et al., 1999; Fischer et al., 2014; Mamlok et al., 1986).

The presence of microcephaly was also indicated in a case that evaluated 6 siblings who were diagnosed with cblC (Waheed et al., 2021a). Each of the siblings had a different age of onset and had different disease progression, severity, and outcome. Specifically, “sibling 2” had an early onset type of cblC. Clinical diagnosis is difficult in this case because the subject contained a compound heterozygous mutation in the MMACHC and CD320 genes (Waheed et al., 2021b). In an additional case report, an asymptomatic mother, carrying two mutations in MMACHC, gave birth to a male proband carrier with a homozygous mutation in MMACHC at c.658_660delAAG (p.Lys220del). The patient presented with microcephaly and several neurological abnormalities (Liu et al., 2023). In an additional cohort of eight patients, three individuals had congenital microcephaly and dysmorphic facial features. Among those features were an abnormal formation of the cupid bow, upper lip, and bilateral epicanthal folds (Andersson et al., 1999). No variants were indicated by Andersson and colleagues.

Marfan features have also been indicated in cblC disorder. Marfan features include a tall and slender build, overcrowded teeth, an abnormally curved spine, and long and narrow facial features. Such features were reported in two siblings diagnosed with cblC syndrome. The female patient presented with Marfan features, characterized as increased arm-span, arachnodactyly, joint hyperlaxity, and scoliosis (Heil et al., 2007). In contrast, the male patient was described to have dysmorphic presentations, but without further detail (Heil et al., 2007). Therapy consisting of folic acid, vitamin B₆, l-carnitine, and intramuscular Cbl resulted in a clear improvement of biochemical parameters (Heil et al., 2007). Disconnecting Marfan features from facial features was difficult when assessing causality and it remains unclear if the case from Heil et al. falls into our current category of dysmorphic features.

A case study by Bassim et al. (2009) reported enamel defects and hypoplasia (Bassim et al., 2009). In an uncommon case, a two-year-old female patient was diagnosed with both cblC disease and Mowat-Wilson syndrome (MWS); two independent hereditary conditions, representing the first description of cblC disease and MWS in a patient (Liu et al., 2022b). Physical examination revealed a head circumference of 44.7 cm, sparse eyebrows and hair, square skull, prominent forehead, wide eye distance, esotropia, epicanthus, wide nose bridge, widely spaced teeth, prominent round nose tip, M-shaped upper lip, triangular-shaped pointed chin, and ear folds mark (Liu et al., 2022b). However, the authors did note that facial abnormalities could not be fully explained by cblC disease. Nevertheless, they believed that the heterogeneity of clinical manifestations is mainly related to the mutation of MMACHC and interestingly, the described dysmorphic features resembled those observed across multiple reports without indicated variants as summarized above.

3.3. cblD

cblD results in isolated homocystinuria (cblD-homocystinuria) caused by mutation in the regions encoding both the C or N-terminal regions of the MMADHC protein but can cause combined methylmalonic aciduria and homocystinuria (cblD-combined) after nonsense, splice site, or truncating variants in the C-terminal region of the protein (Abu-El-Haija et al., 2018). A male patient exhibited craniofacial anomalies characterized by a Marfan appearance associated with the p.Arg250X (c.748C > T) variant (Coelho et al., 2008). Goodman and colleagues described Marfanoid features in an additional patient (Goodman et al., 1970) and a parental report of microcephaly was associated with the pTyr249Cys (c.746A > G) variant (Atkinson et al., 2014). These are indicated in Table 2. The facial phenotypes in cblD have not been comprehensively indicated clinically and therefore, there is a lack of understanding as to their penetrance, prevalence, or whether there is any direct causality with mutations in MMADHC. Additionally, the extent of the phenotypes in these patients are unclear because Marfan features span body type, chest, spine, joints, face, and mouth. We don’t have an indication of which of these features are directly attributed to mutations in MMADHC. At the time of this review, there were 46 pathogenic variants (excluding copy number changes) of MMADHC in the ClinVar database with very limited association of dysmorphic features in the literature. As such we cannot confidently link mutations in MMADHC with facial features.

3.4. cblE

cblE is caused by mutations in the MTRR gene. There are currently 81 pathogenic variants of MTRR in the ClinVar database, excluding copy number variants. From these 81 variants, we identified only two associated with microcephaly; homozygous inheritance of c.903 + 469T > C (p.?) or the same variant as a compound heterozygous inheritance with p.Val488_Lys519del (c.15574_1557+3del7) (Zavadáková et al., 2005) (Table 2). There was no direct link with facial features. We conclude a very low probability of facial features in cblE.

3.5. cblF

cblF disorder is caused by mutations in the LMBRD1 a gene on chromosome 6q13 that encodes a lysosomal membrane protein that facilitates the export of Cbl into the cytoplasm (Table 1; Fig. 2). The majority of mutated alleles carry the deletion c.1056delG (p.Leu352fsX18) (Gailus et al., 2010; Oladipo et al., 2011). The mutation affects the availability of Cbl, making it insufficient to carry out the synthesis AdoCbl and MeCbl, ultimately, resulting in accumulation of byproducts causing MMA and HC (Oladipo et al., 2011). Only a small number of cases have been reported in the literature, but some case reports describe subjects with craniofacial abnormalities (Table 2) (Constantinou et al., 2015). A patient was reported to have facial asymmetry, low-set ears, and prognathism (Braz et al., 2022). These phenotypes are very similar to those described in cblC disorder and this overlap may suggest that LMBRD1 is required for facial development. Additionally, a clinical case from a female patient with developmental and speech delay, short stature, microtia, microcephaly, dysphagia, and abnormal dentation (Altawil et al., 2020). Rutsch et al., described two patients to have minor facial abnormalities. The female patient presented with bifid incisors and the male patient displayed pegged teeth. Additional dysmorphic features included a metopic ridge, midline cleft soft palate, low-set ears, broad nasal tip, and micrognathia (Rutsch et al., 2009). Bifid incisors were also report by Rosenblat and colleagues (Rosenblatt et al., 1986). In addition, Oladipo et al. observed a ridged metopic suture with trigonocephaly, asymmetric and cup-shaped ears, micrognathia/retrognathia, a high arched palate, left torticollis, and up-slanting palpebral fissure (Oladipo et al., 2011). Lastly, a female patient was described with epicanthal folds, small low-set ears, thin upper lip, high palate (Shih et al., 1989). We identified 4 unique pathogenic variants in the LMBRD1 gene associated with dysmorphic features and atypical dental features, with one variant having been associated with cleft palate (Constantinou et al., 2015). This may indicate that the release of Cbl from the lysosome is a critical requirement for facial development. There are only 28 pathogenic variants in the ClinVar database, excluding copy number variants, and we found 6 variants individually or in a compound heterozygous inheritance model associated with dysmorphic features in the literature (Table 2). This equates to roughly 14.2% of variants associated with facial features.

3.6. cblG

In our analysis of cblG, we observed two variants associated with dysmorphic features; these are p.Asn900IlefsX2 (c.2699del) and p.Pro1173Leu (c.3518C > T) (Huemer et al., 2015; Komulainen-Ebrahim et al., 2017). Interestingly, the p.Pro1173Leu was independently reported in a patient with microcephaly (Kasapkara et al., 2019) without any indication of dysmorphic features. Therefore, the role of this variant in facial development is unclear. These are summarized in Table 2. Microcephaly was also reported in a pair of siblings with an intronic variant IVS-166A > G in a compound heterozygous inheritance with c.2112delTC (Wilson et al., 1998). The intronic variant results in a 165bp insertion at junction 339/340 while the c.2112 variant results in the deletion of two nucleotides and a predicted premature frameshift. There are 56 pathogenic variants in the MTR gene in the ClinVar database, excluding copy number variants. Our analysis found very few variants associated with dysmorphic or facial features (approximately 3.5% of all variants in ClinVar). Consequently, we cannot confidently state that MTR is required for facial development at this time.

3.7. cblJ

ABCD4 is an ATP-binding cassette (ABC) transporter that has been classified as a member of the D subfamily of peroxisomal ABC half transporters and mutations in ABCD4 cause cblJ (Coelho et al., 2012). Mutations were identified in two subjects, each subject carrying two independent variants, each on a different allele. Subject 2 carried two mutations that affect consensus splice sites c.542 + 11G > T and c.1456G > T. the c.542 + 11G > T variant results in an in-frame deletion of amino acids 39 (p.Asp143_Ser181del) and 43 (p.Gly443_Ser485del) (Coelho et al., 2012) while the c.1456G > T variant results in the p.Gly486Cys change. Micrognathia and hypertelorism were reported (Table 2). Micrognathia phenotypes overlap with dysmorphic features present in cblF, which is anticipated because these two proteins interact at the lysosome to facilitate Cbl release (Deme et al., 2014; Fettelschoss et al., 2017; Kawaguchi et al., 2016). There are only 16 total pathogenic variants of ABCD4 in the ClinVar database, which is likely the reason so few (2) of them were associated with dysmorphic features. Despite the small number of variants, our analysis suggests approximately 12.5% of identified variants are associated with facial features, thus we suggest a strong link between cblJ and facial development. We excluded all cases with copy number gain/loss documented in ClinVar.

3.8. cblX

Methylmalonic acidemia and homocysteinemia cblX type (cblX) is an X-linked recessive disorder characterized by multiple congenital anomalies including craniofacial abnormalities. Mutations in the gene encoding the transcriptional cofactor Host Cell Factor C1 (HCFC1) cause cblX syndrome. HCFC1 is a transcriptional co-regulator that interacts with multiple transcription factors to regulate different biological processes, such as cell proliferation, migration, and cell death. Yu et al. reported 13/17 individuals with cblX and a subset of them had facial phenotypes (Yu et al., 2013). ClinVar reports 6 pathogenic missense variants in HCFC1, however, this number is an underestimate when compared to those reported in the literature (Castro and Quintana, 2020; Gérard et al., 2015; He et al., 2023; Huang et al., 2012; Hussain et al., 2024; Jolly et al., 2015; Koufaris et al., 2016; Wongkittichote et al., 2021; Yu et al., 2013). In addition, some variants reported in ClinVar as pathogenic are of unknown significance because they lack validation. Facial phenotypes across known variants that cause cblX or intellectual disability have been reviewed elsewhere (Castro and Quintana, 2020). HCFC1 has been shown to bind to the MMACHC promoter (Dejosez et al., 2010) and mutations in HCFC1 preferentially reduce the expression of MMACHC (Yu et al., 2013) in patient fibroblasts. Thus, the biochemical manifestations of cblX are considered MMACHC dependent. HCFC1 binds to >5000 downstream promoters (Michaud et al., 2013) and therefore, it is unclear if facial phenotypes in cblX are MMACHC dependent. Interestingly, mutation of the THAP11 and ZNF143 genes cause cblX-like syndromes (Pupavac et al., 2016; Quintana et al., 2017). HCFC1 interacts with these proteins to regulate MMACHC and other gene expression (Dejosez et al., 2010; Vinckevicius et al., 2015). Coincidentally, both THAP11 and ZNF143 have been implicated in facial development (Perez et al., 2023; Quintana et al., 2017). Collectively, these data support a role for MMACHC in facial development.

3.9. Unifying phenotypes across complementation groups

Across all complementation groups the biochemical accumulation of metabolites is observed. However, the severity and extent of this phenotype does display some heterogeneity. The penetrance of facial features across each group and within each complementation group differs. This may be partially attributed to the fact that dysmorphic features were not considered hallmark phenotypes in the past and documentation is minimal. In our genotype: phenotype analysis, we observed overlapping Marfan features in cblC and cblD, which result from mutations in MMACHC and MMADCH genes. Microcephaly was a unifying phenotype observed in all complementation groups, indicating a clear role for Cbl in brain development. We only found one report of severe facial phenotypes (cleft lip) and that was in cblF disorder. Thus, it was much more common to observe mild to moderate facial phenotypes. Dysmorphic features were primarily observed in cblC and cblF. In these two complementation groups, we found large low set ears, dental features that include wider teeth and bifid incisors, high arched palate, and epicanthal folds. cblF and cblJ were the only two complementation groups to present with micrognathia.

4. The neural crest and facial development: lessons from animal models

Neural crest cells (NCCs) are a multi-potent progenitor cell which arise from the dorsal end of the neural tube during closure. NCCs produce multiple differentiated cell types including cartilage and bone of the viscerocranium, which comprises the jaw and other structures formed from the pharyngeal arches. Pharyngeal arches are transient embryonic structures that organize and produce specific facial features. For example, cells from pharyngeal arch 1 produce the paired maxillary and mandibular prominences (Cordero et al., 2011). These paired prominences represent 4 of the 5 facial prominences present at 5 weeks gestation (Baxter and Shroff, 2011). Thus, analysis of the development of cranial NCCs, a specific subset of NCCs that produce cartilage and bone of the head, neck, and face, could yield critical insight into the mechanisms by which vitamin B₁₂ metabolism regulates facial development.

Currently, animal models exist for some complementation groups. The zebrafish has made major headway in studying craniofacial anomalies that are present in disorders of Cbl metabolism, specifically those observed in complementation groups of cblC and cblX. A study in 2014 (Summarized in Fig. 3), analyzing cblX related phenotypes, revealed craniofacial abnormalities after knockdown of hcfc1b, one of the two zebrafish orthologs of HCFC1 (Quintana et al., 2014). Analysis of NCC development demonstrated normal expression of sox10 at early times points but reduced expression of sox10 at later time points. Sox10 (nomenclature is zebrafish specific within this section) protein is a Sry-box transcription factor that is expressed in pre and post-migratory NCCs (Aoki et al., 2003). Sox10 regulates specification, is necessary for the delamination of NCCs from the neural tube, and for the differentiation of many neural crest lineages (Schock and LaBonne, 2020). sox10 expression is stimulated by the function of FoxD3 protein (Dottori et al., 2001). FoxD3 is a forkhead box family transcription factor that regulates NCC development. Reduced expression of sox10 in the zebrafish models of cblX is also associated with reduced sox9 expression, a marker of cranial NCC differentiation whose expression and function is required for expression of col2a1 (Schock and LaBonne, 2020). Additional differentiation markers were not comprehensively studied and therefore, an exact cellular mechanism has not been identified, but proliferation of Sox10+ (through transgenic analysis) cells was reduced. Markers associated with early NCC development were also monitored after knockdown of hcfc1b in zebrafish. These include expression of dlx2a and prdm1a (Fig. 3). dlx2a is a zebrafish ortholog of DLX2, a homeobox transcription factor that regulates specification, migration, and maintenance of cranial NCCs (Sperber et al., 2008). Prdm1 is a member of the PRDM family, which are a group of epigenetic remodeling proteins (Hohenauer and Moore, 2012). Mutation or knockdown of prdm1a causes abnormal NCC and facial development in zebrafish (Birkholz et al., 2009; Hernandez-Lagunas et al., 2005). Loss of prdm3 and prdm16 were recently associated with defects in chondrocyte stacking and polarity (Shull et al., 2022) which is interesting because mutation of the zebrafish mmachc gene (cblC) causes a very similar phenotype (i.e. abnormal chondrocyte polarity) (Paz et al., 2023). Mechanisms in Prdm3/6 deficient mice and zebrafish are linked to regulation of Wnt signaling, which is yet to be explored in zebrafish with mutations in mmachc.

Fig. 3. Summary of neural crest cell analysis in zebrafish after knockdown of hcfc1b.

Knockdown of hcfc1b in zebrafish causes craniofacial phenotypes. (A–C) Expression of dlx2a (purple) and prdm1a (turquoise) were performed at 24 h post fertilization (hpf). Reduced prdm1a was observed in morphant animals (A′-B′). Red arrow indicates reduced prdm1a expression. At later stages (3 days post fertilization-dpf), a loss of Sox10+ (magenta) cells was observed alongside reduced sox9 expression (light blue) (D′-F′). A comparison of observed phenotypes is shown in schematic form. Red arrows indicate regions with no expression relative to non-injected controls (D–F). abbreviations: pq-palatoquadrate, m-Meckel’s cartilage, ch-ceratohyal

As it relates to knockdown of hcfc1b and facial phenotypes, ectopic expression of human MMACHC restores facial phenotypes and the knockdown of zebrafish mmachc results in similar craniofacial phenotypes (Quintana et al., 2014). Further, Quintana et al. demonstrated that knockdown of the Hcfc1 (zebrafish nomenclature) interacting protein Thap11 also resulted in craniofacial phenotypes (Quintana et al., 2017). As a follow-up to these studies, Chern et al. generated the Hcfc1^p.A115V/Y and Ronin^F80L/F80L mouse knock-in models, which represent the equivalent human syndromes (Chern et al., 2022). These mutant mice demonstrate hallmark metabolic toxicity (Chern et al., 2020). Other identified phenotypes include intrauterine growth restriction, hydrocephalus, anemia, and congenital heart malformations. Facial phenotypes were described for both knock-in alleles and could not be restored by expression of Mmachc (Chern et al., 2020), a finding that appears to be distinct from zebrafish (Quintana et al., 2014). These genetic mouse knock-in models led to the hypothesis that abnormal ribosomal protein expression (Chern et al., 2022) underlies the phenotypes in cblX and cblX-like mice. However, restoration of facial phenotypes through gain/loss of ribosomal proteins has not been performed.

Chern and colleagues also characterized a mutant null allele of Mmachc. A failure of palate fusion was observed in homozygous mutants, although the allele was embryonic lethal (Chern et al., 2020). Comprehensive characterization of NCC development was not performed. There is a critical need to analyze NCC markers at unique stages of NCC development. In 2023, Paz et al. identified chondrocyte intercalation defects in the development of major structures of the viscerocranium in a zebrafish with a nonsense mutation in mmachc. Sloan and colleagues had previously characterized the metabolic and ocular phenotypes present in this allele (Sloan et al., 2020). Early NCC development was not affected according to barx1 and sox10 expression, though prdm1a expression was not evaluated. barx1 is expressed in the craniofacial mesenchyme (Tissier-Seta et al., 1995) and is a marker for migratory NCCs (Sperber et al., 2008).

The phenotypes described in zebrafish (mmachc) were mild to moderate and more consistent with the phenotypes present in humans with cblC relative to the phenotypes observed in the murine null allele. Interestingly, the facial phenotypes in the zebrafish mutant of mmachc can be restored with a vitamin B₁₂ binding deficient variant (p.Gly147Asp) of cblC. These data suggest that the vitamin B₁₂ binding domain may be dispensable for facial development. However, we did find the p.Gly147Ala variant of human MMACHC associated with dysmorphic features in our literature review (Table 2), which could indicate that multiple mechanisms are at play and that unique non-synonymous missense changes, even when located at the same residue, can disparately regulate facial development.

It continues to be difficult to decipher whether the metabolic accumulation in Cbl disorders contributes to the onset or formation of dysmorphic features. In the Hcfc1^p.A115V/Y and Ronin^F80L/F80L mouse models there is metabolite accumulation associated with the craniofacial phenotypes (Chern et al., 2022). However, that is not the case in zebrafish (Quintana et al., 2014). Therefore, in zebrafish there is a disconnect between facial features and metabolite accumulation. Metabolic accumulation does occur in zebrafish with mutations in the mmachc gene, but metabolites were not monitored prior to 7 days post fertilization (Sloan et al., 2020). Facial phenotypes in zebrafish mutants of mmachc are present as early as 4 days post fertilization. Consequently, these phenotypes may occur prior to the onset of a metabolic phenotype. This disconnect between metabolic accumulation and facial phenotypes is not only observed in zebrafish. In cblX syndrome, the metabolic and biochemical phenotypes can be mild and the degree of metabolic accumulation is not a very good indicator for the severity or presence of other phenotypes (Yu et al., 2013). Additionally, facial dysmorphia was not observed across all complementation groups and therefore, it is likely that additional mechanisms are at play.

5. Final perspective and future directions

The processing of Cbl is a multi-step pathway including intake, absorption, and processing by various enzymes to yield active forms of Cbl, namely AdoCbl and MeCbl. Mutations in the vitamin B₁₂ processing pathway cause inborn errors of Cbl metabolism. At least 5 (including cblX and cblX-like) complementation groups are associated with dysmorphic features, but the frequency of identified variants in cblG is very low complicating interpretations. We acknowledge that more complementation groups may be associated with facial phenotypes, but because facial phenotypes are mild to moderate, they may not have been reported consistently in the literature. The most common inborn error of Cbl metabolism is cblC. It is highly related to cblX and cblX-like disorders because these disorders are caused by mutations in the upstream regulators of MMACHC expression (Pupavac et al., 2016; Quintana et al., 2017; Yu et al., 2013). All 3 complementation groups (cblC, cblX, and cblX-like) are associated with facial phenotypes, presumably due to MMACHC specific mechanisms when considering results in animal models. These data are supported by germline mutants in the zebrafish mmachc gene, which have mild to moderate facial phenotypes. cblX like zebrafish, namely the knockdown of hcfc1b, do not have accumulation of metabolites (Quintana et al., 2014) and a vitamin B₁₂ binding mutant of MMACHC can restore facial development in mmachc mutant zebrafish (Paz et al., 2023; Quintana et al., 2014). These data collectively suggest that metabolite accumulation and vitamin B₁₂ binding may not be the primary or only mechanism regulating facial development in Cbl disorders. Future directions require comprehensive characterization of cellular and molecular phenotypes after genetic manipulation of individual enzymes associated with facial features. An understanding of when and where specific NCC markers are abnormally expressed may indicate putative pathways that are Cbl dependent. For example, literature has established a comprehensive understanding of the timing and function of many signaling pathways. One example is the Wnt signaling pathway, in which there are different functions for this pathway during NCC induction versus differentiation.

The role of Cbl metabolism could be time and cell type specific. Therefore, additional studies that characterize the temporal cellular and molecular mechanisms associated with individual enzymes or proteins are required. We posit that additional studies into the expression and function of Prdm proteins in Cbl deficient animals are warranted since chondrocyte defects were observed in both zebrafish mmachc mutants and Prdm (mice and zebrafish) mutants. It remains to be seen if single cell RNA sequencing approaches can parse out the heterogeneity of cellular phenotypes after Cbl deficiency.

Finally, our literature analysis of patient variants and complementation groups suggests that there are at least 3 Cbl processing enzymes definitively associated with “facial/dysmorphic” phenotypes in patients. Some of these complementation groups have very few pathogenic variants but a higher percentage of variants associated with facial features (cblF and cblJ). It is important to note that the number of pathogenic variants we report from ClinVar do carry some limitations. For example, not all the variants are associated with a publication that clearly indicate the number of patients that carry each specific allele, and many do not have functional analysis associated with them. There are many putative pathogenic variants listed that are deposited from large consortiums or clinical diagnoses. These variants are not necessarily confirmed to be pathogenic, but based on the type of variant (nonsense, frameshift) are classified as pathogenic due to gene function and/or other variants in the same class (i.e. nonsense) known to cause disease. Moreover, some of the alleles are inherited as compound heterozygous variants skewing the number of patients in which specific alleles definitively cause disease. Other variants are reported as part of a cohort of patients without a clear indication of the actual number of diagnosed patients carrying each allele. Nonetheless, we emphasize facial features in 3 groups: cblC, cblJ, and cblF. Interestingly, the 3 proteins affected in these groups interact with each other at the lysosome to facilitate Cbl release (Deme et al., 2014). They are LMBRD1, ABCD4, and MMACHC. Future studies that characterize this interaction in the context of facial and NCC development are justified. Moreover, there are no animal models for ABCD4 or LMBRD1 that have evaluated facial development. Those that exist are embryonic lethal (Buers et al., 2016) or have not been characterized for facial dysmorphia (Choi et al., 2019).

5.1. Concluding remarks

We have used the ClinVar database to identify Cbl complementation groups associated with dysmorphic features, microcephaly, and Marfan features. Some complementation groups have many pathogenic variants (50–150), yet very few are associated directly with facial features. In those cases, we primarily found the phenotype of microcephaly, which we do not consider a craniofacial phenotype. Other complementation groups have very few overall pathogenic variants but have clear indications of dysmorphic features. To our knowledge, this is the first comprehensive summary of genotype: phenotype correlations as it relates to craniofacial phenotypes. Our clinical literature review is supported by available animal models of specific complementation groups. These collective analyses support a function for Cbl metabolism or specific Cbl processing proteins in NCC and facial development. These findings are particularly important because maternal Cbl intake has been associated with severe facial phenotypes and a comprehensive understanding could help to prevent craniofacial abnormalities.