Clinical and genetic characterization of congenital disorders of glycosylation in 20 Chinese patients
Authors: Peiwei Zhao, Li Tan, Qingjie Meng, Lei Zhang, Yufeng Huang, Xiankai Zhang, Yanqiu Hu, Shiqiong Zhou, Xuelian He
Published: 2025

ABSTRACT
BackgroundCongenital disorders of glycosylation (CDG) are a complex and heterogeneous family of rare metabolic diseases that affect protein and lipid glycosylation and glycosylphosphatidylinositol synthesis. These disorders can affect multiple organs, leading to a broad spectrum of symptoms that vary among different CDG subtypes and between individuals with same type of CDG. This study aimed to investigate the genetic variants, molecular etiologies, and clinical features of 20 Chinese patients diagnosed with CDG.ResultsUsing whole-exome sequencing (WES), functional prediction tools, Sanger sequencing, and segregation analysis, we identified variants in several genes: ALG2 (3 patients), DPM2 (3 patients), PMM2 (3 patients), and ALG13 (2 patients). Additionally, variants in COG5, COG6, MOGS, DPM3, ALG1, ALG3, ALG11, SSR4 and SLC35A2 each were observed in single case. In total, 28 distinct variants were identified, 11 of which were previously unreported. Genotype-phenotype correlations revealed notable findings: variants in the N-terminus of ALG2 before the intramembrane domain were associated with congenital myasthenic syndromes (CMS), whereas those in the C-terminus caused ALG2-CDG; DPM2-CDG patients with variants in transmembrane region 1 exhibited more severe phenotypes; male patients with hemizygous variants in SLC35A2 demonstrated milder phenotypes compared to those with mosaic variants.ConclusionsThis findings expand the spectrum of known clinical presentations and genetic variants in CDG, and establish possible genotype-phenotype correlations of several pathogenic genes, emphasizing the need for functional studies to unravel the underlying mechanisms.

Background

Glycosylation is the predominant form of post-translational modification for proteins and lipids, and glycosylationcan be classified into N-glycosylation, O-glycosylation, glycosphingolipid and glycosylphosphatidylinositol (GPI)-anchor glycosylation, according to the attached amino acids or residues. N-linked glycosylation refers to the covalent attachment of a glycan to nitrogen atom of asparagine residues whileas O-linked glycosylation linked to the hydroxyl group of side chains from serine or threonine residues [1]. The N-glycans are initially synthesized as a branched structure on a lipid anchor in the endoplasmic reticulum (ER), and are critically involved in the processes such as protein folding, trafficking, and signal transduction. Therefore, glycosylationcan plays an important role in various physiological phenomena such as development, immune recognition, and neurogenesis [1, 2].

Congenital disorders of glycosylation (CDG), are a clinically and genetically heterogeneous group of inherited metabolic disorders caused by defects in glycosylation pathways [3, 4]. Glycosylation is ubiquitous, occurring in every cell and organism, with over half of all cellular proteins undergoing glycosylation. Consequently, CDG often present highly diverse clinical presentations and involve multisystems [5, 6]. Typical clinical features of CDG include neurological and developmental disabilities, such as psychomotor retardation/cognitive impairment, epilepsy, hypotonia, and ataxia. Other systems or organs, including hepatic, gastrointestinal, hormonal, and immune systems as well as the heart, eyes, and skeleton, are usually affected. In addition, the severity of symptoms ranges widely, from perinatal death or miscarriage to mild adult-onset forms [7]. Glycosylation is a multicompartmental process occurring in the ER, with the help of precursors synthesized in the cytoplasm, followed by trimming and maturation in the Golgi apparatus, and the process requires lots of glycosyltransferases and glycosidases [8]. According to the mechanism and location of molecular defects, CDG are grouped into disorders of N-linked glycosylation (i.e. PMM2-CDG, ALG12-CDG, SSR4-CDG), disorders of GPI biosynthesis (i.e. PIGN-CDG, PIGA-CDG), disorders of Golgi transport (i.e. SLC35A2-CDG, SLC35C1-CDG), disorders of Golgi trafficking (i.e. COG6-CDG, COG4-CDG), etc [9]. More than half of CDG are due to defects in the N-glycosylation pathway, the remaining ones affecting O-glycosylation pathway or multiple glycosylation pathways. Serum transferrin and apolipoprotein C-III, can be used as a diagnostic biomarkers, for most N-linked glycosylation disorders and mucin core1 O-glycosylation defects [10]. However, these related approaches are labor-intensive and time-consuming, and are unsuitable for screening for all CDG. In addition, serum glycosylation spontaneously or gradually normalizes over time, which makes the diagnosis more challenging (notably in adult individuals) [6]. Therefore, a timely and comprehensive detection method, next-generation sequencing (NGS), has offered great additional value, and rapidly expanded both the discovery of novel glycosylation-related disorders and the unraveling of many unsolved cases.

With the advent of NGS in clinical practice, the number of identified CDG has exponentially increased, to date, over 180 subtypes of CDG have been identified [11]. However, due to the rarity, novelty, clinical heterogeneity of CDG, patients with this disorder often face significant delays in diagnosis and limited treatment options, resulting in poor outcomes. In this study, we included a total of 20 patients whose clinical features and genetic data were consistent with the diagnosis of CDG by retrospectively analyzing all patients who underwent whole exome sequencing (WES) from August 2018 to March 2025 in Wuhan Children’s Hospital. Among these 20 patients, we identified 28 variants in several genes: ALG2 (3 cases), DPM2 (3 cases), PMM2 (3 cases) and ALG13 (2 cases), with the remaining patients carrying variants in COG5, COG6, MOGS, DPM3, ALG1, ALG3, ALG11, SSR4 and SLC35A2, 11 of which are novel and have not been reported. Our study provide valuable insights into CDG genetic diversity and offer foundational clues for future diagnosis and therapeutic advancements.

Materials and methods

Study subjects

As mentioned above, we retrospectively analyzed a total of 4912 patients who were suspected of genetic disorders and underwent WES from August 2018 to March 2025 in Wuhan Children’s Hospital, and identified 20 patients suspected or diagnosed of having CDG, along with their healthy family members. All patients were born into families with healthy parents and exhibited normal at birth. Six families underwent prenatal diagnosis following a confirmed diagnosis. Detailed medical and clinical data, including family history, MRI neuroimaging, electroencephalography (EEG), and other relevant examinations, were collected. The study was approved by the ethics committee of the Institutional Review Board of Wuhan Children’s Hospital, Tongji Medical College, Huazhong University of Science & Technology, with the approval code 2021R060-E1. Written informed consent was secured from all participants whose data are presented individually in accordance with the principles outlined in the Declaration of Helsinki.

Whole-exome sequencing, copy number variation sequencing and data analysis

WES and subsequent data analysis were conducted with the help of the third party medical laboratory (Chigene Lab, Beijing China). For WES, the target DNA fragments were enriched by hybridization to construct an exome library (IDT The xGen Exome Research Panel v1.0). High-throughput sequencing was performed using an Illumina NovaSeq 6000 sequencer. Single-nucleotide and indel variants were identified using the Genome Analysis Toolkit (GATK). Paired-end alignment was performed using the Burrows-Wheeler Aligner (BWA). Moreover, ANNOVAR software annotated and filtered the variants. SNPs and indels were filtered and screened according to sequence depth and variant quality. The variants were annotated using the OMIM, ClinVar and HGMD databases. Candidate variants were confirmed by Sanger sequencing using self-designed primers in the patients and their parents.

To predict the pathogenicity of the variant, we used several bioinformatics software and tools.

At first, variants were preferentially selected for further analysis and validation with their minor allele frequency (MAF) < 0.01 in the ExAc browser (https://exac.broadinstitute.org) and gnomAD (https://gnomad.broadinstitute.org) database. Then, single nucleotide variants (SNV) and short indel candidates were identified. The SIFT utilities were used to forecast the change in protein structure. Conservation of each amino acid change was calculated using PhyloP2 and MutationTaster (http://www.mutationtaster.org/) algorithms were used to predict the effects of variants on protein function. All variants were named according to the guidelines of the Human Genome Variation Society (http://www.hgvs.org/) and were described and classified based on the standard guidelines of American College of Medical Genetics and Genomics (ACMG) [12]. According to the ACMG guidelines, variants are classified as pathogenic, likely pathogenic (LP) and variant uncertain significance (VUS).

The copy number variation sequencing (CNVseq) methodology employed in this study is a low-coverage whole-genome sequencing (WGS) strategy to achieve comprehensive genomic coverage. This approach enables the detection of large-scale CNVs across the genome.

WES detection has its advantages over single nucleotide variations and small insertion/deletion variations (< 20 bp), however, WES does not cover 100% the exome, thus variations within deep intronic and regulatory regions cannot be detected. In addition, WES is not validated for the detection of structural variations (SVs), including inversions and translocations. Due to the deepth of sequencing, low-level mosaicism (< 10%) may be missed.

In vitro splicing assay

To investigate the effect on the splicing process of the COG5 gene, minigene splicing assay was performed based on the pcDNA3.1(+) expression vector. Fragments of COG5 gene, including exon7-exon10 was amplified and connected to construct minigene vectors. HEK 293T cells were cultured in 6-well plates and transfected with wildtype and mutant plasmid using Lipofectamine 3000 (Invitrogen) according to the manufacturer’s instructions. Cells were harvested 48 h later, RNA was extracted from each group using TRIzol reagent, COG5 was amplified by PCR using site specific primers and the PCR products were separated by 2% agarose gel electrophoresis and sequenced on ABI 3500DX equipment.

Genotype-phenotype associations

The novel candidate variants in this study or previously reported variants to cause CDG disease was summarized and reviewed and adapted to the clinical symptoms of each patient. They were also determined using several databases such as Human Gene Mutation Database, PubMed (https://www.ncbi.nlm.nih.gov/pubmed/), and Online Mendelian Inheritance in Man (https://omim.org/) to analysis the association between genotype and phenotype.

Results

Overall characteristics and molecular diagnosis

This study included 20 CDG patients (13 males and 7 females) from 18 non-consanguineous families (Fig. 1; Table 1). Clinical symptoms were presented for 20 cases and cataloged using Human Phenotype Ontology (HPO) terms (Table 1). The overall median age of symptom onset was 6.0 months (range: 1.0–48.0 months), and the median age at diagnosis was 1.45 years (range: 4 months to 20 years). The most common symptoms were developmental delay (100%, 19/19), followed by facial dysmorphia (63.2%, 12/19), seizures(60%, 12/20), abnormal liver function (52.6%, 10/19), and hearing or ophthalmological problems (47.4%, 9/19) (Fig. 2).

Fig. 1Pedigrees of 18 families (20 patients) with CDG. Squares and circles represent males and females, respectively, while diamonds indicate cases of prenatal diagnosis. A slash through a symbol denotes deceased individuals

Pedigrees of 18 families (20 patients) with CDG. Squares and circles represent males and females, respectively, while diamonds indicate cases of prenatal diagnosis. A slash through a symbol denotes deceased individuals

Table 1Clinical manifestation of patients in this cohortFamily NO.Patient NO.SexAge of onsetAge of diagnosisFirst symptomsDysmorphic featuresDevelopmental delayGrowth retardationMicrocephalyHypotoniaRecurrent infectionHepatobiliary abnormalitiesSkeletal abnormalitySeizureHearing abnormality/ OphthalmologicalBrain MRICDG typeOther findingsFamily 1P1M3 M8 MEpileptic seizuresnoYesnonononononoInfantile spasmNoWidth of cerebral sulcusCDG Ii ALG2-CDGserious language delayFamily2P2M4 M4 MEpileptic seizuresnoYesyesnononoElevated in AST, ALTnoInfantile spasmNonoCDG Ii ALG2-CDGserious language delayP3M3 M7 MEpileptic seizuresNANANANANANANANATreatment-resistant epilepsyNANACDG Ii ALG2-CDGDied at 7 MFamily 3P4F4Y11YAbnormal walking postureBow shaped eyebrows, strabismusMild language, motor delayNoNohypertoniaNoNoNoAbnormal EEG, mildStrabismusNOCDG Iu DPM2-CDGCK increased; Inguinal herniaFamily 4P5FNA20YMild intellectual disabilityFull moon face, broad noseYesNoNoNoNoNoNoNonoNACDG Iu DPM2-CDGCK increasedP6F6 M8YEpileptic seizuresAbnormal eyebrows, eye distanceLanguage, motor delayNoNoNoNoNoNoYesNoEnlargement of the ventricular systemCDG Iu DPM2-CDGCK increasedFamily 5P7M1 M16 MGlobal Developmental DelayAbnormal eye distance; flat nose;Global Developmental DelayYesNoyesNoNoNoNoStrabismusnoCDG Ir SSR4-CDGIntrauterine growth retardationFamily 6P8M18 M35 MLanguage, motor delayMicrognathia, Downward tilt of eyesLanguage, motor delayShort stature(<-3sd)YesNoNoElevated in ASTPectus carinatumAbnormal EEGHyperopiaDelayed myelination of the brainCDG IIm SLC35A2-CDGnoFamily 7P9M6 M1YDevelopmental delayStrabismus; Abnormal eyebrowsYesHeight weight(<-3SD)nohypertoniaNoElevated in AST, ALTnoYESStrabismus, NystagmusCerebellar atrophyCDG Ia PMM2-CDGdecreased myodynamia; Decreased antithrombinFamily 8P10M3 M6 MDevelopmental delayNoLanguage, motor delayNANoNoYesElevated in AST, ALTNoAbnormal EEG; Febrile convulsionStrabismusCerebellar atrophyCDG Ia PMM2-CDGactivated partial thromboplastin time increasedFamily 9P11M17 M19 MDevelopmental delayNoLanguage, motor delayNANoNoNoElevated in AST, ALTNoNoNoCerebellar atrophyCDG Ia PMM2-CDGactivated partial thromboplastin time increasedFamily 10P12F6 M14 MEpileptic seizuresNoYesNoNoNoYesNoNoInfantile spasmNoWidening of bilateral lateral fissuresxtra cranial spaceALG13-CDGnoFamily 11P13M17 M19 MEpileptic seizuresNoLanguage, motor delayNonoNoNoIncreased r-GTNoInfantile spasmNoRight choroidal cystALG13-CDGnoFamily 12P14M7 M3Y8MMotor delayAbnormal eye distance; flat noseYesYesnoMildNAElevated in AST, ALTScoliosisNAMyopia nystagmusCerebellar atrophyCDG IIi COG5-CDGSmall penis and testiclesFamily 13P15F8 M20 MPsychomotor delayNarrow forehead, hypertelorism, bulbous noseYesYesYesNAYesCholestasis, abnormal liver functionAdducted thumbs and camptodactylyAbnormal EEGNoLesions of bilateral temporal, occipital, and parietal lobesCDG IIl COG6-CDGatrial septal defectFamily 14P16M4 M1YEpileptic seizuresMicrognathia; abnormal eye distanceYesYesYeshypertoniaNoElevated in AST, ALTNoTreatment-resistant epilepsynoSmall volume of white matterCDG Ik ALG1-CDGDied at 14 MFamily 15P17F1 M4 MEpileptic seizuresFull moon face, broad nose, narrow foreheadYesYesnoYesLow levels of IgG, IgA, IgMHepatomegaly, elevated hepatic enzymesnoInfantile spasmLow vision, small palpebral fissureThin corpus callosum; widecerebral sulcusCDG IIb MOGS-CDGHeart PFO/ASDFamily 16P18M3 M4 MEpileptic seizuresNarrow forehead, downturned corners of the mouth, and a bulbous noseYesYesYeshypertoniaNononoepileptic encephalopathyHearing lossreduced frontal volumeCDG Ip ALG11-CDGabsent speechFamily 17P19MNA11YLanguage delay; unilateral hearing lossCongenital facial paralysis, low-set eyesLanguage, motor delayNoNoNANoNoNoyesAmblyopia in both eyes, hearing lossEnlargement of the ventricular systemCDG Id ALG3-CDGunilateral facial palsyFamily 18P20F6 M7YDevelopmental delay; hypotoniaNoYesNoNoyesNoNoNoYesnocerebral white matter demyelinationDPM3-CDGCK increasedNA: Not Available

Table 1Clinical manifestation of patients in this cohortFamily NO.Patient NO.SexAge of onsetAge of diagnosisFirst symptomsDysmorphic featuresDevelopmental delayGrowth retardationMicrocephalyHypotoniaRecurrent infectionHepatobiliary abnormalitiesSkeletal abnormalitySeizureHearing abnormality/ OphthalmologicalBrain MRICDG typeOther findingsFamily 1P1M3 M8 MEpileptic seizuresnoYesnonononononoInfantile spasmNoWidth of cerebral sulcusCDG Ii ALG2-CDGserious language delayFamily2P2M4 M4 MEpileptic seizuresnoYesyesnononoElevated in AST, ALTnoInfantile spasmNonoCDG Ii ALG2-CDGserious language delayP3M3 M7 MEpileptic seizuresNANANANANANANANATreatment-resistant epilepsyNANACDG Ii ALG2-CDGDied at 7 MFamily 3P4F4Y11YAbnormal walking postureBow shaped eyebrows, strabismusMild language, motor delayNoNohypertoniaNoNoNoAbnormal EEG, mildStrabismusNOCDG Iu DPM2-CDGCK increased; Inguinal herniaFamily 4P5FNA20YMild intellectual disabilityFull moon face, broad noseYesNoNoNoNoNoNoNonoNACDG Iu DPM2-CDGCK increasedP6F6 M8YEpileptic seizuresAbnormal eyebrows, eye distanceLanguage, motor delayNoNoNoNoNoNoYesNoEnlargement of the ventricular systemCDG Iu DPM2-CDGCK increasedFamily 5P7M1 M16 MGlobal Developmental DelayAbnormal eye distance; flat nose;Global Developmental DelayYesNoyesNoNoNoNoStrabismusnoCDG Ir SSR4-CDGIntrauterine growth retardationFamily 6P8M18 M35 MLanguage, motor delayMicrognathia, Downward tilt of eyesLanguage, motor delayShort stature(<-3sd)YesNoNoElevated in ASTPectus carinatumAbnormal EEGHyperopiaDelayed myelination of the brainCDG IIm SLC35A2-CDGnoFamily 7P9M6 M1YDevelopmental delayStrabismus; Abnormal eyebrowsYesHeight weight(<-3SD)nohypertoniaNoElevated in AST, ALTnoYESStrabismus, NystagmusCerebellar atrophyCDG Ia PMM2-CDGdecreased myodynamia; Decreased antithrombinFamily 8P10M3 M6 MDevelopmental delayNoLanguage, motor delayNANoNoYesElevated in AST, ALTNoAbnormal EEG; Febrile convulsionStrabismusCerebellar atrophyCDG Ia PMM2-CDGactivated partial thromboplastin time increasedFamily 9P11M17 M19 MDevelopmental delayNoLanguage, motor delayNANoNoNoElevated in AST, ALTNoNoNoCerebellar atrophyCDG Ia PMM2-CDGactivated partial thromboplastin time increasedFamily 10P12F6 M14 MEpileptic seizuresNoYesNoNoNoYesNoNoInfantile spasmNoWidening of bilateral lateral fissuresxtra cranial spaceALG13-CDGnoFamily 11P13M17 M19 MEpileptic seizuresNoLanguage, motor delayNonoNoNoIncreased r-GTNoInfantile spasmNoRight choroidal cystALG13-CDGnoFamily 12P14M7 M3Y8MMotor delayAbnormal eye distance; flat noseYesYesnoMildNAElevated in AST, ALTScoliosisNAMyopia nystagmusCerebellar atrophyCDG IIi COG5-CDGSmall penis and testiclesFamily 13P15F8 M20 MPsychomotor delayNarrow forehead, hypertelorism, bulbous noseYesYesYesNAYesCholestasis, abnormal liver functionAdducted thumbs and camptodactylyAbnormal EEGNoLesions of bilateral temporal, occipital, and parietal lobesCDG IIl COG6-CDGatrial septal defectFamily 14P16M4 M1YEpileptic seizuresMicrognathia; abnormal eye distanceYesYesYeshypertoniaNoElevated in AST, ALTNoTreatment-resistant epilepsynoSmall volume of white matterCDG Ik ALG1-CDGDied at 14 MFamily 15P17F1 M4 MEpileptic seizuresFull moon face, broad nose, narrow foreheadYesYesnoYesLow levels of IgG, IgA, IgMHepatomegaly, elevated hepatic enzymesnoInfantile spasmLow vision, small palpebral fissureThin corpus callosum; widecerebral sulcusCDG IIb MOGS-CDGHeart PFO/ASDFamily 16P18M3 M4 MEpileptic seizuresNarrow forehead, downturned corners of the mouth, and a bulbous noseYesYesYeshypertoniaNononoepileptic encephalopathyHearing lossreduced frontal volumeCDG Ip ALG11-CDGabsent speechFamily 17P19MNA11YLanguage delay; unilateral hearing lossCongenital facial paralysis, low-set eyesLanguage, motor delayNoNoNANoNoNoyesAmblyopia in both eyes, hearing lossEnlargement of the ventricular systemCDG Id ALG3-CDGunilateral facial palsyFamily 18P20F6 M7YDevelopmental delay; hypotoniaNoYesNoNoyesNoNoNoYesnocerebral white matter demyelinationDPM3-CDGCK increasedNA: Not Available

Clinical manifestation of patients in this cohort

Thin corpus callosum; wide

cerebral sulcus

NA: Not Available

Fig. 2Proportion of individuals presenting with specific symptoms at diagnosis in our cohort

Proportion of individuals presenting with specific symptoms at diagnosis in our cohort

By using WES, we identified 28 disease-causing variants in 13 genes among the 20 patients. Among these variants, 17 previously reported and 11 novel variants. A summary of detected variants is provided in Table 2.

Table 2Variant information of disease-causing genes was detected in the studyFamilyNo.Gene and transcriptVariantProtein changers IDSIFTPolyphenCADD scoreREVELgnomADACMG classACMG criteriaReferencesFamily 1P1ALG2 NM_033087c.422 A > Tp.D141VNAdamaging0.99928.30.9456.20E-07LPPM2-supporting, PP3-strong, PP4This workc.751 C > Tp.R251Crs568029420tolerated0.97325.70.4910.000002LPPM2-supporting, PP4,PM3,PP1,PP530,945,278Family 2P2ALG2 NM_033087c.935_937delp.L312delrs757232575NANANANANALPPM2-supporting, PP4,PM4-supporting, PP1,PM330,945,278c.751 C > Tp.R251Crs568029420tolerated0.97325.70.4910.000002LPPM2-supporting, PP4,PM3,PP1,PP530,945,278P3ALG2 NM_033087c.935_937delp.L312delrs757232575NANANANANALPPM2-supporting, PP4,PM4-supporting, PP1,PM330,945,278c.751 C > Tp.R251Crs568029420tolerated0.97325.70.4910.000002LPPM2-supporting, PP4,PM3,PP1,PP530,945,278Family 3P4DPM2 NM_003863c.176T > G(homo)p.L59RNAdamaging0.99924.80.656NAVUSPM2-supporting, PP3This workFamily 4P5DPM2 NM_003863c.197G > A (homo)p.G66ENAdamaging0.999320.837NALPPM2-supporting, PP3-moderate, PS3,PP437,152,991P6DPM2 NM_003863c.197G > A (homo)p.G66ENAdamaging0.999320.837NALPPM2-supporting, PP3-moderate, PS3,PP437,152,991Family 5P7SSR4 NM_006280Loss exon4-6NANANANANANANAPPVS1, PM2-supporting, PP4This workFamily 6P8SLC35A2 NM_005660c.902 C > Tp.T301INAdamaging0.88723.40.4480VUSPM2-supportingThis workFamily 7P9PMM2 NM_000303c.430T > Cp.F144Lrs150719105damaging0.98824.90.9560.0002PPM2-supporting, PS1, PM5, PM3_Strong, PP3_Strong34,671,977c.308G > Tp.C103FNAdamaging0.999320.977NALPPM3,PM2-supporting, PP3-strong, PP411,058,895Family 8P10PMM2 NM_000303c.140 C > Tp.S47Lrs138306798damaging0.995340.8880.00004LPPM3,PM2-supporting, PP3-modetare, PP433,413,482c.406T > Gp.C136GNAdamaging0.98326.60.966NALPPM2-supporting, PM3,PP3-strongThis workFamily 9P11PMM2 NM_000303c.430T > Cp.F144Lrs150719105damaging0.98824.90.9560.0002PPM2-supporting, PS1, PM5, PM3_VeryStrong, PP3_Strong34,671,977c.391 C > Gp.P131Ars1274547742damaging0.97526.20.905NALPPM3.PM2-supporting, PP3-moderate, PM59,140,401Family 10P12ALG13 NM_001099922c.320 A > Gp.N107Srs398122394damaging0.85824.20.277NALPPS1,PM5,PM2-supporting, PP436,930,724Family 11P13ALG13 NM_001257231c.999G > Cp.K333NNAdamaging0.553250.06NAVUSPM3.PM2-supportingThis workFamily 12P14COG5 NM_001161520c.1039 C > Tp.Q347*NANANA41NANAPPVS1, PM2-supporting, PP4This workc.928 + 3 A > GNANANANANANA0.00001LPPS3,PM2-supporting, PM3,PP4This workFamily 13P15COG6 NM_020751c.1843 C > Tp.Q615*NANANA47NANAPPVS1, PM2-supporting, PP434,331,832c.428G > Tp.S143INAdamaging0.269240.119NAVUSPM2-supporting, PM3, PP434,331,832Family 14P16ALG1 NM_019109c.328 C > Ap.Q110KNAdamaging0.93924.90.75NALPPM2-supporting, PP3,PM3,PP4,PP535,715,422c.863–2 A > GNANANANANANA0.00003PPVS1, PM2-supporting, PP4This workFamily 15P17MOGS NM_006302c.1694G > Ap.R565Qrs777861500damaging0.9987.5270.2550.000019LPPM2-supporting, PP4,PS331,925,597c.1619G > Ap.R540HNAdamaging0.999320.4460.000008LPPM2-supporting, PP4,PS331,925,597Family 16P18ALG11 NM_001004127c.1403G > Ap.R468Hrs1204420316damaging1340.8460.000008PPM2-supporting, PM5,PS3,PP4,PP3-modetare39,260,222c.1307G > Tp.G436VNAdamaging128.70.917NAPPM3,PM2-supporting, PP3-modetare, PP4,PS339,260,222Family 17P19ALG3 NM_005787c.716 C > Tp.A239VNAdamaging1330.80VUSPM2-supporting, PP3-moderate, PP4This workc.845 C > Tp.A282VNAtolerated0.38622.70.2580.0009VUSPM2-supporting, PP3-moderate, PP4This workFamily 18P20DPM3 NM_018973c.124 C > Gp.P42Ars745692004damaging129.60.9310.000007LPPM3.PM2-supporting, PP3-moderate, PP431,469,168c.254T > Ap.L85*rs121908155NANA42NANAPPVS1, PM2-supporting, PP431,469,168

Table 2Variant information of disease-causing genes was detected in the studyFamilyNo.Gene and transcriptVariantProtein changers IDSIFTPolyphenCADD scoreREVELgnomADACMG classACMG criteriaReferencesFamily 1P1ALG2 NM_033087c.422 A > Tp.D141VNAdamaging0.99928.30.9456.20E-07LPPM2-supporting, PP3-strong, PP4This workc.751 C > Tp.R251Crs568029420tolerated0.97325.70.4910.000002LPPM2-supporting, PP4,PM3,PP1,PP530,945,278Family 2P2ALG2 NM_033087c.935_937delp.L312delrs757232575NANANANANALPPM2-supporting, PP4,PM4-supporting, PP1,PM330,945,278c.751 C > Tp.R251Crs568029420tolerated0.97325.70.4910.000002LPPM2-supporting, PP4,PM3,PP1,PP530,945,278P3ALG2 NM_033087c.935_937delp.L312delrs757232575NANANANANALPPM2-supporting, PP4,PM4-supporting, PP1,PM330,945,278c.751 C > Tp.R251Crs568029420tolerated0.97325.70.4910.000002LPPM2-supporting, PP4,PM3,PP1,PP530,945,278Family 3P4DPM2 NM_003863c.176T > G(homo)p.L59RNAdamaging0.99924.80.656NAVUSPM2-supporting, PP3This workFamily 4P5DPM2 NM_003863c.197G > A (homo)p.G66ENAdamaging0.999320.837NALPPM2-supporting, PP3-moderate, PS3,PP437,152,991P6DPM2 NM_003863c.197G > A (homo)p.G66ENAdamaging0.999320.837NALPPM2-supporting, PP3-moderate, PS3,PP437,152,991Family 5P7SSR4 NM_006280Loss exon4-6NANANANANANANAPPVS1, PM2-supporting, PP4This workFamily 6P8SLC35A2 NM_005660c.902 C > Tp.T301INAdamaging0.88723.40.4480VUSPM2-supportingThis workFamily 7P9PMM2 NM_000303c.430T > Cp.F144Lrs150719105damaging0.98824.90.9560.0002PPM2-supporting, PS1, PM5, PM3_Strong, PP3_Strong34,671,977c.308G > Tp.C103FNAdamaging0.999320.977NALPPM3,PM2-supporting, PP3-strong, PP411,058,895Family 8P10PMM2 NM_000303c.140 C > Tp.S47Lrs138306798damaging0.995340.8880.00004LPPM3,PM2-supporting, PP3-modetare, PP433,413,482c.406T > Gp.C136GNAdamaging0.98326.60.966NALPPM2-supporting, PM3,PP3-strongThis workFamily 9P11PMM2 NM_000303c.430T > Cp.F144Lrs150719105damaging0.98824.90.9560.0002PPM2-supporting, PS1, PM5, PM3_VeryStrong, PP3_Strong34,671,977c.391 C > Gp.P131Ars1274547742damaging0.97526.20.905NALPPM3.PM2-supporting, PP3-moderate, PM59,140,401Family 10P12ALG13 NM_001099922c.320 A > Gp.N107Srs398122394damaging0.85824.20.277NALPPS1,PM5,PM2-supporting, PP436,930,724Family 11P13ALG13 NM_001257231c.999G > Cp.K333NNAdamaging0.553250.06NAVUSPM3.PM2-supportingThis workFamily 12P14COG5 NM_001161520c.1039 C > Tp.Q347*NANANA41NANAPPVS1, PM2-supporting, PP4This workc.928 + 3 A > GNANANANANANA0.00001LPPS3,PM2-supporting, PM3,PP4This workFamily 13P15COG6 NM_020751c.1843 C > Tp.Q615*NANANA47NANAPPVS1, PM2-supporting, PP434,331,832c.428G > Tp.S143INAdamaging0.269240.119NAVUSPM2-supporting, PM3, PP434,331,832Family 14P16ALG1 NM_019109c.328 C > Ap.Q110KNAdamaging0.93924.90.75NALPPM2-supporting, PP3,PM3,PP4,PP535,715,422c.863–2 A > GNANANANANANA0.00003PPVS1, PM2-supporting, PP4This workFamily 15P17MOGS NM_006302c.1694G > Ap.R565Qrs777861500damaging0.9987.5270.2550.000019LPPM2-supporting, PP4,PS331,925,597c.1619G > Ap.R540HNAdamaging0.999320.4460.000008LPPM2-supporting, PP4,PS331,925,597Family 16P18ALG11 NM_001004127c.1403G > Ap.R468Hrs1204420316damaging1340.8460.000008PPM2-supporting, PM5,PS3,PP4,PP3-modetare39,260,222c.1307G > Tp.G436VNAdamaging128.70.917NAPPM3,PM2-supporting, PP3-modetare, PP4,PS339,260,222Family 17P19ALG3 NM_005787c.716 C > Tp.A239VNAdamaging1330.80VUSPM2-supporting, PP3-moderate, PP4This workc.845 C > Tp.A282VNAtolerated0.38622.70.2580.0009VUSPM2-supporting, PP3-moderate, PP4This workFamily 18P20DPM3 NM_018973c.124 C > Gp.P42Ars745692004damaging129.60.9310.000007LPPM3.PM2-supporting, PP3-moderate, PP431,469,168c.254T > Ap.L85*rs121908155NANA42NANAPPVS1, PM2-supporting, PP431,469,168

Variant information of disease-causing genes was detected in the study

c.176T > G

(homo)

The patients were followed from 6 months to 5 years in outpatient of Wuhan children’s hospital. During this time, two patients died due to recurrent intractable epilepsy associated with their diseases (P3 and P16). Prenatal diagnosis was performed in six families (F4, F7, F10, F13, F14 and F15) using amniotic fluid cells.

Patients with variants in the ALG2 gene

Variants in the ALG2 gene (NM_003087) are associated with ALG2-CDG, also known as CDG Ii (OMIM#607906), which is characterized by neurological symptoms such as convulsive syndrome, epilepsy, axial hypotonia, mental and motor regression [13]. Additionally, ALG2 variants can cause congenital myasthenic syndrome type 14 (ALG2-CMS, OMIM#616228), which presents within the first decade of life, progresses slowly [14].

In this study, three patients (P1, P2 and P3) from 2 families with compound heterozygous ALG2 variants were included. The initial symptom of these 3 patients was epileptic seizures. All of them exhibited seizures, mental and developmental regression, severe speech defects, and abnormal liver function. Genetic analysis identified compound heterozygous variants (c.422 A > T (p.D141V) and c.751 C > T (p.R251C) in P1, and variants (c.751 C > T and c.935_937del (p.L312del)) in P2 and P3. According to standard guidelines of ACMG, c.751 C > T was classified as LP, while c.422 A > T and c.935_937del as LP. The potential effects of these variants were predicted using a three-dimensional structural model, indicating Protein stability being reduced compared to the wild type (lower panel in Fig. 3). Based on typical clinical manifestations and genetic findings, these patients were diagnosed with ALG2-CDG.

Fig. 3(A) Distribution of ALG2 variants and genotype-phenotype analysis. Variants associated with CDG are highlighted in red, while those linked to CMS are marked in black. IM: intramembrane. (B) Predicted effects of ALG2 variants identified in this study on protein stability using a three-dimensional structural model. WT: wild type

(A) Distribution of ALG2 variants and genotype-phenotype analysis. Variants associated with CDG are highlighted in red, while those linked to CMS are marked in black. IM: intramembrane. (B) Predicted effects of ALG2 variants identified in this study on protein stability using a three-dimensional structural model. WT: wild type

Patients with variants in the DPM2 gene

DPM2-CDG (OMIM#615042) is an extremely rare subtype of CDG. In this study, P4 presented with seizures, followed by neurological and motor developmental delay, and elevated creatine kinase (CK) levels. Genetic analysis identified a homozygous variant c.176T >G (p.L59R) in exon 4 of DPM2 (NM_003863.4), inherited from heterozygous parents. According to standard guidelines of ACMG, this variant was classified as VUS. Two additional cases, P5 and P6, had been previously reported [15]. Together with our case in this study, only seven cases have been identified [16, 17].

Patients with variants in SSR4 gene

SSR4-CDG is a rare and relatively mild subtype of CDG (also known as CDG Iy), predominantly affecting males [18]. Key clinical features include developmental delay, speech delay, intellectual disability, muscular hypotonia, microcephaly, skeletal abnormalities, and distinct facial features [19]. In this study, we identified a de novo hemizygous deletion spanning exons 4 to 6 of SSR4 in P7, a 16-month-old boy who had healthy parents, and the deletion was classified as pathogenic. He was born at 39 weeks following a pregnancy complicated by intrauterine growth retardation, and present with global developmental delay, speech delay, cognitive impairment, hypotonia, febrile seizures, feeding difficulty, and mild facial dysmorphisms (large ears, smooth and long philtrum, abnormal eye distance, prominent nose). A literature review revealed 11 point variants and 7 fragment deletions in SSR4 from 23 SSR4-CDG patients(Fig. 4B). High degree of phenotypic heterogeneity across individuals was found and no clear genotype-phenotype correlation has been established.

Fig. 4(A) Distribution of DPM2 variants and genotype-phenotype analysis. Patients with variants highlighted in red exhibited milder symptoms, while those with variants in black experienced severe symptoms and died before age 3 years. (B) Distribution of SSR4 variants, including fragment deletions. TM: transmembrane

(A) Distribution of DPM2 variants and genotype-phenotype analysis. Patients with variants highlighted in red exhibited milder symptoms, while those with variants in black experienced severe symptoms and died before age 3 years. (B) Distribution of SSR4 variants, including fragment deletions. TM: transmembrane

Patients with variants in SLC35A2 gene

SLC35A2-CDG, an X-linked dominant disorder, typically manifests during infancy. The condition is characterized by microcephaly, psychomotor development delay, epileptic seizures, and hypotonia [20]. Male patients usually carry mosaic variants [21]. A hemizygous variant (c.902 C >T, p.T301I) in was SLC35A2 was identified in P8. He was a 35-month-old boy with short stature (<-3sd), mild language and motor developmental delay, mild facial dysmorphisms, pectus carinatum, and elevated aspartate transaminase (AST) levels. According to the ACMG guidelines and clinical manifestations, this variant was classified as VUS. A literature review identified four male cases with hemizygous variants in SLC35A2 [22–24], and their clinical features included short stature, abnormal liver function, language delay, intellectual disability, mild facial abnormalities, skeletal deformities and epileptic seizures [23]. In addition, male patients with mosaic SLC35A2 variants were reported to exhibit more severe symptoms, including epileptic encephalopathy or drug-resistant focal epilepsy [25, 26]. Based on these findings, we propose that male patients with hemizygous SLC35A2 variants generally present with milder phenotypes compared to those with mosaic variants.

Patients with variants in the PMM2 gene

PMM2-CDG (CDG Ia, OMIM#212065) is the most common CDG with more than 1000 cases recorded worldwide [27, 28]. Three patients (P9, P10, and P11) presented with developmental delay and hypotonia. They also had abnormal liver functions, evidenced by elevated AST and alanine aminotransferase (ALT). Brain MRI showed cerebellar atrophy and white matter demyelination. Compound heterozygous variants in PMM2 (NM_000303.3) were identified in all three cases. The variant c.406T >G (p.C136G) in P10 was first reported in this work and classified as LP (PM2-supporting, PM3, PP3-strong) according to ACMG guidelines.

Patients with variants in the ALG13 gene

ALG13-CDG (OMIM#300884) is a rare X-linked CDG, which affects the N-linked glycosylation pathway [29]. Affected individuals typically exhibit refractory seizures, psychomotor development delay, poor or absent speech, hypotonia, facial dysmorphism, and intellectual disability [30]. Epileptic spasm is a common presenting symptom of ALG13-CDG.

Two unrelated patients (P12 and P13) with infantile spasms associated with hypsarrhythmia on EEG as the initial symptom was identified (Table 1). Both patients currently exhibit delayed language and motor development. In addition, P16 has the characteristic of recurrent infections. Brain MRI showed a right choroidalcyst and widening of bilateral lateral fissures intracranial space. A de novo heterozygous variant c.320 A > G (p.N107S) in ALG13 (NM_001099922.3), a known hotspot pathogenic variant, was identified in P12. In P13, a hemizygous variant (c.999G > C, p.K333N) was found and classified as a VUS, inherited from the patient’s unaffected mother.

Patients with variants in COG5 and COG6 genes

The COG (conserved oligomeric Golgi) complex is composed of 8 subunits, including lobe A (from subunit COG1 to COG4) and lobe B (from subunit COG5 to COG8) [31]. Genetic defect in all subunits disrupts the function of the COG complex and have been reported to be associated with different types of CDG. In our study, two patients (P14 and P15) had genetic variants in COG5 (c.1039 C >T (p.Q347*), c.928 + 3 A >G) and COG6 (c.1843 C >T (p.Q615*), c.428G >T(p.S143I)), respectively, the patient P15 had been reported before [32]. Their clinical phenotypes were shown in Table 1 and similar to previously reported cases.

Minigene splicing assay was performed to investigate the effect of the variant (c.928 + 3 A > G) in patient 14 on the splicing of COG5. Wild-type and variant plasmids were transfected in 293T cells and the result demonstrated that the splice variant (c.928 + 3 A > G) resulted in exon skipping (Fig. 5), and this variant was classified as LP according to the ACMG guidelines.

Fig. 5Minigene splicing assay analyses of COG5 variant c.928 + 3 A > G. (A) Pattern map of splice mutation location. (B) Agarose gel images of RT-PCR products for HEK293T cells and Sanger sequencing of the RT-PCR products from each clone. This variant resulted in a smaller amplifed fragment due to exon 8 skipping. M, DNA marker, WT, wild-type; Mut, mutation

Minigene splicing assay analyses of COG5 variant c.928 + 3 A > G. (A) Pattern map of splice mutation location. (B) Agarose gel images of RT-PCR products for HEK293T cells and Sanger sequencing of the RT-PCR products from each clone. This variant resulted in a smaller amplifed fragment due to exon 8 skipping. M, DNA marker, WT, wild-type; Mut, mutation

Patients with variants in ALG1 gene

ALG1-CDG is an autosomal recessive disease with severe multiorgan involvement (OMIM 608540) caused by pathogenic variants in ALG1 [33]. The most common phenotypic manifestations are developmental delay, intellectual disability, failure to thrive, hypotonia, and epilepsy. P16, 1-year-boy, had compound heterozygous variants in ALG1 gene (NM_019109, c.328 C >A(p.Q110K), and c.863–2 A >G), and these two variants were classified as LP and P according to ACMG guidelines. This boy had drug-resistant epilepsy, facial dysmorphisms and abnormal liver function and died at 14 month old. Prenatal diagnosis was performed for this family when his mother was pregnant and they got a healthy baby.

Discussion

Despite global recognition, CDG is under reported in China, with limited case studies available. This is first comprehensive report from China to analyze 20 diagnosed CDG cases, highlighting clinical manifestations such as seizures, motor delays, facial dysmorphism, and growth retardation, supported by findings like abnormal MRI neuroimaging (e.g., cerebellar atrophy), elevated CK levels, and abnormal liver function. This study underscores the importance of integrating genomic tools and systematic evaluations to improve CDG diagnosis and characterization.

A recent large cohort study of 280 individuals with genetic data consistent with a CDG diagnosis found that (s) developmental delay was the most frequent presenting symptom (77%), followed by hypotonia (56%) and feeding difficulties (42%) [34]. Similarly, in our cohort, development delay was the most common initial symptom, observed in all individuals by the time of enrollment, followed by facial dysmorphism (64.7%), abnormal liver function (52.6%), and seizures (50%). The broad differential diagnosis associated with these symptoms, reflecting significant clinical heterogeneity, continues to pose a challenge for timely CDG diagnosis.

We found 28 distinct variants, including 11 novel variants, in 13 different genes: ALG1, ALG2, ALG3, ALG11, ALG13, COG5, COG6, MOGS, DPM2, DPM3, SSR4, SLC35A2, and PMM2. The majority of variants were missense (78.5%), followed by nonsense (10.7%), splice site (7.1%), and frameshift deletions (3.6%) (Table 2). According to ACMG classification guidelines, 8 variants were classified as pathogenic, 14 as LP, and 6 as VUS. Autosomal recessive inheritance was observed in 80% (16/20) of cases and 20% (4/20) displayed X-linked inheritance.

The most common variants were in genes associated with disorders of N-linked protein glycosylation, including PMM2 (3 patients from 3 families), ALG2(3 from 2 families), ALG13 (2 patients), ALG1 (1), ALG3 (1), ALG11 (1), MOGS (1), and SSR4 (1), followed by genes causing disorders of dolichol metabolism (DPM2 and DPM3) and disorders of Golgi trafficking and transport (COG5, COG6, and SLC35A2), which is consistent with previous report [34]. This study significantly expands the current mutational spectrum of CDG, and also revealed considerable genetic heterogeneity, reflecting the complex molecular underpinnings of CDG.

In this study, we conducted a correlation analysis between genotype and phenotype by retrieving all previously published cases, including our patients, for these genes associated with uncommon CDGs.

ALG2 variants could develop ALG2-CDG or ALG2-CMS, by reviewing literature, a total of 9 cases of ALG2-CMS and 5 cases of ALG2-CDG were reported from 7 articles [13, 14, 35–39]. Together with our study, 11 variants in ALG2 have been identified. we noticed that variants located before the intramembrane region (N-terminal, before 158aa) are associated with ALG2-CMS, while variants occurring after the intramembrane region are linked to ALG2-CDG, which presents with more severe clinical symptoms (Fig. 3). ALG2 takes part in the fourth and fifth step of lipid-linked oligosaccharide synthesis, with two different mannosyltransferase activities. Specifically, Alg2 adds both α 1,3-and α1,6-mannose onto ManGlcNAc2–Pdol to form the trimannosyl core Man3GlcNAc2-PDol [30]. The Alg2 conserved C-terminal EX7E motif, the N-terminal cytosolic tail, and 3G-rich loop motifs, are essential for these enzymatic activities, both in vitro and in vivo [40]. Variants in these regions impair the enzyme’s function, which likely contributes to the more severe clinical manifestations observed in ALG2-CDG compared to the relatively milder phenotype of CMS caused by ALG2 variants.

DPM2-CDG is an extremely rare form of CDG, and the main clinical manifestations of these patients include hypotonia, intractable epilepsy, muscle damage, elevated serum CK levels, and microcephaly [15]. To date, including our cases, only 7 cases from four families have been identified [15–17]. DPM2 contains two putative transmembrane domains (amino acid residues 11–31 and 49–69) [16]. Variants in the first domain, such as Phe21 and Tyr23 substitutions, impair its ability to bind with DPM1 [41], suggesting that the first transmembrane region is crucial for maintaining the stability of the complex. Our analysis suggested variants affecting the first transmembrane region or disrupting it were associated with more severe symptoms, and three patients with such variants died before the age of 3 years, in contrast, variants in the second transmembrane region resulted in less severe manifestations (Fig. 4A), however, further study with more cases is needed to confirm this correlation.

SSR4-CDG is ultra-rare X linked, comparably mild subtype of CDG, presenting mostly in males. We reported a 16-month-old boy, who had most of the key symptoms of SSR4-CDG, developmental delay, speech delay, cognitive impairment, feeding difficulties, and muscular hypotonia. In addition, this patient had febrile seizure and ocular abnormality. However, he had no obvious microcephaly, and his facial feature is not as typical as described previously, which could be due to the young age.

SLC35A2, located on chromosome Xp11.23, encodes UGT-1 and is part of the SLC35 family of nucleotide sugar transporters. De novo variants in this gene are associated with SLC35A2-CDG, often with epileptic encephalopathy as a prominent feature [20]. Somatic SLC35A2 variants have also been reported in patients with focal seizures and suspected focal cortical dysplasia [21]. Furthermore, recent studies have linked brain somatic variations in SLC35A2 to mild malformation of cortical development with oligo dendroglial hyperplasia in epilepsy [42]. While SLC35A2-CDG typically shows a strong gender bias, with most cases occurring in females, there have been reports of male patients with a milder phenotype, including minor neurological involvement and growth deficiency [23, 24]. By reviewing literature, we postulate that males with hemizygous variants in SLC35A2 may present with a milder phenotype compared to those with mosaic variants in brain, although this requires further research. We slao found that the mosaic variants carried by male patients were mostly frameshift or nonsense, whereas all variants found in five male patients including our patient were missense. Loss of function variants may be tolerated in males when in the somatic state, but likely embryotic lethal in the germline, and this may be athe reason for gender bias in SLC35A2-CDG. Our study has some limitations. First, due to the small number of patients enrolled, we were unable to analyze more specific genotype-phenotype correlations. The rarity of CDG and the limited screening of this condition in China may partly explain this limitation. Second, changes in glycan composition in patient serum were not analyzed by MALDI-TOF MS due to the lack of necessary equipment. In addition, we identified some variants of VUS, functional studies were not conducted. Further research with larger patient cohorts and long-term follow-up would enhance our understanding of these diseases, which is critical for elucidating pathogenic mechanisms, improving the genotype-phenotype correlation, refining genetic counseling, and developing new treatments.

Conclusion

We identified 28 disease-causing variants, including 11 novel variants, using WES in 20 CDG patients. Our findings expand the spectrum of known variants and their related clinical phenotypes in Chinese CDG patients and establish possible genotype-phenotype correlations of several genes.