Recognizable phenotypes in CDG

Abstract

Pattern recognition, using a group of characteristic, or discriminating features, is a powerful tool in metabolic diagnostic. A classic example of this approach is used in biochemical analysis of urine organic acid analysis, where the reporting depends more on the correlation of pertinent positive and negative findings, rather than on the absolute values of specific markers. Similar uses of pattern recognition in the field of biochemical genetics include the interpretation of data obtained by metabolomics, like glycomics, where a recognizable pattern or the presence of a specific glycan sub-fraction can lead to the direct diagnosis of certain types of congenital disorders of glycosylation. Another indispensable tool is the use of clinical pattern recognition–or syndromology–relying on careful phenotyping. While genomics might uncover variants not essential in the final clinical expression of disease, and metabolomics could point to a mixture of primary but also secondary changes in biochemical pathways, phenomics describes the clinically relevant manifestations and the full expression of the disease. In the current review we apply phenomics to the field of congenital disorders of glycosylation, focusing on recognizable differentiating findings in glycosylation disorders, characteristic dysmorphic features and malformations in PMM2-CDG, and overlapping patterns among the currently known glycosylation disorders based on their pathophysiological basis.

Introduction

A phenotype is the set of all observable characteristics of an organism. The etymology of the word stems from the Greek word “phainein”, (to show) and “typos” (type), or “visible type”. Phenotyping is the oldest practice used by physicians to diagnose disease. Syndromology is a specific way of phenotyping, using a group of signs and symptoms that occur together to characterize a particular abnormality or condition. Patients born with an inborn error of metabolism can present with a genetic syndrome, showing a set of concurrent findings that form an identifiable pattern, even if the symptoms vary in severity. Congenital disorders of glycosylation (CDG) belong to a rapidly expanding group of metabolic disorders, frequently present as “syndromes” (Jaeken and Peanne, 2017). In fact, they were initially discovered as a novel disease group based on careful phenotyping of the most frequent form, now known as PMM2-CDG, where patients show dysmorphic facial features–strabismus, inverted nipples, abnormal fat distribution–and a variable combination of endocrine abnormalities and coagulation defects, making the disorder recognizable. Since the description of PMM2-CDG there have been more than 130 CDG subtypes described, and the number of disorders is rapidly growing. A few of these disorders are treatable, therefore early diagnosis is very important (Dorre et al., 2015, Riley et al., 2017). Some subtypes are detectable by transferrin isoform screening or other forms of glycosylation analysis, helping the diagnostic crusade (Buczkowska et al., 2015, Bengston et al., 2016). But is there a specific clinical pattern in CDGs, which could point the clinician towards the correct diagnosis? In the current era of next-generation sequencing, is there a role for the clinician, or will the art of phenotyping succumb to molecular techniques in the –omics era? In this mini-review we structured all subtypes of CDGs described to date according to their biochemical nosology, and then attempt to establish a correspondence between certain subtypes with overlapping phenotypes and their biochemical basis. Finally, we review the existing literature on the phenotypic features associated with PMM2-CDG, and present data on computer-aided phenotyping by means of facial recognition analysis. We propose that the art of phenomics, either as performed by clinicians or with the help of currently-available technologies, continues to be very much alive, and rather than being supplanted by other –omics, it’s complementary to these.

Methods

A literature review was performed for each individual disorder of glycosylation, and a table was created with all subtypes of glycosylation disorders described to date (see Table 1). Each subtype of CDG was individually evaluated, and a concise summary was performed on those disorders that were felt to have clinical, imaging or laboratory findings that allow recognition and distinction from other disorders.

Table 1.

Current list of subtypes of congenital disorders of glycosylation

Disease	Defective gene				Gene locus OMIM#		Defective protein	Localization of defect	Inheritance			Paper with initial molecular characteriza tion (PMID)
N-linked glycosylation (effects)
Interconversion of monosaccharides)
PMM2-CDG	PMM2				601785		Phosphomannomutase	Cytosol	AR			9140401
MPI-CDG	MPI				154550		Phosphomannose isomerase	Cytosol	AR			9525984
N-glycan lipid-linked oligosaccharide assembly
DPAGT1-CDG; CMS13	DPAGT1				191350		GlcNAc-1-P transferase	ER (cytosolic side)	AR			12872255
ALG13-CDG in males; EEIE36 in females	ALG13				300776		UDP-GlcNAc transferase	ER (cytosolic side)	XL			22492991
CMS15	ALG14				612866		UDP-GlcNAc transferase	ER (cytosolic side)	AR			23404334
ALG1-CDG	ALG1				605907		β1–4 Man-transferase	ER (cytosolic side)	AR			14709599, 14973778, 14973782
ALG2-CDG; CMS14	ALG2				607905		α1–3/6 Man-transferase	ER (cytosolic side)	AR			12684507
ALG11-CDG	ALG11				613666		α1–2 Man-transferase	ER (cytosolic side)	AR			20080937
RFT1-CDG	RFT1				611908		Man5GlcNAc2-PP-Dol flippase	ER	AR			18313027
ALG3-CDG	ALG3				608750		α1–3 Man-transferase	ER (luminal side)	AR			10581255
ALG9-CDG	ALG9				606941		α1–2 Man-transferase	ER (luminal side)	AR			15148656
ALG12-CDG	ALG12				607144		α1–6 Man-transferase	ER (luminal side)	AR			11983712
ALG6-CDG	ALG6				604566		α1–3 Glc-transferase	ER (luminal side)	AR			10359825
ALG8-CDG	ALG8				608103		α1–3 Glc-transferase	ER (luminal side)	AR			12480927
Glycan transfer to nascent protein
TUSC3-CDG (MRT7/MRT22)		TUSC3			601385		OST subunit	ER	AR			18455129, 18452889
DDOST-CDG		DDOST			614507		OST subunit	ER	AR			22305527
SST3A-CDG		STT3A			601134		OST subunit	ER	AR			23842455
SST3B-CDG		STT3B			608605		OST subunit	ER	AR			23842455
SSR4-CDG		SSR4			300090		Translocon-associated protein, delta subunit	ER	AR			24218363
N-glycan processing
MOGS-CDG		MOGS		601336			α1–2 glucosidase I	ER	AR			10788335
Polycystic kidney disease 3		GANAB		104160			α1–3 glucosidaseII subunit alpha	ER	AD			27259053
Polycystic liver disease 1		PRKCSH		177060			α1–3 glucosidase II subunit beta	ER	AD			12529853, 12577059
MAN1B1-CDG (MRT15)		MAN1B1		604346			α1–2 mannosidase I	ER	AR			21763484
MGAT2-CDG		MGAT2		602616			β1–2 GlcNAc-transferase II	Medial-Golgi	AR			8808595
B4GALT1-CDG		B4GALT1		137060			β1–4 Gal-transferase	Trans-Golgi	AR			11901181
O-linked glycosylation defects
O-Man glycosylation
MDDGA1, MDDGB1, MDDGC1		POMT1		607423			Protein O-Man-transferase	ER	AR			12369018
MDDGA2, MDDGB2, MDDGC2		POMT2		607439			Protein O-Man-transferase	ER	AR			15894594
MDDGA3, MDDGB3, MDDGC3, RP76		POMGNT1		606822			β1–2 GlcNAc-transferase	Golgi	AR			11709191
MDDGA8		POMGNT2		614828			β1–4 GlcNAc-transferase	ER	AR			22958903
MDDGA11		B3GALNT2		610194			β1–3 GalNAc-transferase II	ER	AR			23453667
MDDGA12, MDDGC12		POMK		615247			Protein O-mannose kinase	ER	AR			23519211
MDDGA7, MDDGC7		ISPD		614631			Methyl erythritol 4-phosphate cytidylyltransfer ase	Cytosol	AR			22522420, 22522421
MDDGA4, MDDGB4, MDDGC4		FKTN		607440			Fukutin	Golgi	AR			9690476
MDDGA5, MDDGB5, MDDGC5		FKRP		606596			Fukutin-related protein	Golgi	AR			11592034
MDDGA10		TMEM5		605862			β1–4 xylosyltransferase	Golgi	AR			23217329
MDDGA13		B4GAT1		605581			β-1,3 glucuronyltransferase I	Golgi	AR			23359570
MDDGA6, MDDGB6		LARGE1		603590			β1–3 GlcA-transferase/α1–3 Xyl-transferase	Golgi	AR			12966029
O-Xyl glycosylation
Desbuquois dysplasia 2			XYLT1	608124			Xyl-transferase 1	Golgi	AR		23982343
Spondyloocular syndrome			XYLT2				Xyl-transferase 2	Golgi	AR		26027496
Progeroid EDS 1 (Larsen of Reunion Island syndrome)			B4GALT7	604327			β1–4 Gal-transferase I	Golgi	AR		10473568, 10506123
SEMDJL Beighton type (progeroid EDS 2)			B3GALT6	615291			β1–3 Gal-transferase II	Golgi	AR		23664117
Larsen-like syndrome			B3GAT3	606374			β1–3 GlcA-transferase I	Golgi	AR		21763480
MHE type 1			EXT1	608177			β1–4 GlcA-transferase II/α1–4 GlcNAc-transferase II	Golgi	AD		7550340
MHE type 2; Seizures, scoliosis and macrocephaly syndrome			EXT2	608210			β1–4 GlcA-transferase II/α1–4 GlcNAc-transferase II	Golgi	AD; AR		8782816; 26246518
Immunoskeletal dysplasia with neurodevelopmental abnormalities			EXTL3				α1–4 GlcNAc-transferase I and II	Golgi	AR		28132690, 28148688
Temtamy preaxial brachydactyly syndrome			CHSY1	608183			β1–3 GlcA-transferase/β1–4 GalNAc-transferase	Golgi	AR		21129728
SED with congenital joint dislocations (AR Larsen syndrome, SED Omani type, humerospinal dysostosis)			CHST3	603799			GalNAc-6-O-sulfotransferase	Golgi	AR		15215498
EDS musculocontract ural type 1			CHST14	608429			GalNAc-4-O-sulfotransferase	Golgi	AR		20004762
EDS musculocontractural type 2			DSE	605942			Dermatan sulfate epimerase	Golgi	AR		23704329
Desbuquois dysplasia			CSGALNACT1	616615			β1–4 GalNAc-transferase I	Golgi	AR		27599773
Macular corneal dystrophy			CHST6	605294			GlcNAc-6-O-sulfotransferase	Golgi	AR		11017086
Desbuquois dysplasia 1			CANT1	613165			UDP-Galnucleotidase	ER and Golgi	AR		19853239
O-GalNAc glycosylation
Hyperphosphatemic familial tumoral calcinosis			GALNT3	601756			Polypeptide GalNAc-transferase	Golgi	AR		15133511
Tn polyagglutination syndrome			C1GALT1C1	300611			Core 1 β1–3 galactosyltransferase chaperone	ER	Somatic		16251947
O-GlcNAc glycosylation
MRX106		OGT		300255			O-GlcNAc transferase	Nucleus and cytosol	XL		28302723, 28584052
Adams-Oliver syndrome 4		EOGT		614789			EGF-domain O-GlcNAc transferase	ER	AR		23522784
MRT12; EIEE15		ST3GAL3		606494			α2–3 Sia-transferase	Golgi	AR		21907012; 23252400
O-Glc glycosylation
Dowling-Degos disease 4; LGMD2Z		POGLUT1		615618			Protein O-glucosyltransferase	ER	AD; AR		24387993; 27807076
O-Fuc glycosylation
Dowling-Degos disease 2		POFUT1		607491			Protein O-fucosyltransferase	ER	AD		23684010
Spondylocostal dysostosis type 3		LFNG		602576			β1–3 GlcNAc-transferase	Golgi	AR		16385447
Peters-Plus syndrome		B3GALTL		610308			β1–3 Glc-transferase	ER	AR		16909395
GPI biosynthesis defects
MCAHS2 (GPIBD4, EEIE20)			PIGA	311770		GlcNAc-transferase complex, catalytic subunit		ER (cytosolic side)	XL		22305531
MRT62 (GPIBD16)			PIGC	601730		GlcNAc-transferase complex		ER (cytosolic side)	AR		27694521
PIGQ-CDG			PIGQ	605754		GlcNAc-transferase complex		ER (cytosolic side)	AR		24463883
EIEE55 (GPIBD14)			PIGP	605938		GlcNAc-transferasecomplex		ER (cytosolic side)	AR		28334793
HPMRS6 (GPIBD12)			PIGY	610662		GlcNAc-transferase complex		ER (cytosolic side)	AR		26293662
CHIME syndrome (GPIBD5)			PIGL	605947		GlcNAc-PI de-N-acetylase		ER (cytosolic side)	AR		22444671
HPMRS5 (GPIBD11)			PIGW	610275		Inositol acyltransferase		ER (luminal side)	AR		24367057
GPIBD1			PIGM	610273		Man-transferase 1		ER (luminal side)	AR		16767100
HPMRS1 (GPIBD2)			PIGV	610274		Man-transferase 2		ER (luminal side)	AR		20802478
MCAHS1 (GPIBD3)			PIGN	606097		EtNP-transferase 1		ER (luminal side)	AR		21493957
HPMRS2 (GPIBD6)			PIGO	614730		EtNP-transferase 3		ER (luminal side)	AR		22683086
MRT53 (GPIBD13)			PIGG	616918		EtNP-transferase 2		ER (luminal side)	AR		26996948
MCAHS3 (GPIBD7)			PIGT	610272		GPI transamidase		ER (luminal side)	AR		23636107
GPIBD15			GPAA1	603048		GPI anchor attachment protein 1		ER (luminal side)	AR		29100095
MRT42 (GPIBD9)			PGAP1	611655		Inositol deacylase		ER (luminal side)	AR		24784135
HPMRS4 (GPIBD10)			PGAP3	611801		Phospholipase A2		Golgi	AR		24439110
HPMRS3 (GPIBD8)			PGAP2	615187		GPI-anchorremodeling protein		Golgi	AR		23561846, 23561847
Glycolipid glycosylation
Amish infantile epilepsy syndrome (salt and pepper syndrome)			ST3GAL5	604402		α2–3 Sia-transferase		Golgi	AR			15502825
SPG26			B4GALNT1	601873		β1–4 GalNAc-transferase		Golgi	AR			23746551
NOR polyagglutination syndrome			A4GALT	607922		α1–4 Gal transferase		Golgi	AD			22965229
Disorders of multiple pathways
Monosaccharide synthesis
CMS12			GFPT1	138292		Glutamine:F6P amidotransferase		Cytosol	AR			21310273
GNE myopathy; sialuria			GNE	603824		UDP-GlcNAc 2-epimerase/ManNac kinase		Cytosol	AR; AD			11528398; 10330343
SEMD, Camera-Genevieve type			NANS	605202		N-acetylneuraminic acid-9-phosphate synthase		Cytosol	AR			27213289
Monosaccharide interconversion
PGM1-CDG			PGM1	612934		Phosphoglucomutase		Cytosol	AR			19625727
Immunodeficiency 23			PGM3	172100		GlcNAc phosphate mutase		Cytosol	AR			24589341, 24698316, 24931394
Severe congenital neutropenia 4			G6PC3	611045		Glucose-6-phosphatase		ER	AR			19118303
Dolichol biosynthesis
RP59			DHDDS	608172		Cis-isoprenyl transferase		ER (cytosolic side)	AR			21295283
NUS1-CDG			NUS1	610463		Cis-isoprenyl transferase subunit		ER	AR			25066056
SRD5A3-CDG			SRD5A3	611715		Polyprenol reductase		ER	AR			20637498
DOLK-CDG			DOLK	610746		Dolichol kinase		ER	AR			17273964
Dolichol-P-sugar biosynthesis and utilization
DPM1-CDG			DPM1	603503		Dol-P-Man synthase		ER (cytosolic side)		AR		10642597, 10642602
DPM2-CDG			DPM2	603564		Dol-P-Man synthase		ER		AR		23109149
DPM3-CDG			DPM3	605951		Dol-P-Man synthase		ER		AR		19576565
MPDU1-CDG			MPDU1	604041		Dol-P-Man flippase		ER		AR		11733564
Nucleotide-sugar synthesis
CAD-CDG (EEIE50)			CAD	114010		Carbamoyl-phosphate synthetase 2, aspartate transcarbamylase, and dihydroorotase		Cytosol		AR		25678555
Alacrima, achalasia, and mental retardation syndrome			GMPPA	615495		GDP-mannose pyrophosphoryla se α subunit		Cytosol		AR		24035193
MDDGA14, MDDGB14, MDDGC14			GMPPB	615320		GDP-mannose pyrophosphoryla se β subunit		Cytosol		AR		23768512
Transporters
SLC35A1-CDG		SLC35A1		605634		CMP-Sia transport		Golgi		AR		15576474
SLC35A2-CDG (EIEE22)		SLC35A2		314375		UDP-Galactose transport		Golgi		AR		23561849
SLC35A3-CDG		SLC35A3		605632		UDP-GlcNAc transport		Golgi		AR		24031089
SLC35C1-CDG		SLC35C1		605881		GDP-Fuc transport		Golgi		AR		11326279, 11326280
Schneckenbecke n dysplasia		SLC35D1		610804		UDP-GlcA/UDP-GalNAc transporter		Golgi		AR		17952091
SLC39A8-CDG		SLC39A8		608732		Cation transporter		Plasma membrane		AR		26637978, 26637979
Vesicular traffickiilg
COG1-CDG		COG1		606973		Conserved oligomeric Golgi complex subunit 1		Vesicular membrane/cytosol	AR			16537452
COG2-CDG		COG2		606974		Conserved oligomeric Golgi complex subunit 2		Vesicular membrane/cytosol	AR			24784932
COG4-CDG		COG4		606976		Conserved oligomeric Golgi complex subunit 4		Vesicular membrane/cytosol	AR			19494034
COG5-CDG		COG5		606821		Conserved oligomeric Golgi complex subunit 5		Vesicular membrane/cytosol	AR			19690088
COG6-CDG (Shaheen syndrome)		COG6		606977		Conserved oligomeric Golgi complex subunit 6		Vesicular membrane/cytosol	AR			20605848
COG7-CDG		COG7		606978		Conserved oligomeric Golgi complex subunit 7		Vesicular membrane/cytosol	AR			15107842
COG8-CDG		COG8		606979		Conserved oligomeric Golgi complex subunit 8		Vesicular membrane/cytosol	AR			17220172
Autoimmune interstitial lung, joint, and kidney disease		COPA		601924		COP-I subunit alpha		Vesicular membrane/cytosol	AD			25894502
Primary microcephaly		COPB2		606990		COP-I subunit beta-2		Vesicular membrane/cytosol	AR			29036432
Rhizomelic short stature with microcephaly, micrognathia, and developmental disability		ARCN1		600820		COP-I subunit delta		Vesicular membrane/cytosol	AD			27476655
Craniolenticulos utural dysplasia (Boyadjiev-Jabs syndrome)		SEC23A		61051		COPII		Vesicular membrane/cytosol	AR			16980979
Congenital dyserythropoietic anemia type II; Cowden syndrome 7		SEC23B		610512		COPII component		Vesicular membrane/cytosol	AR; AD			19561605, 19621418; 26522472
Cole-Carpenter syndrome 2		SEC24D		607186		COPII component		Vesicular membrane/cytosol	AR			25683121
Chylomicron retention disease		SAR1B		607690		COPII GTPase		Vesicular membrane/cytosol	AR			12692552
Achondrogenesis IA		TRIP11		604505		Golgi microtubule-associated protein 210		cis-Golgi	AR			20089971
SED tarda		TRAPPC2		300202		Subunit of TRAPP tethering complex		Vesicular membrane/cytosol	XL			10431248
TRAPPC6B-CDG		TRAPPC6 B		610397		Subunit of TRAPP tethering complex		Vesicular membrane/cytosol	AR			28626029
MRT13		TRAPPC9		611966		Subunit of TRAPP tethering complex		Vesicular membrane/cytosol	AR			20004763, 20004764, 20004765
LGMD2S		TRAPPC11		614138		Subunit of TRAPP tethering complex		Vesicular membrane/cytosol	AR			23830518
TRAPPC12-CDG		TRAPPC12		614139		Subunit of TRAPP tethering complex		Vesicular membrane/cytosol	AR			28777934
Golgi homeostasis
Autosomal recessive cutis laxa type IIA (wrinkly skin syndrome)		ATP6V0A2		611716		Subunit of vacuolar ATPase		Vacuolar membrane	AR			18157129
Immunodeficiency 47		ATP6AP1		300197		Subunit of vacuolar ATPase		Vacuolar membrane	XL			27231034
X-linked mental retardation, Hedera type; X-linked parkinsonism with spasticity		ATP6AP2		300556		Subunit of vacuolar ATPase		Vacuolar membrane	XL			15746149
Autosomal recessive cutis laxa type IID		ATP6V1A		607027		Subunit of vacuolar ATPase		Vacuolar membrane	AR			28065471
Autosomal recessive cutis laxa type IIC		ATP6V1E1		108746		Subunit of vacuolar ATPase		Vacuolar membrane	AR			28065471
TMEM199-CDG		TMEM199		616815		Assembly factor for vacuolar ATPase		Vacuolar membrane	AR			26833330
CCDC115-CDG		CCDC115		613734		Assembly factor for vacuolar ATPase		Vacuolar membrane	AR			26833332
X-linked myopathy with excessive autophagy		VMA21		300913		Assembly factor for vacuolar ATPase		Vacuolar membrane	XL			23315026
TMEM165-CDG		TMEM165		614726		Ca2+/H+ antiporter		Golgi	AR			22683087
Unknown
Catel-Manzke syndrome		TGDS		616146		TDP-Glc 4,6-dehydratase		Unknown	AR			25480037

Abbreviations: AD, autosomal dominant; AR, autosomal recessive; CMS, congenital myasthenic syndrome; CS, chondroitin sulfate; DS, dermatan sulfate; EDS, Ehlers-Danlos syndrome; EGF, epidermal growth factor-like; EIEE, early infantile epileptic encephalopathy; EtNP, ethanolamine phosphate; F6P, fructose 6-phosphate; GPIBD, glycosylphosphatidylinositol biosynthesis defect; HPMRS, hyperphosphatasia with mental retardation syndrome; HS, heparin sulfate; KS, keratan sulfate; LGMD, limb-girdle muscular dystrophy; MCAHS, multiple congenital anomalies-hypotonia-seizures syndrome MDDG, muscular dystrophy-dystroglycanopathy; MHE, multiple hereditary exostoses; MRT, mental retardation; Neu5Ac, N-acetylneuraminic acid; PI, phosphatidylinositol; OST, oligosaccharyltransferase; RP, retinitis pigmentosa; SED, spondyloepiphyseal dysplasia; SEMDJL, spondyloepimetaphyseal dysplasia with joint laxity; SPG, spastic paraplegia; TSR, thrombospondin type 1 repeats; XL, X-linked.

Next, a literature review was performed on congenital malformations in PMM2-CDG, using a PubMed search. PubMed database search was performed using the following terms: Congenital disorder(s) of glycosylation, type Ia OR PMM2 OR PMM2-CDG OR phosphomannomutase 2 deficiency OR Jaeken syndrome OR CDG-Ia OR carbohydrate deficient glycoprotein syndrome plus Anomaly OR malformation OR Malformation OR dysmorphic. All the publications were included in the search from the time of the disorder discovery, i.e. 1980. Only patients with confirmed PMM2 mutations or enzymatically-confirmed phosphomannomutase 2 deficiency were included.

A total of 165 papers were reviewed. Reported cases were collected and summarized in an Excel table (Supplementary Table 1). In total, 561 patients were reported in the literature with confirmed PMM2-CDG.

Finally, we performed facial recognition analysis using the Facial Dysmorphology Novel Analysis (FDNA, Boston MA, USA) technology, which allows for assessment of subtle craniofacial dysmorphic features and facial pattern recognition. The input consisted of 53 facial frontal photos from 50 unaffected controls, as well as photos 18 patients with confirmed PMM2-CDG. A binary comparison was then performed between groups, and the capacity to distinguish them was assessed by measuring the Area Under the Curve (AUC) of the Receiver Operating Characteristic (ROC) curve, which plots the true positive rate as function of the true negative rate. An AUC of 1 indicates perfect accuracy, while an AUC of 0.5 is the performance obtained by a totally random system (as for example a coin toss). Cross validation was performed by recurrently and randomly splitting the data into training sets, each set contained half of the photos. This random split was repeated 10 times (Lumaka et al. 2017).

Results

Recognizable phenotypes in congenital disorders of glycosylation

(Type 1 serum transferrin N-hypoglycosylation pattern: *, type 2 N-hypoglycosylation pattern: **, O-linked hypoglycosylation#, GPI-anchoring deficiency ##)

MPI-CDG*

MPI-CDG presents with liver disease (hepatomegaly and hepatopathy, cirrhosis), coagulopathy (increased bleeding tendency and risk for thrombotic episodes), hyperinsulinemic hypoglycemia, and gastrointestinal symptoms (chronic diarrhea, protein-losing enteropathy). It is treatable by mannose (200 mg/kg 4–6 times a day). Some patients need liver transplantation. There is normal intellectual development.

PMM2-CDG*

PMM2-CDG is the most common N-glycosylation abnormality. Strabismus, large ears, thin upper lip, inverted nipples, abnormal fat distribution and developmental delay (intellectual disability in most cases) are characteristic. Coagulation abnormalities, thrombotic events, stroke-like episodes, hepatopathy, endocrine abnormalities are not detectable in all patients, although hypogonadism is prevalent. Cerebellar hypoplasia can be present as early as the neonatal period.

GFPT1-CDG, DPAGT1-CDG*, ALG14-CDG, ALG2-CDG *

All of these subtypes can present with disordered neurotransmission in the form of myasthenia. In addition, DPAGT1-CDG and ALG2-CDG can present with a multisystemic disorder.

RFT1-CDG*, ALG11-CDG*

Both subtypes are associated with developmental delay, seizures and intellectual disability, all rather non-specific findings, but unlike other CDG subtypes, they are frequently accompanied by sensorineural hearing loss.

ALG6-CDG* and ALG8-CDG*

Patients with ALG6-CDG have developmental disability, proximal muscle weakness, seizures and in most cases ataxia. Brachydactyly and finger malformations have been described in several patients. ALG8-CDG is also accompanied by brachydactyly, as well as intellectual disability and dysmorphic features including low-set ears, hypertelorism and macroglossia.

ALG3-CDG*, ALG9-CDG* and ALG12-CDG*

The severe form of ALG9-CDG (Gillessen-Kaesbach-Nishimura syndrome) is accompanied by polycystic kidneys, hepatic fibrosis, congenital heart disease, and characteristic skeletal changes including mesomelia, a round pelvis, shortened sacrosciatic notch and ovoid ischia, hypomineralization of skull, cervical vertebral bodies, and pubic rami, and thick occipital bone. Similar skeletal anomalies were described in ALG3-CDG and ALG12-CDG.

MGAT2-CDG**

MGAT2-CDG is a disorder with recognizable facial features (long eyelashes, prominent nasal bridge, underdeveloped alae nasi, low-hanging columella, thin vermillion of the upper lip, thick vermillion of the lower lip). This disease is also characterized by severe growth delay, mental disability with absent speech and in some patients, radio-ulnar synostosis.

MAN1B1-CDG**#

MAN1B1-CDG is a Golgi CDG with intellectual disability and, in most patients, facial dysmorphism (widely spaced eyes with down-slanting palpebral fissures, long ears, underdeveloped naso-labial fold, thin vermillion of the upper lip) and truncal obesity.

SRD5A3-CDG*

SRD5A3-CDG is a syndromal form of a dolicholphosphate synthesis defect; patients have eye malformations (cataract, retinal anomalies, glaucoma, visual loss), cerebellar ataxia and in about one third of the cases ichthyosiform skin anomalies. Developmental disability is mostly severe.

DOLK-CDG*

DOLK-CDG is usually associated with a multisystem phenotype, with neurologic, endocrine and coagulation abnormalities. A discriminative feature is dilated cardiomyopathy. Milder cases benefit from early cardiac transplantation.

DPM1-CDG*, DPM2-CDG* and DPM3-CDG*

All of these subtypes lead to alpha-dystroglycan hypoglycosylation, elevated creatine kinase levels and muscle dystrophy. Unlike other phenotypes of alpha-dystroglycanopathies, patients with either of these three forms of CDG also show abnormal transferrin glycosylation in serum.

COG1-CDG**

COG1-CDG patients show variable phenotypic features and variable severity. Some patients have been described with a costo-cerebro-mandibular-like syndrome.

COG7-CDG**

COG7-CDG is recognizable in most patients based on microcephaly, growth impairment, adducted thumbs, ventricular septal defect and episodes of hyperthermia.

SLC35C1-CDG**

SLC35C1-CDG is a disorder with frequent infections, neutrophilia and distinct facial features including brachycephaly, low insertion of the anterior hairline, coarse features, puffy eyelids, flat nasal bridge, large tongue, long upper lip with everted lower lips. Patients also have a rare blood type, the Bombay blood group. The immune phenotype improves with oral fucose supplementation.

ATP6V0A2-CDG**#

ATP6V0A2-CDG is also called autosomal recessive cutis laxa type 2A (ARCL2A). Generalized congenital sagging skin, cutis laxa, abnormal fat distribution, motor developmental disability, skeletal abnormalities, short stature, and joint luxation are characteristic. Intellectual development can be normal. Cutis laxa improves with age. Some patients have cobblestone-like brain dysgenesis and seizures.

ATP6V1A-CDG and ATP6V1E1-CDG**#

ATP6V1A-CDG and ATP6V1E1-CDG patients are very similar to those diagnosed with ATP6V0A2-CDG. They have generalized cutis laxa, motor and intellectual developmental disability, cardiac involvement and hypercholesterolemia. Cutis laxa improves with age.

ATP6AP1-CDG and ATP6AP2-CDG**#

ATP6AP1-CDG and ATP6AP2-CDG both present with early liver disease (cholestasis, fibrosis, cirrhosis). Frequent infections are mostly associated with hypogammaglobulinemia. Early symptoms include cutis laxa, which disappears with age.

TMEM199-CDG and CCDC115-CDG**#

TMEM199-CDG and CCDC115-CDG also present with early liver disease (cholestasis, fibrosis, cirrhosis) as well as hypercholesterolemia, low serum ceruloplasmin and developmental disability. These conditions can mimic Wilson disease, given the liver disease with hypoceruloplasminemia. Liver transplantation is a therapeutic option.

TMEM165-CDG**#

TMEM165-CDG is a recognizable spondyloepimetaphyseal skeletal dysplasia with extreme osteopenia, short stature, hyperinsulinism, growth hormone deficiency and variable intellectual disability. Galactose treatment (1g/kg/day) showed biochemical improvement of glycosylated proteins in blood.

SLC39A8-CDG**#

SLC39A-CDG is a multisystem disorder with seizures. Several patients show a skeletal dysplasia with rhizomelic shortening and dwarfism. Not all patients have abnormal transferrin isoform analysis. Laboratory examinations, however, reveal low blood manganese and zinc concentrations, with increased urinary concentration from renal wasting. Patients show clinical and biochemical improvement on galactose (1g/kg/day) and individually dosed oral manganese therapy.

PGM1-CDG*/**

PGM1-CDG is a disorder with midline malformations (Pierre-Robin sequence, bifid uvula, cleft palate), hepatopathy, muscle weakness, cardiomyopathy and hypoglycaemia. This is an unusual CDG in that the type 2 transferrin pattern hides a combined type 1/type 2 defect. Galactose treatment (1g/kg/day) showed biochemical improvement of blood glycoproteins and liver function tests.

GMPPA-CDG

GMPPA-CDG is accompanied by alacrima, achalasia, and intellectual disability. Although NGLY1 deficiency can have alacrima and intellectual disability, patients don’t have achalasia. Triple A syndrome can present alacrima and achalasia, but patients also have adrenal insufficiency, not seen in GMPPA-CDG.

PIGL-CDG##

PIGL-CDG is also known as CHIME syndrome, an acronym that describes its salient features of ocular Colobomas, congenital Heart disease, early-onset Ichthyosis, Mental retardation and Ear anomalies (conductive hearing loss).

PIGM-CDG##

PIGM-CDG is characterized by portal vein thrombosis and seizures. The seizures can improve with sodium butyrate, which increases transcription.

Phenotype of PMM2-CDG

The following congenital anomalies were reported in patients with confirmed PMM2-CDG.

Strabismus

Strabismus was described in 290 of 395 patients (73 %). In the available case series of 10 or more patients, the frequency of strabismus has been variably reported as 36 of 56 patients, or 64 % (Erlandson et al. 2001), 8 of 10 patients (Thompson et al. 2013) and 16 of 20 (de Lonlay et al. 2001), or 80 %, 25 of 28 patients, or 89 % (Monin et al. 2014), 12 of 13, or 92 % (Serrano et al. 2015), and 17 of 17 (Honzík et al. 2012), 23 of 23 (Kjaergaard et al. 2001), or 33 of 33 patients (Grünewald et al. 2001).

Inverted nipples

Inverted nipples were seen in 183 of 343 patients (53%). In the available case series of 10 or more patients, the frequency of inverted nipples has been variably reported as 5 of 20 patients, or 25 % (Barone et al. 2015), 3 of 11 adult patients, or 27 % (Monin et al. 2014), 9 of 26, or 35 % (Briones et al. 2002), 36 % out of 66 patients (Pérez et al. 2011), 10 of 26, or 38 % (Grünewald et al. 2001), 7 of 10 patients (Thompson et al. 2013), 10 of 13, or 77 % (Imtiaz et al. 2000), 16 of 17 patients, or 94 % (Honzík et al. 2012)–although it disappeared over time in 6 of those 16 patients–and in 23 of 23 patients (Kjaergaard et al. 2001), although it disappeared over time in 2 of them. It can sometimes be unilateral (van de Kamp et al. 2007; Léticée et al. 2010).

Abnormal fat distribution

Abnormal fat distribution was reported in 168 out of 357 patients (47 %). In the available case series of 10 or more patients, the frequency of abnormal fat distribution has been variably reported as 5 of 20 patients, or 25 % (Barone et al. 2015), 27% of 66 patients (Pérez et al. 2011), 7 of 19 adult patients, or 37 % (Monin et al. 2014), 10 of 26, or 38 % of patients (Grünewald et al. 2001), 4 of 10 (Thompson et al. 2013), and 21 of 23, or 91 % (Kjaergaard et al. 2001), but it disappeared over time in 9 of those 21 patients. A specific review of abnormal fat distribution found it in 24 of 32 patients, or 75 % (Wolthuis et al. 2013).

Dysmorphic facial features

Dysmorphic features have been described in large series in 29 % of 66 patients (Pérez et al. 2011), 11 of 20, or 55 % (de Lonlay et al. 2001), and 24 of 37, or 65 % of patients (Barone et al. 2015), including a prominent forehead, large ears and ear lobules, thin upper lip, prominent jaw, and long and slender fingers and toes. The prominent jaw develops over time, as the mandible can start as retrognathic (Al-Maawali et al. 2014). Almond-shaped eyes have also been frequently described, including in 7 of a series of 13 patients (Serrano et al. 2015). Other features reported in more than single patients include skin/peau d’orange in 7 (Clayton et al. 1992; Vabres et al. 1998; de Michelena et al. 1999; Stark et al. 2000; Noelle et al. 2005; van de Kamp et al. 2007), dysplastic ears (cupped, anteverted or crumpled) in 7 patients (Clayton et al. 1992; Garel et al. 1998; Tayebi et al. 2002; Rudaks et al. 2012; Resende et al. 2014; Serrano et al. 2015), low-set ears in 6 (Clayton et al. 1992; Imtiaz et al. 2000; Al-Maawali et al. 2014; Serrano et al. 2015), long philtrum in 6 (Harding et al. 1988; Pavone et al. 1996; Morava et al. 2004; van de Kamp et al. 2007; Kasapkara et al. 2017), high-arched palate in 5 (Tayebi et al. 2002; Bubel et al. 2002; Işıkay et al. 2014; Kasapkara et al. 2017), long face in 3 (de Michelena et al. 1999; Tayebi et al. 2002), narrow/short palpebral fissures in 3 (Truin et al. 2008; Işıkay et al. 2014), epicanthal folds in 2 (Tayebi et al. 2002; Enns et al. 2002), prominent nose in 2 (Barone et al. 2008; Thong et al. 2009), and anteverted nares in 2 (Tayebi et al. 2002; Morava et al. 2004). Other features variably reported included hypertelorism in 2 (Edwards et al. 2006; Kasapkara et al. 2017), but hypotelorism in 2 (Pavone et al. 1996); downslanted palpebral fissures in 3 (Clayton et al. 1992; Bubel et al. 2002), while others had upslanted palpebral fissures (Kjaergaard et al. 2001; Brum et al. 2008); flat nasal bridge in 5 (Harding et al. 1988; Tayebi et al. 2002; Thong et al. 2009; Al-Maawali et al. 2014; Kasapkara et al. 2017) but high nasal bridge in 4 (Artigas et al. 1998; Antoun et al. 1999; Laplace et al. 2003). A coarse face has been described in older adults (Bortot et al. 2013; Barone et al. 2015).

Congenital cardiac malformations

Conotruncal heart defects were described in a few patients, including tetralogy of Fallot in 3 patients (Vermeer et al. 2007; Romano et al. 2009; Al Teneiji et al. 2017), while common arterial trunk (also known as truncus arteriosus) was reported in 2 (Romano et al. 2009; Serrano et al. 2015, Marques-da-Silva et al. 2017).

Congenital airway and lung malformations

No malformations have been reported so far in PMM2-CDG patients.

Congenital liver and bile duct malformations

No liver and bile duct malformations have been reported so far in PMM2-CDG patients. Fibrotic changes and bile duct dilatation described in a few cases appeared later during the course of disease compared to other CDGs, like MPI-CDG (Marques-da-Silva et al. 2017)

Congenital urogenital malformations

Hydrocele, inguinal hernia and/or cryptorchidism was seen in 10 of 12 boys in one series (Kjaergaard et al. 2001), while cryptorchidism was described in 4 other patients (Veneselli et al. 1998; Garel et al. 1998; Truin et al. 2008; Resende et al. 2014). The most commonly reported malformations were multicystic kidneys, reported in seventeen patients (thirteen detected by ultrasound, three by renal biopsy and one by autopsy) followed by enlarged kidney size in four patients, and hydronephrosis in one patient (Schiff et al. 2017, van de Kamp et al. 2007, Hertz-Pannier et al. 2006, De Lonlay et al. 2001).

Facial recognition analysis of PMM2-CDG

For the facial phenotypic analysis of patients with PMM2-CDG, a composite image of unaffected controls vs that of patients can be found in figures 2A–B. Facial recognition analysis allowed for the discrimination of patients with PMM2-CDG vs controls, with an AUC of 0.84 (p value 0.014, see figures 2C–D).

Figure 2.

Diagram on the different steps in lipid-linked oligosaccharide assembly.

Discussion

Here we define the PMM2-CDG phenotype, which is recognizable in at least 50% of the patients based on the presence of strabismus and inverted nipples and abnormal fat distribution. However, not only the inverted nipples and abnormal fat distribution, but also the facial phenotype seems to be recognizable among patients with PMM2-CDG. In particular, commonly reported facial dysmorphic features included almond-shaped eyes, large ears in relation to the (variably) decreased head circumference, a thin vermillion of the upper lip, short nose with long, smooth philtrum, and over time also prognathism. Although not reported in the literature, deep-set eyes and high-arched eyebrows are also often found in the authors’ experience. Another dysmorphic finding frequently reported in the literature is arachnodactyly.

It is easy to understand why patients with mutations in genes involved in similar pathways would have similar symptoms. For example, it is understandable that mutations in 12 genes involved in the O-glycosylation of alpha-dystroglycan would all lead to a similar phenotype of muscular dystrophy, since glycosylation is essential for the anchoring function of dystroglycan (Endo 2015). However, by the same token it is hard to explain how mutations in certain genes involved in N-glycosylation could lead to such discrepant phenotypes. Since the transfer of the lipid-linked oligosaccharide (LLO) from the dolichol carrier to the nascent protein in the ER happens en bloc, in theory a deficiency of any of the 14 enzymes involved in the synthesis of the LLO would lead to a similar clinical presentation, as all would lead to the same deficient N-glycosylation of the nascent protein. However, this is not the case, and the phenotypes appear to be clustered instead. For example, a defect in the synthesis of UDP-GlcNAc (as in GFPT1-CDG) or proteins participating in the addition of the first or second GlcNAc to the LLO (DPAGT1- and ALG14-CDG, respectively) all cause a similar phenotype of congenital myasthenic syndrome (Wu et al. 2003; Senderek et al. 2011; Cossins et al. 2013). A defect in the addition of the fourth and fifth mannoses (ALG11-CDG) and of the immediate next step of flipping the LLO from the cytoplasmic to the luminal side of the ER (RFT1-CDG) both lead to sensorineural hearing loss, a finding that is otherwise quite rare in other types of CDG (Vleugels et al. 2009a; Rind et al. 2010; Regal et al. 2015). Defects in the addition of the last four mannoses in the luminal side of the ER (ALG3-CDG, ALG9-CDG and ALG12-CDG) are associated with very characteristic skeletal changes, defined by mesomelic brachymelia, round pelvis, shortened sacrosciatic notch and ovoid ischia, hypomineralization of the skull, cervical vertebral bodies, and pubic rami, and a thick occipital bone (Tham et al. 2016). These skeletal findings are not similar to those seen in any of the other CDGs. Finally, defects in the next two steps in the synthesis of the LLO, meaning the addition of the first (ALG6-CDG) and second (ALG8-CDG) Glc lead to brachydactyly (Höck et al. 2015; Morava et al. 2016), while the most common type of CDG is associated with the opposite phenotype of long and slender fingers and toes. See figure 2 for a diagram depicting the aforementioned enzymatic steps.

This clustering of phenotypes is puzzling, since again all of these steps occur prior to the en bloc transfer of the glycan chain to the protein. The explanation for this phenotype clustering can thus not solely be related to protein hypoglycosylation. One possibility is that the accumulated LLOs themselves, or modifications of these, could in some way be toxic. For example, the cytosolic accumulation of Man3GlcNAc2-PP-Dol or Man4GlcNAc2-PP-Dol could lead to damage of the inner ear cells, while ER accumulation of Man5GlcNAc2-PP-Dol to Man8GlcNAc2-PP-Dol could lead to a reproducible skeletal dysplasia, and ER accumulation of Man9GlcNAc2-PP-Dol or GlcMan9GlcNAc2-PP-Dol would lead to brachydactyly as opposed to arachnodactyly. Alternatively, it is possible that the pathogenesis of these disorders is not only related to protein hypoglycosylation, but also to protein misglycosylation, with truncated LLOs, since it is known that truncated LLOs can indeed be transferred to the nascent protein, either with decreased efficiency (Krasnewich at el. 1995, Panneerselvam & Freeze 1996; Vleugels et al. 2009b). Even in the latter scenario, proteins that are misglycosylated with specific truncated LLOs as above could lead to the clustered phenotypes described.

It is clear that our understanding of the pathophysiologic mechanisms leading to the different phenotypes seen in CDG patients is still poor. However, the rapid scientific advances that ushered in the next-generation sequencing should not deter us from careful patient phenotyping, the oldest tool in the physician’s armamentarium. Next-generation sequencing should lead to next-generation phenotyping, and the different –omics should be complementary with the process of phenomics.

Figure 1.

Facial recognition analysis of patients with PMM2-CDG.

A: Composite photo created by averaging the extracted mathematical information of the photos in the control cohort. B: Composite photo obtained from images of patients with PMM2-CDG. C: Score distribution of the binary comparison between unaffected controls and patients with PMM2-CDG. D: ROC curve with pertinent statistics obtained after conducting 10 random splits (true positive rate on the ordinate, true negative rate on abscissa).