How to find and diagnose a CDG due to defective N-glycosylation

Most CDG with a N-glycosylation defect are multi-organ diseases with neurological involvement. Only few of these CDG are known as non-neurological disorders, such as MPI-CDG (de Lonlay and Seta 2009), DPM3-CDG (Lefeber et al. 2009) and SEC23B-CDG (Schwarz et al. 2009). Nearly all organs have been reported to be affected in CDG (Jaeken 2010). One should thus screen for CDG (i) in any unexplained neurological syndrome, particularly when associated with other organ disease, (ii) and also in any unexplained syndrome even without neurological involvement.

Experience gained in the last decade has shown that a number of clinical clues might alert for a CDG and that some of these clues might also provide a hint for a specific CDG subtype, either in the CDG-I group (Morava et al. 2008) or in the CDG-II group (Mohamed et al. 2011). These specific features include also the well known abnormality in fat distribution or cutis laxa. A brief overview of alerting clinical features is presented in Table 1 and currently known CDG subtypes are tabulated in Table 2. A laboratory finding to be added to this list is syndromic factor XI deficiency in CDG-I, and in an unknown number of CDG-II defects.

Since its introduction in 1984, isoelectric focusing (IEF) of serum transferrin is still the method of choice for the diagnosis of N-glycosylation disorders associated with sialic acid deficiency (Jaeken et al. 1984). More recently, HPLC (Helander et al. 2001) and capillary zone electrophoresis (Carchon et al. 2004) have been introduced as well as mass spectrometry (Babovic-Vuksanovic and O'Brien 2007). In the absence of a protein polymorphism, the isoelectric point of transferrin is determined only by its glycosylation status. A protein polymorphism can be identified by incubation of the sample with commercial sialidase, by analysis of different proteins such as thyroxine-binding globulin (TBG) or by analysis of parental serum samples. In human plasma, the tetrasialotransferrin fraction is the most abundant form. Children younger than 1 month can show mildly elevated underglycosylated transferrin isoforms as compared to older children. Also, the diagnosis of CDG might be missed in very young children, in whom the profile becomes abnormal after 1–2 months of age. A few CDG patients have been described with a normal profile particularly in adolescent and adult age (Vermeer et al. 2007). Before proceeding with further diagnostics, it is important to rule out secondary causes of abnormal transferrin profiles, such as fructosemia, galactosemia, alcohol abuse and bacterial sialidase.

In general, two diagnostic types of abnormal profiles can be distinguished (Fig. 1a): (i) a type 1 pattern in CDG-I, characterized by an increase of di- and/or asialotransferrin; (ii) a type 2 pattern in CDG-II, characterized by an increase of tri-, di-, mono-and/or asialotransferrin.

The type 1 pattern (CDG-I) points to an assembly or transfer defect of the dolichol-linked glycan (in the cytosol or ER glycosylation pathway). Measurement of phosphomannomutase activity in fibroblasts or leukocytes is the next step for further diagnostics because PMM2-CDG (CDG-Ia) is by far the most frequent N-glycan assembly defect. In case of a purely hepato-intestinal clinical presentation, the activity of phosphomannose isomerase should be measured for diagnosis of MPI-CDG (CDG-Ib), which is a treatable disease. A normal activity of these enzymes necessitates analysis of the lipid-linked oligosaccharides (LLO; more specifically dolichol-linked oligosaccharides) in fibroblasts or other more specific assays to identify the known (or still unknown) CDG-I subtype. The recently identified SRD5A3-CDG could be diagnosed directly by analysis of plasma polyprenols (Morava et al. 2010).

The type 2 pattern (CDG-II) indicates a processing defect after glycan transfer in the ER or during Golgi glycosylation. Not all processing defects can be picked up by transferrin IEF. Patients with GCS1, SLC35C1 (fucose has no electric charge) and SLC35A1 defects (formerly CDG-IIb, -IIc and CDG-IIf) have been described with a normal transferrin IEF profile. Type 2 patterns can show an asialo type 2 profile with increased asialo- to trisialotransferrin, or a disialo type 2 profile with increased disialo- and trisialotransferrin. The next step in the identification of the CDG-II subtypes is mass spectrometry of isolated serum N-glycans (Guillard et al. 2011). This will permit to identify isolated N-glycosylation defects as MGAT2-CDG (CDG-IIa) or B4GALT1-CDG (CDG-IId) and secondary causes, but most CDG-II types are associated with an aspecific glycan profile. In the latter situation, the possibility of an associated mucin type O-glycosylation defect should be investigated by isoelectrofocusing of serum apolipoprotein C-III (APOC3). This protein contains a single core 1 mucin type O-glycan on Thr-74. Abnormal profiles (Fig. 1b) include an increased monosialo APOC3 with decreased disialo APOC3 (a so-called APOC3-1 profile) or an increased asialo APOC3 (a so-called APOC3-0 profile) (Wopereis et al. 2005). Young children in the first months of life can present with an APOC3-2 profile in which an increase of disialo APOC3 is observed. In our experience, the latter pattern can also be observed in patients with severe liver pathology.

Currently, no genetic glycosylation disorders have been described with an isolated abnormality of the APOC3 profile. However, several defects are known with a combined N- and O-glycosylation defect. In general, these defects are characterized by abnormal functioning of the secretory pathway, such as abnormal retrograde trafficking or abnormal acidification. If the patient presents with clinical symptoms of cutis laxa, mutation analysis of the ATP6V0A2 gene is indicated. It should be noted that ATP6V0A2-CDG patients younger than 6 months of age can present with normal transferrin glycosylation, while apolipoprotein C-III isofocusing is abnormal. In all other cases, we recommend to look for a defect in one of the COG (conserved oligomeric Golgi complex) subunits by mutation analysis of the COG1-8 subunit genes.

The emerging innovative next-generation sequencing techniques like whole-exome-sequencing will provide new opportunities to unravel genetic causes for glycosylation disorders, such as the CDG-II trafficking defects, but also the many glycosylation defects that cannot be identified via transferrin screening.