Three unreported cases of TMEM199-CDG, a rare genetic liver disease with abnormal glycosylation

The differential diagnoses of elevated transaminases in children include a large variety of infectious, autoimmune, metabolic, genetic, gastrointestinal, and extrahepatic disorders [1]. Recently a novel disease with chronically elevated serum transaminases and low serum ceruloplasmin, caused by mutations in the gene encoding the transmembrane protein TMEM199 [2]. These patients also showed deficient transferrin (TF) glycosylation with a Type 2 congenital disorder of glycosylation (CDG) pattern. TMEM199 is the human orthologue of yeast Vma12p, which, together with Vma21p and Vma22p, constitute a complex that chaperones the assembly of the V₀ domain of the Vacuolar H+ ATPase (V-ATPase) [3]. This proton pump is responsible for the acidification of the endosomes of the secretory pathway, including the lysosome and the Golgi apparatus. Failure to acidify the Golgi apparatus may affect the complex Golgi-located glycosylation machinery, leading to a glycosylation defect [4]. A decade ago we published a report on four children with glycosylation deficiencies with liver disease [5], but without known genetic cause. Two of these children were later diagnosed with phosphoglucomutase-1 CDG (PGM1-CDG; OMIM 614921) [6]. In this report, we present a long-term follow-up on the other two children, and present data on a third, unrelated child, all diagnosed with TMEM199-CDG (OMIM 616829).

The patients/their legal guardians gave informed consent for this study.

Patient 1 and patient 2 (siblings) are healthy young adults with normal physical and psychological development. They were investigated for hypertransaminasemia and slight liver enlargement at 2 years of age. They also presented with increased values of total and LDL serum cholesterol, creatine kinase (CK), and alkaline phosphatase (ALP). Serum ceruloplasmin and serum copper were low in spite of a normal urinary excretion of copper at basal level and after penicillamine. Liver histology showed mild periportal fibrosis and focal steatosis. The liver copper content was mildly increased, but molecular studies ruled out Wilson’s disease [9]. Serum, liver and brain (MRI) did not show increased metal storage, thus ruling out aceruloplasminemia. Serum TF glycosylation analysis was consistent with a Type 2 CDG [5]. At a recent follow-up (at age 27 and 24 years, respectively) both patients were clinically well, and there was no sign of progressive liver, bone or neuromuscular disease, nor any signs of malignant tumors. Exome-sequencing was performed identifying compound heterozygous mutations in TMEM199 (c.13-14delTT, c.92G > C), causing a frameshift with a premature stop (p.Ser4Serfs*30) and a missense change (p.Arg31Pro), respectively. The Genome Aggregation Database (gnomAD) (http://​gnomad.​broadinstitute.​org/) (10.19.2017 ver2.0) [10] of 123,136 exome and 15,496 whole-genome sequences contains 10 heterozygotes from 121,705 apparently healthy individuals (0 homozygotes) for the p.Arg31Pro exchange, but no alleles carrying the c.13-14delTT, yielding a carrier frequency of 0.000082 and 0, respectively). Western Blot analysis could not detect the TMEM199 protein in patient fibroblasts, but showed a strong signal in both control fibroblasts and in HepG2 control hepatocytes (Fig. 1).

Patient 3 is a clinically healthy 2-year-old child who, at 10 months of age, was diagnosed with mild hypertransaminasemia as an incidental finding. Liver ultrasonography revealed a mild steatosis and the follow-up showed fluctuating values of serum transaminases, and elevated ALP and CK. Other liver tests were normal except antithrombin-III activity of 68% and low ceruloplasmin (8 mg/dL, normal range 20–46). Mass spectrometric analysis of serum TF glycosylation showed a pathological pattern (CDG-II, Fig. 2a and b) similar to Patient 1 and Patient 2 [5]. We therefore performed direct Sanger sequencing of TMEM199 for Patient 3 and found the same set of compound heterozygous mutations as seen in Patient 1 and Patient 2. The two families have no known common ancestors, but both originate from the same region of Campania, Southern Italy. Table 1 summarizes clinical and laboratory features of the three patients.

The causes of elevated transaminases are many, spanning from transient, harmless conditions to ones associated with a progressive, life-threatening course [1]. Here we present three patients with TMEM199-CDG, a recently described genetic cause of hypertransaminasemia [2]. They all carried the same set of compound heterozygote mutations, despite no known common family ancestors. One of the mutations, c.92G > C, was previously reported by Jansen et al. [2], whereas the other one is pathogenic due to its early frameshift/termination. Using Western Blot analysis in patient fibroblasts, it was shown that there is no measurable amount of TMEM199 protein, indicating that the missense mutation causes a diminished production or a rapid degradation of the resulting protein. A protein product from the c.13-14delTT allele would not be picked up by the antibody used in the Western Blot, however, given its early termination, we think it is very unlikely that such a protein would be functional at al. The clinical and biochemical findings in our patients were consistent with the previously published cases, and included, apart from hypertransaminasemia, deficient TF glycosylation and low ceruloplasmin, slightly increased copper content of the liver, mild non-progressive liver fibrosis and normal (except one patient [2]) psychomotor development. The only patient that has been diagnosed with symptoms from the nervous system, including benign hypotonia and psychomotor developmental issues [2], has a homozygous mutation (c.92G > C) in TMEM199, which could indicate consanguinity and thus a possible second mutation as a cause of these symptoms. Notably, the clinical course for two of these patients has been stable over two decades, with no deterioration of the liver pathology or development of other symptoms. TMEM199 is the human orthologue of yeast Vma12p, which acts as a chaperone in the formation of the V-ATPase, the proton pump responsible for the acidification of the vesicles of the secretory pathway [3]. Why TMEM199-CDG mainly presents with hepatopathy is enigmatic. Mutations in a number of genes involved in the formation of the V-ATPase (either encoding subunits or assembly factors) cause a plethora of symptoms with either systemic or organ specific features [11‐15]. In several of these disorders, pathological TF glycosylation (Type 2 patterns) is detected, as in patients with mutations in ATP6V0A2 (OMIM 219200) [14], ATP6V1A (OMIM 617403) and ATP6V1E1 (OMIM 617402) [15], ATP6AP1 (OMIM 300972) [13] and CCDC115 (OMIM 616828) [11]. This was also the case in our patients as well as in the four previously published TMEM199-CDG patients [2]. The glycosylation pattern indicates a Golgi related problem (CDG-II) stressing the need for a tight pH regulation in the Golgi stacks for proper function of the glycosylation enzymes. Disturbances in copper metabolism in patients with CDG-II [16] as well as in V-ATPase deficiencies [11, 13, 15] have been reported. These include hepatopathy with low serum ceruloplasmin and low serum copper. In both TMEM199-CDG and CCDC115-CDG, a modest accumulation of copper in the liver is also seen (this report and [11]. The mechanisms behind these disturbances are at the moment unclear, but potentially involve at least partial loss of either or both of the copper transporting proteins ATP7A and ATP7B. In Wilson’s disease (ATP7B deficiency) [17], export of copper from the hepatocyte (and neuron) is deficient. This causes low serum ceruloplasmin, low serum copper but accumulation of copper in the liver, similar to the TMEM199-/CCDC115-CDG patients. However, ATP7B is not glycosylated [18] and its transfer of copper is not pH dependent [19], why a potential link between V-ATPase malfunction and loss of ATP7B mediated copper transportation might thus reside in the synthesis or degradation of the ATP7B protein. This has not yet been investigated. In the case of ATP7A related disorders (Menkes disease or occipital horn syndrome), the uptake of intestinal copper is deficient, leading to low serum copper, low ceruloplasmin but also low copper in liver tissue [20]. ATP7A is, in contrast to ATP7B, glycosylated [18] and deficient glycosylation could potentially affect its function. Also, a very recent study showed that several COG subunits (involved in Golgi homeostasis and glycosylation) are within the interactome of ATP7B, linking Golgi homeostasis, glycosylation and intracellular copper homeostasis [21]. The patients with TMEM199-CDG described so far have displayed a strikingly similar clinical presentation, with hepatic features resembling other V-ATPase chaperone deficiency syndromes, however without associated symptoms such as hypogammaglobulinemia [13], epilepsy and cognitive impairment [11, 13] noted in the related syndromes.

Interestingly, it was recently published that loss of TMEM199 (and CCDC115) rescues the angiogenic factor HIF-1α from proteasome-mediated degradation under normoxic conditions by intracellular iron depletion [3]. If this is true also in man, it may cause a chronically elevated HIF-1α activity in our patients, potentially increasing their risk of developing cancer. However, in the present adult patients, no suspicion of malignancy has yet occurred.

Three novel TMEM199-CDG cases show a biochemical and clinical phenotype in accordance with the previously described patients. In our patients, the disease course seems stable and mild over several decades, why it is possible that an early diagnosis of TMEM199-CDG may allow for the avoidance of patient and parental worries and unnecessary repeat blood sampling. There is, however, a need for more patient descriptions of this rare condition to better understand the disease course and further elucidate whether TMEM199-CDG will remain a liver disease only.