Galactose Supplementation in Patients With TMEM165-CDG Rescues the Glycosylation Defects

Abstract

Context:

TMEM165 deficiency is a severe multisystem disease that manifests with metabolic, endocrine, and skeletal involvement. It leads to one type of congenital disorders of glycosylation (CDG), a rapidly growing group of inherited diseases in which the glycosylation process is altered. Patients have decreased galactosylation by serum glycan analysis. There are >100 CDGs, but only specific types are treatable.

Objective:

Galactose has been shown to be beneficial in other CDG types with abnormal galactosylation. The aim of this study was to characterize the effects of galactose supplementation on Golgi glycosylation in TMEM165-depleted HEK293 cells, as well as in 2 patients with TMEM165-CDG and in their cultured skin fibroblast cells.

Design and Setting:

Glycosylation was assessed by mass spectrometry, western blot analysis, and transferrin isoelectrofocusing.

Patients and Interventions:

Both unrelated patients with TMEM165-CDG with the same deep intronic homozygous mutation (c.792+182G>A) were allocated to receive d-galactose in a daily dose of 1 g/kg.

Results:

We analyzed N-linked glycans and glycolipids in knockout TMEM165 HEK293 cells, revealing severe hypogalactosylation and GalNAc transfer defects. Although these defects were completely corrected by the addition of Mn²⁺, we demonstrated that the observed N-glycosylation defect could also be overcome by galactose supplementation. We then demonstrated that oral galactose supplementation in patients with TMEM165-deficient CDG improved biochemical and clinical parameters, including a substantial increase in the negatively charged transferrin isoforms, and a decrease in hypogalactosylated total N-glycan structures, endocrine function, and coagulation parameters.

Conclusion:

To our knowledge, this is the first description of abnormal glycosylation of lipids in the TMEM165 defect and the first report of successful dietary treatment in TMEM165 deficiency. We recommend the use of oral d-galactose therapy in TMEM165-CDG.

We studied the effects of oral D-galactose therapy in patients with TMEM165-CDG. We found that abnormal glycosylation and clinical parameters could be improved by galactose treatment.

In 2012, we reported pathogenic variants in TMEM165 in 5 patients with abnormal N-glycans partially lacking galactose and sialic acid, establishing TMEM165 deficiency as a new subtype of congenital disorder of glycosylation (CDG; TMEM165-CDG) (1) (phenotypic OMIM #614727). The most severely affected individuals present with severe skeletal symptoms (1, 2). TMEM165 is highly conserved during evolution, but its biological function remains controversial. We recently demonstrated that TMEM165 deficiency disrupts Golgi manganese (Mn²⁺) homeostasis, resulting in Golgi glycosylation defect and hypoglycosylation, which can be subsequently rescued by Mn²⁺ supplementation (3). Detailed structural N-linked glycan analysis in TMEM165-deficient cells showed a dramatic increase of agalactosylated glycan structures. This indicated that a lack of TMEM165 impairs the function of β-1,4-galactosyltransferase, a Mn²⁺ Golgi-dependent glycosyltransferase required for the biosynthesis of sialylated complex N-glycan structures (4–6).

Currently, only a handful of CDG subtypes shows improved glycosylation following dietary supplementation with monosaccharides. MPI-CDG was the first CDG subtype discovered to be treatable with a few grams of oral mannose (7). Another monosaccharide, d-galactose, has been shown to benefit patients with PGM1-CDG (8).

Interestingly, galactose has also been successful for treating patients with another CDG subtype that is caused by pathogenic variants in SLC39A8, a Zn²⁺/Mn²⁺ transporter. Patients with defective SLC39A8 present with severe hypogalactosylation of serum transferrin (8), demonstrating the crucial requirement of adequate Golgi Mn²⁺ homeostasis in normal Golgi glycosylation processes. Although it is clear that galactose therapy in these patients rescued the impaired galactosylation, the mechanism remained poorly understood.

In this study, we investigated whether galactose supplementation could be beneficial in TMEM165 deficiency. The aim of this work was to characterize the effects of galactose supplementation on Golgi glycosylation in TMEM165-depleted HEK293 cells, as well as in 2 TMEM165-CDG patients and their cultured skin fibroblast cells.

Patients and Methods

Cell culture and transfections

All cell lines were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (Lonza, Basel, Switzerland) at 37°C in humidity-saturated 5% CO₂ atmosphere.

Skin fibroblast cultures

Skin fibroblasts derived from skin biopsy specimens were maintained in Eagle's minimum essential medium (American Type Culture Collection, Manassas, VA) in a humidified 37°C incubator. Culture medium was supplemented with 10% fetal bovine serum and a 1% combination of 100 U penicillin/0.1 mg/mL streptomycin. TMEM165-deficient fibroblasts were cultured in the presence of additional galactose in the medium.

In vitro galactose feeding

Eagle’s minimum essential culture medium was supplemented with 0, 0.75, 2, 5, or 10 mM d-galactose 18 to 24 hours after seeding the cells. d-Galactose powder (Sigma-Aldrich, St. Louis, MO) was dissolved in phosphate-buffered saline (PBS) to create a stock solution of 100 mM, which was subsequently diluted to the desired concentration in the culture medium. Cells were fed for 5 to 7 days, with culture medium refreshed every 2 days. On the final day of the feeding experiment, cells were harvested by scraping.

ICAM-1: immunofluorescence staining

Immunohistochemistry was performed on cells seeded on glass coverslips and cultured for 5 days. Samples were fixed with 4% formaldehyde (Sigma-Aldrich) for 10 minutes, and permeabilized with 0.1% saponin (weight-to-volume ratio) and 0.1% bovine serum albumin (weight-to-volume ratio). They were blocked in 5% [volume-to-volume ratio (v/v)] normal donkey serum (Jackson ImmunoResearch Laboratories, West Grove, PA) and incubated with primary antibody against ICAM-1 (1:750; catalog no. MA5-13021; Thermo Fisher Scientific, Waltham, MA) at 4°C for 3 nights, followed by incubation with biotinylated donkey anti-mouse immunoglobulin G antibody (1:800; catalog no. 715-065-151; Jackson ImmunoResearch Laboratories) and streptavidin-conjuguated Cy3 (1 μg/mL; catalog no. 016-160-084; Jackson ImmunoResearch Laboratories) at room temperature for 1.5 hours. The samples were mounted with VectorShield Mounting Medium containing 4′,6-diamidino-2-phenylindole (Vector Laboratories, Burlingame, CA). Images were obtained using an Olympus BX51 System microscope with an Olympus DP80 color camera (Olympus, Tokyo, Japan). Particle analysis and manual counting were performed with Fiji (http:imagej.net/) using the Cell Counter plug-in (version 2.0.0-rc-49/1.51a) (9).

LAMP2: actin western blot

Cells were scraped in Dulbecco’s PBS and then centrifuged at 4500 rpm for 3 minutes. Supernatant was discarded and cells were resuspended in radioimmunoprecipitation assay buffer (Tris/HCl, 50 mM; pH 7.9; NaCl, 120 mM; NP40 0.5%; EDTA, 1 mM; Na₃VO₄, 1 mM; NaF, 5 mM) supplemented with a protease cocktail inhibitor (Roche Diagnostics, Penzberg, Germany). Cell lysis was done by passing the cells several times through a syringe with a 26G needle. Cells were centrifuged for 30 minutes at 20,000g. The supernatant containing protein was estimated with the micro BCA Protein Assay Kit (Thermo Fisher Scientific). Total protein lysate (20 µg) was put in NuPAGE lithium dodecyl sulfate sample buffer (Invitrogen), pH 8.4, supplemented with 4% β-mercaptoethanol (Fluka). Samples were heated for 10 minutes at 95°C and then separated on 4% to 12% Bis-Tris gels (Invitrogen) and transferred to nitrocellulose membrane Hybond ECL (GE Healthcare, Little Chalfont, United Kingdom). The membranes were blocked in blocking buffer (5% milk powder in tris-buffered saline and Polysorbate 20 (TBS-T; 1X TBS with 0.05% Tween20 for 1 hour at room temperature, then incubated overnight with the primary antibodies in blocking buffer, and washed 3 times for 5 minutes in TBS-T. The membranes were then incubated with the peroxidase-conjugated secondary goat anti-rabbit or goat anti-mouse antibodies (1:10,000 dilution; DakoSanta Clara, CA) in blocking buffer for 1 hour at room temperature and later washed 3 times for 5 minutes in TBS-T. Signal was detected with chemiluminescence reagent (ECL 2 Western Blotting Substrate; Thermo Fisher Scientific) on imaging film (GE Healthcare).

ICAM-1: β-integrin western blot

Protein expression of ICAM-1 in patients and healthy control subjects was assessed by western blotting. Hydrophobic protein was extracted and enriched by CelLytic MEM Protein Extraction kit (Sigma-Aldrich). Enriched hydrophobic protein (25 μg) of enriched hydrophobic protein was resolved on 10% Bis-Tris Gel (Invitrogen) at 200 V for 55 minutes. After sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), samples were transferred to a 0.45-μm nitrocellulose membrane at 30 V for 2.5 hours on ice. Both running and transfer buffer were supplemented with antioxidant per manufacturer’s recommendation (Invitrogen). The membrane was rocked for 1 hour in membrane blocking solution (Invitrogen) at room temperature and incubated overnight with the primary antibodies at 4°C, followed by six 10-minute washes with 1× PBS supplemented with 0.2% Tween 20 (Sigma-Aldrich). Incubation with secondary antibodies at room temperature for 1 hour was followed by six 10-minute washes with 0.2% Tween-PBS. Signal was detected by Licor Odyssey CLx IR Imaging System (LI-COR Biosciences, Lincoln, NE). Signal intensity was quantified by Odyssey software (version 2.0).

Primary antibodies were polyclonal rabbit anti-ICAM-1 (1:4000; catalog no. sc-7891; Santa Cruz Biotechnology, Dallas, TX) and monoclonal mouse anti-integrin-β1 (catalog no. sc-374429; Santa Cruz Biotechnology). Secondary antibodies were DyLight 800-conjugated goat anti-rabbit (1:10000; catalog no. SA5-35571; Thermo Fisher Scientific) and DyLight 680-conjugated goat anti-mouse (1:10,000; catalog no. SA5-10082; Thermo Fisher Scientific). Antibodies were diluted in membrane blocking solution.

Extraction of glycolipids and mass spectrometry analysis

Cells were washed twice with ice-cold PBS; resuspended in 500 µL of water, and sonicated on ice. The material was then dried under N₂ and extracted by CHCl₃/CH₃OH (2:1, v/v), CHCl₃/CH₃OH (1:1, v/v) and CHCl₃/CH₃OH/H₂O (1:2:0.8, v/v/v). The supernatants were pooled, dried, and subjected to a mild saponification in 0.1M NaOH in CHCl₃/H₂O (1:1, v/v) at 37°C for 2 hours and then evaporated to dryness. Samples were purified by using a C18 cartridge (Waters, Milford, MA) equilibrated in a CH₃OH/H₂O solvent. The glycosphingolipids (GSLs) were eluted by CH₃OH and CHCl₃/CH₃OH (1:1, v/v) and CHCl₃/CH₃OH (2:1, v/v). Permethylation of the freeze-dried GSL and matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS) of permethylated GSL were performed as described elsewhere (10).

Extraction of cellular and sera N-glycans and mass spectrometry analysis

Cells were sonicated in extraction buffer (25 mM Tris, 150 mM NaCl, 5 mM EDTA, and 1% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate [CHAPS], pH 7.4) and then dialyzed in 6- to 8-kDa cut-off dialysis tubing in an ammonium bicarbonate solution (50 mM, pH 8.3) for 48 hours at 4°C and lyophilized. The proteins/glycoproteins were reduced and carboxyamidomethylated, and then underwent sequential tryptic and peptide N-glycosidase F digestion and Sep-Pak purification. Permethylation of the freeze-dried glycans and MALDI-TOF-MS of permethylated glycans were performed as described elsewhere (10). For sera N-glycan analysis, 20 µL of serum was dried and denatured using β-mercaptoethanol and SDS at 100°C during 20 minutes. After adding Nonidet P40, the N-deglycosylation was performed using peptide N-glycosidase F PNGase F. After purification, the permethylation of the freeze-dried glycans and MALDI-TOF-MS of permethylated glycans were performed as described elsewhere (10).

Patients

In 2 previously described patients, a galactose supplementation trial was initiated (patient 1 and patient 3) (11). Both unrelated patients have the same deep intronic homozygous mutation (c.792+182G>A). The protocol was approved by the institutional review board (IRB; Tulane University Hayward Genetics Center IRB no. 517339-4; Clinicaltrials.gov NCT02955264).

Patient 1 (a 9-year-old boy) has dysmorphism and a severe multisystem disease with neurologic, pulmonary, gastrointestinal, endocrine, and hematological involvement (Table 1), as well as severe skeletal involvement with generalized osteopenia; hypoplasia of the skull base; anterior beaking of T12–L2; broad radial, ulnar, femoral, and tibial metaphyses; and underdeveloped epiphyses.

Table 1.

Hematological and Clinical Parameters in Patients With TMEM165 Receiving Galactose Therapy

	Patient 1				Patient 2
	Baseline	Week 6	Week 12	Week 18	Baseline	Week 6	Week 12	Week 18
Creatine kinase, IU/L (nl, <190 IU/L)	1470a	820a	710a,b	1184a	1223a,c	1253a	1186a	904a
PT, s (nl, 9.4–12.5 s)	12.7a	11.5d	11.3d	12.1d	13a,e	12.4	13.3a	13.7a
APTT, s (25–36.5 s)	35.0	35.8	31.5	27.2	40.2a,e	25.4d	38a	34.2a
Factor IX, % (nl, 70%–130%)	59.2a	63.1a	66.1a	65.3a	75e	74.8	67a	71.8
Factor XI, %(nl, 70%–130%)	83	84.9	89.4	82.7	89.4 §	64.8a	51a	55.1a
Antithrombin, % (nl, 70%–130%)	78	96	88	88	61a,e	87d	79d	85d
LH, U/L (nl, 1.7–8.6 U/L)	<0.1a	<0.1a	<0.1a	0.1a	2.0	3.1	1.8	2.3
FSH, U/L (nl, 1.2–7.7 U/L)	0.5a	0.6a	0.3a	0.5a	0.8a	0.8a	0.6a	0.6a
IGF-1, µg/L (nl, 47–251 µg/L)	<35a	56d	47d	64d	123	103	125	118
IGFBP3, µg/L (nl, 1995–4904 µg/L)	1848a	2172d	2149d	2129d	3219	2494	2931	Pending
TSH, mIU/L (nl, 0.27–4.2 mIU/L)	5.33a	4.54a	2.98d	4.17a	5.37a,e	4.58a	4.26a	5.83a
Free T4, pmol/L (nl, 12–22 pmol/L)	20.5	19	21.6	22.4a	20.5e	18.7	17.8	21.1
AST, U/L (<37 U/L)	407a,b	409a,b	549a,b	494a,b	319a,e	287a	265a	245a
ALT, U/L (nl, <51 U/L)	78a	75a	131a	94a	66a,e	56a	51d	43d
HDL cholesterol, U/L (nl, >40 U/L)	45	45	38a	41	27a,e	24a	23a	24a

Abbreviations: ALT, alanine aminotransferase; APTT, activated thromboplastin time; AST, aspartate aminotransferase; FSH, follicle-stimulating hormone; HDL, high-density lipoprotein; LH, luteinizing hormone; nl, normal limit; PT, prothrombin time; TSH, thyroid-stimulating hormone.

^aOutside reference range.

^bHemolysis.

^c5 April 2016.

^dCorrected.

^e16 October 2015.

Patient 2 (a 25-year-old man) has a similar multisystem involvement but to a lesser degree. He has significant endocrine and coagulation anomalies (Table 1). He can walk (with crepitus) and has a normal language (11).

d-Galactose supplementation in patients

All procedures were approved by the Tulane University School of Medicine IRB (Tulane IRB no. 517339-3). Two patients with genetically confirmed TMEM165 defect in this observational pilot study received oral d-galactose supplementation (Necese) over 18 weeks. d-Galactose intake was increased in increments, as follows: weeks 0 to 6, 0.5 g/kg/d; weeks 7 to 12, 1.0 g/kg/d; weeks 13 to 18, 1.5 g/kg/d. The purpose of the escalating dosing schedule was to minimize gastrointestinal irritation and metabolic side effects. Laboratory studies in blood (i.e., glucose, lactic acid, ammonia, and galactose-1-phosphate) and in urine (i.e., galactitol excretion) were conducted before the study and at every time point during the study to monitor tolerability to increasing intake of galactose.

Patients were instructed to continue their regular diet in addition to daily oral galactose supplementation. The maximum daily dose of galactose either patient received was 50.0 g (this amount is within the recommended daily intake). Prior studies investigating focal segmental glomerulosclerosis have demonstrated 50.0 g/d of galactose can be safely consumed and tolerated by patients (12).

Before beginning galactose therapy, clinical and metabolic baselines were established. At each time point in the study, each patient underwent a clinical examination, including an assessment for clinical changes, adverse effects, or any concerns regarding the therapy. Blood analysis was performed every 6 weeks to assess glycosylation (isoelectric focusing of serum transferrin and glycomic analysis in blood by mass spectrometry). Additionally, known glycoproteins, such as thyroid hormones, growth hormone, and coagulation and anticoagulation factors, were measured. Other biochemical parameters, including liver function enzymes and creatine kinase, were monitored. Urinary galactose levels were measured as a safety parameter.

Results

Galactose supplementation rescued galactosylation deficiency in TMEM165-depleted cells

Because galactose supplementation in patients with PGM1-CDG and SLC39A8-CDG led to substantial galactosylation improvement, we first investigated the effects of galactose supplementation on Golgi glycosylation in TMEM165-depleted cells. Therefore, TMEM165 knockout (KO) HEK293 cells were generated by the genomic editing technique CRISPR-Cas9 (Supplemental Fig. 1), followed by studies on the steady-state glycosylation status of LAMP2, which is an extensively N-glycosylated lysosomal resident protein. Control and KO TMEM165 HEK293 cells were cultured without or with galactose supplementation in the culture medium (Fig. 1). Consistent with previous reports, a dramatic change in the LAMP2 gel mobility was observed in TMEM165 KO HEK293 cells, indicating a severely defective glycosylation [Fig. 1(A)] (3). As expected, MnSO₄ and MnCl₂ treatment completely restored normal LAMP2 gel mobility in TMEM165 KO HEK293 cells. Remarkably, 1 mM galactose supplementation achieved the same effect. Other monosaccharides, such as GlcNAc and GlcNH₂, were also tested, but none had any effect on the gel mobility of LAMP2 in TMEM165 KO HEK293 cells. These results strongly suggest that d-galactose is the sole monosaccharide capable of rescuing the abnormal LAMP2 gel mobility.

Figure 1.

Galactose specifically suppresses the observed LAMP2 glycosylation defect in TMEM165 KO cells. (A) Steady-state cellular level and gel mobility of LAMP2. Control and TMEM165 KO HEK293 cells were cultured in absence or presence of galactose (1 mM), MnCl₂ (100 µM), or MnSO₄ (100 µM) for 18 hours; cell lysates were prepared, subjected to SDS-PAGE, and western blot with the indicated antibodies. (B) Steady-state cellular level and gel mobility of LAMP2. Control and TMEM165 KO HEK293 cells were cultured in absence or presence of galactose (1 mM), glucosamine (5 mM), or N-acetylglucosamine (5 mM) for 36 hours. Cell lysates were prepared, subjected to SDS-PAGE, and western blot with the indicated antibodies.

To further characterize the effect of galactose on N-glycosylation, mass spectrometry analysis of total N-glycans was performed in control and TMEM165 KO HEK293 cells that were treated without or with galactose (Fig. 2). Consistent with previous studies, a pronounced hypogalactosylation was seen in TMEM165 KO HEK293 cells, with the accumulation of agalactosylated glycan structures detected at mass-per-charge (m/z) ratios 1836, 2081, and 2326. Although these structures were also found in control cells, the level was detected on a much lower level that is consistent with normal metabolic intermediates generated during normal glycosylation. Of note, the structures detected at m/z ratios greater than 3095 were completely absent in TMEM165 KO HEK293 cells. Altogether, the accumulating glycan structures observed in TMEM165 KO HEK293 cells pointed to a severe galactosylation defect. Although galactose treatment slightly decreased the processing intermediates in control HEK293 cells, its effect on the galactosylation defect observed in TMEM165 KO HEK293 cells was much more pronounced, as indicated by the decreased abundance of the structures with m/z ratios of 1836, 2081, and 2326. After galactose supplementation, N-glycosylation completely normalized. This demonstrated that galactose treatment not only considerably improved N-glycosylation in TMEM165 KO HEK293 cells but also increased N-glycosylation in control cells.

Figure 2.

Galactose supplementation suppresses the observed galactosylation defect on N-glycans in TMEM165 KO cells. (A) MALDI-TOF-MS spectra of the permethylated N-glycans from control HEK293 cells. (B) MALDI-TOF-MS spectra of the permethylated N-glycans from control HEK293 cells treated with galactose (1 mM, 36 hours). (C) MALDI-TOF-MS spectra of the permethylated N-glycans from TMEM165 KO HEK293 cells. (D) MALDI-TOF-MS spectra of the permethylated N-glycans from TMEM165 KO HEK293 cells treated with Galactose (1 mM, 36 hours). The symbols representing sugar residues are as follows: closed square, N-acetylglucosamine; open circle, mannose; closed circle, galactose; open diamond, sialic acid; and closed triangle, fucose. Linkages between sugar residues have been removed for simplicity.

Next, we investigated whether other glycoconjugates were also affected and potentially rescued by galactose supplementation in TMEM165 KO HEK293 cells. Therefore, glycolipids were analyzed by MALDI-MS. As expected, the difference in gangliosides’ profiles between control cells and TMEM165 KO HEK293 cells was clear. Although control cells expressed a complex pattern of GSLs, including GM3 (m/z 1372/1484), GM2 (m/z 1617/1729), GM1 (m/z 1821/1933), GD2 (m/z 1978/2090), and GD1 (m/z 2182/2294) species, KO TMEM165 HEK293 cells only showed traces of GM3 and GM2 species (Fig. 3). This indicated a severe glycolipid glycosylation defect that was secondary to a galactosylation defect. Although galactose supplementation did not change the glycolipid glycosylation profile in control cells, glycolipid glycosylation was found partially restored in TMEM165 KO HEK293 cells upon galactose supplementation. The GM3 species were certainly restored but, surprisingly, the GM2 species were not. This result strongly suggested (1) a lack of TMEM165 affected the β-1,4 N-acetylgalactosaminyltransferase (GM2/GD2 synthase) and (2) that galactose supplementation could not rescue the activity of the β-1,4 N-acetylgalactosaminyltransferase. If this was true, then Mn²⁺ treatment would be expected to rescue both GM3 and GM2 levels in TMEM165 KO HEK293 cells. To test this hypothesis, glycolipid glycosylation was analyzed in TMEM165 KO HEK293 cells treated with Mn²⁺. Although KO cells presented a drastic change of GM3 and GM2 species compared with control cells, Mn²⁺ supplementation rescued the synthesis of both GM3 and of GM2 species (Fig. 4). This clearly showed that the transfer of GalNAc on glycolipids was also impaired in the absence of TMEM165.

Figure 3.

Galactose supplementation only partially suppresses the observed glycosylation defect on glycolipids in TMEM165 KO cells. (A) MALDI-TOF-MS spectra of permethylated glycolipids from control HEK 293 cells. (B) MALDI-TOF-MS spectra of permethylated glycolipids from control HEK 293 cells treated with galactose (1 mM, 36 hours). (C) MALDI-TOF-MS spectra of permethylated glycolipids from TMEM165 KO HEK293 cells. (D) MALDI-TOF-MS spectra of permethylated glycolipids from TMEM165 KO HEK293 cells treated with galactose (1 mM, 36 hours). The symbols representing sugar residues are as follows: yellow closed square, N-acetylgalactosamine; blue circle, glucose; yellow circle, galactose; open diamond, sialic acid. Linkages between sugar residues have been removed for simplicity. All GSLs are present as d18:1/C16:0 and d18:1/C24:0 ceramide isomers.

Figure 4.

MnCl₂ supplementation totally suppresses the observed glycosylation defect on glycolipids in TMEM165 KO cells. (A) MALDI-TOF-MS spectra of permethylated glycolipids from control HEK293 cells. (B) MALDI-TOF-MS spectra of permethylated glycolipids from control HEK293 cells treated with MnCl₂ (100 µM, 36 hours). (C) MALDI-TOF-MS spectra of permethylated glycolipids from TMEM165 KO HEK293 cells. (D) MALDI-TOF-MS spectra of permethylated glycolipids from TMEM165 KO HEK293 cells treated with MnCl₂ (100 µM, 36 hours). The symbols representing sugar residues are as follows: yellow closed square, N-acetylgalactosamine; blue circle, glucose; yellow circle, galactose; open diamond, sialic acid. Linkages between sugar residues have been removed for simplicity. All GSLs are present as d18:1/C16:0 and d18:1/C24:0 ceramide isomers.

ICAM-1 levels increased significantly in vitro on galactose treatment of TMEM165-CDG patient cells

Immunohistochemistry showed a significant decrease in glycosylated intercellular adhesion molecule 1 (ICAM-1) in cultured skin fibroblasts derived from patient 1 carrying the homozygous c.792+182G>A mutation (Supplemental Fig. 2). ICAM-1 has eight N-glycosylation sites and its cell surface expression is known to be markedly diminished in CDG cells. This finding is consistent with the global decrease of cell surface glycoprotein associated with TMEM165 deficiency. Following five days of galactose supplementation in culture media, a slight but significant improvement of ICAM-1 protein expression was observed by immunohistochemistry (Supplemental Fig. 2). This supports the beneficial effect of galactose supplementation in TMEM165 deficiency. The up-regulation of ICAM-1 cell surface expression was significant at 2.0 and 10 mM d-galactose supplementation, with a 2.5-fold and 3.5-fold increase (P = 0.006 and 0.02), respectively. ICAM-1 western blot after 7 days of galactose supplementation also showed a 3.9-, 2-, and 2.8-fold increase at 0.75, 2.0, and 5 mM, respectively (data not shown).

Galactose supplementation is well tolerated and improves biochemical parameters in patients with TMEM165-CDG

Both patients were compliant with d-galactose supplementation during the study. Neither patient had diarrhea or vomiting during the trial. No other adverse effects or galactosuria were reported.

Improved well-being and increased activity levels were reported in both patients. Both, although carrying the same homozygous TMEM165 variant, showed substantial variability between laboratory parameters. The few overlapping abnormalities included increased baseline levels of blood creatine kinase, transaminases, prothrombin time, and thyroid-stimulating hormone. Furthermore, patient 1 had decreased factor IX activity and IGF1 and IGFBP3 levels. These levels were normal in the 16-year-older, adult patient, patient 2. He had decreased activated thromboplastin time (APTT) and antithrombin. Manganese levels were normal in blood samples.

Upon administration of oral d-galactose, there was either a prospective improvement or no substantial effect on the different laboratory parameters. Positive effects of galactose were observed for APTT, antithrombin, IGF1, IGFBP3, and alanine aminotransferase levels in either of the patients. Factor IX deficiency did not normalize in patient 1, but its improvement was apparently sufficient to normalize APTT. Alanine aminotransferase, but not aspartate aminotransferase, improved in patient 2. During the 18 weeks of d-galactose supplementation, there was no effect on creatine kinase, thyroid-stimulating hormone, and cholesterol levels in either patient. To normalize parameters, 1 g/kg/d was sufficient, and no additional benefit was seen from increasing the dosage to 1.5 g/kg/d (Table 1).

Galactose treatment improves N-glycosylation in patients with TMEM165-CDG

To follow changes in N-glycosylation during galactose supplementation serum N-glycan analysis was performed by mass spectrometry. Figure 5 shows the structures of the N-linked glycans on serum glycoproteins from 2 patients before and after galactose therapy. Before galactose treatment, a strong presence of undersialylated and undergalactosylated glycans (Hex5HexNAc4 was at m/z 2040, and Hex4HexNAc4 was at m/z 1836) was observed. Interestingly, under galactose therapy, a consistent decrease in the abundance of the structures m/z 1836 (−14%) and 2040 (−14%) was seen. These results demonstrate the positive effects of the galactose therapy on glycosylation. To confirm these results, analysis was performed of the serum transferrin isoelectric focusing. Patient 1, the most severely affected, presented no substantial changes in 2-, 3-, and 4-sialo transferrin glycoforms under galactose therapy. However, a 15% decrease in 1-sialo and an 8% increase in 5-sialo transferrin glycoforms were observed. For the second patient, a 36% decrease in 1-sialo, a 55% decrease in 2-sialo, a 5% increase in 4-sialo, and a 24% increase in 5-sialo transferrin glycoforms were observed under galactose therapy. Altogether, these results highlight that galactose supplementation reduces the glycosylation defect initially observed in patients deficient in TMEM165.

Figure 5.

Galactose therapy ameliorates the N-glycosylation in individuals with TMEM165deficiency. (A and C) MALDI-TOF-MS spectra of the permethylated N-glycans from sera of 2 patients with TMEM165 deficiency who carry the homozygous c.792+182G>A mutation. (B and D) MALDI-TOF-MS spectra of the permethylated N-glycans from sera after 18 weeks of galactose therapy. The symbols representing sugar residues are as follows: closed square, N-acetylglucosamine; open circle, mannose; closed circle, galactose; open diamond, sialic acid; and closed triangle, fucose. Linkages between sugar residues have been removed for simplicity.

Discussion

TMEM165 deficiency is a subtype of CDG, a group of inherited diseases with impaired glycan biosynthesis (13–16). Here, we demonstrate that Golgi hypoglycosylation due to deficiency in TMEM165, a putative Golgi ion transporter and regulator in Golgi Mn²⁺ homeostasis, improves on galactose supplementation (3).

Recently, mutations have been described in another solute carrier transporter causing CDG, namely, the electroneutral Mn²⁺/HCO₃⁻² and Zn²⁺/HCO₃⁻² transporter encoded by SLC39A8 (17). This clearly demonstrates the crucial requirement of ion transporters to provide divalent cations for the activity of Golgi glycosyltransferases. Interestingly, Park and collaborators (17) have shown that oral galactose supplementation resulted in complete normalization of glycosylation in SLC39A8-CDG. Dietary supplementation of galactose has already been reported to be successful in patients with PGM1 deficiency and SLC35A2 deficiency (18–20). We wanted to test whether galactose treatment is beneficial in both TMEM165 KO cells and in patients with TMEM165-CDG. Oral galactose treatment of our patients was well tolerated and compliance was well achieved. No galactosuria occurred on incremental dosage. Manganese treatment was not attempted, because of normal blood levels of manganese and potential toxicity.

Galactose supplementation improved several, but not all, laboratory results in our patients (Table 1); specifically, coagulation parameters, IGF1, and IGFBP3 were improved in patient 1. The interindividual variability between patients (already at baseline) possibly was due to their different ages (16 years apart). Although most patients with CDG who underwent a trial of d-galactose supplementation in the past because of galactosylation defect showed some biochemical and/or clinical improvement after 2 to 3 months, we cannot rule out that a longer dietary period could have more obvious beneficial effects in patients with TMEM165-CDG. Consistent with these clinical observations, serum N-glycan analysis by mass spectrometry showed improved galactosylation in both patients under galactose therapy.

At the cellular level, a strong galactosylation defect was observed in both N-glycoproteins and glycolipids before treatment. Mn²⁺ treatment completely suppressed the glycosylation defect. However, differences were observed under galactose treatment, resulting in a complete normalization of N-linked glycans and a partial normalization of glycolipids. In TMEM165 KO cells, Mn²⁺ supplementation completely rescued the transfer of both Gal and GalNAc on glycolipids, whereas the galactose treatment only rescued the transfer of galactose residues on glycolipids. This shows that TMEM165 deficiency not only impairs the activities of Golgi galactosyltransferases but also those of Golgi GalNAc transferases. Altogether, this reinforces the idea that the observed impaired Golgi glycosylation in TMEM165-deficient cells results from Golgi Mn²⁺ disturbances. How does galactose supplementation rescue the galactosylation defect? Galactose is known to be transported into the cells by several glucose transporters such as GLUT1, GLUT3, and GLUT4, the insulin-independent transporter (21). Within the cytosol, galactose is rapidly phosphorylated into galactose-1-phosphate. UDP-galactose (UDP-Gal) is then generated by the galactose-1-phosphate uridyltransferase that catalyzes the transfer of a UMP group from UDP-glucose to galactose-1-phosphate (22). Thus, galactose supplementation might be able to increase the intracellular level of UDP-Gal. Translocation of UDP-Gal from the cytosol to Golgi lumen occurs through the UDP-Gal transporter (UGT), which is an antiporter that mediates the exchange of cytosolic UDP-Gal and luminal UMP across the Golgi membrane. Two Golgi UDP-Gal transporters have been identified so far, UGT1 and UGT2 (23).

Although these transporters could play a role in the UDP-Gal Golgi luminal increase after galactose treatment, the import is strictly dependent on a UMP exchange (24, 25). In galactose-treated cells, we can reasonably hypothesize that the Golgi luminal UMP concentration, due to impaired Golgi galactose transporter activities, is too low to create a gradient sufficient to import UDP-Gal. Similarly, in a human disorder of UDP-Gal transport (SLC35A2 defect) treatment with galactose restores glycosylation and improves clinical symptoms in patients (19).

Another possibility is that cellular galactose entry is linked to Mn²⁺ entry. In that case, the glycosylation normalization observed under galactose treatment would be linked to a corresponding increase in cellular Mn²⁺. Our results show that this is unlikely, because glycolipid biosynthesis was still altered after galactose treatment in TMEM165-deficient cells.

The UGT transporters probably do not play any role in this process, because galactose treatment of CHO-Lec8 cells defective in UDP-Gal import in the Golgi increases galactosylation (19). One can hypothesize that UDP-Gal reaches the Golgi by using an endoplasmic reticulum (ER) transporter. Although no UDP-Gal transporters have clearly been identified in the ER so far, there are some promising candidates. In Arabidopsis thaliana, it was shown that AtUTr1 is a UDP-Gal/ UDP-Glc transporter transporting UDP-Glc 200 times faster than UDP-Gal (26). Another interesting potential candidate is the UDP-galactose transporter-related protein 1 SLC35B1 that localizes to the ER membrane. It is possible that the cytosolic UDP-Gal increase favors the ER UDP-Gal entry by such transporters and its transport toward the Golgi via the secretory pathway (27). Besides, it has also been shown that there is an association of the Golgi UDP-galactose transporter with UDP-galactose:ceramide galactosyltransferase, localized in the ER. This allows UDP-galactose entry in the ER that can be transported to the Golgi apparatus (28).

Park et al. (17) showed a robust effect of galactose supplementation in SLC39A8-deficient patients with manganese deficiency as an underlying cause of their glycosylation defect. As in TMEM165, supplementing manganese completely rescued the defect while there was only a partial restitution on galactose treatment. In this process, the most likely enzymatic step is β-1,4-galactosyltransferase, a manganese-dependent enzyme essential for glycan synthesis. Along these lines, we hypothesize that in TMEM165-CDG, the improvement due to galactose therapy is through a combination of increased cellular galactose uptake (via GLUT1), increased transport of UDP-Gal to the Golgi, and an increased activity of intraluminal galactosyltransferases, such as B4GALT1.

In conclusion, the beneficial effects of oral galactose supplementation on glycosylation adds TMEM165 deficiency to the list of CDG treatable by galactose.

Abbreviations:

APTT

activated thromboplastin time

CDG

congenital disorders of glycosylation

ER

endoplasmic reticulum

GSL

glycosphingolipid

IRB

institutional review board

KO

knockout

MALDI-TOF-MS

matrix-assisted laser desorption/ionization time of flight mass spectrometry

m/z

mass-per-charge

TBS-T

tris-buffered saline and Polysorbate 20

UDP-Gal

UDP-galactose

UGT

UDP-Gal transporter

v/v

volume-to-volume ratio.