A Double-Blind, Placebo-Controlled, Crossover Study of Magnesium Supplementation in Patients with XMEN Disease

Abstract

X-linked MAGT1 deficiency with increased susceptibility to Epstein-Barr virus (EBV) infection and N-linked glycosylation defect (XMEN) disease is an inborn error of immunity caused by loss-of-function mutations in the magnesium transporter 1 (MAGT1) gene. The original studies of XMEN patients focused on impaired magnesium regulation, leading to decreased EBV-cytotoxicity and the loss of surface expression of the activating receptor “natural killer group 2D” (NKG2D) on CD8⁺ T cells and NK cells. In vitro studies showed that supraphysiological supplementation of magnesium rescued these defects. Observational studies in 2 patients suggested oral magnesium supplementation could decrease EBV viremia. Hence, we performed a randomized, double-blind, placebo-controlled, crossover study in 2 parts. In part 1, patients received either oral magnesium l-threonate (MLT) or placebo for 12 weeks followed by 12 weeks of the other treatment. Part 2 began with 3 days of high-dose intravenous (IV) magnesium sulfate (MgSO₄) followed by open-label MLT for 24 weeks. One EBV-infected and 3 EBV-naïve patients completed part 1. One EBV-naïve patient was removed from part 2 of the study due to asymptomatic elevation of liver enzymes during IV MgSO₄. No change in EBV or NKG2D status was observed. In vitro magnesium supplementation experiments in cells from 14 XMEN patients failed to significantly rescue NKG2D expression and the clinical trial was stopped. Although small, this study indicates magnesium supplementation is unlikely to be an effective therapeutic option in XMEN disease.

Introduction

X-linked MAGT1 deficiency with increased susceptibility to Epstein-Barr virus (EBV) infection and N-linked glycosylation defect (XMEN, formerly known as X-linked immunodeficiency with magnesium defect, EBV infection, and neoplasia) disease is a primary immunodeficiency (PID) caused by loss-of-function (LOF) mutations in the magnesium transporter 1 (MAGT1) gene [1]. Loss of MAGT1 predominantly affects immunity although other organs are affected [2–5]. Clinical manifestations include recurrent sinopulmonary and ear infections, lymphadenopathy (LAD) and/or splenomegaly, persistent cutaneous viral infections, and autoimmune (AI) cytopenias [5]. Most patients have transient elevations in their liver enzymes with preserved hepatic function [5]. EBV-associated lymphoproliferative disease (LPD) and lymphoma are the most common cause of morbidity and mortality but EBV-naïve children have a similar immunophenotype to EBV⁺ patients [1, 5]. Invariably, XMEN patients have decreased surface expression of the activating receptor “natural killer group 2D” (NKG2D) on CD8⁺ cytotoxic T lymphocytes (CTL) and natural killer (NK) cells [5].

XMEN patient lymphocytes have decreased surface expression of NKG2D and CD70, which are required for EBV control [5–11]. Early in vitro studies showed that supplementation of culture medium with supraphysiologic concentrations of magnesium sulfate (MgSO₄) led to a partial but significant recovery of NKG2D expression and better cytotoxic function in both XMEN NK and CD8⁺ T cells [6, 7, 12]. Observational studies in two XMEN patients suggested that oral magnesium l-threonate (MLT) supplementation resulted in improved NKG2D expression and decreased EBV viral loads (EBV-VL) with stronger and faster responses observed in one patient treated with continuous intravenous (IV) MgSO₄ for 3 days [7]. However, minimal or no increase in NKG2D expression has also been observed by our group and others in a few XMEN patients after open-label administration of oral Mg²⁺ supplementation [1, 4]. To rigorously assess the safety and efficacy of Mg²⁺ supplementation in XMEN disease, we designed this double-blind, placebo-controlled, crossover study.

Recently, MAGT1 has been shown to be a critical non-catalytic subunit of the oligosacchryltransferase (OST) complex that carries out asparagine (N)-linked glycosylation (NLG) [3, 5, 13, 14]. We recently showed that the decreased expression of NKG2D and CD70 in XMEN lymphocytes is the result of impaired glycosylation, making XMEN disease a congenital disorder of glycosylation (CDG) [5, 7, 15]. However, Mg²⁺ is critical for proper NLG and hypomagnesemia decreases MAGT1-dependent glycosylation [6]. Whether magnesium supplementation can rescue glycosylation defects in XMEN patients is unclear.

Methods

Study Design

From May 2016 through February 2019, we conducted a single-center, two-part, phase 1 and 2 trial at the National Institutes of Health (NIH) Clinical Center (CC) to evaluate the safety and efficacy of Mg²⁺ supplementation in XMEN (Figs. 1 and 2, Table 1, and Supplement 1). Part 1 was a randomized, double-blind, placebo-control crossover study; part 2 was an open-label trial. We divided patients into 2 cohorts based on their baseline blood EBV-VL. Patients in cohort 1 (high EBV group) had EBV-VL ≥ 5000 copies/mL or log10 ≥ 3.7 IU/mL; cohort 2 (low/no EBV group) was defined as EBV-VL < 5000 copies/mL or log10 < 3.7 IU/mL and all enrolled patients who met this standard were EBV naïve. The trial began with a 2-week washout in which patients avoided magnesium-containing supplements or vitamins. In part 1, we randomly assigned patients to receive escalating doses of either oral MLT or placebo for 12 weeks followed by crossover treatment for another 12 weeks (Fig. 1 and Table 1). Successful completion of the trial was defined as a ≥ 0.5-log reduction in the absolute number of EBV-infected B cells (#EBV⁺Bc) or a ≥ twofold increase in NKG2D surface expression on CTLs during MLT treatment versus placebo for EBV-positive or EBV-naïve patients, respectively. Patients who did not meet these criteria underwent a second 2-week washout and transition to part 2 in which they were hospitalized to receive 3 days of IVMgSO₄ followed by escalating doses of oral MLT for 24 weeks (Fig. 1). The study end was defined as the last visit of the last patient.

Fig. 1.

Study design. A Diagram of the timeline for the XMEN clinical trial

Fig. 2.

Screening, randomization, and follow-up. A Consort diagram of the XMEN clinical trial

Table 1.

Dose-escalation scheme

Patient weight	Starting dose		2 weeks after start		4 weeks after start		6 weeks after start
	Elemental Mg2+ per day	No. capsules per day	Elemental Mg2+ per day	No. capsules per day	Elemental Mg2+ per day	No. capsules per day	Elemental Mg2+ per day	No. capsules per day
< 30 kg	96 mg	2	144 mg	3	192 mg	4	240 mg	5
≥ 30 kg	192 mg	4	288 mg	6	336 mg	7	384 mg	8

No of capsules:	Dosed as:
2	1 morning and 1 evening
3	2 morning and 1 evening
4	2 morning and 2 evening
5	2 morning, 1 afternoon, 2 evening, at ~ 8-h intervals
6	2 morning, 2 afternoon, 2 evening, at ~ 8-h intervals
7	3 morning, 2 afternoon, 2 evening at ~ 8-h intervals
8	3 morning, 2 afternoon, 3 evening, at ~ 8-h intervals

Study Oversight

This study followed the principles of the Declaration of Helsinki, the International Conference on Harmonization (ICH)-Good Clinical Practice (GCP) guidelines, and the National Institute of Allergy and Infectious Diseases (NIAID) guidelines. The protocol and consent forms were approved by an independent Institutional Review Board (IRB) at the NIAID/NIH. An independent Data and Safety Monitoring Board (DSMB) performed regular safety surveillance. This trial is registered at the US National Institutes of Health (ClinicalTrials.gov) #NCT02496676.

Study Population

Patients were eligible for participation if they were 6 years or older, had molecularly confirmed XMEN disease, and were willing to restrict Mg²⁺-containing supplements to the study agent (Supplement 1). Exclusion criteria included chemotherapy or radiotherapy for lymphoma within 12 months prior to enrollment, rituximab exposure within 6 months prior to enrollment, advanced heart block, renal insufficiency, hypermagnesemia, and a history of hypersensitivity to any of the study agents (Supplement 1). Patients were recruited by physician referral and from other studies approved by NIH IRBs. All patients or their legal representatives provided written informed consent prior to randomization. Control samples were unmatched healthy volunteers under NIH IRB-approved protocols.

Study Interventions

Participants self-administered MLT and placebo. MLT is an over-the-counter oral Mg²⁺ supplement and was purchased from commercial sources. Each 670-mg capsule of MLT contained 48 mg (4 mEq) of elemental Mg²⁺. Matching placebo capsules comprised of Avicel PH102 and starch 1500 LM were produced by the NIH CC Investigational Drug Management Research Section (IDMRS). The dose-escalation scheme was based on the patient’s weight at baseline and tolerance to the study drug (Table 1). Tolerance was assessed every 2 weeks when the dose was escalated. Participants who experienced intolerance were reverted to the last tolerated dose and remained at that dose for the remaining duration of part 1. At the beginning of part 2, patients received IV MgSO₄ at 30 mg/kg/day divided into 3 daily doses (every 8 h) for 3 consecutive days prior to restarting escalating doses of oral MLT up to the highest dose tolerated from part 1.

Randomization

Patients were randomly allocated to arm 1 or arm 2 using a permuted block design with a block size of 20 (Fig. 2). The study statistician provided the randomized list of arm assignments. Participants, clinicians, and researchers were unaware of the trial group assignments and blinded from potentially revealing laboratory results.

Study Outcomes

The primary endpoint for EBV-positive patients was the difference between the #EBV⁺Bc as determined by EBV fluorescent in situ hybridization (FISH) after 12 weeks of MLT versus placebo. For EBV-naïve patients, the primary endpoint was an NKG2D expression change in CD8⁺ T cells. Key secondary endpoints were the incidence and severity of adverse events (AEs) and improvement in EBV viral load (cohort 1) or NKG2D expression (cohort 2) in part 2.

Cells

Peripheral blood mononuclear cells (PBMCs) were isolated by Ficoll-Hypaque density gradient centrifugation and cryopreserved in liquid nitrogen in RPMI-1640 medium with 10% dialyzed fetal bovine serum (FBS) and 7.5% dimethyl sulfoxide (DMSO).

Flow Cytometry

NKG2D surface expression on CD8⁺ T cells and CD16⁺CD56⁺ NK cells was performed as previously described [5]. In brief, cryopreserved PBMCs were thawed and incubated with human Fc receptor blocking solution (Human TruStain FcX, Biolegend) and viability dye (Zombie Green, Biolegend) in phosphate-buffered saline (PBS) at 25 °C for 20 min. Cells were stained in ice-cold FACS buffer (PBS, 2% FBS, and 1% sodium azide) with fluorochrome-conjugated antibodies on ice for 30 min. The following antibodies (Biolegend) were used: anti-CD3 (UCHT1), CD4 (OKT4), CD8 (RPA-T8), CD16 (3G8), CD56 (HCD56), and either CD314 (NKG2D) (1D11) or mouse IgG1κ isotype control. T cell blasts were processed similarly and included an anti-CD70 (113–16) antibody. Cells were washed three times in FACS buffer prior to acquisition on Fortessa or LSRII instruments (BD Biosciences) and analyzed with FlowJo.

Immuno-FISH

PBMCs were resuspended in PBS at 10⁶ cells/mL and incubated with human Fc receptor blocking solution and viability dye (LIVE/DEA Fixable Yellow Dead Cell Stain Kit, ThermoFisher Scientific) for 20 min. Cells were washed in PBS and stained with biotinylated anti-CD20 (2H7) antibody in FACS buffer for 30 min at 25 °C. After two washes, cells were incubated with fluorescently labeled streptavidin for 20 min at 25 °C. The cells were then fixed with 5% acetic acid-4% paraformaldehyde in PBS for 20 min and permeabilized with 0.5% Tween 20 in PBS for 10 min at 25 °C. Cells were then hybridized with a fluorescein-labeled EBV-encoded small RNA (EBER) PNA probe (Agilent) in Hybridization Buffer (37.5 mM NaCl, 6.25 mM EDTA, 62.5 mM Tris–HCl pH 7.5, and 37.5% formamide) at 56 °C for 90 min. Cells were washed twice with 0.5% Tween 20 at 56 °C. The signal was amplified using the Alexa Fluor 488 Signal-Amplification Kit for Fluorescein and Oregon Green Conjugated Probes (ThermoFisher Scientific). Samples were acquired on FACSymphony A3 (BD Biosciences) and analyzed with FlowJo.

Determination of Transferrin and Apolipoprotein CIII (Apo-CIII) Isoforms

Serum transferrin (sTf) and Apo-CIII isoforms were identified by immunoaffinity liquid chromatography and electrospray mass spectrometry by Mayo Clinic Laboratories (Rochester, MN) as previously described [5, 16].

In Vitro Mg²⁺ Supplementation

Cryopreserved PBMCs were thawed and washed in pre-warmed complete RPMI medium (cRPMI): RPMI-1640 medium (Lonza) with 10% FBS, 2 mM glutamine, penicillin, and streptomycin (100 U/mL each, Invitrogen). Cells (1 × 10⁶/mL) were stimulated with anti-CD3 (HIT3a, Biolegend) and anti-CD28 (CD28.8, Biolegend) antibodies at 1 μg/mL each in cRPMI with or without the indicated amount of MgSO₄. After 2 days, cells were supplemented with recombinant human Interleukin-2 (rhIL-2, R&D) at 100 U/mL. Cells were then maintained in cRPMI plus rhIL-2 with or without supplemental MgSO₄.

Sample Size and Study Analysis

We set a target sample size of 10 patients for each cohort. With n = 10, our study would have a power of 90% to detect a 1.15 standard deviation change by a two-sided paired t-test with α = 0.05. To assess the fold change in NKG2D, the mean fluorescence intensity (MFI) after Mg²⁺ supplementation was normalized to the placebo (for part 1) or baseline II (for part 2) value for that patient. We then performed a two-sided, one-sample t-test with a known mean of 1 to measure the 95% confidence interval (CI).

In vitro Mg²⁺ supplementation experiments were analyzed by Bayesian estimation supersedes the t-test (BEST) analysis [17, 18]. BEST yields a probability distribution of the effect size, giving the probability of multiple alternative hypotheses (e.g., Mg²⁺ supplementation causes an increase in NKG2D expression and Mg²⁺ supplementation causes at least a twofold increase in NKG2D expression). The distribution median is the average response of the sample, and the 95% high density interval (HDI) has a 95% chance of capturing the true average response. Like the t-test, a one-sample BEST analysis measured fold change in NKG2D expression after Mg²⁺ supplementation. To assess changes in CD70 expression, a two sample BEST analysis compared the difference in CD70⁺ T cells with and without Mg²⁺ supplementation. Analysis was conducted in R (version 4.03, Supplement 2).

Results

Patients

Eight patients were assessed for eligibility. One EBV-naïve patient was excluded because of age below the minimum requirement, and one EBV-positive patient was excluded because of recent rituximab therapy for refractory AI cytopenias. A third EBV-positive patient declined to participate (Fig. 2). Of the 5 patients who were enrolled and underwent randomization, two were EBV-positive (XMEN01 and XMEN03) and 3 EBV-naïve (XMEN02, XMEN04, and XMEN05). The baseline demographic and clinical characteristics of the patients are summarized in Table 2, and clinical vignettes are provided in Supplement 3. XMEN03 was removed from the trial because he developed diarrhea and did not tolerate the lowest dose of the study drug (placebo) during the first period of part 1 (Fig. 2). After completing part 1, XMEN05 was removed from the trial due to the asymptomatic elevation of liver enzymes during IV MgSO₄ treatment in part 2 (Fig. 2). XMEN01, XMEN02, and XMEN04 completed both part 1 and part 2 of the trial. Inability to obtain a clinical-grade placebo prevented enrollment of additional patients.

Table 2.

Patient characteristics at baseline

	XMEN01	XMEN02	XMEN03	XMEN04	XMEN05
Age	21	9	14	8	12
Race	White	White	Hispanic	White	White
Weight (kg)	89.8	32.7	102.2	21.7	30.9
Height (cm)	175	129.5	153	124	145.8
BMI	28.6	19.5	43.5	14.1	14.5
EBV PCR (Log10)	4.53	Undetectable	4.85	Undetectable	Undetectable
AST (0–40)†	20	26	42	43	22
ALT (5–30)†	32	40	28	30	29
Total magnesium (0.66–1.07)*	0.85	0.86	0.83	0.78	0.82
Ionized magnesium (0.44–0.59)*	0.55	0.54	0.5	0.49	0.47
Creatinine (0.67–1.17)#	0.87	0.37	0.51	0.4	0.47

BMI, body mass index. Normal reference values are shown in parenthesis.

^†U/L

^*mmol/L

^#mg/dL

Efficacy

Because improved cytotoxic function against EBV-infected cells may cause transient increases in blood EBV DNA due to the release of virus from dying cells, we used a flow cytometric FISH assay to directly quantify the #EBV⁺Bc and %EBV⁺Bc. In patients with active EBV infection, the number of EBER-positive cells by FISH correlates well with the EBV-VL determined by PCR [19, 20]. XMEN01, the only EBV-infected patient to complete the study did not have a ≥ 0.5-log reduction in either the #EBV⁺Bc or the EBV-VL in either part 1 or 2 (Fig. 3A, B).

Fig. 3.

Primary Outcomes. A The percent and count of EBV⁺ B cells in XMEN01, as determined by EBV fluorescent in situ hybridization (FISH). B The EBV viral load by polymerase chain reaction (PCR) in XMEN01. Quantification of the flow cytometry mean fluorescence intensity (MFI) of NKG2D on CD8⁺ T cells (C) and NK cells (D) throughout part 1, the double-blinded, placebo-controlled crossover part of the clinical trial. NKG2D staining results from part 2, the open-label portion of the study is shown for CD8⁺ T cells (E) and NK cells (F). Measurements are shown during baseline (blue), on oral magnesium (green), on placebo (magenta), on IV magnesium (dark green), and post-study (gray). Healthy controls (HCs) are shown in black. In C–F, the dotted line shows the average MFI of the baseline for that part of the trial

In week 24 (while on placebo), XMEN01 developed acute infectious gastroenteritis (due to enteropathogenic E. coli and rotavirus A), and his EBV-VL and inflammatory markers transiently increased. He then had a minor (≤ 0.5 log) decrease in the number of EBV-infected cells while on IV MgSO₄ (Fig. 3A). However, the #EBV⁺Bc was high to begin part 2 and the decrease returned the number to baseline levels seen for part 1. Conversely, the part 2 baseline EBV-VL was 1 log lower than any other value but returned to levels seen throughout part 1 for the remainder of the trial (Fig. 3B).

Regardless of EBV status, none of the patients who completed part 1 experienced a ≥ two fold increase in NKG2D on CTLs or NK cells with oral MLT compared with placebo (Fig. 3C, D). However, we noticed a mild, 20% increase in the MFI of NKG2D on CTLs (95% CI: 1.02–1.38) but no change in NK cells (95% CI: 0.80–1.67). There was no change in NKG2D on CTLs (95% CI: 0.79–1.23) or NK cells (95% CI: 0.49–1.79) after 24 weeks of oral MLT treatment or during the 3-day infusion of IV MgSO₄ in part 2 (Fig. 3E, F).

We used the carbohydrate-deficient transferrin (CDT) test on cryopreserved serum as a surrogate for NLG function. Transferrin has two NLG sites carrying two negatively charged sialic acids each [21]. Defects in the assembly or transfer of the glycan precursor produce a Type-I CDT pattern, and those affecting downstream processing result in a type-II pattern [22]. XMEN patients have a sTf glycosylation pattern with increased mono-oligo/di-oligo ratio (type-I CDT), a normal or mildly elevated tri-sialo/di-oligo ratio (Type-II CDT), and a reduced a-oligo/di-oligo. This pattern was not changed with Mg²⁺ supplementation (Supplementary Fig. 1). In summary, none of the 4 patients completing the trial showed improvement in the clinical markers of XMEN disease.

Safety

Although ionized Mg²⁺ and total Mg²⁺ concentrations were slightly higher while on Mg²⁺ supplementation, these values were not close to the upper limit of normal. Urinary magnesium was predictably higher during Mg²⁺ supplementation (Fig. 4A). These results suggest that most of the Mg²⁺ is excreted, reducing the risk of side effects from hypermagnesemia, and that oral magnesium supplementation was difficult to achieve.

Fig. 4.

Limitations of supplementation. A Concentration of total blood, ionized blood, and urinary magnesium throughout the trial. Normal ranges for total and ionized blood magnesium are shown in gray. Measurements are shown during baseline (blue), on oral magnesium (green), on placebo (magenta), and on IV magnesium (dark green). Levels of alanine transaminase (ALT) (B) and aspartate transaminase (AST) (C) in the blood throughout the clinical trial. Normal ranges are shown in gray

We used the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE, version 4) to evaluate the severity of AEs [23]. There were no serious adverse events (SAE). The most common dose-escalation limiting AEs were gastrointestinal complaints (Table S1). Only one subject (XMEN05) developed grade 3 AEs possibly related to Mg²⁺ supplementation. He experienced asymptomatic elevation of the liver enzymes aspartate transaminase (AST) and alanine aminotransferase (ALT) with the highest doses of oral MLT during part 1. After the wash-out period, his liver enzymes decreased but again rose to grade 3 levels with IV MgSO₄, and the patient was removed from the protocol (Fig. 4B). After 3 weeks off the trial, his liver enzymes had declined to mildly above normal and within the range typically seen in XMEN patients (Fig. 4B) [5]. Two other patients developed grade 3 AEs while on placebo: XMEN01 developed vomiting and XMEN03 developed diarrhea (Table S1). Together, the safety data indicates that magnesium supplementation is generally well tolerated.

In Vitro Data

Recently, the model for disease pathogenesis in XMEN has incorporated new information that demonstrates impaired MAGT1-dependent glycosylation [3, 5, 6, 14]. Given this and the lack of efficacy in the first 4 subjects who completed the clinical trial, we reevaluated the in vitro effects of culturing cells from XMEN with supplemental Mg²⁺. T cell blasts from 14 XMEN patients with unique MAGT1 mutations and 6 healthy controls (HC) were assessed for NKG2D and CD70 surface expression by flow cytometry, respectively. After 7 days without Mg²⁺ supplementation, XMEN had 9% (HDI: 8.1–9.6%) of the NKG2D expression on CTLs seen in HCs (Fig. 5A, S2A) and 17% (HDI: 8.2–26.0%) fewer CD70⁺ T cells (Fig. 5B, S3B). With Mg²⁺ supplementation, a small but statistically significant increase in NKG2D was seen in both HC and XMEN (Fig. 5A). BEST statistical analysis showed a 100% probability of Mg²⁺ causing an increase in NKG2D but a 0% probability of this increase being more than twofold (Figure S2C–D). The magnitude of the change was 22% (HDI: 15–29%) in HC and 10% (HDI: 5–15%) in XMEN (Figure S2C–D). No change in CD70 expression was seen with Mg²⁺ supplementation in either HC or XMEN (Fig. 5B, S2E–F). A titration of Mg²⁺ supplementation showed the NKG2D increase is dose dependent with a larger change seen at 10 mM Mg²⁺ and no change at 1 mM Mg²⁺ (Figure S3A). No change was seen in CD70 expression in either CD4⁺ or CD8⁺ T cells at any concentration of Mg²⁺ (Figure S3B–C).

Fig. 5.

In vitro Mg²⁺ supplementation. Representative flow cytometric histogram of NKG2D staining on CD8⁺ T cells (A) and CD70 staining on pan T cells (B) from a healthy control (HC) and an XMEN patient. Cells were cultured in complete RPMI (cRPMI) supplemented with either 0 mM (black for HC and red for XMEN) or 5 mM Mg²⁺ (blue for HC and orange for XMEN). Isotype (NKG2D) or fluorescence minus one (CD70) is shown in gray. Quantification of 6 HCs and ≥ 14 XMEN is shown on the right. Error bars are mean ±S.D. Statistical analysis: Bayesian estimation supersedes the t-test (BEST) analysis. ***P(difference) ≥ 99.5%. ****P(difference) > 99.9%; n.s., not significant

Discussion

Mg²⁺ supplementation was considered an attractive therapy for XMEN disease because it is a readily available over-the-counter supplement and could distribute throughout the body and potentially rescue extra-hematopoietic manifestations [7]. However, the initial data were uncontrolled and limited and a placebo-controlled, double-blind study was essential to make a definitive recommendation regarding Mg²⁺ supplementation [7]. Unfortunately, in this current trial, we found that Mg²⁺ supplementation had no beneficial effect on the primary endpoints in either EBV-positive or EBV-negative patients. The only EBV-positive patient who completed the trial did not achieve a 0.5-log reduction in #EBV⁺Bc or EBV-VL. None of the patients who completed part 1 or part 2 of the trial showed a twofold increase in NKG2D on CD8⁺ T or NK cells. Except for asymptomatic elevations in liver enzymes with both MLT and IV MgSO₄ administration in one EBV-naïve XMEN patient, both MLT and IV MgSO₄ were generally well tolerated. It remains unclear if the AST/ALT elevations were induced by Mg²⁺ or were transient elevations seen with the mild liver disease found in XMEN patients [1, 5]. Thus, while the trial shows that Mg²⁺ supplementation is a safe intervention, evidence for effectiveness is lacking.

The lack of response in the first 4 participants in the clinical trial along with the recent discovery that MAGT1 can be a subunit of the OST complex involved in the glycosylation of select glycoproteins, including NKG2D and CD70, prompted us to more rigorously evaluate the effects of Mg²⁺ supplementation on T cells in vitro. We supplemented T cells from 14 XMEN patients, including cells from the same XMEN patients previously reported to have had a positive in vitro and in vivo response to magnesium supplementation [7]. In contrast with our previous study but consistent with some recent case reports, we did not observe a rescue of NKG2D or CD70 expression on T cells from XMEN patients cultured in media supplemented with MgSO₄.

Although Mg²⁺ supplementation in vitro did induce a 10% increase in NKG2D expression in XMEN cells, this same shift was also seen in HC. It is important to note that XMEN cells began with ~ 9% of the NKG2D expression of HC so with the shift, they have ~ 10% of HC. Such a small increase in NKG2D, in our judgment, may not be functionally significant since a 20% increase in NKG2D expression was seen in the CTLs of XMEN patients in part 1 of the trial but there was no change in clinical presentation. Furthermore, the increase was seen when supplementing T cell media with 5 or 10 mM MgSO₄, which is above the upper limit of Mg²⁺ achievable in serum. Recently, we showed that deprivation of Mg²⁺ in T cell culture causes a decrease in glycosylation of MAGT1-dependent glycoproteins [6]. It is possible that the increase in NKG2D expression seen here indicates that standard T cell media is relatively deficient in magnesium. If this is the case, Mg²⁺ likely has a role in NLG independent of MAGT1 since the shift was seen both in HC and XMEN, which is consistent with the recent cryoelectron microscopy structure of the OST complex showing a Mg²⁺ ion in the active site, not contacting MAGT1 [13].

We also evaluated the effect of Mg²⁺ supplementation in the glycosylation status of sTf for all the patients in the clinical trial and observed that magnesium supplementation did not rescue the glycosylation defect in sTf based on the CDT analysis.

One important limitation of our study was the low number of patients who completed the clinical trial, including only one EBV-positive patient. The study size was limited by the rarity of XMEN disease, lack of available placebo, and inability to include patients younger than 6 years. Additionally, the Mg²⁺ levels in the blood were not always in the upper limit of normal with Mg²⁺ supplementation.

The combination of (i) lack of efficacy in the enrolled patients, (ii) failure to rescue the abnormal glycosylation status of sTf, and (iii) inability to rescue NKG2D on T cell blasts in vitro led us to close the trial prematurely. Our conclusion from the data collected is that Mg²⁺ supplementation is unlikely to be an effective therapeutic intervention in XMEN disease. Although only 4 patients completed the clinical trial, our results from 14 different XMEN patients’ cells in vitro increase the power of our study and reinforces our conclusions.

Gene therapy is currently being pursued as an alternative to Mg²⁺ for treating XMEN disease. Hematopoietic stem cell transplant (HSCT), a standard treatment option for some PIDs, has a high mortality in XMEN patients due to post-transplant bleeding and is undertaken with extreme caution [1]. Recently, preclinical studies testing MAGT1 mRNA electroporation in PBMCs suggest the possibility of autologous blood transfusions with corrected lymphocytes [24]. However, this method involves a specialized care center, is transient, and only corrects the lymphocyte deficiencies of XMEN disease. Recent studies have shown widespread expression of MAGT1 suggesting that correcting the genetic defect only in hematopoietic cells may not reverse the deficiency in all appropriate cell types [6]. However, repair of hematopoietic cells may prevent high EBV-associated mortality in XMEN patients.

Conclusion

In our sample of 4 patients, there was no benefit from Mg²⁺ supplementation and it is probably unlikely to be effective for the management of XMEN disease. The recent discovery that MAGT1 is a facilitator of NLG offers new therapeutic targets aimed at restoring the glycosylation defect involved in this disorder.

Data Availability

Data are available on written request to the corresponding authors.