High-dose N -acetyl cysteine reduces GFAP levels: Exploratory results from a pilot double-blind,placebo-controlled randomized trial

Introduction

Multiple sclerosis (MS) is a chronic autoimmune demyelinating condition that is characterized by neuroinflammation and neurodegeneration. ¹ Current disease-modifying therapies have a modest effect on the prevention of disability progression independent of relapse activity (PIRA), suggesting that, although residual inflammatory processes exist in non-active progressive forms of MS (PMS), mechanisms such as neuronal damage associated with chronic demyelination, metabolic failure, and oxidative stress are potentially more critical determinants.^1,2

Oxidative injury is mediated by endogenous accumulation of reactive oxidative species (ROS) released by innate central nervous system (CNS) immune cells, iron deposition in the CNS parenchyma, and/or reduction of antioxidant synthesis leading to neurodegeneration and demyelination. ² Glutathione (GSH), a major intrinsic antioxidant, has been shown by brain magnetic resonance imaging (MRI) spectroscopy to be decreased in the gray matter of MS patients compared to healthy controls. ³ N-acetyl cysteine (NAC) directly scavenges ROS and may restore neuronal GSH by providing the rate-limiting substrate for its intracellular synthesis. ³ As NAC crosses the blood–brain barrier at high oral doses, it is being tested as a promising add-on therapy in PMS patients.^4,5

In this study, we investigated the effect of high-dose oral NAC on 20 serum biomarkers in PMS patients in a pilot phase 2 randomized, placebo-controlled, double-blinded, clinical trial (NCT02804594).

Methods

Individuals with PMS aged 18–75 years, with at least 1 year since progressive symptom onset, a Modified Fatigue Impact Scale (MFIS) of more than 38, and an Expanded Disability Status Scale (EDSS) 2.0–6.5, were randomized 2:1 to NAC 1250 mg TID or placebo for 4 weeks. Further details on the trial protocol are provided in Krysko et al. ⁴

Serum biomarkers were assessed utilizing a multi-protein assay panel, based on Olink proximity extension assay methodology.^6,7 This platform enabled measurement of an a priori defined set of biomarkers, including neurofilament light chain (NfL), glial fibrillary acidic protein (GFAP), and myelin oligodendrocyte glycoprotein (MOG), providing mechanistic insights into pathways of potentially averted tissue damage, as well as interleukin-12 subunit beta (IL-12B), a surrogate marker of NAC effect. Additional proteins were available for post hoc evaluation. Changes in levels from baseline to week 4 between NAC and placebo arms were compared with the Wilcoxon rank-sum tests, as pre-specified for efficacy outcomes using a modified intention-to-treat analysis. The last observation carry-forward was performed to handle one missing follow-up sample in the NAC arm.

A sensitivity analysis to further adjust efficacy outcomes for age was conducted using permutation-based analysis of covariance (ANCOVA). Since this was an exploratory investigation, analyses were not adjusted for multiple comparisons. All statistical analyses were conducted in R version 4.3 (R Foundation, Austria) using two-sided tests.

Results

Fifteen individuals were randomized to receive either NAC (n = 10) or placebo (n = 5), with one subject assigned to the placebo arm not providing blood samples and not included in this study, resulting in a final placebo sample of four subjects. Baseline characteristics between the treatment arms were well balanced except for age. In the NAC arm, 80% of participants were female, with a median age of 51.0 years (interquartile range (IQR) = 44.50–56.75), whereas in the placebo arm, 75% of participants were female, with a higher median age of 63.5 years (IQR = 52.50–71.30).

Changes in GFAP levels, a marker of astrocytic damage, were significantly different between treatment groups, with levels decreasing in NAC while increasing in the placebo-treated participants (median change: −10.1 pg/mL vs. 39.5 pg/mL, respectively; p = 0.008) (Figure 1 and Supplemental Figure S1). NfL, a marker of neuro-axonal damage, remained stable over time with no evidence of level changes between the two arms (median change: 0.25 pg/mL vs. 0.61 pg/mL, respectively; p = 0.777). MOG, a marker of myelin injury, tended to decrease when comparing the difference in changes between NAC and placebo arms, especially after age adjustment (median change: 0.15 pg/mL vs. 4.8 pg/mL, respectively; unadjusted p = 0.106, adjusted p = 0.057). Additional differences in change between NAC and placebo were observed in IL-12B (median change: −6.5 pg/mL vs. 17.2 pg/mL, respectively; p = 0.054) and sCD6 levels (median change: 14.3 pg/mL vs. −19.9 pg/mL; p = 0.054) approached but did not meet the significance threshold. Results for the full panel of proteins are provided in Supplemental Figure S2 and were robust to age adjustment (data not shown).

OPEN IN VIEWER

Figure 1.

Changes in serum biomarker levels over 4 weeks between NAC and placebo treatment arms: (a) GFAP, (b) NfL, (c) MOG, (d) IL-12B, and (e) sCD6.

Discussion

In this study, a 4-week course of high-dose oral NAC in patients with PMS led to a decrease in markers of ongoing astrocytic and myelin damage, suggesting a biologically plausible impact on mechanisms implicated in PIRA. This trial was too short in duration to evaluate the impact of NAC on clinical PIRA, but previous observations have shown that GFAP levels, a marker of astrocytic damage, are 50% higher in patients experiencing PIRA than in matched stable MS patients. ⁸ Furthermore, NfL levels, a marker of neuro-axonal damage, were no longer predictive of PIRA after adjustment for GFAP, ⁸ indicating NfL may be less suited for identifying progression risk than for relapse activity. Therefore, the observed reductions in GFAP levels alongside stable NfL levels are consistent with a potential delay in biological processes linked to disability progression.

This biomarker profile is consistent with proposed mechanisms of disease progression and NAC treatment effect. Post-mortem studies suggest that axons remain relatively preserved while macrophages phagocytose reactive astrocytes and oligodendrocytes at the margins of expanding brain lesions. ⁹ Myelin debris becomes susceptible to peroxidation by ROS as they accumulate in a CNS environment of impaired microglial clearance, characteristic of PMS, contributing to experimental autoimmune encephalomyelitis severity.^1,10 Directly counteracting this process, NAC has been shown to reverse lipid peroxidation and ameliorate disease severity in mice. ¹⁰ Our findings of relative reductions in MOG and GFAP levels in NAC recipients support the hypothesis that this pathway may contribute to the progression of disability in humans.

Furthermore, inhibition of IL-12B assembly has been previously shown in mice exposed to NAC, ¹¹ which, in keeping with decreased IL-12B levels in NAC-treated subjects, may indicate a direct effect on activated myeloid cells in PMS. Finally, even though sCD6 results from proteolytic cleavage from its membrane-bound form following T-cell activation, its functional role and relationship to NAC treatment remain uncertain. ¹²

In summary, this exploratory study suggests NAC may be a promising agent to delay MS progression. However, given the relatively small sample size and short follow-up period, our findings should be considered hypothesis-generating and require confirmation. The ongoing NACPMS trial (NCT05122559) will further elucidate the potential of NAC as a neuroprotective agent in PMS.

Supplemental Material