Axonal dysfunction is associated with interferon-γ levels in childhood-onset systemic lupus erythematosus: a multivoxel magnetic resonance spectroscopy study

Abstract

Objective

Axonal/neuronal damage has been shown to be a pathological finding that precedes neuropsychiatric manifestations in SLE. The objective of this study was to determine the presence of axonal dysfunction in childhood-onset SLE patients (cSLE) and to determine clinical, immunological and treatment features associated with its occurrence.

Methods

We included 86 consecutive cSLE patients [median age 17 (range 5–28) years] and 71 controls [median age 18 (5–28) years]. We performed proton magnetic resonance spectroscopic imaging using point resolved spectroscopy sequence over the superior–posterior region of the corpus callosum and signals from N-acetylaspartate (NAA), choline-based (CHO), creatine-containing (Cr), myo-inositol (mI), glutamate, glutamine and lactate were measured and metabolites/Cr ratios were determined. Complete clinical, laboratory and neurological evaluations were performed in all subjects. Serum IL-4, IL-5, IL-6, IL-10, IL-12, IL-17, TNF-α and INF-γ cytokine levels, antiribosomal P protein antibodies (anti-P) and S100β were measured by ELISA using commercial kits. Data were compared by non-parametric tests.

Results

NAA/Cr ratios (P = 0.035) and lactate/Cr ratios (P = 0.019) were significantly decreased in cSLE patients when compared with controls. In multivariate analysis, IFN-γ levels [odds ratio (OR) = 4.1; 95% CI: 2.01, 7.9] and depressive symptoms (OR = 1.9; 95% CI: 1.1, 3.2) were associated with NAA/Cr ratio. Increased CHO/Cr was associated with the presence of cognitive impairment (OR = 3.4; 95% CI: 2.034, 5.078; P < 0.001). mI/Cr ratio correlated with cumulative glucocorticoids dosage (r = 0.361, P = 0.014).

Conclusion

NAA and CHO ratios may be useful as biomarkers in neuropsychiatric cSLE. Longitudinal studies are necessary to determine whether they predict structural damage.

Rheumatology key messages.

• Childhood-onset systemic lupus erythematosus patients have axonal dysfunction associated with IFN-γ levels and depressive symptoms.

• Increased choline (CHO) levels in ¹H-MRS are associated with cognitive impairment and glucocorticoids use.

• Axonal dysfunction can be an early predictor of neuropsychiatric damage in childhood-onset systemic lupus erythematosus.

Introduction

Neuropsychiatric manifestations are frequently observed in childhood-onset SLE (cSLE) and are associated with worse prognosis and greater cumulative damage [1–4]. Previous studies have shown that neuronal damage precedes neuropsychiatric manifestations in cSLE and is related to inflammation, ischaemia and the effects of autoantibodies [5, 6].

In adult-onset SLE, axonal damage has been shown to occur before structural abnormalities [7, 8]. In these studies, proton magnetic resonance spectroscopy (¹H-MRS) was used to detect metabolites associated with axonal damage to detect early involvement of the CNS when structural MRI showed no focal changes [7–9]. The main brain metabolites commonly evaluated in ¹H-MRS are N-acetyl aspartate (NAA), considered a marker for assessing the viability and neuronal function, creatine (Cr), a cellular energy metabolism and cell marker, and choline (CHO), specifically phosphocholine (PCH) and glycerophosphorylcholine (GPC), which are precursors for biosynthesis and degradation of the cell membrane. Other markers of interest in SLE include the myo-inositol (mI), lactate (Lac), glutamate (Glu) and glutamine (Gln) [7–9]. However, no study has analysed axonal damage in cSLE through proton magnetic resonance spectroscopic imaging (¹H-MRSI) so far. Therefore, this study aimed to determine the presence of ¹H-MRSI changes in cSLE and to determine clinical, immunological and treatment features associated with its occurrence.

Methods

Subjects

Consecutive cSLE patients followed at a tertiary Pediatric Rheumatology Unit of the University of Campinas, Brazil were invited to take part in this cross-sectional study. Patients were included in the present study if they (i) fulfilled at least four of the ACR criteria [10, 11]; (ii) were below 18 years of age at disease onset; and (iii) had no infection at the time of study inclusion. We excluded patients who had a contraindication for MRI, such as the presence of a pacemaker or metal clips, and patients who were using a dental appliance, had claustrophobia or who had another disease that made it difficult to interpret the results (neonatal anoxia, development delay due to non-SLE causes, uraemia or metabolic abnormalities).

Controls matched by age, sex, self-declared ethnicity and sociodemographic characteristics were included as a control group. None of the controls had any history of delayed neurological development or chronic disease, including autoimmune diseases or diseases associated with brain damage.

This study was approved by the ethics committee at our institution (University of Campinas CAAE: 82701517.0.0000.5404), and informed written consent was obtained from each participant and/or legal guardian. A total of 86 cSLE patients and 71 controls were included in the study.

Clinical and neurological features

A complete clinical and neurological evaluation was performed in all individuals. Past clinical and neuropsychiatric manifestations for the SLE patients were retrieved by careful review of the medical charts.

Clinical evaluation and disease activity

The clinical manifestations were evaluated according to the definitions established by the ACR [10, 11]. Disease activity was measured by the SLEDAI 2000 at the time of MRI [12]. Active disease was considered when scores were >3 points [13]. Adjusted mean SLEDAI scores over time were calculated by careful review of the medical charts and previous exams [14].

Cumulative damage evaluation

Cumulative SLE-related damage was identified using the Systemic Lupus International Collaborating Clinics (SLICC)/ACR Damage Index (SDI) at the time of MRI. Damage was considered to be present if scores were ≥1 [15].

Neuropsychiatric evaluation

Neurological and psychiatric involvement was defined according to the ACR and Childhood Arthritis and Rheumatology Research Alliance (CARRA) guidelines for neuropsychiatric involvement [16, 17].

Mood and anxiety evaluation

All subjects completed the Beck Depression Inventory (BDI) [18] and Beck Anxiety Inventory (BAI) [19] at study entry and the day of MRI. For patients under 16 years of age, the Children's Depression Inventory (CDI) was applied [20]. BAI/BDI scales comprise 21 items, each describing a common symptom of depression or anxiety. The respondent is asked to rate how much he/she has been bothered by each symptom over the past month on a 4-point scale, ranging from 0 to 3. The items are summed to obtain a total score that can range from 0 to 63. The cut-offs used for the BDI are as follows: 0–13: no/minimal depressive symptoms; 14–19: mild depressive symptoms; 20–28: moderate depressive symptoms; and 29–63: severe depressive symptoms. The cut-off used for presence of depressive symptons was ≥14. For the BAI, the cut-offs are as follows: 0–7: no/minimal level of anxiety symptoms; 8–15: mild anxiety symptoms; 16–25: moderate anxiety symptoms; and 26–63: severe anxiety symptoms. The cut-off used for presence of anxiety symptons was ≥8. The CDI is a 27-item self-report inventory of childhood depression that taps a variety of depressive symptoms. Items assess negative mood, interpersonal difficulties, negative self-esteem, ineffectiveness and anhedonia in children of 7–17 years of age. Each item offers respondents three alternatives scored 0, 1 or 2 and accordingly raw scores range from 0 to 54. The cut-off used for the CDI was 17 (20).

Cognitive assessment

Cognitive function was assessed by the Wechsler Intelligence Scale (WIS). The WIS is applicable in children (WISC-48 III—aged ≤16 years and 9 months) [21] and adults (WAIS-III—aged ≥16 years and 10 months) [22]. This method consists of 15 subtests grouped into two sets: verbal and performance. Cognitive dysfunction was defined as a cognitive function with Z mean score ≤−2 s.d. or two or more functions with Z mean score between −1 and −2 s.d. [23].

Treatment with glucocorticoids

The treatment prescribed at study entry, as well as previously prescribed treatment, was carefully evaluated through medical chart review. The current and cumulative prednisone doses were calculated. Doses of oral and parenteral glucocorticoids were converted to equivalent doses of prednisone.

MRI and ¹H-MRSI data acquisition

All included subjects had MRI/¹H-MRSI performed. High-resolution images were obtained in a 3T Philips Achieva (Best, Netherlands) with acquisitions in the coronal, sagittal and axial planes and ¹H-MRSI acquisitions. The ¹H-MRSI data were converted to SDAT/SPAR format for compatibility with LCModel Version 6.3 program [24]. Structural T1 and T2 images were reviewed to determine whether abnormalities (e.g. white matter lesions) were present in the region of interest (ROI) where ¹H-MRSI was done. ¹H-MRSI was acquired using point resolved spectroscopy sequence (repetition time = 2000 ms, echo time = 144 ms, number of excitations = 90). The ROI analysed by the MRI scanner software was the superior–posterior region at the level of the corpus callosum (white matter) and volume of interest (100.52 × 84.74 × 16 mm; grid of 16 × 13 spectra: 208 spectra) (Fig. 1). Prior to the acquisition, a localized shimming at the ROI was performed to ensure adequate field homogeneity followed by water suppression adjustment. We evaluated all the bounded regions (208 voxels) in all subjects and the estimated concentrations were obtained. Cr was used as the common denominator for all the compounds included, since this compound does not change significantly in most pathological conditions, and is considered an internal reference [25]. The absolute quantification of Cr was not different between patients and controls (P = 0.723).

Fig. 1.

Placement of grid spectra (proton magnetic resonance spectroscopic imaging) over the superior-posterior white matter region at the level of corpus callosum level

Laboratory evaluation and cytokine assays

ANA were measured by indirect immunofluorescence using HEp-2 cells as the substrate and were regarded as positive if higher than 1:40. Double-stranded DNA (dsDNA) antibodies were determined by indirect immunofluorescence using Crithidia as the substrate and were considered positive if higher than 1:20. Precipitating antibodies to ENA, including Ro (SSA), La (SSB) and Sm, were detected using a standardized ELISA method and were considered positive if higher than 1:80. IgG- and IgM-isotype anticardiolipin antibodies (aCL) were measured using an ELISA. Lupus anticoagulant (LA) activity was detected by coagulation assays in platelet-free plasma obtained by double centrifugation.

Blood samples were collected from all participants at the time of MRI, allowed to clot for 30 min at room temperature, and centrifuged at 3000 rpm for 15 min. Sera were separated as soon as possible from the clot of red cells after centrifugation to avoid TNF-α production by blood cells, which could falsely increase its values. Sera were kept in aliquots at −80°C until the time of assay. None of the samples were taken during an episode of acute or chronic infection. Assays were performed in duplicate. Commercially available kits from R&D Systems (Minneapolis, MN, USA) were used for the measurement of serum IFN-γ, TNF-α, IL-4, -5, -6, -10, -12 and -17 by ELISA. The minimum detectable levels for cytokines were <2.0 pg/ml for IL-12, 7 pg/ml for TNF-α, 3.9 pg/ml for IL-4, 0.29 pg/ml for IL-5, 0.039 pg/ml for IL-6. The assay sensitivity was <2 pg/ml for IFN-γ, <1 pg/ml for IL-10 and 15 pg/ml for IL-17. Anti-ribosomal P proteins (anti-P antibodies) and S100β were also evaluated by ELISA, following the manufacturer’s instructions. All measurements were made on a single occasion to minimize the intra-assay variability.

Statistical analysis

Statistical analyses were performed using SYSTAT software (version 12; Systat Software, Inc., San Jose, CA, USA) and SPSS^® 21.0 (IBM Corp., Armonk, NY, USA). Fisher's exact test was used to compare categorical variables. For normally distributed variables, we used Student’s t-test. For non-parametric variables, we used the Mann–Whitney U-test. We used the Mann–Whitney U-test for comparison of ¹H-MRS parameters (i.e. NAA/Cr) and independent groups, i.e. groups (SLE patients/controls), disease activity (SLEDAI ≥ 3/SLEDAI < 3), cumulative damage (SDI ≥ 1/SDI = 0), sex (male/female) and medication (taking any medication/without medication). The correlations between continuous variables (i.e. cytokine values) and ¹H-MRS parameters were explored by Spearman’s rank correlation. Additionally, we performed univariate linear regression, using ¹H-MRS parameters as dependent variable (i.e. NAA/Cr). The Bonferroni method was applied to univariate analysis to reduce type I errors. Multivariate analysis was performed using a linear regression model (with backward elimination model), using ¹H-MRS parameters (i.e. NAA/Cr) as dependent variable. All variables that were significant in the univariate analyses were included, tested against each metabolite separately. For all analyses, P < 0.05 was considered statistically significant.

Results

Demographic data

We included 86 consecutive cSLE patients. Seventy-six (88.37%) were female, with a median age of 17 (range 5–28) years. The median disease duration was 5 (range 1–16) years. The control group comprised 71 controls (64 women) with a median age of 18 (range 5–28) years. According to the self-declared ethnicity, 68 (79.1%) patients were white, 2 (2.3%) black and 16 (18.6%) of mixed race. The patients and controls were statistically comparable in terms of age, sex and self-declared race distribution.

Clinical and immunological SLE features are shown in Table 1. Presentation at day of MRI exam and cumulative neuropsychiatric manifestations are shown in Table 2.

Table 1.

Demographics of 86 childhood-onset SLE patients

Variable	Distribution
Disease duration, median (range), years	5 (1–16)
Adjusted mean SLEDAI over time, median (range)	2.6 (0–12.5)
SDI score
Range	0–2
≥1 point, n (%)	23 (26.74)
0 point, n (%)	63 (73.26)
SLEDAI score at the day of MRI, median (range) ≥3 points, n (%)	8 (4–24)   29 (33.7)
Prednisone on exam day     In use, n (%)     Not in use, n (%)     Dose on exam day, median (range), mg	72 (83.73)   14 (16.27)   37.5 (0–80)
Cumulative prednisone dose, median (range), mg/kg	428 (0–1220)
ANAs, n (%)	86 (100)
Anti-DNA antibodies (dsDNA), n (%)	54 (68.79)
Lupus anticoagulante, n (%)	35 (40.69)
Anticardiolipin antibodies, n (%)	25 (29.06)
Anti- Sm antibodies, n (%)	24 (24.41)
Anti- Ro (SS-A), n (%)	14 (16.27)
Anti-La (SS-B), n (%)	6 (6.97)
Anti-P ribosomal antibody, n (%)	11 (12.79)
Antiphospholipid syndrome, n (%)	5 (8.81)

SDI: SLICC/ACR Damage Index.

Table 2.

Neuropsychiatric manifestations in 86 childhood-onset systemic lupus erythematosus patients

Neuropsychiatric manifestation	Cumulative, n (%)	At MRI date, n (%)
Headache	54 (62.79)	7 (8.14)
Movement disorder	3 (3.48)	—
Seizure disorders	15 (17.44)	1 (1.16)
Cerebrovascular disease	9 (10.46)	—
Acute confusional state	2 (2.32)	—
Aseptic meningitis	1 (1.16)	—
Cranial neuropathy	1 (1.16)	—
Psychosis	3 (3.48)	—
Polyneuropathy	1 (1.16)	—
Demyelinating syndrome	0 (0)	—
Symptoms of anxiety	Not tested	30 (34.88)
Symptoms of depression	Not tested	8 (9.3)
Cognitive dysfunction	Not tested	19 (22.09)

Cytokines

Anti-P antibodies were observed in 11 (12.79%) cSLE patients and in none of the controls (Table 1). We observed significantly higher levels of IL-4, IL-6, IL-10 and IL-12 in sera of cSLE patients when compared with controls (Table 3).

Table 3.

Levels of cytokines and S100B observed in childhood-onset SLE patients and controls

	Level, median (range), pg/ml
Cytokine	cSLE	Control	P-value
IL-4	0.20 (0.02–4.14)	0.11 (0.01–1.20)	0.001*
IL-5	3.74 (0.61–29.52)	5.61 (0.06–10.30)	0.101
IL-6	1.27 (0.34–10.52)	0.91 (0.40–8.83)	0.001*
IL-10	2.15 (0.50–26.93)	0.98 (0.18–9.54)	0.004*
IL-12	3.63 (0.05–6.30)	1.35 (0.14–8.09)	0.034*
IL-17	36.86 (17.30–105.92)	32.61 (22.14–41.07)	0.442
TNF-α	1.82 (0.38–23.04)	1.91 (0.48–8.97)	0.082
INF-γ	40.63 (4.09–64.76)	24.63 (0.49–75.35)	0.331
S100 β	27.65 (2.40–24.23)	19.84 (0.63–66.09)	0.102

^*cSLE: childhood-onset SLE. Significant difference P <0.05.

¹H-MRSI univariate analyses

No structural abnormality was observed in the ROI by visual analysis. When comparing cSLE patients with controls, we observed a significant decrease in NAA/Cr (P = 0.035). There was no significant difference between patients and controls for choline compounds (GPC/Cr: P = 0.249; PCh/Cr: P = 0.689; mI/Cr: P = 0.389; Glu/Cr: P = 0.316; and Gln/Cr: P = 0.482) (Fig. 2).

Fig. 2.

Box plot of evaluated metabolites: cSLE patients and controls. Data are presented as box plots, where the boxes represent the 25th–75th percentiles, the lines within the boxes represent the 50th percentile, and the lines outside the boxes represent the minimum and maximum values. Circles represent outliers. cSLE: childhood-onset SLE.

There was an association between reduction of the NAA/Cr ratio and the presence of depressive symptoms (P = 0.021), chorea (P = 0.039), psychosis (P = 0.015), presence of cognitive impairment (P = 0.003) and current use of prednisone (P = 0.001). The NAA/Cr ratio correlated with SLEDAI scores (r = −0.320, P = 0.03), IL-10 levels (r = 0.307, P = 0.045), IL-12 levels (r = 0.370, P = 0.012) and IFN-γ levels (r = 0.426, P = 0.046). There was no association of Lac/Cr with any of the included variables.

Increased CHO/Cr ratio was associated with the presence of cerebrovascular disease (P = 0.046) and presence of cognitive impairment (P = 0.001). The CHO/Cr ratio correlated with cumulative glucocorticoids dose (r = 0.391, P = 0.005). The Glu/Cr ratio was associated with hypocomplementaemia (P = 0.023) and correlated with IL-10 levels (r = 0.348, P = 0.044), IFN-γ levels (r = 0.834, P = 0.0001) and total glucocorticoids dosage (r = 0.367, P = 0.011). The Gln/Cr ratio was associated with chorea (P = 0.018) and correlated with the adjusted mean SLEDAI score over time (r = 0.388, P = 0.014) and IL-12 levels (r = 0.617, P = 0.001). The mI/Cr ratio correlated with cumulative glucocorticoids dosage (r = 0.361, P = 0.014).

Multivariate analysis

In multivariate analysis, the NAA/Cr ratio was associated with IFN-γ levels [odds ratio (OR) = 4.1; 95% CI: 2.01, 7.9] and depressive symptoms (OR = 1.9; 95% CI: 1.1, 3.2). Increased CHO/Cr was associated with the presence of cognitive impairment (OR = 3.4; 95% CI: 2.034, 5.078; P < 0.001). The Glu/Cr ratio and Gln/Cr ratio were not significantly associated with any of the variables in the multivariate analysis (P = 0.7).

Discussion

This is the first study to assess levels of metabolic changes in white matter in cSLE through ¹H-MRSI. We observed that the NAA/Cr ratio (P = 0.035) and Lac/Cr ratio (P = 0.019) were significantly decreased in cSLE patients when compared with controls. In multivariate analysis, IFN-γ levels and depressive symptoms were associated with the NAA/Cr ratio. Increased CHO/Cr was associated with the presence of cognitive impairment.

Reduced NAA/Cr ratio has been observed in studies analysing adult-onset SLE (aSLE) [7, 26–30]. NAA is considered a marker for assessing neuronal viability and function [31]. Previous studies analysing ¹H-MRS in aSLE have found that a reduction of NAA/Cr was associated with disease activity [7], the presence of cerebral atrophy [29] and neuropsychiatric manifestations [28, 30]. In our univariate analysis, we observed similar associations. However, our study is the first to analyse, additionally, the association of ¹H-MRSI metabolites and cytokine levels.

Reduction of the NAA/Cr ratio reflects axonal dysfunction that may precede neuronal loss [29, 30]. In our patients, no structural abnormality was observed on visual analysis, suggesting that the reduction of NAA/Cr preceded neuronal loss. However, this dysfunction can, in part, be transient as shown in longitudinal studies in which NAA/Cr increased when disease activity improved [7]. Longitudinal studies are necessary to evaluate this association and determine whether these patients have greater cerebral volume loss than cSLE patients with normal NAA/Cr ratio. Interestingly, NAA/Cr was associated with IFN-γ levels. Studies in animal models of lupus have shown that type I IFN stimulates microglial activation and synapse loss and that blocking type I IFN prevents these abnormalities [32]. Type I IFN and type II IFN (IFN-γ) share some similarities in downstream signalling, and these two IFNs are correlated in many SLE patients [33]. Thus, type I IFN could also be playing a role in this phenomenon. We did not measure type I IFN in this study, but that would be an important future direction.

We did not observe a difference in choline-based compounds when compared with controls. However, higher CHO/Cr ratio in cSLE patients was associated with cognitive impairment. Increased CHO/Cr ratio has been shown to predict structural white matter abnormalities in aSLE [8]. CHO, especially PCH and GPC, are precursors for biosynthesis and degradation of the cell membrane [34, 35]. Therefore, we hypothesize that these patients may progress to a more severe axonal loss and cerebral atrophy than cSLE with a lower CHO/Cr ratio.

We observed that glucocorticoids were associated with increased CHO/Cr, Glu/Cr and mI/Cr ratios in univariate analysis. Previous studies have shown that exposure to glucocorticoids can cause neuronal damage, dendritic atrophy, impairment in neurogenesis and synaptic plasticity [36]. Intravenous glucocorticoids, on the other hand, have been associated with increased white matter [23]. In our cohort, all patients received oral glucocorticoids during the course of the disease, so we cannot assess this possibility. The cumulative and daily dose of glucocorticoids are high in our cohort. Patients are often referred to us already on high dose of glucocorticoids used for several weeks. In some regions, glucocorticoid sparing drugs, such as azathioprine and hydroxychloroquine, take 1–2 months to be regularized. Micophenolate distribution is limited to a few distribution points in the region, due to difficult access. We are actively working to modify this scenario and reduce glucocorticoid dose.

We found no significant difference between patients and controls in the analysis of mI/Cr, Glu/Cr and Gln/Cr ratios. Few studies have evaluated mI/Cr levels in SLE, and in the majority levels are similar between patients and controls [27, 37, 38]. The mI metabolite is considered a glial marker due to its higher concentrations in astrocytes compared with other cell types [34, 35]. When its elevation is associated with a reduction of the NAA peak, it becomes an important biomarker of dementia [39, 40]. Despite the high frequency of cognitive impairment in SLE, the frequency of dementia is relatively low, which could explain the mI levels within the normal range [41].

Glu/Cr and Gln/Cr ratios are often elevated in pain conditions such as fibromyalgia and positively correlated with the severity of the pain [42]. None of our patients had associated fibromyalgia. Previous studies have found a decrease in the Glu/Cr ratio in frontal white matter and a correlation with memory functions in the hippocampus of aSLE patients [43]. Recent studies have demonstrated changes in the glutamate correlating to memory dysfunction in SLE patients without neuropsychiatric symptoms [43]. Glutamine and glutamate may be additional biomarkers for cerebral involvement in SLE patients as these early metabolic changes occur in the brain of SLE patients before neurological and imaging manifestations become apparent [43].

We chose to select the white matter superiorly of the corpus callosum to study metabolic changes, where no structural abnormalities were observed. This area is predominantly white matter, with little interference from the cerebrospinal fluid or bones.

The metabolite concentration levels in the brain are heavily dependent on the MR systems, ¹H-MRSI technique and acquisition parameters [44]. During the first 3 years of brain development, metabolite concentrations vary primarily due to the myelination process [45]. The levels of the metabolites NAA and Cr are at their lowest, but increase within the first 2 years of life [46]. The metabolic changes for children between the age of 7 and 16 are quite similar to those of an adult, and the pathological variations are identical to those observed in an adult with the same condition [47]. As a result of changes in metabolites according to the age of the individuals, seen in both total brain volume and specific tissue-type volume loss associated with ageing, tissue segmentation should be considered essential for ¹H-MRS studies in populations that span a wide age range, which was not so in our case, as it included only young and young adults individuals [48].

Cr is relatively constant throughout the brain and tends to be resistant to pathological conditions. It is often considered an internal reference and the results are presented as ratios [49–51]. It is also recommended to use Cr to correct artefacts and signal intensity variation due to the lack of homogeneity of the magnetic field [52].

The study has limitations. We performed a cross-sectional study, with a small number of some individual neuropsychiatric manifestations and analysed only a single area of the brain. Further studies with a larger sample size and longitudinal follow-up are needed before a definitive conclusion can be drawn.

Functional studies have shown that disruption of the blood–brain barrier appears before the presence of brain atrophy and is associated with the presence of cognitive impairment and antineuronal antibodies [5, 6]. In addition, axonal/neuronal damage has been shown to be a pathological finding and may precede neuropsychiatric manifestations in SLE. This may explain the neuropsychiatric symptoms and cognitive dysfunction, even in the absence of systemic disease activity [5, 6].

Conclusion

We suggest that neuronal metabolites evaluated by MRS may be useful biomarkers for brain disease in cSLE, as metabolic abnormalities may occur before clinical manifestations and structural changes. The identification and quantification of biomarkers associated with neuronal injury in cSLE may help not only the diagnosis of injury but also the prognosis of CNS involvement.

Data availability statement

Data available on request.