Tissue Sodium Content in Patients with Systemic Lupus Erythematosus: Association with Disease Activity and Markers of Inflammation

Abstract

Objectives:

Sodium is stored in skin and muscle and plays an important role in immune regulation. In animal models, increased tissue sodium is associated with activation of the immune system, and high salt intake exacerbates autoimmune disease and worsens hypertension; however, there is no information about tissue sodium and human autoimmune disease. We hypothesized that muscle and skin sodium content is (1) higher in patients with systemic lupus erythematosus (SLE) than control subjects, and (2) associated with blood pressure, disease activity, and inflammation markers (IL-6, IL-10 and IL-17A) in SLE.

Methods:

Lower leg skin and muscle sodium content was measured in in 23 patients with SLE and 28 control subjects using ²³Na⁺ magnetic resonance imaging. Demographic and clinical information was collected from interviews and chart review, and blood pressure was measured. Disease activity was assessed by SLE Disease Activity Index (SLEDAI). Plasma inflammation markers were measured by multiplex immunoassay.

Results:

Muscle sodium content was higher in patients with SLE [18.8 (16.7–18.3) mmol/L] than control subjects [15.8 (14.7–18.3) mmol/L], p<0.001; skin sodium content was also higher in SLE patients than controls, but not statistically significantly. Among patients with SLE, muscle sodium was associated with SLEDAI and higher concentrations of IL-10 after adjustment for age, race, and sex. Skin sodium was significantly associated with systolic blood pressure, but this was attenuated after covariate adjustment.

Conclusion:

Patients with SLE had higher muscle sodium content than control subjects. In patients with SLE, higher muscle sodium content was associated with higher disease activity and IL-10 concentrations.

Multiple recent insights implicate salt in the regulation of immune function. First, sodium (Na⁺) and water balance are not perfectly coupled; therefore, the body stores substantial amounts of Na⁺ in tissues such as the skin and muscle in the absence of changes in serum Na⁺ concentrations and without fluid accumulation.[1, 2] Second, immune cells are involved in the regulation, storage and clearance of Na⁺ in tissue.[3, 4] Third, increased skin Na⁺ content lead to activation of the immune system [5, 6] and a high salt diet exacerbates autoimmune disease in animals and increases pro-inflammatory cytokines in humans.[5] Fourth, skin and muscle Na⁺ content can be measured reproducibly in humans and skin Na⁺ content is associated with hypertension.[2]

Systemic lupus erythematosus (SLE) is a prototypical autoimmune disease that is also characterized by an increased prevalence of hypertension.[7, 8] In a murine model of lupus, a high salt diet increased disease markedly.[9] Moreover, in patients with SLE, restricted sodium intake was associated with reduced Th17 cells and increased regulatory T cells, suggesting that low sodium diet may reduce the inflammatory response in SLE.[10]

However, despite the established relationship between salt, inflammation, and hypertension in other settings, there is no information about tissue sodium content in SLE. In this pilot study, we hypothesized that muscle and skin sodium content is higher in patients with SLE than control subjects, and that higher tissue sodium content is associated with hypertension, lupus disease activity, and markers of inflammation.

METHODS

Subjects

This cross-sectional study included 23 patients with SLE and 28 control subjects 18 years of age or older. Patients with SLE met American College of Rheumatology revised classification criteria for SLE.[11] Control subjects did not have SLE or other autoimmune diseases. Subjects who could not provide informed consent or undergo an MRI scan were excluded. Study participants with SLE were recruited from the Vanderbilt University Medical Center Rheumatology clinic. Control subjects were recruited from the Vanderbilt community and surrounding areas. The Vanderbilt University Medical Center Institutional Review Board approved the study (IRB# 110365) and all participants provided written informed consent. Patients did not contribute to study design, data interpretation, or manuscript development.

²³Na MRI

The MRI methods and quantification of tissue sodium quantification using MRI images have been described previously. [12] Briefly, tissue Na⁺ content in the lower leg was measured with a ²³Na⁺ knee-coil at 3.0 Tesla with a magnetic resonance imaging scanner (Philips). Phantoms containing aqueous solutions with 10, 20, 30, and 40 mmol/L NaCl were used as calibration standards. Results are reported in mmol/L.

Clinical characteristics:

Patient assessment included a detailed review of medical records, a standardized interview, physical examination, and laboratory testing. Demographic and clinical information, including details about medications and co-morbidities, were obtained directly from subjects and their medical records. In particular, previous and current use of non-steroidal anti-inflammatory medications, corticosteroids, and other immunosuppressive drugs were recorded.[13]

Disease activity was assessed using the Systemic Lupus Erythematosus Disease Activity Index (SLEDAI). For the SLEDAI score, we obtained the clinical criteria at the time of the study entry and the laboratory criteria within 14 days of the study, including C-reactive protein (CRP) and complement C3 and C4 proteins. The Systemic Lupus Collaborating Clinics (SLICC) score was used to assess SLE disease damage.[14] The modified health assessment questionnaire (mHAQ) was used to assess functional status.[15]

Renal function was assessed by serum creatinine and estimated the glomerular filtration rate using the Modification of Diet in Renal Disease formula.[16]

Nursing staff obtained blood pressure as part of a routine clinical visit for SLE patients and study staff measured blood pressure during the enrollment visit for controls. Hypertension was defined by use of antihypertensive medications and/or blood pressure ≥140/90 mmHg.

Inflammation marker panel:

Plasma samples were used for an inflammation multiplex immunoassay consisting of 92 inflammation-related markers (Inflammation panel; Olink Bioscience, Uppsala, Sweden). The association between inflammatory markers and tissue sodium was assessed, with a prespecified interest in IL-6, IL-10 and IL-17A, since these have been linked with lupus activity [17–19] and tissue sodium or sodium intake [6, 20] in previous studies. Panel inflammation markers were excluded from analysis if >25% of subjects had concentrations below the limit of detection.

Sample Size Calculation

Sample size estimation was based on the outcome of skin Na⁺ concentration using PS-Power for an unpaired t-test with a 2-sided test at a significance level of 5%.[21] Based on our previous findings, the standard deviation (SD) of skin Na⁺ concentration was approximately 6.0 mmol/L and this difference appeared clinically important.[2] A sample size of at least n=22 in each group would provide 90% power to detect a difference of 6.0 mmol/L.

Statistical Analyses

Continuous and categorical variables were expressed as median and interquartile range (IQR) or count and percentages respectively. Between-group comparisons were assessed using Wilcoxon or Chi-square tests for continuous or categorical variables, as appropriate. Tests of hypotheses concerning within-group comparisons were performed using Wilcoxon rank test for continuous covariates, or Chi-square test for categorical covariates. Spearman correlations were used to assess the relationship between continuous variables and multivariate linear models were used to assess whether associations were independent of preselected covariates. Statistical significance was set to p-values < 0.05 and 95% confidence intervals for the prespecified comparisons of primary interest: skin and muscle Na⁺ content in patients with SLE compared to controls, and the association between skin and muscle Na⁺ content and IL-6, IL-10 and IL-17A concentrations in patients with SLE. Exploratory analyses were performed assessing the association between tissue sodium and clinical features, as well as the full panel of inflammation markers to detect any novel relationships. Enhanced heat maps with hierarchical clustering represented by dendrograms were constructed with tissue sodium, clinical and inflammation marker data using Spearman rho correlation coefficients. All analyses were performed using Stata Statistical Software: Release 14. College Station, TX: StataCorp LP or RStudio version 3.5.1.

RESULTS

Patient characteristics:

Patients with SLE (n=23) and controls (n=28) were of similar age, race, sex, and BMI (Table 1). Ten of the 23 patients (43%) with SLE were diagnosed with lupus nephritis. Among these 10 patients with lupus nephritis, two had active urine sediment or proteinuria at the time of the study. Renal function, measured by serum creatinine concentrations and estimated glomerular filtration rate, was within the normal range in patients with SLE and did not differ significantly from that of the control group. The prevalence of HTN was higher in the SLE group (34.8%) compared to controls (7.1%). In patients with SLE, disease activity was stable and organ damage low. The median (interquartile range) SLEDAI score in patients with SLE was 4 (2–6) and the median SLICC score was 0.5 (0–1).

TABLE 1:

Baseline characteristics

	Controls (n=28)	SLE (n=23)	p-value
Age (Years)	35 (28 – 52)	37 (31 – 53)	0.78
Sex, Female, n (%)	24 (85.7)	20 (87.0)	0.61
Race, White, n (%)	21 (75.0)	16 (69.6)	0.89
SLEDAI score	-	4 (2 – 6)	-
SLICC score	-	0.5 (0 – 1)	-
BMI (kg/m2)	24.8 (22.6 – 27.5)	26.5 (23.2 – 33.8)	0.19
Creatinine (mg/dL)	0.8 (0.7 – 0.9)	0.8 (0.6 – 0.9)	0.99
eGFR (mL/min)	90 (85 – 105)	96 (81 – 111)	0.51
Na (serum, mEq/L)	141 (140 – 141)	139 (138 – 141)	0.08
K (serum, mEq/L)	4.1 (3.7 – 4.5)	4.2 (3.7 – 4.4)	0.95
Hypertension, n (%)	2 (7.1)	8 (34.8)	0.02
Diuretics use, n (%)	1 (3.6)	1 (8.7)	0.89
Angiotensin Converting Enzyme Inhibitor use, n (%)	1 (3.6)	5 (21.7)	0.05
Angiotensin Receptor Blocker use, n (%)	0 (0.0)	3 (13.0)	0.10
Diabetes mellitus, n (%)	0 (0.0)	1 (4.4)	0.45
Smoking status (Ever), n (%)	20 (71.4)	13 (61.9)	0.52
NSAIDs use, n (%)	10 (35.7)	12 (52.2)	0.19
Corticosteroid use, n (%)	1 (3.7)	13 (56.5)	<0.001
C-reactive Protein (mg/dL)	-	2.9 (1.7 – 12.6)	-

Data are presented in medians, interquartile ranges, counts and percentages for continuous and categorical characteristics respectively. Fisher exact test and Wilcoxon rank-sum test were used to compare between study groups for continuous and categorical characteristics, respectively.

Tissue sodium content in SLE and control subjects

Patients with SLE had higher muscle sodium content [18.8 (16.7–18.3) mmol/L] than control subjects [15.8 (14.7–18.3) mmol/L], p<0.001 (Figure 1). Skin sodium content was not significantly different in patients with SLE [15.6 (13.4–17.7) mmol/L] and control subjects [13.7 (11.9–16.8) mmol/L], p=0.16 (Supplementary Figure 3). Some medications, particularly corticosteroids, diuretics and those medications altering the renin-angiotensin-aldosterone system can alter sodium retention. However, among patients with SLE and control subjects no difference in skin or muscle sodium content was observed comparing those taking versus not taking corticosteroids and comparing those taking versus not taking a diuretic, angiotensin-converting enzyme inhibitor, or angiotensin-receptor blocker (all p>0.05).

FIGURE 1:

Tissue sodium content in patients with SLE and control subjects

Association of tissue sodium with blood pressure in patients with SLE

Among patients with SLE, based on hierarchical clustering analysis, skin sodium content segregated best with blood pressure (Figure 2). There was a significant association between skin sodium content and systolic blood pressure (ρ=0.4, p=0.03) (Figure 3); however, this association did not remain significant after adjustment for age, race, and sex (p=0.18). Muscle sodium content was not significantly associated with blood pressure (Table 2).

FIGURE 2.

Heatmap with hiierarchical clustering of tissue sodium and disease related measurements among patients with SLE

FIGURE 3:

Correlation between skin sodium and systolic blood pressure in SLE

TABLE 2:

Correlations between tissue sodium, select disease related measurements and blood pressure in SLE

	Muscle sodium		Skin sodium
	rho	p-value	rho	p-value
Age (Years), correlation	0.1	0.56	0.3	0.13
Office SBP, correlation	0.01	0.96	0.4	0.03
Office DBP, correlation	−0.002	0.99	0.3	0.11
SLEDAI score, correlation	0.4	0.07	−0.3	0.12
SLICC score, correlation	−0.1	0.56	−0.1	0.61
C3, correlation	−0.4	0.07	0.3	0.13
C4, correlation	0.0	0.98	0.7	0.001
eGFR (mL/min), correlation	−0.3	0.17	−0.4	0.08
CRP (mg/dL), correlation	−0.1	0.52	−0.1	0.54
	Muscle sodium		Skin sodium
	β-coefficient	p-value	β-coefficient	p-value
Age (Years), adjusted	0.1	0.41	0.05	0.31
Office SBP adjusted	−0.1	0.83	0.1	0.18
Office DBP adjusted	0.1	0.9	0.1	0.44
SLEDAI score adjusted	0.6	0.006	−0.2	0.42
SLICC score adjusted	−0.5	0.44	−0.5	0.31
eGFR (mL/min) adjusted	−0.1	0.02	−0.01	0.58
CRP (mg/dL), adjusted	−0.04	0.59	−0.03	0.59

Unadjusted analyses are presented with Spearman rho.

Adjusted analysis is presented with β-coefficients from linear regressions adjusted for age, race and sex, except “Age (Years)”, adjusted for race and sex only. Tissue sodium was the dependent variable.

Association of tissue sodium with clinical features in patients with SLE

Among patients with SLE, based on hierarchical clustering analysis, muscle sodium content segregated best with SLE-related factors (Figure 2). For example, among patients with SLE muscle sodium content was associated significantly with disease activity as measured by the SLEDAI score (p=0.006), and was inversely associated with glomerular rate filtration (p=0.02), after adjustment for age, race and sex (Table 2).

Association of tissue sodium and prespecified plasma inflammation markers in patients with SLE

Several plasma inflammatory markers differed among patients with SLE and control subjects (Supplementary Table 1). Among the prespecified cytokines, IL-6 (p<0.001) and IL-10 (p<0.001) were significantly increased, and IL-17A was not significantly different in patients with SLE compared to control subjects. In SLE, muscle sodium was associated with plasma IL-10 concentrations after adjustment for age, race and sex (Supplementary Table 3). Otherwise, neither skin nor muscle sodium was associated with the other prespecified plasma cytokine concentrations after adjustment for age, race and sex.

Exploratory analysis of tissue sodium and inflammation marker panel in SLE

An exploratory analysis was performed including all measured plasma inflammatory markers to determine other potential associations with skin and muscle sodium (Supplementary Table 2). Among patients with SLE based on hierarchical clustering analysis (Figure 4), a cluster of inflammatory markers including beta nerve growth factor (B-NGF), C-C motif chemokine ligand 23 (CCL23), interleukin 10 (IL-10), programmed death-ligand 1 (PD-L1), monocyte chemotactic protein 3 (MCP-3), vascular endothelial growth factor A (VEGFa), fibroblast growth factor 23 (FGF-23), C-X-C motif chemokine 10 (CXCL10), fibroblast growth factor 21 (FGF-21), colony stimulation factor-1 (CSF-1) and interleukin 8 (IL-8) clustered in positive association with muscle sodium but not with skin sodium. Among these, CCL23 (ρ=0.6, p=0.002), IL-10 (ρ =0.4, p=0.04), PD-L1 (ρ =0.4, p=0.04), VEGFa (ρ =0.5, p=0.01), FGF-23 (ρ =0.5, p=0.02), FGF21 (ρ =0.7, p=0.0001), CSF-1 (ρ =0.6, p=0.003), and IL-8 (ρ =0.5, p=0.01) were significantly positively associated with muscle sodium.

FIGURE 4.

Heatmap with hierarchical clustering of tissue sodium and plasma inflammatory markers among patients with SLE

Sensitivity analyses excluding men and women over 60 years of age

Both age and sex have been shown to influence the content of sodium in tissue[2, 22] and could obscure associations with tissue sodium. Therefore, in a sensitivity analysis we excluded men (3 and 4 in the SLE and control groups, respectively) and women over 60 (none in the SLE group and 3 in the control group). Within women under 60, the difference in skin sodium between study groups becomes significant (15.4 mmol/L vs 13.6 mmol/L in the SLE and control groups respectively, p-value=0.04. Supplementary Figure 1), and the correlation between skin sodium and systolic blood pressure in SLE became stronger (Supplementary Figure 2).

DISCUSSION

The main findings of this pilot study are that: 1) muscle sodium content was higher in patients with SLE than control subjects; and 2) there was an association between muscle sodium content with lupus disease activity, as quantified by the SLEDAI score, and IL-10 blood concentrations. Although not statistically significant, our data suggest that patients with SLE had a higher skin sodium content than controls, and that skin sodium was positively associated with systolic blood pressure in patients with SLE.

We propose three potential mechanisms explaining our findings of increased muscle sodium content in patients with lupus. First, and consistent with previous findings in other inflammatory conditions, that tissue sodium accumulates in patients with SLE due to the chronic inflammatory nature of the disease. Second, because aging is associated with increased sodium deposition; increased accumulation of tissue sodium in patients with SLE could be a consequence of premature aging. Third, because muscle sodium content increases in patients with some forms of myopathy or muscle injury,[23] it is possible that higher muscle sodium in patients with lupus indicates subclinical myopathy. Finally, we cannot exclude the possibility that the medications for the treatment of SLE and hypertension may have affected tissue sodium content in patients with SLE. Corticosteroids can increase sodium retention while diuretics and agents that block the renin-angiotensin-aldosterone system increase sodium losses; however, we found no significant difference in the tissue sodium based on use of these medications, though sample size was small.

Our findings that muscle sodium content was associated with higher IL-10 concentrations are also of interest. IL-10 is a B-cell stimulator and its concentrations are increased in patients with SLE; [24] furthermore, higher IL-10 concentrations are associated with higher lupus disease activity and with lupus nephritis.[25] Consistent with that notion, we found that higher lupus disease activity was associated with higher sodium muscle content independent of age, race and sex. Our data also suggest an inverse correlation between muscle sodium content and renal function; however, because almost all patients with SLE enrolled in this study had normal renal function and normal urine sediment at the time of the study, additional studies are necessary to determine the association between muscle sodium with lupus nephritis.

Our goal for the exploratory analysis examining the association of a large panel of inflammatory markers with tissue sodium, was to identity novel potential links between muscle sodium and disease activity in SLE. Most of the markers which clustered in association with muscle sodium, such as CCL23 [26], IL-10 [27], PD-L1 [28], VEGFa [29], FGF23 [30], CSF-1 [31], and IL-8 [29, 30, 32] are produced by macrophages and/ or dendritic cells and many of these also have important roles in monocyte and dendritic cell chemotaxis, differentiation and/or survival. This is consistent with prior studies showing that tissue macrophages and dendritic cells [33, 34] create an inflammatory response to and store tissue sodium.

Because salt alone can enhance IL-17A expression in TH17 cells [5] and activate dendritic cells [34] to promote T cell production of IL-17A and IFN-gamma, it is possible that salt-induced IL-17A and IFN-gamma could provide a direct link to SLE disease activity. However, we did not observe a significant positive correlation between IL-17A and tissue sodium in the SLE patients. We speculate that the relationship between muscle sodium and SLE disease activity may be more related to direct influences of activation of the macrophages and dendritic cells.

Body sodium content can be mobilized.[35] This is important because recent data suggest that salt intake could affects both body sodium content and lupus activity. In MRL/lpr mice, a high-salt diet increased the frequency of lupus characteristics -in particular nephritis- and was associated with increased mortality.[9] Alterations of sodium storage and electrolyte balance may have an even greater role than dietary intake. For example, hyperaldosteronism and local infection cause reversible tissue sodium accumulation with immune activation.[12, 36, 37] Thus, whether diet interventions aiming to decrease total body sodium content will affect the development of lupus or whether low-salt diet will decrease disease activity in patients with lupus is to be determined.

This study has some limitations. First, because this was a pilot study, the numbers are small, and power is limited; thus, we cannot exclude a true association between hypertension, IL-17, and tissue sodium in patients with SLE. Second, it is cross-sectional, so rather than causal inference, we present associations. Third, patients with lupus had mild disease and low organ damage scores; thus, it is possible that more robust differences could be observed in a study involving lupus patients with more severe disease.

In conclusion, this pilot study suggests that patients with lupus have higher tissue sodium content and that muscle sodium is associated with higher disease activity and serum IL-10 concentrations.

Key Messages:

What is already known about this subject?

Tissue sodium plays an important role in immune and blood pressure regulation.

What does this study add?

This is the first study showing that muscle sodium content is higher in patients with lupus than in controls and is associated with lupus disease activity.

How might this impact on clinical practice or future developments?

Future developments after this pilot study include testing whether interventions aiming to decrease tissue sodium content could affect the development of lupus, decrease lupus disease activity, and decrease cardiovascular risk in patients with lupus.