Serum levels of S100B are decreased in chronic starvation and normalize with weight gain
S100B protein is mainly synthesized in glial cells and modulates the balance between cell proliferation and differentiation in neurons and glial cells. However, S100B is not CNS-specific since its production was detected in numerous non-cerebral tissues e.g. adipocytes. In this study we investigated the influence of chronic fasting and subsequent weight gain on serum levels of S100B in patients with anorexia nervosa. We found that nutritional status was an important factor influencing serum levels of S100B.
KeywordsEating disorder Fasting Starvation S100 Anorexia nervosa
In the brain, S100B protein is synthesized in glial cells and modulates the balance between cell proliferation and differentiation in neurons and glial cells. High CSF and serum concentrations of S100B are correlated with multiple forms of CNS damage as well as psychiatric diseases, such as schizophrenia, major depression and mania (Rothermundt et al. 2003; Sen and Belli 2007). Initially, it was assumed, that S100B was CNS-specific, but production of S100B was detected in numerous non-cerebral tissues e.g. adipocytes, smooth muscle cells and lymphocytes (Steiner et al. 2007). The specificity of S100B levels as a reflection of CNS injury is compromised by the findings that extra-cranial injuries can lead to elevations in the absence of brain injury (Anderson et al. 2001).
Existing data indicate that in rat fasting leads to an immediate increase in S100B followed by a decrease in S100B serum levels if negative energy balance is maintained. Netto et al. (2006) demonstrated an increase in serum S100B that was not accompanied by changes in CSF S100B after short-term fasting (48 h) in rats. Consistently, Leite et al. (2004) observed an increment in the extracellular content of S100B in astrocyte cultures 1 h after addition of the ketone body ß-hydroxy-butyrate, but a decrease 24 h later. Similarly, low CSF levels of S100B were reported in rats as a neurochemical reaction to 6 weeks of a ketogenic diet that mimics the physiological effects of fasting (Ziegler et al. 2004).
Patients with acute anorexia nervosa (AN) are characterized by semistarvation, chronic fasting, a low BMI and low body fat content as indicated by hypoleptinemia. In general, these changes resolve with normalization of the nutritional status and weight-rehabilitation (Holtkamp et al. 2003). Thus, AN may represent a condition to investigate the influence of chronic fasting and subsequent weight gain on S100B levels in humans. We hypothesized that serum concentrations of S100B are decreased in patients with acute AN and normalize with weight-rehabilitation.
The study group consisted of a consecutive sample of 28 medication-naïve girls (age 15.3 ± 1.5 years) with a DSM-IV diagnosis of AN that were admitted in the department of child and adolescent psychiatry at the university clinic of Aachen, Germany. Diagnosis was made by senior child and adolescent psychiatrists (K.H., B.H-D.) and by a structured interview (SIAB, Fichter et al. 1998). Patient BMI and serum levels of leptin and S100B were investigated before the start of re-feeding at the second day after admission to inpatient treatment (T1) and after 12 weeks of weight gain (T2) and compared with those of 19 female (age 15.0 ± 1.2 years), normal-weight, healthy controls. Blood sampling was performed at 8 am after an overnight fast. Serum leptin levels were measured using a sensitive RIA (Mediagnost, Tübingen; intra-assay variance: <5%). S100B concentrations were determined by applying the LIAISON Sangtec 100 assay (AB Sangtec Medical, Bromma, Sweden, intra-assay variance: 2.8–6.4% at 0.11–18.4 μg/L), a quantitative automated luminometric immunoassay. Body composition was assessed with body-impedance analysis in control subjects and in 19 of the 28 patients (BIA, Nutri4, Data Input GmbH, Darmstadt, Germany). Routine assessment of dietary intake revealed a caloric intake between 400–1,000 kcal/d the week before admission. Within the first 4 weeks of treatment, patients obtained a protein rich diet (20%). The study was approved by the local ethics committee and participants as well as their parents gave written informed consent.
For normally distributed variables, Pearson correlation was used, for abnormally distributed variables we used Spearman ‘rho. The partial correlations procedure was used to control for the effect of age in the relationship between S100B levels, leptin levels and BMI, respectively. Linear regression analysis was performed to examine the relationship between the change in S100B levels between T1 and T2 (ΔS100B) and Δleptin (leptin change between T1 and T2) and ΔBMI (BMI change between T1 and T2). Significance was defined as P ≤ 0.05. Data are presented as mean ± SD.
We investigated the effect of semistarvation and chronic fasting on serum S100B in humans. We found serum levels of S100B to be decreased in the acute stage of AN and to be normalized after weight gain. Changes in S100B were significantly explained by changes in leptin levels in a regression model. This points to an influence of nutrition related factors on the production and/or secretion of S100B in humans. The reliability of our data is enhanced by the fact that we could replicate the results of the two studies investigating the relationship between age and serum S100B (Portela et al. 2002; Gazzolo et al. 2003).
The conclusions drawn from this explorative study are of a preliminary nature because we were not able to differentiate whether the decrease and normalization in serum S100B levels was from a cerebral or peripherical (or both) source. So far, there is not sufficient basic research to predict the mechanism underlying altered S100B production or secretion during AN. However, since BMI and body fat mass were not significantly correlated with S100B levels either in patients or controls, the size of the fat mass in itself does not seem to be a determinant of the serum S100B concentration. During re-feeding of AN subjects, correlations between leptin levels and BMI or body fat mass typically decline, indicating variability of leptin secretion from adipocytes in response to weight gain (Holtkamp et al. 2003). Since the change in leptin levels and the change in S100B levels were significantly correlated, a similar mechanism may be presumed for S100B secretion from adipocytes.
However, future studies are necessary to investigate possible mechanisms underlying the effect of short-term and chronic fasting on serum S100B levels. The following facts should be considered. (1) The relationship between S100B levels and of milder forms of fasting, which would be more common in potential patient groups of S100B sampling, should be established. (2) During starvation ketogenesis is activated. Ketone bodies like β-hydroxybutyrate are able to decrease the extracellular content of S100B in astrocyte cultures (Leite et al. 2004), and ketogenic diet fed rats have low levels of S100B in cerebrospinal fluid (Ziegler et al. 2004). Levels of ketone bodies are elevated in acute AN, and quickly normalize during re-feeding (Pirke et al. 1985). Thus, high levels of ketone bodies may be responsible for the reduced S100B concentration in our patients at the time of admission. However, the effect of ketone bodies on the release of S100B from adipocytes has not been investigated so far. (3) Serotonin acts on cell growth and differentiation either by acting through specific receptor systems or by promoting the release of glial S100B (Azmitia 2002). Malnutrition-induced reduction in serotonin activity during acute AN ceases with weight rehabilitation (Kaye at al. 1988). Thus, the reported decrease in S100B levels upon admission and the normalization during re-feeding in our AN subjects may be caused by changes in serotonin activity. (4) Physical training leads to an elevation in S100B serum levels (Dietrich et al. 2003; Schulpis et al. 2007). Patients with AN typically pass through recurrent changes in physical activity during the course of their illness. Thus, changes in levels of physical activity may account for the alterations in S100B levels found in our patients.
In summary, we demonstrated that long-term fasting in girls with anorexia nervosa leads to decreased serum S100B levels. Caution is needed when interpreting the results of serum S100B in conditions where undernutrition may be present. In children and adolescents, age-matched reference values are needed to correctly interpret S100B data.