*3.2. Body Composition, Physical Activity, Energy Expenditure, and Psychometric Characteristics of the Study Population*

Data on body composition were available for 191 (57 AN, 71 OB, 63 NW), accelerometric data for 121 (27 AN, 47 OB, 47 NW), and psychometric questionnaires for 218 (GAD-7), 209 (PHQ-9), 214 (PSQ-20), and 195 (EDI-2) of the 219 women (Table 2).

Patients of all three groups showed significant differences from each other in terms of fat mass, fat free mass, body cell mass, and total body water, with the highest values in OB and the lowest in AN (*p* < 0.001; Table 2). Extracellular mass differed between AN and OB (*p* < 0.001), as well as NW and OB (*p* < 0.001), with higher values in OB (Table 2). Lower values were observed for phase angles in AN compared to OB (*p* < 0.001) and NW (*p* < 0.001; Table 2).


**Table 2.** Endocrine parameters, body composition, physical activity, energy expenditure, and patientreported outcomes of the study populations.

Data are expressed as mean ± standard deviation and range in parentheses. Differences between groups were assessed using a Kruskal–Wallis test for non-parametric data and ANOVA for parametric data. Negative fat mass values in bioelectrical impedance analysis are possible in severely underweight patients, due to the manufacturer's algorithms being calculated primarily for normal weight subjects. Significant differences (without correction for multiplicity) between AN and OB groups are displayed as \* (*p* < 0.05), or \*\*\* (*p* < 0.001); between AN and NW groups as # (*p* < 0.05), ## (*p* < 0.01), or ### (*p* < 0.001) and between NW and OB groups as + (*p* < 0.05), ++ (*p* < 0.01), or +++ (*p* < 0.001). Abbreviations: AN, anorexia nervosa; EAT, exercise-related activity thermogenesis (energy expenditure of more than three metabolic equivalents of a task); EDI-2, Eating Disorder Inventory-2; GAD-7, Generalized Anxiety Disorder-7; MET, metabolic equivalents of tasks; NEAT, non-exercise-related activity; NW, normal weight; OB, obesity; PHQ-9, Patient Health Questionnaire-9; PSQ-20, Perceived Stress Questionnaire-20; TEE, total energy expenditure.

With regard to physical activity, patients with OB performed less steps per day than NW *(p* < 0.05) and AN (*p* < 0.001; Table 2). All three groups differed from each other concerning MET per day and TEE, with the highest MET levels in AN followed by NW and OB (*p* < 0.001); and the highest TEE in OB, followed by NW and AN (*p* < 0.01; Table 2). In NW subjects, NEAT was lower (*p* < 0.001) and EAT higher (*p* < 0.001) than in OB. NEAT was also lower in NW than in AN (*p* < 0.001; Table 2). EAT levels were lower in OB in comparison to AN and NW (*p* < 0.001; Table 2).

As shown in Table 2, there were no differences between groups regarding anxiety (GAD-7; *p* > 0.05) and perceived stress (PSQ-20 including all subscales; *p* > 0.05), while patients with AN exhibited higher depression scores than NW (PHQ-9; *p* < 0.01).

Regarding EDI-2 total scores, patients with AN and OB did not differ from each other (*p* > 0.05) but displayed higher scores than NW subjects (*p* < 0.001; Table 2). All three groups differed from each other in the EDI-2 subscale "body dissatisfaction", with the highest scores in OB followed by AN and NW (*p* < 0.001; Table 2) and in "interoceptive awareness", with the lowest scores in NW (*p* < 0.001 vs. AN and OB) and the highest in AN (*p* < 0.05 vs. OB; Table 2). Moreover, the NW group showed the lowest, and AN and OB groups similar scores for the subscales "drive for thinness", "bulimia" (*p* < 0.001), "ineffectiveness" (*p* < 0.01), "maturity fears", and "interpersonal distrust" (*p* < 0.05). Furthermore, patients with AN displayed a higher "perfectionism" level than patients with OB (*p* < 0.05) and NW (*p* < 0.001; Table 2).

*3.3. Spexin Is Negatively Associated with Body Mass Index and Fat Mass but Not with Physical Activity or Energy Expenditure*

The results of correlation analyses of body composition, physical activity, and energy expenditure, as well as psychometric parameters, with SPX are presented in Table 3.

**Table 3.** Correlations of body mass index, body composition, physical activity, energy expenditure, and patient-reported outcomes with spexin.



**Table 3.** *Cont.*

Correlations were assessed using Pearson's or Spearman's analyses. Significant correlations are indicated in bold. Abbreviations: EAT, exercise-related activity thermogenesis (energy expenditure of more than three MET); EDI-2, Eating Disorder Inventory-2; GAD-7, Generalized Anxiety Disorder-7; MET, Metabolic equivalents of tasks; NEAT, non-exercise-related activity; PHQ-9, Patient Health Questionnaire-9; PSQ-20, Perceived Stress Questionnaire-20; TEE, Total energy expenditure.

In the whole study population, the mean plasma SPX concentration was 0.436 ± 0.153 ng/mL (range: 0.092–1.035 ng/mL). SPX levels were found to be significantly higher in AN than OB (*p* < 0.05) and did not differ between the other groups (*p* > 0.05; Table 2). This was reflected by a negative correlation of peripheral SPX with BMI in the whole study group (*r* = −0.149; *p* = 0.027; Figure 1A). Plasma SPX was also negatively associated with absolute (*r* = −0.149; *p* = 0.04; Figure 1B) and relative (*r* = −0.159; *p* = 0.028; Table 3) fat mass. No relationships were observed between circulating SPX and other parameters of body composition, as measured by bioelectrical impedance analysis (*p* > 0.05; Table 3).

**Figure 1.** Correlations of spexin with (**A**) BMI and (**B**) fat mass (kg) in the whole study population. Negative fat mass values in bioelectrical impedance analysis are possible in severely underweight patients, due to the manufacturer's algorithms being calculated primarily for normal weight subjects. Abbreviation: BMI, body mass index.

We observed no associations between SPX and all measured parameters of physical activity (steps/day, MET/day; *p* > 0.05) and energy expenditure (TEE, EAT, NEAT; *p* > 0.05; Table 3).

#### *3.4. SPX Is Not Correlated with Depressiveness, Anxiety, Perceived Stress, and Eating Disorder Pathology in the Whole Study Group*

No significant associations between SPX and anxiety (GAD-7), depressiveness (PHQ-9), eating disorder pathology (EDI-2), and perceived stress (PSQ-20) total score were observed

**Figure 2.** Correlations between (**A**) spexin and anxiety (GAD-7), (**B**) perceived stress (PSQ-20 total score), (**C**) depressiveness (PHQ-9), (**D**) and eating disorder pathology(EDI-2 total score). Abbreviations: EDI-2, Eating Disorder Inventory-2; GAD-7, Generalized Anxiety Disorder-7; PHQ-9, Patient Health Questionnaire-9; PSQ-20, Perceived Stress Questionnaire-20.

#### **4. Discussion**

The current study investigated the relationship between plasma SPX levels and objectively assessed PA, body composition, and patient-reported outcomes in a group of hospitalized adult women over a wide BMI range. In the whole study population, we showed a weak but statistically significant negative correlation between SPX and BMI, as well as SPX and body fat mass. However, we could not observe any relationships between SPX and parameters of physical activity and SPX and depressiveness, anxiety, stress, and eating disorder psychopathology.

The negative correlation between SPX and body fat mass is in line with earlier observations in adult individuals with obesity, in which a negative association between SPX and body fat percentage was reported [38]. Consistent with this, significantly lower SPX levels were reported in children with high compared to normal fat mass [39]. In one study, no correlation was found between SPX and body fat percentage [12], although this could be explained by the fact that the participants did not differ as much in BMI and body fat percentage as in our study. SPX is reduced after glucose load [4] and leads to lipolysis [5], which could explain the observed negative correlation of SPX with BMI and fat mass, and this would point toward SPX being responsible for the reduction in fat mass and not vice versa and could help to understand SPX's decrease in the peripheral circulation.

We did not observe any correlations between SPX and eating disorder symptoms as measured by EDI-2. As already mentioned in the introduction, several studies in animal models showed that SPX affects the levels of anorexigenic and orexigenic hormones and its effects result in suppression of food intake through a decrease of orexigenic peptides (AgRP, NPY, orexin) [13,15,40] or upregulation in mRNA expression of anorexigenic peptides (CCK, POMC, MCH) [15], predominantly in the hypothalamus. In addition, a fasting period led to a decline in SPX levels in the forebrain [41], and repeated daily intraperitoneal

administration of SPX reduced both the meal size and meal duration, leading to weight loss in animals [6]. However, to date, no study has shown a direct effect of SPX on eating behavior in humans. A study that analyzed the effects of weight gain during inpatient treatment on SPX levels in AN indicated no significant results [42]. Nevertheless, some studies showed a negative association between SPX and leptin [43,44]. Leptin suppresses food intake and leads to weight loss [45]. In blood serum in patients with AN, it is downregulated, which is primarily attributed to the reduced mass of adipose tissue, where peripheral leptin is primarily expressed [46]. Perhaps low leptin also reduces its anorexigenic effects and represents a compensatory mechanism that protects, although insufficiently, against further weight loss [47]. Since both SPX and leptin cause loss of appetite and are simultaneously negatively correlated, it could be claimed that their anorexigenic effects are based on different mechanisms. Given its negative correlation with BMI, SPX, unlike leptin, may act as a driver of weight loss rather than a mere satiety signal and may be one answer to why AN persists or becomes a chronic condition. The fact that SPX was not associated with any scale of the EDI-2 in the present study also suggests that impaired body image or impaired eating habits are not responsible, even partly, for anorexigenic effects of SPX (observed in animals) and the negative association between SPX and leptin (in humans), but that SPX acts predominantly as a brain signal to induce weight loss. Therefore, one might conclude that SPX is not involved in the pathogenesis of AN in terms of impaired eating habits as an expression of eating disorder psychopathology. Longitudinal studies simultaneously measuring leptin and SPX in eating disorders and adjusting for BMI and body fat mass are needed, to further investigate the relationship between SPX and leptin and the function of SPX in eating disorders.

As reported, we also did not observe an association between plasma SPX levels and physical activity, as accelerometrically measured with a SenseWear® armband. This findingdoes not support the findings of a recent study suggesting that SPX could work as an indicator of response to physical activity [9] or correlate with results from an animal study conducted in mice showing increased SPX mRNA expression in muscle tissue and increased concentration in blood serum after exercise [48]. However, a possible explanation for these inconsistent findings could be that, in the present study, PA was measured only cross-sectionally, whereas associations may be detectable only over time or only in subjects responding to exercise. Additionally, SPX might be associated only with voluntary PA, or only in obesity but not in AN and, therefore, might not be associated with the hyperactivity observed in patients with AN. It could also be speculated that the intensity of daily PA, as captured in our study by measuring steps per day, does not increase circulating SPX levels, as in exercise training as reported in men with type 2 diabetes might [49].

We did not identify any association between SPX levels and anxiety (GAD-7), depressiveness (PHQ-9), or perceived stress (PSQ-20). This supports the results of a study conducted in adolescent inpatients with AN, where no associations with depressiveness (measured by BDI), eating disorder symptoms (EAT-26), or obsessive-compulsive disorder symptoms (Y-BOCS) were observed [42]. However, this study did not include any experimental design to investigate the causal relationship between SPX release and psychometric parameters. To date, a possible relationship between (psychological) stress reaction and peripheral SPX has only been demonstrated in animals. One study indicated that overexpression of SPX1 (one of two SPX orthologs occurring in zebrafish) in the dorsal habenula reduced anxiety-associated behaviors in zebrafish [22]. In addition, in mice in which anxiety was induced, SPX mRNA expression was reduced in the hippocampus, whereas CRF mRNA expression was upregulated [21]. Furthermore, CRF treatment has been shown to decrease SPX expression [21], and fish chronically stressed by social defeat exhibited upregulated cortisol and SPX levels in the brain [20]. Thus, studies in animals indicate a role for SPX in the regulation of stress, emotion, and behavior, which may also apply to humans. Therefore, further investigations of changes in peripheral SPX levels following interventions inducing stress or anxiety, e.g., by using stress paradigms such as the Trier social stress test, are needed. However, the investigation of alterations of cerebral SPX

expression in humans would require using molecular imaging for peptide detection (such as the nanoflow liquid chromatography-mass spectrometry in combination with invasive microdialysis or the less invasive, but requiring the use of radioactive substances, positron emission tomography [50]). Nonetheless, at present, the known peptide monitoring methods are very expensive and not widespread. Moreover, the fact that repeated imaging would be needed (at least once before and after the stress test) makes the above-mentioned imaging techniques even more difficult to perform. In our explorative study, we did not find any association between PA and stress, and SPX, nor in the whole study group or in the subgroups (data not shown). Unfortunately, no studies investigated the effects of SPX on stress-mediated PA. Since there is evidence in the literature that stress interferes with PA [51], we suggest conducting an interventional study investigating changes in PA patterns and stress-like-behaviors (e.g., using an elevated-zero-maze test) after peripheral or central SPX injection.

The generalizability of the reported results is subject to certain limitations. First, BIA and the SenseWear® armband device are well-established measures for the determination of body composition and physical activity in clinical practice and research. However, BIA might have limited validity in severely underweight subjects [52], and the SenseWear® armband seems to slightly underestimate step counts, so the results must be interpreted with caution [53]. Second, the present study was conducted under naturalistic conditions; therefore, no healthy control group was employed. Consequently, we were unable to compare study participants with and without mental disorders. However, the included patients showed a wide range of psychological impairment on the different scales, so that circulating SPX levels could be well related to the constructs of depressiveness, anxiety, stress, and eating disorder pathology. In addition, the naturalistic design is also a strength, since it reproduces real-world conditions during inpatient treatment. Third, the naturalistic study design entails heterogeneity with regard to existing comorbidities, which are potential confounders and could therefore have contributed to the weak or absent associations observed. Therefore, future studies with more stratified study populations and an experimental research design should be conducted. Fourth, a cross-sectional study can only show associations and not cause-effect relationships. Therefore, in addition to experimental studies with healthy control groups, longitudinal studies are needed, to further examine changes in peripheral SPX levels over the course of improvements under treatment. Lastly, while studies in animals and humans indicate interrelations between SPX and the reproductive system [54], we did not assess menstrual status, and the intake of estrogen-containing medications was not an exclusion criterion in our female study population.

In this exploratory study, our findings replicate the negative association between SPX and both BMI and body fat. However, using a naturalistic and cross-sectional study design, no associations between circulating SPX and both patient-reported outcomes and PA were observed. Since animal studies indicated a possible effect of SPX on PA, anxiety, depressiveness, stress, and feeding behavior, further research in humans, employing longitudinal and interventional studies in more homogenous and larger study samples, is required.

**Author Contributions:** Conceptualization, A.S. and T.H.; Methodology, A.S. and T.H.; Validation, A.R., P.K., M.R., A.S. and T.H.; Formal Analysis, M.S. and A.S.; Investigation, M.S., A.R., E.W. and S.S.; Resources, A.S. and T.H.; Data Curation, M.S., A.R., E.W. and S.S.; Writing—Original Draft Preparation, M.S.; Writing—Review & Editing, A.R., P.K., M.R. and A.S. and T.H.; Visualization, M.S.; Supervision, A.S. and T.H.; Project Administration, A.S. and T.H.; Funding Acquisition, A.S. and T.H. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was supported by the Charité University Funding UFF 89/441-176 (A.S. and T.H.). We acknowledge financial support from the Open Access Publication Fund of Charité— Universitätsmedizin Berlin and the German Research Foundation (DFG).

**Institutional Review Board Statement:** Investigations were conducted according to the Declaration of Helsinki. The study was reviewed and approved by the Ethikkommission Charité–Universitätsmedizin Berlin. All patients provided written informed consent to participate in this study.

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author. The data are not publicly available due to data privacy.

**Acknowledgments:** We thank Reinhard Lommel and Petra Buße for laboratory work, and Mitsuru Murata and Christina Hentzschel for their help with the organization of clinical assessments.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

### **References**

