Adipokines in depressed women with and without adverse childhood experiences

Background: Adverse childhood experiences (ACE) elevate the risk of both major depressive disorder (MDD) and metabolic diseases. The underlying pathophysiology might include alterations of adipokine levels as a consequence of ACE. In this study, we used a full-factorial design to investigate the levels of select adipokines in women with ACE-only (n = 23), MDD-only (n = 27), ACE + MDD (n = 25) and healthy controls (HC, n = 29) to identify metabolic makers associated with vulnerability and resilience of developing MDD after ACE exposure. Methods: Serum levels of adiponectin, leptin, adiponectin-to-leptin (A/L) ratio, and retinol binding protein 4 (RBP4) were measured using enzyme-linked immunosorbent assay (ELISA). Results: Adiponectin levels did not differ between groups. Individuals with vs. without MDD showed higher leptin serum concentrations. As predicted, A/L ratio indicated lower values in individuals with vs. without ACE. RBP4 showed a more nuanced pattern with reduced levels in the ACE-only and MDD-only groups compared to HC. Furthermore, the ACE-only group showed lower RBP4 concentrations compared to ACE + MDD. These results were not accounted by BMI or medication status. Conclusion: Our results do not support the utility of adiponectin and leptin as predictors of vulnerability or resilience of developing MDD after ACE. In contrast, RBP4 might play a role in resilience towards the development of MDD following ACE. Further research on this more recently discovered adipokine seems warranted.


Introduction
Major depressive disorder (MDD) is a debilitating disease and a leading cause of disease burden worldwide (Whiteford, 2013).Prevalence in women is twice as high as in men (Seedat, 2009) and environmental factors such as sexual, physical or emotional abuse or neglect during childhood show a dose-response relationship with MDD risk, severity and chronicity (Heim and Binder, 2012;Li et al., 2016).Such adverse childhood experiences (ACE) lastingly alter neurobiological stress response systems, most prominently the hypothalamic-pituitary-adrenal (HPA) axis (Kupfer et al., 2012).Individuals with MDD and ACE show chronically increased HPA-axis activity and an altered HPA-axis response to acute psycho-social or endocrine stress (Stetler and Miller, 2011).Reduced glucocorticoid receptor function that may stem from chronic HPA-axis hyperactivity and increased cortisol levels further activate the inflammatory system, which is under inhibitory regulation by cortisol (Miller et al., 2002).Additionally, glucocorticoid resistance elevates adipogenesis (Brown et al., 2004;Hauner et al., 1987), leading to increased body mass (Brown et al., 2004).
Adipose tissue, in turn, acts as an endocrine organ that produces cell signaling proteins, adipokines, which regulate metabolic and inflammatory function (Ouchi, 2011;Pariante, 2017).The role of adipokines in the interplay between poor physical health and MDD etiology has recently started to gain increased attention (Otte, 2016;Gold, 2020;Rosenbaum, 2015).Two well-studied adipokines, adiponectin and leptin, have antagonistic functions.While leptin is positively correlated with increased fat mass and secretion of inflammatory cytokines that promote metabolic syndrome (Srikanthan, 2016;Wang and Scherer, 2016), adiponectin has anti-inflammatory and insulin-sensitizing effects and reduces metabolic risk (Adya et al., 2015).The combined measure of adipokine-to-leptin ratio has been suggested as a reliable marker of adipose tissue dysfunction associated with low-grade chronic inflammation (Frühbeck, 2018).
In patients with MDD, meta-analyses show lowered adiponectin and increased leptin levels compared to healthy controls (Cao, 2018;Carvalho, 2014;Hu et al., 2015), which has been demonstrated both in cross-sectional and prospective longitudinal designs (Pasco, 2008).However, overall findings are heterogeneous and differ as a function of gender, body mass index (BMI) and type of adipokine assay.History of ACE is also associated with lowered adiponectin levels, even after controlling for age, gender, BMI and diagnosis of MDD (Lehto, 2012).However, ACE-induced increase of leptin levels (Joung, 2014) might be primarily driven by increased BMI (Daniels, 2023).Taken together, the above findings suggest that ACE might contribute to a pro-inflammatory phenotype that is more prone to develop mental and physical illness later in life (Coelho, 2014;Danese, 2009;Vuong, 2020).Adipokines might serve as metabolic markers of such dysregulation that link ACE and MDD pathogenesis.However, it is possible that certain adipokine profiles also promote resilience towards developing MDD after ACE exposure.Consistently, a recent study demonstrated the protective role of increased adiponectin in the development of PTSD after trauma (Vuong, 2022).

Present study: Aims and hypotheses
The present study aimed to characterize the adipokine profiles of individuals with and without ACE who did or did not develop MDD later in life.Previous studies associated ACE with serum lipid alterations contributing to persistently increased immunoreactivity in MDD (Péterfalvi, 2019) and increased ex vivo immune responses independent of acute MDD symptoms (Hellmann-Regen, 2019).Here, we targeted adipokine profiles that might be involved in increased inflammation associated with ACE.
Using a full-factorial design, we compared four groups, individuals with ACE who did (ACE+MDD) or did not (ACE-only) develop MDD, MDD patients without ACE (MDD-only) and healthy controls (HC).Our aim was to disentangle the influence of ACE and MDD on adipokine profiles to identify protective or pathogenic adipokine profiles.
Based on previous findings, we hypothesized that individuals with ACE (with and without MDD) would exhibit an inflammatory phenotype characterized by lower A/L ratios compared to the MDD-only and HC group.We also analyzed a more recently discovered adipokine, retinol binding protein 4 (RBP4), which is a vitamin A carrier in the blood but also an adipokine and fatty acid transporter that promotes insulin resistance and metabolic risk (Li, 2018;Christou et al., 2012;Codoñer-Franch, 2016;Zhou, 2017).A recent study showed significantly reduced RBP4 serum levels in depressed patients compared to HC (Yao and Li, 2020).We therefore expected that RBP4 would be lower in both MDD groups compared to HC, irrespective of the presence of ACE.

Sample characteristics
Our total sample N = 104 was powered to detect a medium-size effect, as calculated with G*Power assuming a power of 0.80 and .05alpha-level (Faul, 2009).The four groups were comprised of 29 healthy women with no history of mental disorder or ACE (healthy controls: HC), 25 women with MDD and ACE (MDD+ACE), 27 women with MDD without ACE (MDD-only), and 23 women with ACE but no current or lifetime MDD (ACE-only).ACE was defined as repeated sexual or physical abuse or neglect at least once a month over one year or more before age 18 as assessed by a trained psychologist/psychiatrist in a semi-structured interview based on the German version of the Early Trauma Inventory, ETI (Wingenfeld, 2011).The Childhood Trauma Questionnaire (CTQ) was additionally used to quantify ACE (Wingenfeld, 2010).MDD patients met the DMS-IV diagnosis of current MDD assessed by SCID-I (Wittchen et al).The clinician-rated Montgomery-Åsberg Depression Rating Scale (MADRS) was used to quantify depression severity (Montgomery and Åsberg, 1979).Exclusion criteria for the clinical groups were post-traumatic stress disorder, schizophrenia, schizoaffective disorder, bipolar disorder, depressive disorder with psychotic features, anorexia, and lifetime alcohol or drug dependence.General exclusion criteria included CNS or severe somatic diseases, metabolic or endocrine diseases, autoimmune diseases, current infections, or pregnancy and a BMI higher than 30.For technical reasons, metabolic markers were not available from all participants (see Table 1).This sample corresponds to a previously published sample (Hellmann-Regen, 2019).All individuals provided written informed consent and the study took place at the Department of Psychiatry, Charité -Universitätsmedizin Berlin, Campus Benjamin Franklin.The study was approved by the local ethics committee.

Collection of whole blood
Study participants gave 18 ml of peripheral venous blood at 8 am after an overnight fast.Blood was collected with a vacuum system with two 9 ml serum separation tubes (Vacutainer™, BD Biosciences).The serum was isolated by centrifugation according to the manufacturer's instructions, stored in 1 ml aliquots, and stored at − 80 • C until further analysis.

Statistical analyses
Analyses of demographical group differences were investigated with one-way ANOVA or Chi 2 tests using SPSS (IBM SPSS Statistics 26).For adipokine group analyses, 2 (MDD) × 2 (ACE) ANOVA was calculated, BMI and medication status were entered as covariates of no interest.Significant MDD × ACE interaction effects were followed up by post-hoc t-tests.

Sample characteristics
The groups did not differ in years of education, age or smoking status.No psychotropic medication was taken by the ACE-only and HC group.Medication intake did not differ between the two MDD groups (MDD-only vs. ACE+MDD).A significant group difference in BMI emerged.CTQ scores were higher in the two ACE-exposed groups.Depression severity indexed by MADRS scores was highest in the MDD groups, with lower values for ACE-only and lowest values in HC.See Table 1 for details.

A/L ratio
Adiponectin-over-leptin exhibited a significant main effect of ACE (F (1,73) = 4.56, p <.05) but not MDD (F(1,73) = 2.14, p =.15).Overall, the ACE groups showed lower A/L ratios compared to individuals without ACE.There was a trend towards an ACE × MDD interaction effect (F(1,73) = 3.25, p =.08).Exploratory follow-up t-tests revealed that all clinical groups had lower A/L-ratios compared to healthy controls.

Discussion
Our study examined metabolic variables in individuals with and without ACE who did or did not develop MDD.We aimed to identify the adipokine profiles of ACE and MDD at a cross-sectional level, which may hold future utility to identify subjects at risk or resilience towards development of MDD after ACE exposure in a longitudinal setting.
We did not find differences in serum adiponectin concentrations as a function of ACE and/or MDD.Regarding leptin, a main effect of MDD suggested overall higher levels in MDD patients.Taking the combined measure of adiponectin-to-leptin ratio, a main effect of ACE emerged in line with our hypotheses, suggesting lower A/L ratios in both ACE groups.Thus, individuals who experienced adversity in childhood seem to be characterized by a higher inflammatory status compared to patients without early abuse or neglect.However, a trend towards an interaction effect might suggest that indeed all clinical groups (including MDD-only) showed lower A/L ratios compared to healthy participants.Adiponectin is regulated by leptin, and a reduced ratio is indicative of metabolic alterations, such as increased oxidative stress and low-grade inflammatory processes (Frühbeck, 2018).In our data, clinical groups showed reduced ratios even after controlling for BMI, suggesting that within our cohort of normal-weight women, BMI was not driving these effects.Taken together, profiles of adipokines did not discriminate between individuals who did or did not develop MDD after ACE, which does not support their utility as predictors of the vulnerability or resilience of developing MDD after ACE.
Retinol binding protein 4 (RBP4), a more recently discovered adipokine previously associated with MDD symptomatology, showed a more nuanced pattern.While we could not support our prediction of a main effect of MDD, we observed a main effect of ACE that was further qualified by a significant ACE × MDD interaction.While healthy controls showed the highest RBP4 levels compared to ACE-only and MDDonly individuals, the ACE-only group had even lower levels compared to the ACE+MDD group.This finding is potentially interesting and invites two possible post hoc explanations.Either, it identifies ACE-only as the group that deviates most from healthy controls, suggesting that developing MDD after ACE increases RBP4 levels, thus normalizing metabolic status towards the level of HC.However, this reasoning only applies if the association of RBP4 with pathology is linear in nature.Alternatively, further reduced RBP4 levels in ACE-only might indicate resilience rather than burden.In this case, low or high levels of RBP4 might protect from MDD, while medium levels are associated with acute disease.Such protective or compensatory mechanism of RBP4 levels in the ACE-only group might be related to the role of RBP4 in promoting insulin resistance or adipokine transport (Li, 2018).Based on this novel finding, further studies are needed to confirm the contribution of RBP4 to resilience towards developing MDD after early childhood adversity, and elucidate its interplay with the immune system and the HPA-axis.

Strengths and Limitations
All participants of our study were female, which limits the E. Kulakova et al. generalizability of our findings.The BMI cut-off makes our findings inconclusive regarding possible interactions of adipokines with metabolic alterations associated with adiposity.However, our analysis ruled out that either BMI or medication status were driving the observed effects.This allows to generalize the findings from a naturalistic sample of ACE and MDD patients, who are often characterized by increased BMI and the intake of psychotropic medication.The retrospective assessment of ACE using both a semi-structured clinical interview and a self-report questionnaire introduces the possibility of biases that require future studies using prospective longitudinal designs.Our study was powered to detect medium effect sizes, however, to detect smaller effects, studies with larger sample sizes are warranted.

Conclusion and future directions
Taken together, our study shows alterations of metabolic markers in a clinical population with childhood adversity and major depression.The adiponectin-to-leptin ratio was significantly reduced in the ACE vs. non-ACE groups, suggesting increased inflammatory activity.Importantly, these effects could not be explained by patient BMI or medication status and were observed in a sample of non-obese individuals.RBP4 levels showed an interesting interaction between ACE and MDD: Serum concentrations in the ACE-only group without MDD were lower than in individuals with ACE+MDD, which might suggest a potential mechanism of resilience that warrants further exploration.Janssen, Limes Klinikgruppe, Neuraxpharm, Oberberg Kliniken and Peak Profiling Research Funding from the German Research Foundation (OT 209/7-3; 14-1, 19-1, 21-1, EXC 2049), the European Commission (IMI2 859366), the German Federal Ministry of Education and Research (KS2017-067), and the Berlin Institute of Health (B3010350).EK, KW, LK, JHR and CS report no conflict of interest.

Fig. 1 .
Fig. 1.Average adipokine levels by group measured in serum by ELISA (a) Adiponectin: No group effects.(b) Leptin: Levels showed a main effect of MDD.(c) Adiponectin-to-leptin ratio: Main effect of ACE, however, the ratio was decreased in all clinical groups compared to HC (d) RBP4 (dilution 1:100,000): An ACE × MDD interaction effect indicated that compared to healthy controls, RBP4 levels were decreased for ACE-only and MDD-only individuals.Interestingly, RBP4 discriminated between individuals who did (ACE+MDD) or did not (ACE-only) develop MDD after ACE, with lower levels for individuals who were not affected by MDD despite of ACE experience.Error bars represent standard deviation.Abbreviations: HC = healthy controls, ACE = adverse childhood events, MDD = major depressive disorder, RPB4 = retinol binding protein 4.