Low serum adropin levels predict poor cognitive performance in late middle-aged African Americans
=================================================================================================

* Geetika Aggarwal
* Theodore K. Malmstrom
* Douglas K. Miller
* John E. Morley
* Andrew D. Nguyen
* Andrew A. Butler

## ABSTRACT

Adropin, a secreted peptide that is predominantly expressed in the brain, regulates processes that are important for maintaining cognitive functions. Whether circulating levels correlate with cognitive ability, and the factors regulating brain expression, are not known. Here we report circulating adropin levels correlate with risk of poor cognitive performance (scoring in the lowest quintile for a composite cognitive score) in the African American Health Study (n=352, mean age 56.6yr, STD 4.4yr, range 45 to 65yr). Risk of poor cognitive performance declined with increasing adropin levels (OR=0.715, p<0.01). This relationship is more pronounced in people with impaired glycemic control (OR=0.532, P<0.05). Adropin is highly expressed in cultured human astrocytes, where its expression correlates inversely with age, positively with transcriptomic signatures of energy metabolism and macromolecule synthesis and is suppressed by pro-inflammatory factors. Together, suppression of adropin expression in astrocytes is a plausible link between aging, inflammation, and risk of cognitive impairment.

Keywords
*   Aging
*   Cognitive Function
*   Endocrinology
*   Glucose Metabolism
*   Hormones

## INTRODUCTION

Circulating levels of adropin, a low molecular weight peptide hormone, correlate with vascular tone, glucose metabolism, and hepatic lipid metabolism 1-4. The primary source of adropin in the circulation is not known. However, in humans and nonhuman primates (NHP) the expression of the Energy Homeostasis Associated (ENHO) transcript encoding adropin is higher in brain relative to other tissue samples by orders-of-magnitude 5,6. Recent studies suggest that adropin actions in the brain could delay cognitive decline associated with deteriorating homeostasis in the context of obesity and aging 5,7-9. ENHO expression in tissue samples from human, NHP, and mouse brains correlates with genes implicated in neurodegenerative conditions 5,6. In human postmortem brain samples, strong positive correlations are also observed with gene networks involved in mitochondrial function 5. Positive correlations are also observed with gene networks involved in lipid metabolism, synaptic plasticity, and vascular function 5. Moreover, adropin expression peaks during the first decade of life, coinciding with a time of major brain development 5. These results imply high adropin expression in the human brain might correlate positively with energy utilization and synaptic plasticity which are critical for maintaining brain health 10.

Experiments using rodent models suggest adropin signaling support brain function during the lifespan. Declining adropin expression in brain tissues of aging mice and rats correlates with changes in the expression protein markers of neuroinflammation and oxidative stress 5,9. These correlations could reflect causal relationships, as adropin treatments have a positive impact on cognitive function in preclinical mouse models of natural aging 5. Experiments in mice also suggest protection against cerebrovascular ischemia involving actions on the brain vasculature 8, blood-brain barrier permeability 11, and preservation of endothelial integrity during injury 8,12.

Whether circulating adropin levels correlate with cognitive abilities in humans has not been investigated. Here we report the results from a study comparing serum adropin concentrations in the he African American Health study (AAH). The study’s primary objective was to test the hypothesis that circulating adropin concentrations correlates with cognitive ability. Adropin is highly expressed in astrocytes [https://www.proteinatlas.org/ENSG00000168913-ENHO/single+cell+type/brain](https://www.proteinatlas.org/ENSG00000168913-ENHO/single+cell+type/brain); [https://www.proteinatlas.org/ENSG00000168913-ENHO/tissue](https://www.proteinatlas.org/ENSG00000168913-ENHO/tissue)). The secondary objective investigated factors affecting adropin expression in astrocytes isolated from adult donors.

## RESEARCH DESIGN AND METHODS

### Study participants

All study procedures were reviewed and approved by the Institutional Review Board at Saint Louis University and were performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and later amendments. All participants provided written consent for study, including blood draw.

The African American Health study (AAH) is a population-based cohort of 998 self-identified African American individuals in 2000–2001 in the Saint Louis metropolitan area. Inclusion criteria included self-reported Black or African American race; birth dates between 01/01/1936 and 12/31/1950; standardized Mini-Mental Status Examination (MMSE) scores ≥ 16; and willingness to sign an informed consent form 13-16. Participants completed in-home interviews and assessments at baseline, and at years 3 and 9 Recruitment proportion (participants/enumerated eligible persons) was 76%. Self-reporting bias was minimized using validated and reliable criteria for collecting self-reported data. A subset of participants agreed to donate blood; serum samples were stored at −80 °C for laboratory analyses.

The current study established baseline serum adropin concentrations for 352 of the participants for which serum was still available (**Table S1**) and used baseline cognitive data and serum markers of metabolic homeostasis and immune function.

### Mental status

Cognitive performance was assessed by the mini-mental state examination (MMSE) and one-minute animal naming test (ANT). The MMSE is a screening tool for detecting cognitive deficits in dementia and delirium 17. The ANT assesses memory recall and verbal fluency 18. A composite cognitive test score was calculated using the sum of the Z scores for the results of both tests. MMSE scores of ≤23 or ANT scores of ≤15 predict risk for dementia 18,19. For this study, poor cognitive performance was defined as scores in the lowest quintile (“low\_quint”) for the composite score (MMSE+ANT_low_quint) (**Fig. 1A**). For the composite score, both MMSE and ANT scores were significantly lower in the MMSE+ANT_low_quint group relative to the 2nd to 5th quintile group (**Fig. 1A**). MMSE and ANT scores exhibited a modest correlation (*ρ* =0.333, p<0.001, n=338) (**Fig. 1B**). However, the composite score highly correlated with the results from both tests (MMSE, *ρ* =0.752, p<0.001; ANT, *ρ* =0.832, p<0.001). (**Fig. 1C, D)**.

![Figure 1.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2023/01/11/2021.12.09.21267550/F1.medium.gif)

[Figure 1.](http://medrxiv.org/content/early/2023/01/11/2021.12.09.21267550/F1)

Figure 1. Distribution of cognitive test results between groups in the low_quintiles for MMSE, ANT, or the composite (MMSE+ANT) scores.
\***|, p<0.001 by Mann-Whitney U test between the indicated groups. (A) MMSE and ANT test score data showing mean and distribution in the low_quntile and 2nd-5th quntile groups. The left and middle panels show data for the individual tests. The panel to the right shows the results from the individual MMSE and ANT test score data in the composite score low_quint and 2nd-5th quintile groups. (B) Correlation of MMSE and ANT scores. (C) Correlation of MMSE with the composite (MMSE+ANT) score. (D) Correlation of ANT with the composite (MMSE+ANT) score.

### Laboratory procedures

Serum adropin concentrations were measured using an enzyme immunoassay kit from Phoenix Pharmaceuticals, Inc. (cat. no. EK-032-35) following the manufacturer’s protocol. Assay sensitivity reported by the manufacturer is 0.3 ng/ml, with a linear range from 0.3 to 8.2 ng/ml. Pilot experiments determined a 1:5 dilution of serum produced values within the linear range. Values <0.3 ng/ml were described as ‘low’ (n=1) and arbitrarily assigned a value of 0.15 ng/ml (half of the lowest detectable value). Values >8.2 ng/ml were described as “high” and arbitrarily assigned a value of 8.5 ng/ml (n=2). Adropin assays were performed in triplicates, requiring 15 96-well plates. The initial criterion for inclusion was based on the coefficient of variation (%CV, SD divided by the mean of three replicates). Plate controls included an ‘in-house’ human serum sample, and controls provided by the assay manufacturer. The %CV for the plate controls were 11% and 14%, respectively. For samples with %CV>20, the ‘outlier’ was removed. Data from triplicates with no clear outlier (5 samples) were excluded. The mean, median, and standard deviation (SD) for the 347 samples used in the analysis were 3.17, 3.11, and 1.35 ng/ml, respectively.

The procedures used to measure serum indices of metabolic control and inflammation have been described previously 20,21. Briefly, fructosamine was measured as previously described 22. Soluble Interleukin-6 Receptor (IL-6R) was measured by ELISA (ICN-Biomedicals, Costa Mesa, CA; intra-assay CV 5.0%; interassay CV 5.9%). ELISA kits (BioSource, Camarillo, CA) were used for measuring soluble Tumor Necrosis Factor Receptor-1 (sTNFR1) (intra-assay CV 4.1%, interassay CV 7.3%) and soluble Tumor Necrosis Factor Receptor-2 (sTNFR2) (intra-assay CV 5.1%, interassay CV 8.6%). C-Reactive Protein (CRP) was measured using a High-Sensitivity Enzyme Immunoassay (hsCRP ELISA) (MP Biomedicals, Orangeburg, NY; intra-assay CV 4.5%, interassay CV 4.1%).

### Expression profiling

RNA sequencing (RNA-seq) data was downloaded from published experiments (GSE73721, GSE147870) 23,24. Gene networks correlating with *ENHO* expression were identified using a paired partial correlation analysis as previously described 5,6; donor age was used as a covariate in calculating the correlation matrix. Correlation coefficients of >0.7 or <-0.7 were used as selection criteria. The ToppGene Suite was used to identify gene networks overrepresented in genes identified as correlating positively or negatively with the ENHO transcript encoding adropin.

### Statistical analysis

Data were analyzed using IBM SPSS Statistics version 19.0 (IBM Corp., Somers, NY). Modelling of the relationships between serum adropin concentrations and indices of glycemic control (fructosamine, leptin, adiponectin), lipid metabolism, and inflammation were assessed using bivariate correlations (Spearman’s correlation coefficient, *ρ*) or multiple linear regression. Relationships between variables correlating with serum adropin were also investigated using tertiled data (1st, 2nd, and 3rd tertiles defined by adropin values up to the 33rd percentile, between the 33rd and 67th percentile, and above the 67th percentile, respectively). When comparing groups, analysis of covariance (ANCOVA) was used to control for relevant variables (for example, sex, BMI, and age). This affected the sample size, as cases with missing data were removed from the analysis. Post-hoc tests were used in situations where significance (p<0.05) was observed between >2 groups. Binomial logistic regression determined relationships between serum adropin concentrations and risk of poor cognitive performance, defined as scoring in the low quintile for the composite score.

Raw data are presented as mean ± standard deviation, adjusted data from ANCOVA are reported as estimated marginal means ± standard error.

### Data and resource availability

Access to datasets from the AAH study are available upon request.

## RESULTS

### Low serum adropin concentrations increase risk of poor cognitive performance

Serum adropin concentrations adjusted for age, BMI, and indices of metabolic homeostasis (leptin, adiponectin, fructosamine, and triglycerides) were significantly lower in the MMSE+ANT\_low\_quint group compared to the 2nd-5th quintile group, irrespective of sex (**Table 1**). Modeling of the relationship between serum adropin concentrations and risk for scoring in the low\_quint group using a simple binomial logistic regression also indicated low adropin values associate with poor cognitive performance; increasing serum adropin concentrations reduce odds for scoring in the low\_quint group for the composite score (OR=0.771, p<0.05). When using the scores from the individual tests, serum adropin concentration were significantly lower in MMSE\_low\_quint group relative to the 2nd-5th quintile group, but not for the ANT (**Table 1**). However, serum adropin concentrations did not predict odds of scoring in the low\_quint group for the individual test scores (MMSE_low_quint group, OR=0.869, p=0.148; ANT_low_quint group, OR=0.984, p=0.861).

View this table:
[Table 1.](http://medrxiv.org/content/early/2023/01/11/2021.12.09.21267550/T1)

Table 1. 
Serum adropin concentrations adjusted for age, BMI, leptin, adiponectin, fructosamine, and TG in male and female study participants scoring in the low quintile for the specified tests. Data shown are estimated means ± SE, significant differences between low_quintile and 2nd-5th_quintiles are indicated by asterisks (*p<0.05; \***|, p<0.005).

### Inclusion of cardio-metabolic risk factors strengthens the correlation

Circulating levels of adropin correlate with indices of glucose and lipid metabolism 1-4. Here we also observed correlations between serum adropin concentrations and serum indices of metabolic homeostasis. When separated into three groups (tertiles) ranked by serum adropin concentrations, there were significant differences in the serum concentration of adiponectin, fructosamine, and triglycerides (**Table 2**). These differences reflected weak positive correlations with serum levels of adiponectin (*ρ*=0.198, p<0.001) and fructosamine (ρ=0.128, p<0.05), and a negative correlation with TG (*ρ*=-0.121, p<0.05). While the actual data for cognitive tests was similar between tertiles, the distribution of participants with poor cognitive performance was skewed (**Table 2**). The % scoring in the low_quint for the composite score was 24%, 22%, and 14% for the low, mid, and high adropin tertile groups, respectively.

View this table:
[Table 2.](http://medrxiv.org/content/early/2023/01/11/2021.12.09.21267550/T2)

Table 2. 
Morphometry, serum data, and cognitive test scores for study participants separated in tertiles ranked by serum adropin concentrations. Data shown are mean, SEM and 95% CI for estimated marginal means ± std. error. Serum values are adjusted for age, BMI, and sex; BMI is adjusted for age and sex; cognitive test scores (MMSE, ANT, MMSE+ANT) are adjusted for age and sex.

Sex differences were observed in age-adjusted BMI (32.6±0.44 vs. 28.7±0.62 kg/m2 for females and males, respectively, p<0.001). However, sex and age distributed evenly between adropin tertiles (**Table 2**). BMI was not significantly different between adropin tertiles, although there was a weak inverse correlation when adjusted for age and sex (r=-0.117, p=0.028). Serum adropin concentrations did not correlate with makers of inflammation such as C-reactive protein (CRP) (**Table 2**) or other serum markers (sIL2sR, sIL6sR, sTNFsR1, sTNFsR2; data not shown). High neurofilament light chain (NfL) levels can indicate increased risk of dementia 25, however there was no correlation with adropin (**Table 2**).

Including morphometric (sex, age, and BMI) and metabolic parameters (adiponectin, leptin, TG, and fructosamine) in the binomial logistic regression improved the ability of modelling to predict odds of poor cognitive performance (**Table 3**). For the composite score (MMSE+ANT), serum adropin concentrations were now highly significant predictors of scoring in the low_quint (ORadropin=0.715, p=0.003). Predictably, age was a significant positive predictor of risk (ORage=1.091, p=0.008). High serum adiponectin concentration was also a positive risk factor (ORadiponectin=1.055, p=0.046). However, BMI was not a significant predictor of poor cognitive performance.

View this table:
[Table 3.](http://medrxiv.org/content/early/2023/01/11/2021.12.09.21267550/T3)

Table 3. 
Binomial regression results for equations predicting risk of scoring in the low\_quint for MMSE, or for being in the low\_quint for the composite score (MMSE+ANT). Modelling using serum adropin concentrations only, or including serum adropin concentrations with serum markers of metabolic control and morphometrics are shown.

Applying this approach to scores in the low_quint for the individual tests resulted in a trend for the MMSE (ORadropin=0.826, p=0.060) (**Table 3**). However, there was still no relationship between serum adropin concentration and ANT scores.

### Adropin is strong predictor of poor cognitive performance in situations of impaired glycemic control

Fructosamine (glycated protein) levels in the blood correlate with glucose concentrations over a 2-to 3-week period, with values >287 μmol/L indicating potential incident diabetes 26. Interestingly, binomial regression in participants separated into two groups by fructosamine values (>287 or <286 μmol/L) indicates that the relationship between adropin and cognitive performance may be limited to participants with impaired glucose homeostasis (**Table 4**). For participants with fructosamine values >287 μmol/L, the % scoring in the MMSE+ANT\_low_quint was 47%, 24%, and 11% in the low, mid, and high adropin tertile groups (n=17,21, and 28). For participants with fructosamine values <286 μmol/L the distribution was 20, 22, and 16% (n=94, 92, and 83). A similar difference in the distribution of % scoring in the low_quint was observed using results from the MMSE. The distribution of MMSE_low_quint scores between tertiles was 39%, 18%, and 14% with fructosamine values >287 μmol/L (n=18,22, and 29), and 19%, 20%, and 17% for fructosamine values <286 μmol/L (n=98, 95, and 87).

View this table:
[Table 4.](http://medrxiv.org/content/early/2023/01/11/2021.12.09.21267550/T4)

Table 4. Serum adropin concentrations are a better predictor of cognitive impairment in subjects with serum fructosamine concentrations >287 μmol/L.
The table shows results from a binomial regression for predicting risk of scoring in the low\_quint for the test indicated with serum adropin concentration, sex, and age as covariates. The numbers in parenthesis indicate the distribution of test scores: (#scoring in low_quint/#scoring in 2nd-5th quint)

### Adropin expression in astrocytes correlates with donor age and metabolic activity

The highest level of adropin expression in the human brain is observed in astrocytes, a heterogenous population of cells that interact with neurons to regulate synaptic plasticity 27. In astrocytes isolated from human donors, the range of ENHO (adropin) expression appears to be sex-specific (**Fig. 2A**); male donors exhibited a greater range of expression mirroring the relationship observed in plasma adropin concentrations 28. When expressed as a function of age, removal of a potential outlier suggests the appearance of a decline in expression (**Fig. 2B**). Analysis using age as a covariate suggested a tendency for an effect of sex (age adjusted ENHO expression in males, 3.01±0.351; females, 2.11±0.35; p=0.108).

![Figure 2.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2023/01/11/2021.12.09.21267550/F2.medium.gif)

[Figure 2.](http://medrxiv.org/content/early/2023/01/11/2021.12.09.21267550/F2)

Figure 2. Expression of the adropin transcript in isolated astrocytes.
(A) ENHO expression levels in astrocytes from male or female donors (n=6/group). (B) ENHO expression as a function of donor age. The correlation coefficient shown is after removal of one outlier (open circle); including the outlier results in a trend for an association (ρ=-0.529, p=0.077). (C) Relationship between the pair-correlation matrices for transcripts paired with donor age (y-axis) or ENHO (x-axis). Most transcripts (50%) are in the quadrant that correlate positively with ENHO and negatively with age. (D) Results from the gene enrichment analysis showing cellular pathways correlating positively with expression of ENHO (adropin).

To identify gene networks co-expressed with ENHO, we used a paired regression approach 6. Analysis of the correlation coefficients of all transcripts against age or ENHO suggests a relationship (**Fig. 2C**). Genes correlating positively with ENHO also correlate inversely with donor age. This suggests repression with aging, an effect that has been reported at the level of protein expression in rodent models 5,9.

To identify co-regulated genes, the paired correlation matrix with ENHO as the bait was calculated with age as a covariate. This approach identified 752 genes with a correlation coefficient with ENHO of >0.7 (supplemental data file, **Table S2**). Gene ontology analysis indicated highly significant enrichment for pathways related to macromolecular synthesis and oxidoreductase reactions (**Fig. 2D**, supplemental data file, **Table S3**). Repeating the analysis using genes identified in a simple paired correlation yielded a similar outcome (supplemental data file, **Table S2**). While this approach identified more genes (1627 genes), it did not affect the interpretation with similar metabolic processes showing enrichment, albeit with reduced statistical significance (supplemental data file, **Table S3**). Comparing genes showing an R>-0.7 with ENHO (2276 genes) or an R<-0.7 for age (1627 genes) indicated significant overlap (826 genes) (supplemental data file, **Table S4**). The 826 genes showing overlap were enriched for pathways involved in metabolic processes (protein synthesis/transport and autophagy) (supplemental data file, **Table S4**).

Fewer genes met the selection criteria for correlating negatively with ENHO (420 in age-adjusted matrix, 248 in the simple matrix (supplemental data file, **Table S5**). Enrichment was only observed for the age-adjusted gene set, and primarily identified pathways active in the cell nucleus. Comparison of the genes identified by simple correlations as correlating negative with ENHO and positively with age was also not informative, failing to identify common pathways (supplemental data file, **Table S4)**.

### Regulation of astrocyte adropin expression by inflammation

Adropin expression in the rodent brain is suppressed with aging, correlating with inflammation 5,9. Experiments using astrocytes isolated from either human or mouse donors (GSE147870) suggest ENHO expression is suppressed by the TLR3 agonist Poly I:C, and by TNFα (**Fig. 3A-F**). A two-way ANCOVA (independent variables: treatment, donor species) indicated an inhibitory effect of Poly I:C and TNFα irrespective of the donor species (treatment effect, p<0.05). While adropin stimulates angiogenesis 29, it is not regulated by hypoxia (**Fig. 3**).

![Figure 3.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2023/01/11/2021.12.09.21267550/F3.medium.gif)

[Figure 3.](http://medrxiv.org/content/early/2023/01/11/2021.12.09.21267550/F3)

Figure 3. Expression of adropin is suppressed by inflammatory stressors.
Responses of ENHO expression in astrocyte cultures from human or mouse donors to the TLR agonist Poly I:C (A-C, 200 μg/ml for 72h), TNFα (D-F, 30 ng/ml for 48h), or hypoxia (G-I, 1% O2 for 72h). * p<0.05 vs. control (CTRL).

## DISCUSSION

This investigation began with cross-sectional study testing the hypothesis that serum adropin concentrations correlate with cognitive performance. The study used a well-characterized population-based cohort of African American individuals living in St. Louis, MO. The study rationale was based on human expression profiling data and the results from mouse studies. Expression of the transcript encoding adropin in post-mortem brain samples correlates positively with pathways suspected of influencing risk of cognitive decline (mitochondrial function, synaptic plasticity, vascular function, inflammation) 5,8,9,12,30. Moreover, experiments using C57BL/6J mice suggest high adropin expression prevents or delays cognitive decline with natural aging 5 or obesity-related metabolic dysregulation (hypercholesterolemia, type 2 diabetes) 7.

The main findings of the current study are that the risk of poor cognitive performance declines with increasing serum adropin concentrations. This finding supports the hypothesis that adropin signaling functions, either directly or indirectly, support cognitive function. In this cohort, the relative risk of exhibiting poor cognitive performance decreased by 20-30% for each 1 ng/mL increase in adropin concentration. The relationship between circulating adropin and cognitive ability may be particularly relevant in people with type 2 diabetes. These findings strongly indicate that further study of the potential link between circulating adropin concentrations, cognitive performance, and risk of cognitive decline and dementia are needed. Further studies of the relationship between circulating adropin and cognition in patients with type 2 diabetes may be of particular importance given the relationship between this condition and risk of cognitive decline and aging-related dementias 31.

The mechanisms that could explain the increased risk of cognitive impairment in people with low circulating adropin are unclear and will require further investigation. In this study, low serum adropin concentrations did not correlate with severe obesity, diabetes, or systemic markers of inflammation or neuropathy. Recent experiments using mice and cultured cells indicate that deficits in cerebrovascular blood flow, protection of the blood brain barrier, and synaptic functions could be involved 3,5,7,8,30,32.

There are limitations to the current study. These include the use of a cohort housed in a single geographic location. The study limited to a single racial group (African Americans), albeit from two socioeconomically diverse geographic areas. Another limitation was that the inclusion criteria for the AAH study included a MMSE score of ≧16 13,14. The study population therefore does not include late middle-aged people with dementia. The measurements used to assess glucose control and lipid metabolism were also limited to serum fructosamine and triglyceride levels. In addition, only two tests were used to assess cognitive function, and there was some inconsistency in the correlations between the two tests. These results should therefore be considered preliminary. Future studies are clearly needed to explore the relationships between circulating adropin concentrations and risk of dementia. Studies using subjects with a diagnosis of type 2 diabetes are also needed. However, if the relationship between circulating adropin levels and cognitive ability can be translated to the general population, it could represent an important new clinical metric in an aging population.

The source of circulating adropin remains to be identified, with the brain and (to lesser degree) the liver identified as major sites of expression 5,6,33-35. In a diurnal NHP (baboon), brain adropin expression exhibits a circadian profile with peaks during the daytime 6. Circulating adropin concentrations in another NHP model, the rhesus macaque, also exhibits a diurnal profile with higher levels in the daytime 6. Collectively, these results suggest that adropin could be a circadian-controlled neuropeptide, with levels in the circulation reflecting expression in the nervous system.The initial studies investigating circulating levels of adropin resulted from data suggesting adropin1-76 is a prohormone with a signal peptide (adropin1-33) and a secreted peptide (adropin34-76) 33. Indeed, adropin is often described as a liver-secreted ‘hepatokine’ 2,36. However, a later study refuted the finding that adropin is secreted, describing instead a membrane-bound protein primarily expressed in the brain 34. This conclusion is supported by data from AlphaFold’s transmembrane protein structure database, which predicts a transmembrane region near the N-terminus. However, we have confirmed that adropin is secreted by human iPSCs and is detectable using the same ELISA in the conditioned media (Aggarwal and Nguyen, unpublished data).

There was a weak inverse correlation between serum adropin concentrations and BMI. We also observed a positive correlation between serum adropin and adiponectin concentrations. These outcomes are comparable to the results from a NHP model (rhesus macaque) 6. However, it is also important to note that the relationship between circulating adropin concentrations and obesity in humans is inconsistent and may be age-dependent 28. In humans the consumption of fructose beverages and associated insulin resistance also increase plasma adropin concentrations 37. The positive relationship between adropin and fructosamine values in the current study suggest that it is premature to state that metabolic dysregulation, such as hyperglycemia, commonly associates with lower circulating adropin concentrations.

Our analysis suggested that three factors (serum levels of adropin and adiponectin, age) were risk factors for poor cognitive performance. That increasing age is a risk factor is predictable, while the correlation with adropin supports our hypothesis. The ‘adiponectin paradox’ where high levels of a protein with neuroprotective properties associate with mild cognitive impairment has also been reported elsewhere 38. Serum adropin and adiponectin are likely independent risk factors. While the two hormones exhibit a positive correlation, poor cognitive performance was observed with low adropin or high adiponectin.

Adropin protein levels in the rodent brain are suppressed by aging and metabolic dysregulation 7,9. Here we examined the factors that could regulate adropin expression in the nervous system using data from cultured astrocytes. The pro-inflammatory cytokine TNFα and poly(I:C), a synthetic analog of double-stranded RNA that mimics viral infections, both inhibit ENHO expression in astrocytes. This result is anticipated and suggests that adropin expression is suppressed in inflammatory conditions. On the other hand, hypoxia did not significantly change ENHO expression. Low oxygen supply which could result from vascular injury thus does not appear to have a direct effect on adropin expression.

In human astrocytes, higher adropin expression is associated with pathways involved in energy-demanding cytosolic processes, including the synthesis and turnover of macromolecules dominate. High adropin expression in astrocytes thus appears to correlate with higher metabolic activity and energy demand, with energy production used for the synthesis of macromolecules (RNA, proteins). A similar relationship was observed in RNA-seq data obtained from post-mortem brain samples from people of all ages 5. High adropin expression in the brain may therefore indicate relatively high levels of metabolic activity. Experiments using mice suggest that this relationship could be causal, as adropin treatment enhances the expression of similar pathways 5. Further studies are clearly needed to determine whether circulating adropin levels correlate with brain glucose uptake and blood flow.

In summary, the results of the current study indicate that low circulating adropin concentrations in late middle-aged people associate with poor cognitive function. Further studies are needed to determine whether circulating adropin concentrations correlate with risk for cognitive decline with aging and impaired glucose homeostasis.

## Supporting information

Table S1 [[supplements/267550_file02.pdf]](pending:yes)

supplemental data file, Table S2 [[supplements/267550_file03.xlsx]](pending:yes)

## Data Availability

All data produced in the present study are available upon reasonable request to the authors

## Author contributions

AAB was responsible for data management, data analysis, conceptual development, interpreting data, and preparation of the first draft of the manuscript. GA was responsible for provided data and preliminary data analysis. ADN provided resources and supervised data acquisition. TM was responsible for managing the complete AAH dataset, provided data, and assisted with analysis. DKM and JEM were responsible for conceptual development, design, and funding acquisition for the AAH study. All authors read and approved the manuscript for content. All authors read and approved the manuscript for content.

## Competing interest’s statement

The authors have no competing interests to report.

## ACKNOWLEDGEMENTS

AAB was supported by the NINDS, NIH (R21NS108138). African American Health was supported by a grant from the National Institute on Aging (R01 AG010436 to DKM). AAB and ADN also acknowledge financial support provided by Saint Louis University.

## Footnotes

*   **Disclosures:** The authors declared no conflict of interest

*   The data pertaining to risk of mortality was removed from the manuscript. Cognitive function is assessed using a composite score derived from the Mini-Mental State Exam and Animal Naming Test.

*   Received December 9, 2021.
*   Revision received January 11, 2023.
*   Accepted January 11, 2023.


*   © 2023, Posted by Cold Spring Harbor Laboratory

The copyright holder for this pre-print is the author. All rights reserved. The material may not be redistributed, re-used or adapted without the author's permission.

## REFERENCES

1.  Butler, A. A. & Havel, P. J. Adropin and insulin resistance: Integration of endocrine, circadian, and stress signals regulating glucose metabolism. Obesity (Silver Spring) 29, 1799–1801, doi:10.1002/oby.23249 (2021).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1002/oby.23249&link_type=DOI) 

2.  Mushala, B. A. S. & Scott, I. Adropin: a hepatokine modulator of vascular function and cardiac fuel metabolism. Am J Physiol Heart Circ Physiol 320, H238–H244, doi:10.1152/ajpheart.00449.2020 (2021).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1152/ajpheart.00449.2020&link_type=DOI) 

3.  Jurrissen, T. J. et al. Role of adropin in arterial stiffening associated with obesity and type 2 diabetes. Am J Physiol Heart Circ Physiol 323, H879–H891, doi:10.1152/ajpheart.00385.2022 (2022).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1152/ajpheart.00385.2022&link_type=DOI) 

4.  Fujie, S. et al. Aerobic Exercise Restores Aging-Associated Reductions in Arterial Adropin Levels and Improves Adropin-Induced Nitric Oxide-Dependent Vasorelaxation. J Am Heart Assoc 10, e020641, doi:10.1161/JAHA.120.020641 (2021).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1161/JAHA.120.020641&link_type=DOI) 

5.  Banerjee, S. et al. Adropin correlates with aging-related neuropathology in humans and improves cognitive function in aging mice. NPJ Aging Mech Dis 7, 23, doi:10.1038/s41514-021-00076-5 (2021).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41514-021-00076-5&link_type=DOI) 

6.  Butler, A. A. et al. Low plasma adropin concentrations increase risks of weight gain and metabolic dysregulation in response to a high-sugar diet in male nonhuman primates. J Biol Chem 294, 9706–9719, doi:10.1074/jbc.RA119.007528 (2019).
    
    [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6MzoiamJjIjtzOjU6InJlc2lkIjtzOjExOiIyOTQvMjUvOTcwNiI7czo0OiJhdG9tIjtzOjUwOiIvbWVkcnhpdi9lYXJseS8yMDIzLzAxLzExLzIwMjEuMTIuMDkuMjEyNjc1NTAuYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9) 

7.  Ghoshal, S. et al. Adropin transgenesis improves recognition memory in diet-induced obese LDLR-deficient C57BL/6J mice. Peptides 146, 170678, doi:10.1016/j.peptides.2021.170678 (2021).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.peptides.2021.170678&link_type=DOI) 

8.  Yang, C. et al. Neurovascular protection by adropin in experimental ischemic stroke through an endothelial nitric oxide synthase-dependent mechanism. Redox Biol 48, 102197, doi:10.1016/j.redox.2021.102197 (2021).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.redox.2021.102197&link_type=DOI) 

9.  Yang, C., DeMars, K. M. & Candelario-Jalil, E. Age-Dependent Decrease in Adropin is Associated with Reduced Levels of Endothelial Nitric Oxide Synthase and Increased Oxidative Stress in the Rat Brain. Aging Dis 9, 322–330, doi:10.14336/AD.2017.0523 (2018).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.14336/AD.2017.0523&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=29896421&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F01%2F11%2F2021.12.09.21267550.atom) 

10. Cheng, X. T., Huang, N. & Sheng, Z. H. Programming axonal mitochondrial maintenance and bioenergetics in neurodegeneration and regeneration. Neuron 110, 1899–1923, doi:10.1016/j.neuron.2022.03.015 (2022).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.neuron.2022.03.015&link_type=DOI) 

11. Yang, C., DeMars, K. M., Hawkins, K. E. & Candelario-Jalil, E. Adropin reduces paracellular permeability of rat brain endothelial cells exposed to ischemia-like conditions. Peptides, doi:10.1016/j.peptides.2016.03.009 (2016).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.peptides.2016.03.009&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=27020249&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F01%2F11%2F2021.12.09.21267550.atom) 

12. Dodd, W. S. et al. Adropin decreases endothelial monolayer permeability after cell-free hemoglobin exposure and reduces MCP-1-induced macrophage transmigration. Biochem Biophys Res Commun 582, 105–110, doi:10.1016/j.bbrc.2021.10.032 (2021).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.bbrc.2021.10.032&link_type=DOI) 

13. Miller, D. K. et al. Clinically relevant levels of depressive symptoms in community-dwelling middle-aged African Americans. J Am Geriatr Soc 52, 741–748, doi:10.1111/j.1532-5415.2004.52211.x (2004).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1111/j.1532-5415.2004.52211.x&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=15086655&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F01%2F11%2F2021.12.09.21267550.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000220855300013&link_type=ISI) 

14. Miller, D. K., Wolinsky, F. D., Malmstrom, T. K., Andresen, E. M. & Miller, J. P. Inner city, middle-aged African Americans have excess frank and subclinical disability. J Gerontol A Biol Sci Med Sci 60, 207–212, doi:10.1093/gerona/60.2.207 (2005).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1093/gerona/60.2.207&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=15814864&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F01%2F11%2F2021.12.09.21267550.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000228283600013&link_type=ISI) 

15. Wilson, M. M. et al. Fear of falling and related activity restriction among middle-aged African Americans. J Gerontol A Biol Sci Med Sci 60, 355–360, doi:10.1093/gerona/60.3.355 (2005).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1093/gerona/60.3.355&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=15860474&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F01%2F11%2F2021.12.09.21267550.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000228880900015&link_type=ISI) 

16. Wolinsky, F. D., Miller, D. K., Andresen, E. M., Malmstrom, T. K. & Miller, J. P. Health-related quality of life in middle-aged African Americans. J Gerontol B Psychol Sci Soc Sci 59, S118–123, doi:10.1093/geronb/59.2.s118 (2004).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1093/geronb/59.2.S118&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=15014099&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F01%2F11%2F2021.12.09.21267550.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000220116600013&link_type=ISI) 

17. Cockrell, J. R. & Folstein, M. F. Mini-Mental State Examination (MMSE). Psychopharmacol Bull 24, 689–692 (1988).
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=3249771&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F01%2F11%2F2021.12.09.21267550.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=A1988T349900024&link_type=ISI) 

18. Canning, S. J., Leach, L., Stuss, D., Ngo, L. & Black, S. E. Diagnostic utility of abbreviated fluency measures in Alzheimer disease and vascular dementia. Neurology 62, 556–562, doi:10.1212/wnl.62.4.556 (2004).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1212/WNL.62.4.556&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=14981170&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F01%2F11%2F2021.12.09.21267550.atom) 

19. Anthony, J. C., LeResche, L., Niaz, U., von Korff, M. R. & Folstein, M. F. Limits of the ‘Mini-Mental State’ as a screening test for dementia and delirium among hospital patients. Psychol Med 12, 397–408, doi:10.1017/s0033291700046730 (1982).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1017/S0033291700046730&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=7100362&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F01%2F11%2F2021.12.09.21267550.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=A1982NQ66500021&link_type=ISI) 

20. Nguyen, A. D. et al. Serum progranulin levels are associated with frailty in middle-aged individuals. PLoS One 15, e0238877, doi:10.1371/journal.pone.0238877 (2020).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1371/journal.pone.0238877&link_type=DOI) 

21. Haren, M. T. et al. Predictors of serum testosterone and DHEAS in African-American men. Int J Androl 31, 50–59, doi:10.1111/j.1365-2605.2007.00757.x (2008).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1111/j.1365-2605.2007.00757.x&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=18190426&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F01%2F11%2F2021.12.09.21267550.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000252207600006&link_type=ISI) 

22. Cefalu, W. T., Bell-Farrow, A. D., Petty, M., Izlar, C. & Smith, J. A. Clinical validation of a second-generation fructosamine assay. Clin Chem 37, 1252–1256 (1991).
    
    [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6ODoiY2xpbmNoZW0iO3M6NToicmVzaWQiO3M6OToiMzcvNy8xMjUyIjtzOjQ6ImF0b20iO3M6NTA6Ii9tZWRyeGl2L2Vhcmx5LzIwMjMvMDEvMTEvMjAyMS4xMi4wOS4yMTI2NzU1MC5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 

23. Zhang, Y. et al. Purification and Characterization of Progenitor and Mature Human Astrocytes Reveals Transcriptional and Functional Differences with Mouse. Neuron 89, 37–53, doi:10.1016/j.neuron.2015.11.013 (2016).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.neuron.2015.11.013&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=26687838&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F01%2F11%2F2021.12.09.21267550.atom) 

24. Li, J. et al. Conservation and divergence of vulnerability and responses to stressors between human and mouse astrocytes. Nat Commun 12, 3958, doi:10.1038/s41467-021-24232-3 (2021).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41467-021-24232-3&link_type=DOI) 

25. Khalil, M. et al. Neurofilaments as biomarkers in neurological disorders. Nat Rev Neurol 14, 577–589, doi:10.1038/s41582-018-0058-z (2018).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41582-018-0058-z&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=30171200&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F01%2F11%2F2021.12.09.21267550.atom) 

26. Selvin, E. et al. Fructosamine and glycated albumin for risk stratification and prediction of incident diabetes and microvascular complications: a prospective cohort analysis of the Atherosclerosis Risk in Communities (ARIC) study. Lancet Diabetes Endocrinol 2, 279–288, doi:10.1016/S2213-8587(13)70199-2 (2014).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/S2213-8587(13)70199-2&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=24703046&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F01%2F11%2F2021.12.09.21267550.atom) 

27. Bonvento, G. & Bolanos, J. P. Astrocyte-neuron metabolic cooperation shapes brain activity. Cell Metab 33, 1546–1564, doi:10.1016/j.cmet.2021.07.006 (2021).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.cmet.2021.07.006&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=34348099&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F01%2F11%2F2021.12.09.21267550.atom) 

28. Ghoshal, S. et al. Adropin: An endocrine link between the biological clock and cholesterol homeostasis. Mol Metab 8, 51–64, doi:10.1016/j.molmet.2017.12.002 (2018).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.molmet.2017.12.002&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=29331507&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F01%2F11%2F2021.12.09.21267550.atom) 

29. Lovren, F. et al. Adropin is a novel regulator of endothelial function. Circulation 122, S185–192, doi:10.1161/CIRCULATIONAHA.109.931782 (2010).
    
    [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6MTQ6ImNpcmN1bGF0aW9uYWhhIjtzOjU6InJlc2lkIjtzOjE5OiIxMjIvMTFfc3VwcGxfMS9TMTg1IjtzOjQ6ImF0b20iO3M6NTA6Ii9tZWRyeGl2L2Vhcmx5LzIwMjMvMDEvMTEvMjAyMS4xMi4wOS4yMTI2NzU1MC5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 

30. Yang, C. et al. Therapeutic Benefits of Adropin in Aged Mice After Transient Ischemic Stroke via Reduction of Blood-Brain Barrier Damage. Stroke, doi:10.1161/STROKEAHA.122.039628. (2022).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1161/STROKEAHA.122.039628&link_type=DOI) 

31. Tomic, D., Shaw, J. E. & Magliano, D. J. The burden and risks of emerging complications of diabetes mellitus. Nat Rev Endocrinol, doi:10.1038/s41574-022-00690-7 (2022).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41574-022-00690-7&link_type=DOI) 

32. Fujie, S. et al. Aerobic Exercise Restores Aging-Associated Reductions in Arterial Adropin Levels and Improves Adropin-Induced Nitric Oxide-Dependent Vasorelaxation. J Am Heart Assoc, e020641, doi:10.1161/JAHA.120.020641 (2021).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1161/JAHA.120.020641&link_type=DOI) 

33. Kumar, K. G. et al. Identification of adropin as a secreted factor linking dietary macronutrient intake with energy homeostasis and lipid metabolism. Cell Metab 8, 468–481, doi:10.1016/j.cmet.2008.10.011. PMCID: 2746325 (2008).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.cmet.2008.10.011&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=19041763&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F01%2F11%2F2021.12.09.21267550.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000261379600006&link_type=ISI) 

34. Wong, C. M. et al. Adropin is a brain membrane-bound protein regulating physical activity via the NB-3/Notch signaling pathway in mice. J Biol Chem 289, 25976–25986, doi:10.1074/jbc.M114.576058 (2014).
    
    [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6MzoiamJjIjtzOjU6InJlc2lkIjtzOjEyOiIyODkvMzcvMjU5NzYiO3M6NDoiYXRvbSI7czo1MDoiL21lZHJ4aXYvZWFybHkvMjAyMy8wMS8xMS8yMDIxLjEyLjA5LjIxMjY3NTUwLmF0b20iO31zOjg6ImZyYWdtZW50IjtzOjA6IiI7fQ==) 

35. Banerjee, S. et al. Hepatocyte expression of the micropeptide adropin regulates the liver fasting response and is enhanced by caloric restriction. J Biol Chem 295, 13753–13768, doi:10.1074/jbc.RA120.014381 (2020).
    
    [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6MzoiamJjIjtzOjU6InJlc2lkIjtzOjEyOiIyOTUvNDAvMTM3NTMiO3M6NDoiYXRvbSI7czo1MDoiL21lZHJ4aXYvZWFybHkvMjAyMy8wMS8xMS8yMDIxLjEyLjA5LjIxMjY3NTUwLmF0b20iO31zOjg6ImZyYWdtZW50IjtzOjA6IiI7fQ==) 

36. Meex, R. C. R. & Watt, M. J. Hepatokines: linking nonalcoholic fatty liver disease and insulin resistance. Nat Rev Endocrinol 13, 509–520, doi:10.1038/nrendo.2017.56 (2017).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/nrendo.2017.56&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=28621339&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F01%2F11%2F2021.12.09.21267550.atom) 

37. Butler, A. A. et al. Differential Responses of Plasma Adropin Concentrations To Dietary Glucose or Fructose Consumption In Humans. Sci Rep 5, 14691, doi:10.1038/srep14691 (2015).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/srep14691&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=26435060&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2023%2F01%2F11%2F2021.12.09.21267550.atom) 

38. Formolo, D. A., Cheng, T., Yu, J., Kranz, G. S. & Yau, S. Y. Central Adiponectin Signaling -A Metabolic Regulator in Support of Brain Plasticity. Brain Plast 8, 79–96, doi:10.3233/BPL-220138 (2022).
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.3233/BPL-220138&link_type=DOI)