ABSTRACT
Recent data implicate the secreted peptide adropin in the physiology of aging. In humans, adropin is highly expressed in brain tissues, and correlates positively with transcriptomic signatures of mitochondrial and synaptic functions. Adropin treatment improves performance of old mice in cognitive tests requiring learning and memory. While detected in the circulation of humans, no studies have investigated relationships between adropin, cognitive decline and aging. Here we compared serum adropin concentrations with Mini-Mental State Exam (MMSE) and animal naming test results from a cohort study of African Americans in late-middle age (baseline ages 49-65y, n=357). Using the lowest quintile of the MMSE to identify participants at risk for mild cognitive impairment (MCI) indicated lower serum concentrations (2.95±1.32 ng/ml vs. 3.31±1.56, P<0.05). Grouping into bins using 1-ng/ml increments in serum adropin concentrations further indicated an association between very low serum adropin concentrations and MCI. Using fructosamine as an indicator of moderate-term glucose levels suggested low serum adropin concentration correlate with increased risk of 10-year all-cause mortality in situations of poor glucose control. In summary, these data suggest low circulating adropin concentrations identify late-middle aged people at risk for cognitive impairment, and for all-cause mortality in situations of poor glucose control.
INTRODUCTION
Declining cognitive tasks involving information processing, working memory, flexibility in thinking, and self-control is considered a normal outcome of aging that can be accelerated by poorly controlled diabetes (1; 2). Impaired memory and self-management of daily activities noticeable to the patient and caregivers but insufficient to require managed care is defined as mild-cognitive impairment (MCI) (3; 4). MCI appears to represent the early stages of late-age onset dementias (LOAD), of which Alzheimer’s disease is most prevalent(3; 4). An aging population indicates that managed care of people with LOAD will be a major health care challenge in the 21st century. Treatment options for preventing LOAD or delaying its progression are however limited and controversial (5). Identifying biomarkers that identify elderly individuals at risk of LOAD and developing new treatment strategies to prevent or delay onset remains as a critical unmet need.
Adropin is a pleiotropic secreted peptide with endocrine, autocrine/paracrine and neuropeptide functions(6; 7). Adropin was originally identified as a hepatokine mediating homeostatic responses to variation in dietary macronutrient consumption (8). Subsequent studies suggest it is a stress-responsive hormone regulating glucose metabolism and insulin sensitivity in the liver (9-14), skeletal muscle (15; 16), and cardiac muscle (17-19). Circulating adropin concentrations in humans and nonhuman primates correlate with indices of glucose and cholesterol metabolism (20-24). Circulating adropin concentrations in nonhuman primates also correlate with endocrine and metabolic responses to an obesogenic (high sugar) diet (24). Adropin affects vascular tone by regulating endothelial nitric oxide synthase activity (25-27), and protecting against injury to the endothelium (28). Correlations with vascular endothelial function in humans have also been reported (26; 29; 30). These results suggest adropin is an endocrine link between cellular stress responses and whole-body glucose and vascular homeostasis.
In humans, RNAseq datasets suggest adropin expression is most abundant in astrocytes residing in the central nervous system(24; 31). An emerging literature indicates a role for adropin in neurological aging. Brain adropin expression correlates with gene networks implicated in mitochondrial function, cellular lipid metabolism, inflammation, synaptic plasticity, and vascular function(31). Adropin regulates the blood-brain barrier permeability and may preserve integrity during injury(32-34). Positive relationships are also observed with protein markers of tau-related neuropathology and inflammation in the old-old (age >75y), particularly in people who died without a dementia diagnosis(31). Brain adropin expression also exhibits an inverse relationship with cholesterol synthesis pathways(31). This is an intriguing observation, as plasma adropin and cholesterol concentrations are also negatively correlated(23). This relationship is thought to result from intermediates in cholesterol metabolism inhibiting adropin expression(23).
These physiological processes linked to adropin expression in the nervous system all relate to risk of MCI and LOAD(35; 36). Indeed, gene ontology analysis indicates relationships between expression of the adropin transcript and gene networks associated with neurodegenerative diseases in humans, nonhuman primates, and mice(24; 31; 37). In preclinical mouse models of aging, adropin treatments have a positive impact on cognitive function. Transgenic over expression of adropin in male 18-month-old C57BL/6J (B6) mice, which corresponds to a human age of 56-69y(38), improves cognitive abilities related to spatial learning and memory and novel object recognition(31). Treatment of old B6 mice with synthetic peptide also reverses mild memory impairment observed in middle-aged mice(31). Adropin could thus have clinical potential both as a biomarker of brain health and as a lead for developing therapies against cognitive decline.
No studies on the relationship between serum adropin concentrations and aging have been reported. Here we analyzed serum adropin concentrations in the African American Health (AAH) study. AAH is a prospective analysis of disability and frailty amongst urban-dwelling African Americans over 10y (39). This ancillary study tested the hypothesis that low circulating levels of adropin indicate increased risk of MCI observed with aging. A secondary objective examined relationships between baseline adropin levels and indices of metabolic dysregulation.
RESEARCH DESIGN AND METHODS
Study participants
All study procedures were approved by the Institutional Review Board at Saint Louis University and have therefore been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. The African American Health cohort has been previously described (39; 40). Briefly, the study involved a population-based cohort of 998 self-identified African American individuals aged from 49 to 65 years at baseline. The study began in 2000–2001, and recruited residents in the St. Louis, Missouri metropolitan area. Participants completed in-home interviews and assessments at baseline followed by follow-ups at year 3 and year 9. Laboratory analyses used blood drawn for a subset of participants shortly after the in-home assessment, or during clinical examinations at baseline and 3-year follow-up. The final analytic sample for the current study includes 357 participants for which serum adropin concentrations were measured.
MMSE and ANT tests
The MMSE is used to detect cognitive deficits observed with dementia and delirium, and to measure changes in cognitive function over time(41). Scores of ≤23 are 87% sensitive and 82% specific in detecting dementia and delirium using a psychiatrist’s standardized clinical diagnosis as the criterion(42). One-minute naming tests such as the ANT assess memory recall and verbal fluency; an inability to name 15 or more suggests cognitive impairment and higher risk of dementia(43).
Laboratory procedures
Serum adropin concentrations were measured using an enzyme immunoassay kit from Phoenix Pharmaceuticals, Inc. (cat. no. EK-032-35) per manufacturers recommended protocol. Assays for the other serum measurements were described previously(44; 45).
Statistical analysis
Data were analyzed using IBM SPSS Statistics version 19.0 (IBM Corp., Somers, NY). Adropin tertiles were defined by ranking all individuals by serum adropin low to high. Differences in the incidence of poor glucose control and mortality between adropin tertiles were assessed using chi-square tests. To adjust survival for age and sex, a value of ‘1’ was used to indicate death and ‘0’ indicated survival during the study (e.g., 40 deaths in a group of 100 would produce a mean of 0.4, indicating 40% mortality). A similar approach was used to calculate the prevalence of poorly controlled glucose, with fructosamine values >287 mmol/L assigned a value of ‘1’ and normal values ‘0’. For the analysis of the interaction between adropin and glucose control and 10-year mortality, fructosamine levels (high, normal) and adropin tertile were used as fixed variables. Estimated means were calculated adjusting for BMI, age, and sex as indicated.
Data and resource availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author upon reasonable request.
RESULTS
AAH is comprised of “late middle-age” individuals with above average risk of depressive symptoms and disability(39; 40; 46). Age, sex, and body mass index (BMI) of the 357 participates for which serum adropin concentrations were measured are shown in Table 1.
Serum adropin concentrations in AAH study participants scoring in the lowest quintile for the MMSE at baseline compared to the rest of the cohort. Data shown are mean±std. dev and range (lowest value, highest value). N.S., not significantly different.
Lower serum adropin concentrations in people with MCI
The primary hypothesis is that low circulating adropin levels indicate increased risk of MCI. To address the hypothesis, baseline serum adropin concentrations were compared with baseline results from Mini-Mental State Exam (MMSE). Scores in the lowest quintile (MMSE_low_quint) were used to define study participants with poor cognitive function (Fig. 1).
Distribution of MMSE and ANT scores are shown (A, B). Sample size by sex (C) and serum adropin concentrations (D) for the cohort separated into 1ng/ml bins. M, male; F, female.
The MMSE score for the MMSE_low_quint group ranged from 16 to 26, compared to 27-30 for the other study participants (Fig. 1A). Serum adropin concentrations were significantly lower in the MMSE_low_quint group compared to the rest of the cohort (Table 1). Sex-differences in circulating adropin concentrations have been reported previously(23). A 2-way ANOVA (fixed variables: sex, MMSE_low_quint) again identified a strong trend for effect of the latter (estimated marginal mean in ng/ml and SEM of serum adropin concentrations for MMSE_low_quint, 2.94±0.19, 95%CI 2.575 to 3.30; compared to 3.34±0.10, 95% CI 3.146 to 3.524, P=0.059), but no significant effect of sex (females, 3.14±0.13, 95%CI 2.89 to 3.38; males, 3.14±0.17, 95%CI 2.81 to 3.47), with no interaction between sex and MMSE grouping.
These results were compared with outcomes from a second test of cognitive performance, the animal naming test (ANT). Participants in the lowest quintile (ANT_low_quint) were unable to name more than 13 animals (mean±SD and 95%CI; 11±2 (4-13), range 4-13, n=51) when compared to the upper other participants (21±5 (14-41), range 14-41, n=291, P<0.001 compared to ANT_low_quint). This group also exhibited a trend (p=0.081) for lower serum adropin concentrations (2.97±1.35, (0.74-6.75) vs. 3.28±1.50, (0.03-10.23).
Relationships between circulating adropin levels and indices known to signal risk of metabolic dysregulation (for example, age, obesity, fasting glucose, insulin, leptin, adiponectin, triglycerides) have been reported previously (20-24). Indices of metabolic homeostasis for which data are available (body mass index, age, adiponectin, leptin, and fructosamine) were not significantly different (Table 1). Adjusting for age and BMI in the analysis also did not change the outcome of the 2-way ANOVA. There was no evidence for metabolic dysregulation or inflammation in the ANT_low_quint group (data not shown).
We next used two approaches to further investigate whether high serum concentrations correlate with improved test scores in the MMSE and ANT. First, we ranked participants into tertiles by serum adropin concentrations (low, mid, high) (Table 2). Second, we separated participants into bins defined by 1ng/ml increments in serum adropin concentrations (Fig. 1C, D). An advantage of using tertiles is a larger and consistent sample size (n=119/tertile). However, using tertiles could miss non-linear relationships between serum concentrations and test scores observed with extremely low (or high) values.
Indices of MCI and metabolic homeostasis in study participants separated in tertiles ranked by serum adropin concentrations. Data shown are estimated marginal means ± std. error (95% confidence interval).
Mean values for MMSE and ANT scores were not significantly different between tertiles (data not shown). Differences in baseline adropin concentrations between tertile also did not predict 10y changes in MMSE or ANT scores (data not shown). However, the distribution of MMSE_low_quint and ANT_low_quint scores between tertiles was consistent with high adropin levels reducing risk of cognitive impairment at baseline (Table 2). The highest tertile group exhibit 30% fewer individuals in the low_quint scores. This group also had higher adiponectin and fructosamine, but lower triglycerides, compared to the lowest tertile (Table 2).
Serum adropin concentrations correlated with serum concentrations of adiponectin (positive), fructosamine (positive), and triglycerides (negative) (Table 2). Serum fructosamine concentrations are an independent predictor of incident diabetes and diabetes-related microvascular complications; levels exceeding 287 μmol/L indicate poorly controlled diabetes(47-49). Poorly control glucose homeostasis is also a strong predisposing risk factor for LOAD (2). In the current study participants with fructosamine values exceeding 287 had significantly higher serum adropin concentrations (3.59±1.8, 95%CI 3.23-3.94 ng/ml vs. 3.16±0.09, 2.99 to 3.34 ng/ml, p<0.036, n=69 vs. n=285).
Exploring the relationship between MMSE and ANT test scores and serum adropin concentration and bins in 1 ng/ml further indicated the positive relationship between serum adropin and cognitive performance. Performance in the MMSE was clearly compromised in participants with the lowest serum adropin values (effect of adropin bin for MMSE, P<0.05) (Fig. 2A). Adjusting for BMI and age did not affect the relationship (Fig. 2B). A similar relationship was observed between adropin bin and ANT scores (Fig. 2C, D) (P=0.05). Analysis of other risk factors (indices of inflammation, metabolic homeostasis, sex steroids) failed to identify a defining variable (e.g., higher inflammation) explaining the correlation (Fig. 3).
Low serum adropin concentrations correlate with cognitive impairment in study participants grouped into bins of 1 ng/ml increments in serum adropin concentration. (A) Box and whisker plots of MMSE scores (B) MMSE scores adjusted for age and BMI plotted against serum adropin concentrations. Significance from 0 to 1 ng/ml bin determined by pairwise comparisons indicated by *** (p<0.005), * (p<0.05); (C) Box and whisker plots of ANT scores. (D) ANT scores adjusted for age and BMI plotted against serum adropin concentrations between bins. *** P<0.005 vs. 1 ng/ml bin; * <0.05 vs. 1 ng/ml bin; # P<0.05 vs. 3-4 ng/ml bin.
Serum markers of inflammation and tissue injury (A), gonadotropic axis (B), and metabolic homeostasis (C) for study participants separated into 1ng/ml bins. The data shown are z scores.
Adropin predicts 10y mortality risk in late middle-aged people with poor glucose control
Preclinical experiments using preclinical models suggest high adropin levels could be protective of cognitive decline in situations of metabolic dysregulation(8-10; 15; 24; 31; 37). Here we observed an interaction between serum adropin concentrations and risk of all-cause mortality in study participants with poorly controlled glucose levels, indicated by fructosamine levels exceeding 287 μmol/L (n=69). The distribution of study participants with poorly controlled glucose appeared to be skewed between adropin tertiles (Fig. 4E) and was less frequently observed in the low tertile (17 of 119 participants) when compared to the high tertile (29 of 119). Incidence in study participants in the mid-tertile range was intermediate (23/119). However, risk of all-cause mortality exhibited the opposite relationship. Risk of all-cause mortality was increased 2-fold in participants with fructosamine values >287 μmol/L at baseline (Fig. 4A, B). While the prevalence of poorly controlled glucose homeostasis was increased in the tertile with high serum concentrations (Fig. 4E, Table 2). However, the relative risk of all-cause mortality was lower in this group compared to the lower adropin tertiles (Fig. 4E, F) (Table 3). Risk of all-cause mortality was comparable between adropin tertiles in the whole cohort (Fig. 4C, D).
Increased risk of 10-year all-cause mortality in study participants with poorly controlled glucose (A, B) correlates with baseline serum adropin concentrations (C-F). Data shown are actual values (A, C, E) and %total (B, D, F). Risk of mortality is two-fold higher with dysregulated glucose metabolism (A, B), indicated by fructosamine values >287 μmol/L. Mortality risk is similar between adropin tertiles in all participants (C, D), and in participants with normal fructosamine (E, F). However, while the incidence of poor glucose control increases with higher adropin levels, the risk of all-cause mortality declines. (E, F, P<0.005 using chi-square test). The dashed red reference line shows mortality risk for individuals with normal glucose control. Samples sizes are reported in Table 3.
Interaction between serum adropin concentrations and glucose regulation in 10-year mortality. Data show estimates of %mortality (10y) in individuals with poorly controlled glucose homeostasis (fructosamine levels >287 μmol/L, n=69) compared to individuals with fructosamine levels <287 (n=285) during the AAH study. During the study, 97 participants died and 257 survived. Fructosamine data was not available for 3 study participants. Sample sizes and number of females (F) are indicated. Ranking for adropin tertile was based on the whole cohort.
DISCUSSION
The two major findings of this study indicate serum adropin concentrations are relevant to human aging. First, low circulating adropin concentrations correlate with increased risk of mild cognitive impairment in a late-middle aged cohort. Second, low circulating adropin concentrations could indicate increased risk for all-cause mortality for late-middle aged individuals with poorly controlled glucose.
A limitation of the current study is that it samples from a single geographic location and only included African Americans. However, if either or both relationships are observed in the general population, they could represent an important new metric for the clinical management of an aging population. Moreover, either or both could be amenable to pharmacological interventions using synthetic peptide mimetics to reverse a deficit in adropin activity.
Another limitation was that the AAH study used an MMSE of <15 as an exclusion criterion. The study population therefore excluded late-middle aged people with dementia. Future studies are needed to explore the relationships between circulating adropin concentrations and dementia. These studies could also screen for individuals with extremely low circulating adropin concentrations. The current study suggests that measuring circulating adropin concentrations could be a useful tool in characterizing cognitive functions in the elderly. However, further research is needed to characterize the individuals with very low levels. The use of circulating adropin concentration in defining subcategories of dementia, dementia diagnosis, and risk remains to be explored.
A subset of the cohort had high fructosamine levels indicating poorly controlled glucose levels. In these individuals, low circulating adropin concentrations appeared to predict increased risk of all-cause mortality during the 10y study period. Given the small sample size, these findings should be viewed as preliminary. Further studies using larger sample sizes and cohorts that are more representative of the general population are needed. The study design should also allow for identifying specific causes of mortality indicated by low adropin levels.
Adropin could potentially affect mortality risk in situations of hyperglycemia through mechanisms involving vascular or metabolic homeostasis. Cardiovascular disease is the leading cause of death in people with elevated fasting glucose, although risk of microvascular complications is also increased(50). An emerging literature suggests adropin regulates vascular endothelial function (25; 27; 32). Preclinical studies using mouse models also suggest adropin protects against stroke (34). On the other hand, adropin does not prevent the development of atherosclerosis in mouse models (23). Metabolically, adropin regulates the activities of the principal pancreatic hormones that control the fate of glucose and fatty acids. In the liver, adropin enhances insulin signaling (10; 11) while suppressing glucagon signaling (9). Adropin also regulates insulin signaling and glucose metabolism in skeletal and cardiac muscle (15-17; 19). Clearly, further studies dissecting the association between circulating adropin concentrations and mortality risk in in situations of impaired glucose control are needed.
Further studies are also needed to determine whether circulating concentrations are a useful biomarker to identify middle-aged individuals at risk for LOAD. Investigation of the relationships between circulating adropin concentrations and cognitive functions in other age groups might also be informative. For example, adropin expression in the human brain peaks in the first decade of life, a critical period in brain development (31).
In the current study, serum adropin concentrations were elevated in people with poorly controlled glucose. This result is contradictory, as previous cross-sectional studies comparing groups with or without poorly control glucose reported the opposite(6). Furthermore, in a nonhuman primate model low plasma adropin concentrations correlated with fasting hyperglycemia (24). This discrepancy cannot be explained using the available data. However, it is worth noting that in humans the consumption of fructose beverages and associated insulin resistance increases plasma adropin concentrations, and that the magnitude of the response correlates with pre-existing markers of metabolic dysregulation (21). This observation could imply that circulating adropin levels can show an increase in the early stages of metabolic disease. Data from the current study also suggests that higher circulating adropin levels may protect individuals with poorly controlled glucose homeostasis from mortality and perhaps adverse effects.
We have observed sex-specific differences in plasma adropin concentrations (23). The absence of sex-effect on serum adropin concentrations in the current study could be due to circulating adropin levels declining with age in men (23). On the other hand, other explanations are possible. These include differences in ethnicity, age, health, and socioeconomic status.
In summary, the results of the current study indicate that low circulating adropin concentrations in late middle-aged people associate with poor cognitive function, and for mortality risk in late-middle aged people with poorly controlled glucose. Further investigation is clearly indicated to determine whether the correlations observed in this cohort apply to the general population.
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). AAB and ADN acknowledge financial support provided by Saint Louis University. African American Health was supported by a grant from the National Institute on Aging (R01 AG010436 to DKM).
Footnotes
Abstract edited for clarity. Main text edited to ensure conformity in description of elevated fructosamine levels.
REFERENCES
