Abstract
Background:
Metabolic disorders, including insulin resistance, obesity, and hyperlipidemia occur frequently prior to hyperglycemia in patients with type 2 diabetes mellitus (T2DM) and cause mild cognitive impairment (MCI).
Objective:
We investigated the involvement of resistin in these metabolic abnormalities contributes to MCI in patients with T2DM.
Methods:
A total of 138 hospitalized patients with T2DM were enrolled and categorized into MCI and non-MCI groups according to the Montreal Cognitive Assessment (MoCA) score. Metabolic indicators and cognitive state were assessed, and plasma resistin levels were determined by ELISA.
Results:
The resistin levels and homeostasis model assessment of insulin resistance (HOMA-IR) scores of MCI and gender-stratified subgroups were significantly higher than those of controls without MCI (all p < 0.01). Correlation analysis showed that the resistin level was negatively associated with majority of cognitive domains, e.g., MoCA (r = –0.693, p < 0.001) and Mini-Mental State Examination (r = –0.571, p < 0.001), and was related to HOMA-IR (r = 0.667, p < 0.001) but not to obesity and lipid indices. Multivariable regression analysis indicated that resistin (β= –0.675, p < 0.001) and educational level (β= 0.177, p = 0.003) were independent risk factors of MoCA in patients with T2DM.
Conclusions:
High plasma resistin levels portend the insulin resistance-related susceptibility to early cognitive decline in Chinese patients with T2DM. The involvement of this adipokine in other metabolic disorders leading to diabetic MCI and its clinical value for early disease screening must be further studied.
INTRODUCTION
Type 2 diabetes mellitus (T2DM) is a worldwide epidemic with prevalence in China reaching 10.9% [1] and is an independent risk factor for Alzheimer’s disease (AD) [2] and conversion from mild cognitive impairment (MCI) to AD [3]. As a syndrome of cognitive impairment situated between normal aging and dementia, MCI is also the preclinical state of AD. Evaluation and early identification of MCI is beneficial for AD prevention, especially among patients with diabetes and high risk of AD.
Metabolic disorders, such as insulin resistance (IR) [4], obesity [5], and dyslipidemia [6] are involved in the progression of MCI in patients with diabetes, even before the disease onset. For example, IR usually occurs earlier than T2DM, thus promoting the chronic complications at early stages [7]. IR also impairs cognition and is an independent risk factor of MCI [8]. Dysfunction of insulin signaling may disrupt normal functions in regulating tau protein, amyloid-β, and amyloid- β protein precursor metabolism in neurons [9]. Moreover, some adipocytokines, e.g., adiponectin, leptin, and resistin, may play important roles in these situations [10].
Resistin is an adipocytokine produced by fat cells in mice and expressed mainly in macrophages in humans [11]. This cytokine regulates body weight, blood sugar, lipid levels, and systemic metabolism by modulating insulin sensitivity. The insulin action is improved when neutralizing antibodies consume the circulating resistin [11]. Resistin overexpression induces the development of IR and dyslipidemia [12]. Studies have associated elevated circulating resistin levels with increased risk for T2DM [13], obesity [14], and especially IR [15]. Clinical investigations have reported a positive correlation between resistin level and IR severity [16, 17], and resistin levels are remarkably high in people with prediabetes [18]. General consensus states that resistin affects insulin signaling by interfering with the activation of insulin receptor substrate (IRS)-1 and inducing the expression of the negative mediator of insulin signal cascade regulation [19]. Thus, resistin has an important relationship with IR.
In addition to its contribution to IR, resistin adversely influences brain functions through mechanisms, such as the promotion of inflammatory processes or oxidative stress [20–22]. Some observational studies have reported the negative effects of resistin on cognition [20, 23]. Altered levels of resistin in the central nervous system of obese mice may affect the carbohydrate metabolism of hippocampus, which is related to memory and learning [24], and these influences also represent a feature of MCI [25, 26]. Therefore, resistin is associated with cognitive impairment, especially MCI, and may be involved in the connection between IR and MCI.
Metabolic disorders, especially IR, increase the incidence of AD and MCI, and cognitive performance has a close correlation with resistin level. However, the links between resistin and metabolic imbalances and the early cognitive dysfunction in T2DM remain inconclusive. Comparison and observation of metabolic indicators in diabetic patients with or without MCI can reveal whether resistin is involved in the pathogenesis of MCI caused by metabolic abnormalities such as IR and whether this adipocytokine can be used as a clinical predictor or indicator of MCI in patients with diabetes.
MATERIALS AND METHODS
Study population
The study was conducted at the department of Endocrinology, ZhongDa Hospital, Southeast University. The experimental procedure was approved by the Research Ethics Committee of ZhongDa Hospital affiliated to Southeast University (Registration number: ChiCTR-OCC-15006060). For more information, please see http://www.chictr.org.cn/showprojen.aspx?proj=10536. All patients provided signed, informed consent before their participation in the research.
We recruited 138 hospitalized T2DM patients aged 40–80 years from July 2016 to July 2018. The diagnosis of T2DM was based on the World Health Organization 1999 standard. All the MCI patients satisfied the following diagnostic criteria proposed by the MCI Working Group of the European Consortium on AD: 1) cognitive complaints from family or themselves; 2) decline in cognitive capacity compared with that in the previous years; 3) evidence of cognitive disorder in clinical assessment; 4) no serious restrictions in activities of daily living; and 5) symptoms that had not received the level of dementia [27]. The MoCA score of MCI is less than 26, and plus one point for those with less than 12 years of education [28]. The control groups without MCI were matched for sex, age, weight, height, and educational level. Exclusion criteria were as follows: head trauma, epilepsy, a prior history of stroke, alcohol abuse, severe depression, Parkinson’s disease, central nervous system infection, or other mental or neurological disorders; major medical diseases (e.g., acute cardiovascular disease, cancer, and thyroid functional abnormality) and use of drugs that may affect recognition tests; visual or hearing impairment; serious infection; and other severe acute complications such as diabetic ketoacidosis and hyperosmolar coma.
Collection of clinical data
Demographic variables were recorded, including gender, age, ethnicity, profession, educational level, physical measurements: weight, height, blood pressure, waist circumference (WC), hip circumference (HC), waist-to-hip ratio (WHR; calculated as WC/HC). Body mass index (BMI) was defined as body weight (kg)/ body height2 (m2). History of diabetes and drinking and smoking status were collected. Blood samples were acquired to assess glycated hemoglobin (HbA1c), fasting blood glucose (FBG), 2 h post-meal blood glucose (2 hPG), fasting serum C-peptide (FCP), apolipoprotein A1 (ApoA1), apolipoprotein B (ApoB), triglyceride (TG), total cholesterol (TC), high density lipoprotein (HDL), low density lipoprotein (LDL). HOMA-IR was calculated by fasting serum C-peptide: FBG (mmol/L)×FCP (nmol/L)/22.5 [29]. The continuous glucose monitoring system recorded the largest amplitude of glycemic excursions (LAGE), which was calculated as the fluctuation of blood glucose from peak to valley. The level of education was expressed in terms of years of education and the diagnosis of fatty liver was subject to the report of B-ultrasound. The inspection center of Zhongda Hospital Affiliated to Southeast University implements internal and external quality management procedures directed by the Chinese Laboratory Quality Control.
Neuropsychological tests
All patients received neuropsychological tests, including the general state of mind, memory, attention, executive ability, and visual-spatial function: Montreal Cognitive Assessment (MoCA), Mini-Mental State Exam (MMSE), digit span test (DST), clock drawing test (CDT), logical memory test (LMT), verbal fluency test (VFT), auditory verbal learning test (AVLT), Trail Making Test-A and B (TMT-A and TMT-B), and Stroop color word test (SCWT). The tests were completed by an experienced neuropsychiatrist and all the subjects and the neuropsychiatrist were blinded to the study design.
Measurement of plasma resistin
The 2 mL blood samples were collected at 8 : 00 am in the fasted state and separated using centrifugation at 3000 rpm/10 min. Plasma samples were gathered and freeze in a refrigerator at minus 80°C. Plasma resistin levels were determined with ELISA kits (Biovendor, Brno, Czech Republic). The variation coefficients of intra and interassay were 5.9% and 7.6%, respectively.
Statistical analysis
SPSS 19 (SPSS Inc., Chicago, IL, USA) was used for statistical analysis. Data were showed as mean±standard deviation, median (interquartile range), or percentage. Student’s t test and analysis of variance were used to analyze variables with normally distributed, and non-parametric Mann-Whitney U test was conducted to compare the differences in variables with asymmetrically distributed. The Chi-squared (χ2) test was employed for comparison between non-continuous variables. The correlativities between resistin levels and parameters in clinical and cognitive were assessed by partial correlation analysis. A multivariable regression model was used to identify risk factors for cognitive function in both demographic and clinical factors. Thresholds of statistical significance were set at a corrected two-sided p < 0.05.
RESULTS
Demographic characteristics, clinical characteristics, and cognitive performance
A total of 138 right-handed hospitalized patients with T2DM were recruited, and their baseline characteristics are shown in Table 1. Patients with T2DM and MCI had elevated HbAlc levels (p < 0.05) and significantly low scores in most neuropsychological tests (p < 0.05). The values of HOMA-IR in MCI group were significantly higher than those in group without MCI (p < 0.01). Stratified analysis of gender revealed that men and women showed similar trends, but the difference was statistically significant only in female patients (p < 0.01). The MCI group exhibited significantly higher levels of resistin than the non-MCI group (p < 0.01). The remarkable differences in plasma resistin levels were found in both gender subgroups (p < 0.01).
Demographic, clinical, and cognitive characteristics in all patients with diabetes
Data were presented as n (%), mean±SD, or median (interquartile range). p values were based on Student’s t-test, Mann-Whitney U-test, or Chi-square test for variables and were compared with control. p < 0.05 was considered as a significant difference. p1, MCI versus non-MCI; p2, MCI versus non-MCI in male; p3, MCI versus non-MCI in female. MCI, mild cognitive impairment; WC, Waist-circumference; HC, Hip-circumference; BMI, body mass index; WHR, Waist-to-Hip Ratio; FBG, fasting blood-glucose; 2 h-PG, 2-hour post-meal blood glucose; LAGE, large maximum amplitude of glycemic excursion; HbA1c, glycosylated hemoglobin; FCP, fasting C-peptide; HOMA-IR, homeostasis model of assessment for insulin resistance; LDL, low density lipoprotein; HDL, high density lipoprotein; MoCA, Montreal Cognitive Assessment; MMSE, Mini-Mental State Examination; LMT, Logical Memory Test; AVLT, Auditory Verbal Learning Test; VFT, Verbal Fluency Test; DST, Digit Span Test; CDT, Clock Drawing Test; TMT, Trail Making Test; SCWT, Stroop Color Word Test.
Relationship between plasma resistin levels and clinical indicators
Partial correlation analysis was conducted. After adjustment for gender, age, and educational status, a remarkable negative correlation was found for resistin concentrations and MoCA (r = –0.693, p < 0.001), MMSE (r = –0.571, p < 0.001), LMT (r = –0.371, p < 0.001), VFT (r = –0.368, p < 0.001), DST (r = –0.373, p < 0.001), CDT (r = –0.225, p = 0.009), AVLT-delayed recall (r = –0.301, p < 0.001), AVLT-immediate recall (r = –0.368, p < 0.001), and SCWT-A (number) (r = –0.205, p = 0.018). On the contrary, a significant positive correlation was observed for resistin concentrations and TMT-A (r = 0.341, p < 0.001), TMT-B (r = 0.447, p < 0.001), and SCWT-C (time) (r = 0.421, p < 0.001) (Table 2). Correlation analysis also showed that resistin level was positively related to HOMA-IR score (r = 0.667, p < 0.001) but not with most obesity indices and blood lipid levels (p > 0.05, Table 2).
Correlations of plasma resistin with clinical and cognitive indicators
aPartial correlation analysis adjusted for age, sex, and educational levels. *p < 0.05, **p < 0.01. WHR, Waist-to-Hip Ratio; BMI, body mass index; HbA1c, glycosylated hemoglobin; HOMA-IR, homeostasis model of assessment for insulin resistance; HDL, high density lipoprotein; LDL, low density lipoprotein; MoCA, Montreal Cognitive Assessment; MMSE, Mini-Mental State Examination; LMT, Logical Memory Test; VFT, Verbal Fluency Test; DST, Digit Span Test; CDT, Clock Drawing Test; TMT, Trail Making Test; AVLT, Auditory Verbal Learning Test; SCWT, Stroop Color Word Test.
Simple linear regression in patients with T2DM
A simple linear regression model was established, and all variables with the exception of HOMA-IR scores in Table 1 were included. When MoCA score was considered as a dependent variable, BMI (B = –0.185, p = 0.045), educational level (B = 0.382, p = 0.001), HbA1c level (B = –0.272, p = 0.038), and resistin level (B = –0.313, p < 0.001) were finally introduced to this model. When MMSE score was considered as a dependent variable, educational level (B = 0.237, p = 0.002) and resistin (B = –0.168, p < 0.001) were included in the model (Table 3). This finding revealed that resistin level, educational level, BMI, and HbA1c level were risk factors of MCI in patients with diabetes.
Assessment results of the risk of MCI in a simple linear regression model
*p < 0.05; **p < 0.01. B, regression coefficient; SE, standard error; MoCA, Montreal Cognitive Assessment; MMSE, Mini-Mental State Examination; BMI, body mass index; HbA1c, glycosylated hemoglobin.
Multivariable regression analysis in patients with T2DM
A multivariable linear regression model was used to illuminate independent risk factors that were relevant for MCI. The results showed that the MoCA score was the dependent variable and remarkably positively correlated with educational level (β= 0.177, p = 0.003) but negatively correlated with resistin level (β= –0.675, p < 0.001) (Table 4). These results were similar when MMSE was used as the dependent variable (Table 4).
Assessment results of the risk of having MCI in a multivariable linear regression model
*p < 0.05; **p < 0.01. B, regression coefficient; SE, standard error; MoCA, Montreal Cognitive Assessment; MMSE, Mini-Mental State Examination.
DISCUSSION
Our study suggests that patients with T2DM and MCI displayed remarkably high plasma levels of resistin and HOMA-IR scores, and similar results were obtained for both genders. In addition to old age and low educational attainment, high resistin level was also negatively correlated with most cognitive domains. This adipocytokine was related to IR but not with most obesity indices and blood lipid levels. Thus, resistin was an independent risk factor for MoCA and MMSE scores. High plasma levels of resistin were associated with IR-associated MCI in Chinese patients with T2DM.
Resistin is closely related to cognitive impairment. Patients with AD have elevated resistin levels compared with patients without AD [30, 31]. Altered levels of resistin in the central nervous system of obese mice affect the hippocampal metabolic glucose levels, which are connected with memory and learning capabilities [24] and also represent a feature of MCI [25, 26]. We reached a similar conclusion that high resistin level was negatively correlated with most cognitive domains. On the contrary, some studies have presented different conclusions. Liu found that resistin has a protective effect on AD in vitro studies [32]. Another work confirmed that resistin exerts neuroprotection on cell apoptosis induced by 6-OHDA [33]. Although the reasons for these differences remain to be further analyzed, we suggest that the studies showing the neuroprotection of resistin were limited to in vitro models, whereas population studies did not show similar results. The human internal environment is highly complex that resistin and other factors can simultaneously interact with and restrict each other. Therefore, further study must be conducted to determine whether resistin damages cognitive function or is associated with neuroprotection in cognitive impairment.
IR is one of the most distinctive characteristics of T2DM and emerges prior to dominant T2DM. HOMA-IR, an indicator of insulin sensitivity, is an important indicator for T2DM [34]. Another published research declared that IR is associated with increased AD risk [2] and plays a vital role in the early phase of AD [8]. Our research found that HOMA-IR score in the MCI group was higher than that in the patients without MCI. Although not statistically significant, patients with cognitive impairment were more likely to present abdominal obesity than patients without MCI, indicating a possible increase in visceral fat mass. In addition, MoCA and MMSE scores were negatively correlated with resistin levels, indicating the role of this cytokine in cognitive function impairment.
Consistent with previous studies [16, 17], we found that resistin and IR are closely related. Therefore, we only included resistin as the independent variable in the multivariate regression analysis and did not include HOMA-IR scores due to the strong interference. IR, resistin, and cognitive function have a complex relationship, and HOMA-IR changes may be related to resistin, leading to cognitive impairment. Although the mechanisms underlying this relationship remain a matter of debate, general consensus states that resistin affects insulin signaling by interfering with IRS-1 activation [19]. Emanuelli found that suppressor of cytokine signaling-3, a negative mediator of insulin signal cascade regulation, could be induced by resistin [35]. On the contrary, resistin and the insulin resistance it induces contribute to cognitive impairment [30, 36].
Although resistin can cross over blood-brain barrier, its concentration in cerebrospinal fluid is lower than that in plasma [37, 38]. Hence, resistin levels in peripheral blood may not play an important role in cognitive functions. On the contrary, the pro-inflammatory properties of resistin are remarkable because it can activate inflammatory processes by promoting TNF and IL-6 [39]. High TNF and IL-6 levels are associated with cognitive impairment [40]. In addition, resistin is expressed in the cortex, hypothalamus, and hippocampus to inactivate hypothalamic neurons [41–43]. Thus, we consider that resistin can also cause cognitive impairment through inflammation or inflammation-related insulin resistance. Additional research is needed to explore the underlying molecular mechanisms.
Certain limitations of this study should be noted. First, the study is at preliminary stage, and all participants were Chinese, thus limiting the generalizability of our results. The small sample size may also limit the reproducibility of the findings. Second, healthy volunteers or other patients without T2DM were not enrolled, and most patients are non-obese based on BMI but not on abdominal obesity. Third, data from neuroimaging examination, the pathological changes of MCI in T2DM caused by resistin, and resistin levels in the adipose tissues were not yet obtained. Finally, we did not collect information regarding body fat, C-reactive protein, and other parameters that may also play an important role in the relationship between resistin and MCI. Thus, further studies with large and heterogenous samples, additional related parameters, and different weight statuses are needed for the comprehensive understanding of resistin role in diabetes-related cognitive impairment.
Conclusions
The non-MCI group had lower plasma resistin level and HOMA-IR score than the MCI group. High resistin levels and HOMA-IR scores were associated with impaired cognitive function. Therefore, resistin was an independent variable of MCI in all individuals, suggesting that high plasma resistin levels may be good biomarkers of MCI in patients with T2DM. Further research on large populations is required to corroborate this finding and determine whether plasma resistin can be used as an important marker for the early diagnosis of MCI in patients with diabetes.
Footnotes
ACKNOWLEDGMENTS
This work was partially supported by the National Natural Science Foundation of China (No. 81570732, Shaohua Wang; and No. 81870568, Shaohua Wang). We would like to express our heartfelt gratitude to all participants and the staff of the department of Endocrinology, Affiliated ZhongDa Hospital of Southeast University.
