Abstract
Background:
Plasminogen activator inhibitor 1 (PAI-1) and tissue plasminogen activator (tPA) are involved in the complications of type 2 diabetes mellitus (T2DM) and early pathology of Alzheimer’s disease.
Objective:
This study aimed to investigate the association between plasma PAI-1, tPA/PAI-1 molar ratio, and mild cognitive impairment (MCI) in Chinese T2DM patients.
Methods:
A total of 162 Chinese T2DM patients were recruited and divided into two groups according to the Montreal Cognitive Assessment score. Demographic data were collected, plasma PAI-1 and tPA levels were measured through enzyme-linked immunosorbent assay, tPA/PAI-1 molar ratio was calculated, and neuropsychological test results were examined. The association between PAI-1, tPA/PAI-1 molar ratio, and cognition was analyzed.
Results:
There were 66 diabetic MCI patients and 96 healthy cognition participants (controls). T2DM patients with MCI displayed significantly increased plasma PAI-1 levels (p = 0.016) and decreased tPA/PAI-1 molar ratio (p = 0.021) compared with the controls. High PAI-1 levels and low tPA/PAI-1 molar ratio were associated with MCI in T2DM patients, e.g., plasma level of PAI-1 were negatively correlated (r = –0.343, p = 0.007) with logic memory in T2DM patients with MCI. Linear regression analysis further revealed that PAI-1 concentration was an independent factor of diabetic MCI (p = 0.001).
Conclusions:
High PAI-1 levels and low tPA/PAI-1 molar ratio were significantly correlated with T2DM-associated cognitive impairment, especially memory function, in Chinese patients.
INTRODUCTION
Long-term hyperglycemia can lead to cognitive impairment [1]. The contribution of type 2 diabetes mellitus (T2DM) in cognitive dysfunction is multifactorial. It includes amyloid-β (Aβ) deposition [2], tau protein hyperphosphorylation [3], and damage to the insulin [4] and inflammatory signaling [5] pathways. However, the underlying mechanisms are complex and unclear. Studies have found that fibrinolytic function abnormalities [6] are prevalent in diabetic patients, and fibrinolytic system abnormalities may be involved in Aβ deposition and carotid plaque formation [7, 8], leading to cognitive dysfunction.
Plasminogen activator inhibitor 1 (PAI-1) is synthesized by vascular endothelial cells, hepatocytes, and smooth muscle cells. PAI-1 is closely related to T2DM [9 –11]. Studies [12 –15] have shown that PAI-1 is associated to the complications of diabetes mellitus and cognitive impairment [16]. Suton et al. [16] confirmed that PAI-1 levels in the cerebrospinal fluid are elevated in Alzheimer’s disease (AD) patients and that plasma PAI-1 levels gradually increase as dementia progresses. Their study also demonstrated the significant correlations between plasma PAI-1 levels and cognitive function. In addition, PAI-1 regulates Aβ deposition and phosphorylation of tau proteins via the BDNF-JNK/c-Jun [17] mechanism, which influences cognitive function. Therefore, PAI-1 is thought to be a susceptible factor for AD. The tissue plasminogen activator (tPA) is a physiological agonist in the fibrinolytic system and plays a key role in the regulation of fibrinolysis and clotting. In addition, tPA or tPA/PAI-1 is closely related to diabetes [10, 11] and cognitive impairment [18 –21]. A study [22] has shown that plasmin is activated by tPA and uPA, and it can scavenge amyloid polypeptides. PAI-1 is the major physiological inhibitor of tPA and uPA, and knockdown [23] of PAI-1 can inhibit the degradation of Aβ, thus promoting the deposition of Aβ. Another study found that Aβ41–42 aggregates can activate tPA [24]. Therefore, PAI-1, tPA, and Aβ deposition formed a vicious cycle [25].
For these reasons, we hypothesized that PAI-1 and tPA are closely associated to diabetes-related cognitive impairment. In summary, we believe that plasma PAI-1 and tPA play a key role in diabetes-related cognitive dysfunction.
METHODS
Ethics
This study was conducted at the Endocrinology Division of the ZhongDa Hospital of Southeast University. The Affiliated ZhongDa Hospital of Southeast University Research Ethics Committee approved the study. All participants were Chinese Han provided written informed consent before participation.
Study design and patients
This study was proceeded from June 2015 to June 2017, 162 right-handed T2DM patients (89 men and 73 women, aged 50–75 years) were recruited. Of these participants, 66 patients (including 31 male and 35 female) were diabetic patients with mild cognitive impairment (MCI) and 96 individuals (including 58 male and 38 female) were age-matched diabetic patients with normal cognition. These subjects were diagnosed with T2DM according to the World Health Organization 1999 criteria [26]. 66 patients (31 females, 35 males, mean SD age = 60.55±7.05 years) met the diagnostic criteria for MCI by the European Association for Alzheimer’s Disease MCI Working Group: 1) Complaints of cognition from patients or their families; 2) As reported, the decline in cognitive function relative to the previous capacity has declined in the past year, (CDR score of 0.5); 3) Through clinical evaluation, cognitive disorder were evidenced (impairment in memory or some other cognitive field); 4) Short of major repercussions in daily life; and 5) Cannot reach the level of dementia [27]. The healthy control groups were matched for sex, age, weight, height, and education at the same time. Exclusion criteria in this study include hypoglycemia coma, diabetic ketoacidosis, hypertonic glucose coma, and other acute complications of diabetes, Cerebrovascular accident, Parkinson’s disease, systemic disease such as malignancy, anemia, severe infection; thyroid disease, visual and auditory resolution deficiencies, ethanol, drug abuse or dependence history and depression (within 2 months), anticoagulants, anti-Parkinson disease drugs, benzodiazepines and barbiturates, antiepileptics and so on were used in the first 3 months of screening.
Clinical data collection
All participants enrolled were in the baseline data collection (including: name, gender, age, education level) and a one-to-one neuropsychological battery, and a laboratory testing. The medical histories of patients (the duration of diabetes was calculated by a physician, Hypertension diagnostic grade was derived from Chinese Journal of Hypertension 2004 Hypertension diagnostic criteria), drug use (used to use insulin, metformin, lipid-lowering agents, antiplatelet agents), and physical measurements (including blood pressure, weight, and height) were collected carefully [BMI = weight/height2(kg/m2)]. Hypertension was defined as a systolic blood pressure ≥140 mmHg or a diastolic blood pressure ≥90 mmHg. The fasting blood-glucose (FBG), fasting C-peptide (FCP), glycosylated hemoglobin (HbA1c), serum creatinine, serum uric acid, apolipoprotein A1 (ApoA1), TC, TG, low density lipoprotein (LDL-C), and high density lipoprotein (HDL-C), Plasminogen time (PT), International Normalized Ratio (INR), Fibrinogen (FIB), Activated partial thromboplastin time (APTT), Thrombin time (TT), Antithrombin III (ATIII), Fibrinogen degradation products (FDP), D dimer (DD) were obtained from blood samples, carotid artery plaques were detected by Ultrasound. The test center of Southeast University Affiliated Zhongda Hospital implemented internal and external quality management directed by the Chinese Laboratory Quality Control.
Neuropsychological tests
A series of cognitive testing, including Montreal cognitive assessment (MOCA), Mini-Mental State Exam (MMSE), Auditory Verbal Learning Test (AVLT), Digit Span Test (DST), Trail Making Test-A and B (TMT-A and TMT-B), Logical Memory Test (LMT), Clock Drawing Test (CDT); and Stroop Color Word Test (SCWT) were administered in a certain order to evaluate state of general mental, ability of memory and attention, executive function, and visuospatial function. The assessment of cognitive function mainly includes: overall cognitive function assessment, scene memory, information processing speed, executive function, visual space, among these: overall cognitive function assessment includes MOCA and MMSE; scene memory includes AVLT-D and LMT; information processing speed includes TMTA, SCWT-A and SCWT-B, executive function includes VFT, DST, TMTB, SCWT-C; visual space includes CDT and so on. The tests were completed by an experienced neuropsychiatrist, and all the subjects and the neuropsychiatrist were blinded to the study design.
Measurement of Plasma PAI-1, tPA, Aβ40 and Aβ42 Level
Demographic, clinical, and cognitive characteristics in MCI group compared with the control group
Data are presented as n (%), mean±SD, or median (interquartile range) as appropriate.
The 2 ml fasting elbow venous blood of two groups patients were collected at 8:00 am, anticoagulated by heparin, and centrifuged by low temperature centrifuge 3000 Rpm for 10 min. Plasma samples were stored at –80°C before analysis. The plasma level of PAI-1, tPA, Aβ40 and Aβ42 were measured by enzyme-linked immunosorbent assay (ELISA) kits (Jin Yibai Biological Technology, Nanjing, China) according to the manufacturer’s instructions. Blood samples of all the participants were measured on the same day. The assay was performed in 2 replicates for each sample and the final concentration was calculated by averaging the 2 results. The absorbance (OD) was measured at 450 nm. Intra and interassay coefficients of variance were less than 9% and 11%, tPA/PAI-1 molar ratio was calculated.
Statistical analysis
Relationship of PAI-1 and tPA/PAI-1 with cognitive performances in different groups
Total, all type 2 diabetic patients (n = 162); MCI, mild cognition impairment in type 2 diabetic patients (n = 66); Non-MCI, non-mild cognition impairment in type 2 diabetic patients (n = 96); Aβ40, amyloid-β40; Aβ42, amyloid-β42; MOCA, Montreal Cognitive Assessment; MMSE, Mini-Mental State Examination; LMT, Logical Memory Test; VFT, Verbal Fluency Test; TMTA, Trail Making Test-A; TMTB, Trail Making Test-B; AVLT-I, Auditory Verbal Learning Test-immediate recall; AVLT-D, Auditory Verbal Learning Test-delayed recall; apartial r and p values were obtained after adjustment for age, sex, and education levels. bpartial r and p values were obtained after adjustment for age, sex, education levels, Aβ40, and Aβ42. Significance, p < 0.05.
Statistical analyses were conducted using SPSS version 22.0. All tests were two-sided, and statistical significance was defined as p < 0.05. Data are expressed as mean±SD or medians. Student’s t test and analysis of variance (ANOVA) were conducted to compare normally distributed variables; nonparametric Mann-Whitney U and Kruskal-Wallis tests were employed for asymmetrically distributed variables. Comparison of non-continuous variables between groups was performed using the chi-squared test. Partial correlation analysis and linear regression analysis were performed to explore the relationships between the cognitive measures and demographic characteristics, and the plasma PAI-1 and tPA/PAI-1 levels in all participants or the MCI group. The cutoff value used in this study for suggested MCI was a MoCA score 26, with a one-point adjustment of the total score for subjects with fewer than 12 years of education.
RESULTS
Demographic and clinical characteristics
Table 1 shows the baseline characteristics of the participants. A total of 162 Chinese T2DM patients were recruited in this study. Sixty-six were characterized by MCI and included in the MCI group. Ninety-six had healthy cognition and included in the control group. The MCI and control groups were well matched for age, sex, education level, BMI, history of smoking and drinking, prevalence of hypertension and diabetes, diabetes and hypertension duration, and systolic and diastolic blood pressure. No significant differences between the two groups were found in FPG, HbA1c, FCP, TC, TG, LDL, HDL, APOA1, APOB, LPa, PT, INR, FIB, APTT, TT, ATIII, FDP, DD, plasma tPA, Aβ40, Aβ42, use of insulin, antiplatelet aggregation, and lipid-lowering medications (p > 0.05). Compared with the healthy group, the MCI group exhibited higher levels of plasma PAI-1 (p < 0.05) and tPA/PAI-1 and significantly lower neuropsychological test scores (p < 0.01). The rate of carotid plaque formation was significantly higher in the MCI group than in the control group (p < 0.01). In subgroup analyses of patients with diabetes, we found that between the two groups of males, carotid plaque, tPA/PAI-1 levels, APTT, HbA1C, and neuropsychological test scores were significantly different (p < 0.05). However, in the female subgroup, only carotid plaques and neuropsychological indexes were significantly different between the two groups (p < 0.05).
Correlations of PAI-1 and tPA/PAI-1 levels with cognitive performances
The correlations of PAI-1 and tPA/PAI-1 levels with cognitive performances were analyzed by partial correlation analysis in all diabetic patients after adjusting for age, sex, and education levels and in the MCI group after adjusting for Aβ40, Aβ42, age, sex, and education levels. In all participants, Montreal Cognitive Assessment (MOCA) scores (r = –0.422; p < 0.001), LMT (r = –0.292; p < 0.001), VFT (r = –0.191; p = 0.016), and MMSE (r = –0.376; p < 0.001) were inversely associated with PAI-1. Aβ40 (r = 0.264; p = 0.001), Aβ42 (r = 0.234; p = 0.003), and TMTB (r = 0.168; p = 0.034) were positively correlated with PAI-1. In addition, tPA/PAI-1 levels were related with MOCA scores (r = 0.240; p = 0.002), VFT (r = 0.159; p = 0.045), and MMSE (r = 0.181; p = 0.023). In the MCI subgroup, PAI-1 and tPA/PAI-1 levels were both associated with MOCA scores (r = –0.492, p < 0.001; r = 0.388, p = 0.002), and PAI-1 was also related with MMSE (r = –0.337, p = 0.007), LMT (r = –0.337, p = 0.007), and Aβ40 (r = 0.268; p = 0.034). However, PAI-1 and tPA/PAI-1 levels were neither associated with Aβ42 (r = 0.213, p = 0.094; r = 0.019, p = 0.880). In the control group, only AVLT-I and PAI-1 were related (r = –0.238, p = 0.023) (Table 2).
Linear regression analysis
Assessment results of the risk of having logical memory impairment in a simple linear regression model in T2DM patients
Dependent variable: LMT score; B, regression coefficients; SE, standard error; MCI, mild cognitive impairment; Aβ40, amyloid-β40; Aβ42, amyloid-β42; PAI-1, plasminogen activator inhibitor-1; tPA, tissue plasminogen activator; ATIII, Antithrombin III; Significance, p < 0.05.
First, we used the simple linear regression models to select the so-called independent variables and then a multivariable linear regression model to clarify independent factors that were associated with the prevalence of MCI in diabetic patients. When LMT score was considered as a dependent variable, age, educational level, PT, APTT, INR, sex, smoking, use of insulin, fatty liver, diabetes duration, ATIII, Aβ40, Aβ42, PAI-1, and tPA/PAI-1 were considered as independent variables in multiple step-wise regression analysis. The results indicated that LMT was significantly associated with plasma PAI-1 (B = –0.005, p = 0.001) (Table 3) and that the variables associated with MCI in T2DM included Aβ40, Aβ42, tPA, education level, fatty liver, diabetes duration, ATIII, and sex. Multivariable regression revealed that plasma Aβ42 and PAI-1 levels were associated with MCI in T2DM patients (Table 4), and the residuals were in normal distribution.
Assessment results of the risk of having logical memory impairment in a multivariable linear regression model in patients with T2DM
Dependent variable: LMT score; MCI, mild cognitive impairment; T2DM, type 2 diabetes mellitus; PAI-1, plasminogen activator inhibitor-1; Aβ42, amyloid-β42; Significance, p < 0.05.
DISCUSSION
The main findings of this study were as follows: 1) T2DM patients with MCI displayed significantly increased plasma levels of PAI-1 and lower tPA/PAI-1 levels compared with the controls. 2) In addition to old age and low educational attainment, high PAI-1 and low tPA/PAI-1 levels were associated with MCI in T2DM patients; for example, plasma levels of PAI-1 were negatively correlated with logical memory in T2DM patients with MCI. Taken together, high plasma levels of PAI-1 and low tPA/PAI-1 levels can predict early cognitive deficits in T2DM patients.
As a highly sensitive tool, MOCA can distinguish diabetic MCI from cognitively healthy T2DM patients [28]. In addition, our research suggests that plasma PAI-1 in MCI patients was associated with LMT, which is a standardized assessment of narrative episodic memory [29]. A short story is orally presented, and the examinee is asked to recall the story verbatim. Impairment in logical memory function is an early change in presymptomatic stages of AD [30]. Hyperglycemia and Aβ accumulation also play important roles in the memory ability of diabetic patients [31 –33]. Another study found that LMT is associated with early cognitive impairment in T2DM [34]. In the present study, we applied MOCA as a group variable and LMT for the regression analysis of the corresponding variables. Consistent with previous results [35 –37], our study demonstrated that T2DM-related MCI patients experience cognitive impairment in terms of logical memory. Additionally, plasma PAI-1 level was higher in the MCI group than in the control group [38, 39]. Similar results were observed in studies that compared AD patients with control participants [39, 40]. In some other studies, no statistically significant changes were found in PAI-1 levels in AD [41 –43]. However, in contrast to our results, Trollor et al. [44] found that PAI-1 levels were not increased in amnestic multiple-domain MCI patients compared with individuals who remained cognitively healthy. One plausible explanation is that PAI-1 is also a neurotrophic factor in the central nervous system [45]. A study [44] indicated that PAI-1 levels were higher in cognitively normal and nonamnestic multiple-domain MCI than in amnestic multiple-domain MCI, suggesting that PAI-1 may protect against memory damage. However, participants from the study by Trollor et al. were from the community, and the present study enrolled diabetes patients. Thus, whether plasma PAI-1 serves as a deleterious substance to cognition or reflects neuroprotective response during the progress of cognitive impairment warrants further research.
Accumulation and deposition of Aβ in the brain are a pathological feature of AD; however, the underlying mechanism is unclear. Moreover, the mechanism by which PAI-1 levels reduce Aβ degradation is uncertain. Research shows that PAI-1 contributes to cognitive dysfunction via the BDNF-JNK/c-Jun pathway [17], resulting in the reduction of Aβ degradation. As is well known, PAI-1 exerts its functions mainly by inhibiting the activities of tPA and uPA, thus activating plasminogen. In addition, new evidence [46, 47] shows that plasmin may promote Aβ degradation and that the reduction in Aβ degradation is due to a diminished effect of PAI-1 on tPA and plasminogen after increased Aβ degradation [23]. Data show that the PAI-1 protein is increased in the cerebral cortex but not in the cerebellum of AD patients (non-AD-deposition zone). In summary, PAI-1 is closely related with Aβ in AD patients and T2DM [9 –11], suggesting that mechanisms in diabetic cognitive disorders include abnormalities in fibrinolytic function. In our study, PAI-1 was high in patients with cognitive impairment. In multivariate regression analysis, Aβ and PAI-1 were two independent risk factors for diabetes with cognitive impairment. It is not in conflict with previous research. The mechanisms of PAI-1 diabetes-related cognitive impairment also include vascular plaque formation [48] and inflammatory stress [44, 49]. In the present study, we discovered that the MCI group had a substantially higher prevalence of carotid plaques compared with the control group. This finding was consistent with that in previous reports [8, 48]. Haratz et al. found that carotid plaques are associated with executive function, memory, attention, and global cognitive scores in non-dementia patients with concomitant carotid large-vessel disease [50]. In the mouse model [23], study found increased PAI-1 and decreased tPA levels in mice were related with AD or cognitive impairment. PAI-1 promotes synaptogenesis and protects against Aβ1 - 42-induced neurotoxicity. tPA may also affect synaptic plasticity, excitotoxic neuronal death, and apoptosis [42, 46]. As a matter of fact, cognitive benefits of PAI-1 deletion [23] or inhibition [20] in AβPP-overexpressing mice have already been reported and are associated with amyloid clearance [51]. However, the transport mechanism of PAI-1 in the blood–brain barrier (BBB) is unclear. PAI-1 produced within the brain by microglia [74] and astrocytes may regulate apoptosis and migration of microglia, thereby regulating the transport of PAI-1 in the BBB.
The results of the present study are consistent with previous studies. Low tPA/PAI-1 levels are associated with MCI/AD [20, 21]. The possible mechanism is that the tPA–fibrinolytic proteolytic cascade aids in the clearance of Aβ [20, 46], preventing the progress of AD. Several studies indicated that Aβ aggregates can replace fibrin aggregates in activating tPA and showed that tPA may be activated by Aβ in AD [24, 52]. Later on, brain plasmin was reported to enhance Aβ degradation [47, 53]. Karim et al. [75] demonstrated that blood-derived tPA can reach the brain parenchyma without BBB alteration. In addition, tPA may signal the LRP receptor to open the BBB [54, 55] and recruit and activate the microglia [56, 57]. Combined with current amyloid and tau biomarkers, tPA may have the potential to serve as an alternative marker for AD pathology [58].
In our study, we found that ATIII was statistically significant in univariate regression analysis and considered it to play a role in PAI-1-induced diabetes-related cognitive impairment. ATIII is a serpin that plays an important role in the nervous system. In the study by Kalaria et al. [59], they examined the presence of ATIII in the pathological lesions of AD and suggested that increased ATIII levels are related to astrogliosis and neurofibrillary pathology of AD. They concluded that ATIII may play a role in the pathogenesis of cerebral amyloidosis. In addition, antithrombin inhibits certain proteases, particularly thrombin, which leads to neurite retraction, thus affecting the process of AD [60]. In diabetic patients, fibrinolytic function abnormalities are prevalent [6].
We found that gender did not show any significant association with the incidence of MCI in diabetic patients. However, some studies suggest a higher prevalence of MCI in men [61 –63], others in women [64 –68], while some none at all [69, 70]. The reasons for these differences may be age, races and the degree of enrollment, in medical research [71], gender referred to psychosocial and cultural differences (e.g., education levels). In general, women had a higher incidence of MCI at older ages compared with men. Another study found that the most consistent cross-sectional difference at all ages was that women performed better on verbal memory tasks compared with men [72, 73].
There are some limitations to our research. First, the sample size was relatively small, which may limit our ability to detect the association between inflammation and PAI-1 in T2DM patients without complications. Second, healthy volunteers or other patients without T2DM were not enrolled in this study. This omission may reduce the strength of the results. Third, in this exploratory study, no adjustment to p-values was made for multiple comparisons. Future studies will need to confirm the hypotheses that were generated in this study. Finally, we did not fully analyze the effects of some drugs that affected the levels of inflammation and hemostasis markers. Therefore, continued follow-up of the participants will enable us to better assess these relationships.
Conclusion
Our data indicated that diabetic patients with MCI exhibited significantly increased plasma PAI-1 levels and decreased tPA/PAI-1 molar ratio, and plasma PAI-1 levels were associated with memory function. A high plasma PAI-1 level and a low tPA/PAI-1 molar ratio might be associated with the pathogenesis of MCI in the Chinese T2DM populations. Our study also suggested that plasma PAI-1 levels and tPA/PAI-1 molar ratio might be potential markers of early cognitive impairment in diabetes. Further studies should be conducted to determine whether plasma PAI-1 level and tPA/PAI-1 molar ratio can be used as early biomarkers for the clinical diagnosis of T2DM with MCI.
Footnotes
ACKNOWLEDGMENTS
This work was partially supported by the National Natural Science Foundation of China (No. 81570732, Shaohua Wang), the National Nature Science Youth Foundation of China (No. 81200635, Yang Yuan) and the Jiangsu Provincial Medical Youth Talent (QNRC2016819, Yang Yuan).
