Complement Proteins in Serum Astrocyte-Derived Exosomes Are Associated with Poststroke Cognitive Impairment in Type 2 Diabetes Mellitus Patients

Abstract

Background:

The complement system plays crucial roles in cognitive impairment and acute ischemic stroke (AIS). High levels of complement proteins in plasma astrocyte-derived exosomes (ADEs) were proven to be associated with Alzheimer’s disease. We aimed to investigate the relationship of complement proteins in serum ADEs with poststroke cognitive impairment in type 2 diabetes mellitus (T2DM) patients.

Methods:

This study analyzed 197 T2DM patients who suffered AIS. The Beijing version of the Montreal Cognitive Assessment (MoCA) was used to assess cognitive function. Complement proteins in serum ADEs were quantified using ELISA kits.

Results:

Mediation analyses showed that C5b-9 and C3b in serum ADEs partially mediate the impact of obstructive sleep apnea (OSA), depression, small vessel disease (SVD), and infarct volume on cognitive function at the acute phase of AIS in T2DM patients. After adjusting for age, sex, time, and interaction between time and complement proteins in serum ADEs, the mixed linear regression showed that C3b and complement protein Factor B in serum ADEs were associated with MoCA scores at three-, six-, and twelve-months after AIS in T2DM patients.

Conclusions:

Our study suggested that the impact of OSA, depression, SVD, and infarct volume on cognitive impairment in the acute stage of AIS may partially mediate through the complement proteins in serum ADEs. Additionally, the complement proteins in serum ADEs at the acute phase of AIS associated with MoCA scores at three-, six-, twelve months after AIS in T2DM patients.

REGISTRATION: URL: http://www.chictr.org.cn/,ChiCTR1900021544

Keywords

Keywords: Alzheimer’s disease astrocyte-derived exosomes complement proteins post stroke cognitive impairment type 2 diabetes mellitus

INTRODUCTION

Ischemic stroke is one of the common causes of not only physical disability but also cognitive and emotional impairment. Poststroke cognitive impairment (PSCI) is a common poststroke functional impairment among ischemic stroke survivors and often manifests as executive dysfunction and decreased attention. The epidemiological incidence of PSCI after acute ischemic stroke (AIS) varied widely between different studies due to the different enrollment populations, ranging from 41.8% to 54.9% [1 –4]. In addition, patients with type 2 diabetes mellitus (T2DM) may have an increased risk of PSCI due to changes in brain structure, especially in executive, memory, and attention functions [5].

Previous studies have demonstrated that obstructive sleep apnea (OSA) [6], depression [7], small vessel disease (SVD) [8], and infarct volume [9] are risk factors for PSCI. These factors may influence cognitive function through inflammatory reactions [10 –12]. Chronic inflammation also has been reported to be an important factor associated with PSCI [13]. Astrocytes act as inflammatory cells of the central nervous system and can be activated in pathological conditions such as ischemic stroke [14]. Reactive astrocytes secrete pro-inflammatory cytokines and destroy the blood–brain barrier (BBB), which leads to neuroinflammation and cognitive deficits [15, 16]. As a part of the innate immune system, complement proteins play a critical role in the inflammatory reaction and can be carried by exosomes [17]. Evidence of the involvement of complement system in ischemic stroke and Alzheimer’s disease (AD) has been proven [18, 19]. In the complement cascade reactions, C3b triggers the activation of the complement bypass pathway exclusively by binding with complement protein Factor B, resulting in the production of C3 and C5 invertases. This makes complement protein Factor B a crucial component for activating the complement alternative pathway. C5b-9, known as the membrane attack complex, can accumulate on the cell surface, causing cell lysis [20]. As such, we hypothesis that the impact of OSA, depression, SVD, and infarct volume on cognitive impairment at acute stage of AIS in T2DM patients may be mediate by complement proteins in serum astrocyte-derived exosomes (ADEs) and prospectively explored the associations of C5b-9, C3b and complement protein Factor B in serum ADEs with PSCI in T2DM patients.

MATERIAL AND METHODS

Study participants

The present longitudinal study included survivors of mild first-ever AIS among T2DM patients with a total National Institutes of Health Stroke Scale (NIHSS) ≤10 [21]. Acute ischemic stroke diagnosis according to the World Health Organization criteria and confirmed by magnetic resonance imaging (MRI) [22]. All subjects were prospectively enrolled from consecutive T2DM patients with first-ever mild AIS who were admitted to our center within 72 hours after the onset of symptoms from August 2020 to March 2022. The morning after 3–6 days of hospitalization, peripheral blood samples were collected from all participants by venipuncture of forearm veins following an overnight fast. All blood samples were collected by a technician using the same methods and procedures. The samples were centrifuged at 4000×g for 10 min within 30 min, and the serum was quickly stored in a refrigerator at –80°C until they were assayed.

Patients exhibiting any of the following conditions were excluded from the study: (i) patients presenting with a primary hemorrhagic stroke; (ii) patients presenting with transient ischemic stroke; (iii) patients who suffered from a stroke resulting from trauma or an invasive procedure; (iv) patients presenting with severe aphasia, visual or auditory impairment, or a chaotic conscious state that made them unable to complete neuropsychological assessments; (v) patients with a history of dementia or cognitive impairment (determined by the Informant Questionnaire on Cognitive Decline in the Elderly score≥3.19) [23]; (vi) patients with serious heart diseases (such as recent acute coronary syndrome, life-threatening arrhythmias, and heart failure), liver diseases (e.g., cirrhosis), severe renal failure or those who were receiving hemodialysis; (vii) prescribed medicine that may affect cognitive function; and (viii) patients who were unlikely to participate in the follow-up or be discharged from the hospital. The neurological assessment included a verification of the stroke pathogenesis and the NIHSS. The determination of major cardiovascular risk factors was based on medical history and laboratory findings. The Ethical Committee of Weihai Municipal Hospital approved this study, and all the participants signed an informed consent form.

Clinical data

All participants were interviewed by an experienced physician using a standard questionnaire that included demographic characteristic data (age, sex, body mass index [BMI], education level), lifestyle information (smoking history and alcohol consumption), medical and medication history, and vascular risk factors (hypertension, fasting blood glucose (FBG), glycosylated hemoglobin (HbA1c), low-density lipoprotein cholesterol (LDL), triglycerides, and total cholesterol). BMI was defined as weight (kg) divided by the square of height (cm). Hypertension was defined as systolic blood pressure≥140 mmHg and/or diastolic blood pressure≥90 mmHg after at least 5 min of resting on three different days or previously diagnosed with hypertension. Hyperlipidemia was defined as an LDL-C level≥130 mg/dl, triglyceride level≥150 mg/dl, total cholesterol level≥200 mg/dl, or history of use of lipid-lowering drugs. The NIHSS was used to assess the severity of AIS, and the Essen stroke risk score (ESRS) was used to assess the risk of stroke recurrence.

Neuropsychological assessment

The Mini-Mental State Examination (MMSE) and Beijing version of the Montreal Cognitive Assessment (MoCA) were used to assess comprehensive cognitive function and cognitive subdomains (visuo-executive functions, naming, attention, language, abstraction, delayed recall, orientation) at the acute phase of AIS, and those with a MoCA score < 25 [24] points were defined as acute stage cognitive dysfunction. The Beijing version of the MoCA was also used to assess cognition at three-, six-, and twelve-month follow-ups after AIS; it was considered PSCI if MoCA < 25, and MoCA≥25 indicated no poststroke cognitive impairment (PSNCI). The Hamilton Depression Scale (HAMD-17) was employed to assess the depression symptoms of patients, and depression symptoms were defined by a score≥17 on the HAMD-17. The STOP-Bang questionnaire was used to assess the high risk of OSA, and a high risk of OSA was defined by a total score≥3 on the STOP-Bang.

Magnetic resonance data acquisition

All participants underwent whole-brain MRI and brain magnetic resonance angiography (MRA) with a 3T scanner (Siemens, Erlangen, Germany). The examination sequence included T1-weighted and T2-weighted fluid attenuated inversion recovery (FLAIR) images, apparent diffusion coefficient (ADC) images and diffusion-weighted imaging (DWI) images (repeat time (TR) 4040 ms, echo time (TE) 64 ms, field of view (FOV) 220×220, matrix size = 512×512, section thickness = 5 mm). The scanning parameters of 3D TOF-MRA were TR 21 ms, TE 3.43 ms, FOV 200 mm×180 mm, matrix 313×384, number of excitations 1, and layer thickness 0.7 mm. Infarct volume was semiautomatically quantified on the DWI images using free open-source 3D-Slicer software (http://www.slicer.org) to differentiate between infarct region and normal cerebral tissues. The infarct region was determined by using the “segment editor” in 3D Slicer. MRI was used to semiquantitatively evaluate the markers of SVD, and the quantify criteria of SVD burden are as follows: white matter hyperintensities (WMH) was assessed according to the Fazekas scale proposed by Fazekas. The evaluation criteria of periventricular WMH were as follows: 0 points were scored for no lesion, 1 point was scored for lesions with cap or pencil thin layer, 2 points were scored for lesions with smooth halo, and 3 points were scored for lesions with irregular shape and deep extension of paraventricular WMH. The evaluation criteria for deep WMH are 0 points for non-lesion, 1 point for mottle lesions, 2 points for mottle lesions fused or small patchy lesions, and 3 points for large-area patchy lesions fused. The Fazekas score ranges from 0 to 6 points, which is the sum of paraventricular and deep WMH scores: 0–3 points for mild and moderate WMH, and 4–6 points for severe WMH. Enlarged perivascular spaces (PVSs) were evaluated according to the PVSs visual rating scale, and the level with the largest number of PVSs in the basal ganglia of one side of the brain was counted and rated: no lesions were grade 0, 1–10 were grade 1, 11–20 were grade 2, 21–40 were grade 3, >40 were grade 4, grade 0–2 were rated as mild to moderate PVSs, and grade 3–4 were rated as severe PVSs. According to the overall SVD burden standard proposed by Staals, 1 score was scored if the following items were met: (1) according to the Fazekas scale, the paraventricular WMH was 3 points or the deep cerebral WMH was ≥2 points; (2) the number of vasogenic lacunar infarction≥1; (3) the number of cerebral microbleeds (CMBs) is ≥1; (4) the number of PVSs located in the basal ganglia region is ≥11. The total SVD burden is recorded as 0–4 points, and the higher the score, the more severe of the SVD. In our study, AIS was classified into cortical infarcts, subcortical infarcts and subtentorial infarcts based on imaging data from patients.

Extraction of exosomes and quantification of complement proteins

Isolation of specific ADEs was performed as described in our published protocol [25]. In brief, total exosomes were collected from serum using an ExoQuick exosome precipitate solution (EXOQ, EXOQ20A-1, System Biosciences, USA). ADEs were separated by coimmunoprecipitation using glutamine aspartic acid carrier (GLAST) (ACSA-1) biotinylated antibody (MiltenyiBiotec, 130-118-984, Auburn, CA). According to our previous protocols, we confirmed the success of exosome collection by Western blotting (WB) and transmission electron microscopy (TEM). ADE proteins, Tetraspanning exosome marker CD63 (RayBio, ELH-CD63, Norcross, GA 30092, USA), complement fragment C3b (Abbexa Ltd, abx252114, Cambridge, UK), C5b-9 (Abbexa Ltd, abx054346, Cambridge, UK) and decay acceleration Factor B (Abbxa Ltd, abx055457, Cambridge, UK) were quantified by enzyme-linked immunosorbent assay (ELISA). All protein levels for each sample were normalized to the exosome content of the CD63 exosome marker, and the relative value of CD63 for each sample was used to normalize their recovery. The ELISA microplates were read using a Varioskan LUX 3020 instrument (Thermo Fisher Scientific Oy, Ratastie 2, Vantaa, Finland). Serum samples from all subjects were analyzed by the same technician, who had no knowledge of the clinical data.

Statistical analysis

R software version 4.3.1 and SPSS software version 26.0 were used for statistical analysis, and GraphPad Prism 9.5.1 and R software version 4.3.1 were used for figure preparation. All P values are 2-tailed, and a significance level of 0.05 was used. The Kolmogorov—Smirnov test was applied to ascertain the normality of continuous data. Mean±standard deviation (SD) was used to report continuous variables of normal distribution, and the nonnormal distribution data were represented as a median and interquartile range. The demographic and clinical characteristics of all participants were evaluated by univariate analysis. Analysis of t test or analysis of variance (ANOVA) was used to investigate differences between groups for normally distributed variables, and the nonparametric Mann—Whitney U test or the Kruskal—Wallis variance test was used to examine nonnormally distributed variables. The χ² test was used to analyze intergroup differences in categorical variables. Age wise stratified data from 41 to 90 was used to clarify the characteristics of cognitive impairment and complement proteins in serum ADEs at acute stage of AIS at different age. We further performed mediation analyses to identify whether complement proteins in serum ADEs mediate the effects of OSA, SVD, NIHSS, FBG, HbA1c, depression, and infarct volume on cognitive function and its subdomains during the acute period of AIS in T2DM patients. The mediator models included cognitive function and its subdomains at the acute stage of AIS as dependent variables, OSA and SVD and NIHSS and FBG and HbA1c and depression and infarction volume as independent variables, and the serum ADE levels of C5b-9, C3b and complement protein Factor B as mediator variables. The number of bootstrap samples in all analyses was set to 10000, where each path of the model was controlled for age, sex, BMI, education, hypertension, current smoking, and current alcohol consumption. The “mediation” package in R version 4.3.1 software was used to perform the mediation analysis. We applied mixed linear regression with MoCA scores at follow-up period times as a dependent variable for each complement proteins in serum ADEs. The covariates were the inflammatory biomarkers, time, the interaction between time and inflammatory biomarkers. The participant was included as a random effect. In all models, we performed analyses adjusted for age, sex, time, and interaction between time and complement proteins in serum ADEs. The mixed linear regression models were also used to determine the association between complement proteins in ADEs and the usage conditions of statin by calculating β value and 95% confidence interval (CI).

RESULTS

Baseline characteristics

During the study period, 278 patients with T2DM who experienced mild first-ever AIS were identified as study candidates (Fig. 1). Patients with a history of hemorrhagic stroke (n = 2), dementia or cognitive impairment (n = 14), Parkinson’s disease or parkinsonism (n = 2), or severe organ failure (n = 2) were excluded. We then excluded 29 patients whose MRI examinations or baseline cognitive data were incomplete. We excluded 32 patients whose cognitive assessments were incomplete or who died during follow-up, and 197 patients were included in the final analysis.

Fig. 1

Description of the study population.

A total of 197 participants (95 men and 102 women) were recruited for this study. All participants were classified into 2 groups (cognitive impairment and non-cognitive impairment) based on the MoCA scores at the acute stage of AIS to compare demographic and clinical characteristics. Demographic and clinical characteristics are presented in Table 1. A total of 145 (73.60%) patients had cognitive impairment at the acute stage of AIS. Remarkably, age (p = 0.003), sex (p = 0.009), education (p < 0.001), current alcohol consumption (p = 0.02), HAMD-17 scores (p = 0.001), STOP-Bang scores (p < 0.001), and infarct volumes (p < 0.001) is statistically different between cognitive impairment group and non-cognitive impairment group at acute stage of AIS. BMI, history of ischemic heart disease, history of hypertension, history of hyperlipidemia, the usage condition of statin, current smoking, ESRS, FBG, HbA1c, NIHSS, thrombolysis and infarct location were not significantly different between the two groups. There was not significantly association between the usage of statin, statin medicine type, the intensity of statin and C5b-9, C3b and complement protein Factor B in serum ADEs (Supplementary Table 1).

Table 1

Baseline demographic and clinical characteristics of all patients at the acute phase for AIS

	All patients	Cognitive	Non-cognitive	P
	(n = 197)	impairment	impairment
		(n = 145)	(n = 52)
Age, y, mean±SD	64.60±8.23	65.99±7.65	60.79±8.67	<0.001
Male, n (%)	95 (48.22)	78 (53.79)	17 (32.69)	0.009
Education, y, median (IQR)	7 (5–9)	7 (5–9)	9 (8–12)	<0.001
BMI, Kg/m², mean±SD	26.15±3.42	25.86±3.49	26.81±3.16	0.08
IHD, n (%)	25 (12.69)	21 (14.48)	4 (7.69)	0.31
Hypertension, n (%)	130 (65.99)	92 (63.45)	38 (73.08)	0.21
Current smoking, n (%)	65 (32.99)	43 (29.66)	22 (42.31)	0.1
Current alcohol consumption, n (%)	54 (27.41)	33 (22.76)	21 (40.38)	0.02
ESRS, median (IQR)	3 (2–3)	3 (2–3)	3 (2–3)	0.23
hyperlipidemia, n (%)	69 (35.03)	51 (35.17)	18 (34.62)	0.94
Statin usage, n (%)	196(99.49)	144 (99.31)	52 (100)	1.00
Statin medicineAtovastatin, n (%)	182 (92.39)	134(92.41)	48 (92.31)	1.00
Rosuvastatin, n (%)	14 (7.11)	10(6.9)	4(7.69)	1.00
Medium-intensity of statin dosage, n (%)	184(93.40)	134(92.41)	50(96.15)	0.52
High-intensity of statin dosage, n (%)	12(6.09)	10(6.9)	2(3.85)	0.74
FBG, mmol/L, median (IQR)	8.63(6.85–10.73)	8.64(6.85–10.65)	8.46(6.73–11.94)	0.79
HbA1c, %, mean±SD	8.35±2.20	8.29±2.02	8.54±2.66	0.48
MoCA, mean±SD	20.17±4.84	17.94±3.80	25.9±1.18	<0.001
MMSE, mean±SD	25.01±3.73	23.78±3.59	28.37±1.25	<0.001
Depression, median (IQR)	5 (2–9)	6 (3–10)	3 (0–5.75)	<0.001
OSA, median (IQR)	3 (2–4)	3 (3–4)	2 (1–4)	<0.001
NIHSS, median (IQR)	2(1–4)	2 (1–4)	2 (1–3)	0.14
Thrombolysis, n (%)	20 (10.15)	12 (8.28)	8 (15.38)	0.18
Cortical infarct, n (%)	35(17.77)	22 (15.17)	13 (25)	0.11
Subcortical infarct, n (%)	123(62.44)	91 (62.76)	32 (61.54)	0.74
Subtentorial infarct, n (%)	65(32.99)	44 (30.34)	21 (40.38)	0.19
Infarct volume, mm³, median (IQR)	293.89	563.87	102.55	0.001
	(67.54–948.23)	(78.62–1141.41)	(49.31–429.26)

Midium-intensity is defined as a daily dose of atorvastatin of 10–20 mg or a daily dose of rosuvastatin of 5–10 mg; High-intensity is defined as a daily dose of atorvastatin of 40–80 mg or a daily dose of rosuvastatin of 20–40 mg. Abbreviation: SD, standard deviation; IQR, interquartile range; BMI, body mass index; IHD, ischemic heart disease; ESRS, Essen stroke risk score; FBG, fasting blood glucose; HbA1c, glycosylated hemoglobin; OSA, obstructive sleep apnea; NIHSS, the National Institutes of Health Stroke Scale; MoCA, Montreal Cognitive Assessment; MMSE, Mini-Mental State Examination; AIS, acute ischemic stroke; T2DM, type-2 diabetes mellitus.

Prevalence of PSCI and the levels of complement proteins in serum ADEs in T2DM patients at different ages

As shown in Tables 2 and 3, the mean±SD values of all participants were 1.617±0.589 ng/ml for C5b-9 in serum ADEs, 0.389±0.13 ng/ml for C3b in serum ADEs, and 0.578±0.25 ng/ml for complement protein Factor B in serum ADEs, respectively.

Table 2

Age wise stratified data of cognitive impairment at acute stage and C5b9 and C3b in serum ADEs between AIS patients

	C5b9 (ng/ml), mean±SD				C3b (ng/ml), mean±SD
Age	Total	Cognitive	Total	Cognitive	Non-cognitive	P	Total	Cognitive	Non-cognitive	P
(years)	patients	impairment		impairment	impairment		impairment	impairment
		n (%)		n = 145	n = 52		n = 145	n = 52
All age groups	197	145 (73.60)	1.617±0.589	1.722±0.607	1.324±0.417	<0.001^*	0.389±0.13	0.403±0.136	0.348±0.101	0.002^*
41–50	5	1 (20)	1.025±0.28	–	1.074±0.297	0.509	0.384±0.132	–	0.421±0.12	0.265
51–60	61	35 (57.38)	1.548±0.566	1.701±0.596	1.343±0.456	0.013^*	0.366±0.121	0.383±0.133	0.342±0.099	0.194
61–70	91	76 (83.52)	1.656±0.517	1.721±0.521	0.327±0.348	0.001^*	0.392±0.131	0.405±0.134	0.325±0.089	0.03^*
71–80	35	28 (80)	1.749±0.772	1.84±0.808	1.385±0.492	0.166	0.418±0.144	0.429±0.149	0.373±0.12	0.362
81–90	5	5 (100)	1.405±0.545	1.405±0.545	–	–	0.4±0.099	0.4±0.099	–	–

Abbreviation: ADEs, astrocyte-derived exosomes; AIS, acute ischemic stroke; SD, standard deviation. ^*: P < 0.05.

Table 3

Age wise stratified data of cognitive impairment at acute stage and complement protein Factor B in serum ADEs between AIS patients

Age (years)	Total	Cognitive	Factor B (ng/ml), mean±SD
	patients	impairment	Total	Cognitive	Non-cognitive	p
		n (%)		impairment n = 139	impairment n = 58
All age groups	197	145 (73.60)	0.578±0.25	0.585±0.258	0.559±0.23	0.512
41–50	5	1 (20)	0.671±0.479	–	0.714±0.541	0.744
51–60	61	35 (57.38)	0.536±0.194	0.556±0.185	0.508±0.205	0.338
61–70	91	76 (83.52)	0.582±0.249	0.582±0.263	0.583±0.161	0.983
71–80	35	28 (80)	0.626±0.305	0.631±0.33	0.606±0.192	0.845
81–90	5	5 (100)	0.601±0.202	0.601±0.202	–	–

Abbreviation: ADEs, astrocyte-derived exosomes; AIS, acute ischemic stroke; SD, standard deviation. ^*: P < 0.05.

And there were significant differences in serum ADEs levels of C5b-9 (P < 0.001) and C3b (P = 0.002) between the cognitive impairment group and the non-cognitive impairment group during the acute phase of AIS.

The levels of C5b9 and C3b in serum ADEs were statistically different between the cognitive impairment group and the non-cognitive impairment group at the acute stage of AIS in the 61 to 70 years patients after stratified by age. It was also showed that the level of C5b9 in serum ADEs was significant difference between the cognitive impairment group and the non-cognitive impairment group during the acute phase of AIS in the 51–60 years patients, but there was no significantly different of complement protein Factor B in serum ADEs between the cognitive impairment group and the non-cognitive impairment group at the acute phase of AIS or every age stages.

Mediation analysis

There were significant differences in the ADE levels of C5b-9 (P < 0.001) and C3b (P = 0.002) between the cognitive impairment group and the non-cognitive impairment group during the acute phase of stroke (Figure 2). The mean±SD values of the cognitive impairment group and non-cognitive impairment group were 1.722±0.607 ng/ml and 1.324±0.417 ng/ml for C5b-9, 0.403±0.136 ng/ml and 0.348±0.101 ng/ml for C3b, and 0.585±0.258 ng/ml and 0.559±0.23 ng/ml for complement protein Factor B, respectively.

Fig. 2

The comparison of complement levels in serum ADEs at the acute phase of stroke. A) Levels of C5b9 in serum ADEs between the cognitive impairment and non-cognitive impairment groups at the acute phase of stroke. B) Levels of C3b in serum ADEs between the cognitive impairment and non-cognitive impairment groups at the acute phase of stroke. C) Levels of complement protein Factor B in serum ADEs between the cognitive impairment and non-cognitive impairment groups at the acute phase of stroke. Each point represents the value for a cognitive impairment subject or non-cognitive impairment patient and the horizontal line in point clusters is the mean levels for that group. The significance of difference between values for cognitive impairment and non-cognitive impairment patients were analyzed by the t test. The unit of complement proteins in serum ADEs is ng/ml. ADEs, astrocyte-derived exosome.

We performed a mediation analysis to test whether C5b-9, C3b and complement protein Factor B in serum ADEs mediate the impacts of OSA, SVD, NIHSS, FBG, HbA1c, depression, and infarct volume on cognitive function and its subdomains during the acute phase of AIS in patients with T2DM. Figures 3 and 4 illustrates that the impact of OSA, depression, SVD, and infarct volume on cognitive function during the acute phase of AIS in T2DM patients mediate by the serum ADE levels of C5b-9 and C3b after adjusting for age, sex, education, BMI, hypertension, current smoking, and current alcohol consumption. The mediation effects were all considered a partial mediation in Figs. 3 and 4. C5b-9 in serum ADEs mediate the relationship of OSA, depression, SVD, infarct volume and cognitive function during the acute phase of AIS in T2DM patients, with the mediation proportion varying from 21.7% to 40% (OSA, 21.70% of the total effect; depression, 34.33% of the total effect; SVD, 26.77% of the total effect; and infarct volume, 40% of the total effect) (Fig. 3). We further analyzed whether the association between OSA, depression, SVD, infarct volume and the cognitive subdomains mediate by C5b-9 in serum ADEs. The results showed that C5b-9 in serum ADEs mediate the effects of OSA, depression, SVD, and infarct volume on cognitive subdomains during the acute stage of AIS in T2DM patients, with the mediation proportion varying from < 0.1% to 70.93% (OSA, 14.94% to 41.45% of the total effect; depression, 13.43% to 70.93% of the total effect; SVD, 15.53% to 52.18% of the total effect; infarct volume, all the total effects are < 0.1%). Similarly, the effects of OSA, depression, SVD, and infarct volume on cognitive function during the acute phase of AIS in T2DM patients also mediate by C3b in serum ADEs, with the proportion of mediation varying from 14.6% to 30.71% (OSA, 14.60% of the total effect; depression, 24.62% of the total effect; SVD, 25.36% of the total effect; and infarct volume, 30.71% of the total effect) (Fig. 4). Equally, serum ADE levels of C3b also mediate the effects of OSA, depression, SVD, and infarct volume on cognitive subdomains during the acute stage of AIS in T2DM patients, with the proportion varying from <0.1% to 38.37% (OSA, 9.82% to 21.28% of total effect; depression, 9.28% to 38.37% of total effect; SVD, 11.71% to 33.57% of total effect; and infarct volume, all total effects are <0.1%). However, we did not find statistical interactions of complement protein Factor B in serum ADEs in the associations of OSA, SVD, NIHSS, FBG, HbA1c, depression, and infarct volume with cognitive function and its subdomains during the acute phase of AIS in T2DM patients.

Fig. 3

Mediation analysis. The relationship of OSA (A), depression (B), SVD (C), and infarct volume (D) with cognitive function and its subdomains during the acute stage of acute ischemic stroke mediate by C5b9 in serum astrocyte-derived exosomes. ^*p < 0.05 indicate the mediation pathways are meaningful. The proportions shown in the figure indicate the proportion of mediating factors in the total effect of OSA, depression, SVD, and infarct volume on cognition. The unit of complement proteins in serum ADEs is ng/ml. OSA, obstructive sleep apnea; SVD, small vessel disease; MoCA, Montreal Cognitive Assessment; ADEs, astrocyte-derived exosome.

Fig. 4

Mediation analysis. The relationship of OSA (A), depression (B), SVD (C), and infarct volume (D) with cognitive function and its subdomains during the acute stage of acute ischemic stroke mediate by C3b in serum astrocyte-derived exosomes. ^*p < 0.05 indicate the mediation pathways are meaningful. The proportions shown in the figure indicate the proportion of mediating factors in the total effect of OSA, depression, SVD, and infarct volume on cognition. The unit of complement proteins in serum ADEs is ng/ml. OSA, obstructive sleep apnea; SVD, small vessel disease; MoCA, Montreal Cognitive Assessment.

Results from the mixed linear regression models with MoCA scores as dependent variable and complement proteins as independent variables

The mixed linear regression models showed that after adjusted for age, sex, time and interaction between time and complement proteins in serum ADEs, higher concentrations of C3b in serum ADEs were significantly associated with lower MoCA scores at three- (p = 0001), six- (p < 0.001), and twelve-months (p = 0.009) after AIS, and higher concentrations of complement protein Factor B in serum ADEs were significantly associated with higher MoCA scores at three- (p = 015), six- (p = 0.037), and twelve-months (p = 0.020) after AIS in the context; whereas, we did not find significant association between C5b9 in serum ADEs and MoCA scores at three-, six-, and twelve-months after AIS (Table 4).

Table 4

Mixed linear regression models with MoCA scores as dependent variable and complement proteins as independent variables

MoCA at three months				MoCA at six months				MoCA at twelve months
after AIS (n = 197)				after AIS (n = 197)				after AIS (n = 197)
	β coefficient	95% CI		p	β coefficient	95% CI		p	β coefficient	95% CI		p
C5b9 (ng/ml)	–0.722	–2.026	0.583	0.278	–0.849	–2.154	–0.455	0.202	–0.391	–1.696	0.913	0.556
C3b (ng/ml)	–1.843	–3.221	–0.466	0.001^*	–2.594	–3.972	–1.216	<0.001^*	–2.302	–3.680	–0.924	0.009^*
Factor B (ng/ml)	1.345	0.211	2.479	0.015^*	1.205	0.072	2.339	0.037^*	1.403	0.268	2.536	0.020^*

Adjusting for age, sex, time and interaction between time and complement proteins in serum ADEs. Abbreviation: MoCA, Montreal Cognitive Assessment; ADEs, astrocyte-derived exosomes; AIS, acute ischemic stroke; CI, confidence interval. ^*: P < 0.05.

DISCUSSION

To our knowledge, this is the first study to explore the relationships between C5b-9, C3b, and complement protein Factor B in serum ADEs and PSCI in T2DM patients, and the main findings are as follows: 1) C5b-9 and C3b in serum ADEs may partially mediate the impact of OSA, depression, SVD, and infarct volume on cognitive function and its subdomains at the acute stage of AIS in T2DM patients; and 2) higher concentrations of complement proteins in serum ADEs at the acute stage of AIS were significantly associated with MoCA scores at three-, six-, and twelve-months after AIS in T2DM patients.

The MoCA was used to assess post-stroke cognitive function in 197 patients with T2DM and age stratified the data every 10 years from 40 to 90 years which find an increased prevalence of age-related PSCI in T2DM patients, which is consist with age was a risk factor for cognitive impairment in aged [26]. We also assessed differences in complement proteins in serum ADEs between different age subgroups and showed a gradual age-related increase, which indicate that neuroinflammation induced by complement proteins may increases with age.

From previous studies, we know that there are many risk factors for PSCI, including OSA [6], depression [7], SVD [8], and infarct volume [9]. The inflammatory response induced by complement system activation is associated with early poststroke neurodegeneration [27]. In this study, we found that the ADEs levels of C5b-9 and C3b were significantly higher in the cognitive impairment group during the acute phase of stroke than in the non-cognitive impairment group. Stroke patients with OSA present lower cognitive function and functional status during the rehabilitation phase of stroke recovery [28], and OSA-related cognitive dysfunction has been shown to be associated with inflammation. Our team has previously demonstrated that cognitive dysfunction in OSA patients is associated with elevated C5b-9 and C3b [10]. The mediation analysis of our study further suggests that C5b-9 and C3b in serum ADEs may partially mediate the impact of OSA on cognitive impairment during the acute phase of AIS in T2DM patients. The underlying mechanism may be that chronic intermittent hypoxia in OSA patients induces oxidative stress and other inflammatory responses, damaging vascular endothelial cells and the BBB and ultimately leading to cognitive impairment [29, 30]. Depression, as a highly related factor with PSCI [7], is often considered to be closely related to neuroinflammation, especially the complement system involved in cell-related inflammatory changes in depressive states [31]. Ishii et al. also reported that cerebrospinal fluid C5 was significantly increased in patients with major depression compared with the normal control group [32]. Our study found that C5b-9 and C3b in serum ADEs may partially mediate the impact of depression on cognitive impairment during the acute phase of AIS in T2DM patients through mediation analysis. The potential mechanism may be that elevated inflammatory factors in patients with depression change BBB permeability and trigger the accumulation of neurotoxic byproducts, ultimately leading to cognitive impairment [33, 34]. SVD is a disease characterized by lacunar infarction, CMBs, cerebral atrophy, microinfarcts, PVSs, and WMH, and it plays a necessary role in the pathogenesis of PSCI [8], especially WMH [35] and periventricular white matter damage [36]. Central nervous system glial cells can aggravate white matter injury through the complement signaling pathway after hypoperfusion [37], and periventricular white matter damage is often accompanied by reactive astrocyte hyperplasia and elevated proinflammatory cytokines [11, 38]. Our results also show that C5b-9 and C3b in serum ADEs may partially mediate the impact of SVD on cognitive impairment during the acute phase of AIS in T2DM patients. A larger infarct size has been identified as a more important risk factor for PSCI [39] than a smaller infarct size, and the concentration of activated complement protein after AIS was positively correlated with infarct volume [12]. The mediation analysis results of our study suggested that C5b-9 and C3b in serum ADEs may partially mediate the impact of infarct volume on cognitive impairment during the acute phase of AIS in T2DM patients.

Astrocyte can be activated to A1 type of reactive astrocytes in ischemic stroke, which promote the secretion of pro-inflammatory factors and the permeability of BBB and leads to neuroinflammation and cognitive deficits [14, 15]. In addition, diabetic rats with embolic middle cerebral artery occlusion (MCAO) induced greater neuroinflammation and cognitive impairment than control animal or 60 min MCAO rats [40]. The plasma level of complement C4 is statistically higher in diabetic stroke group than non-diabetic stroke group, and which is associated with patient prognosis of AIS independently [41]. Complement C3 level were also significantly higher in diabetes mice as compared to non-diabetes mellitus mice after transient middle cerebral artery occlusion (TMCAO) and which is seemed dependent on TLR2-NF-κB signaling [42]. Diabetes mellitus has been proven to promote the activation of pro-inflammatory immune cells and causes the body to be in chronic neuroinflammation state for a long time, and when diabetes patients suffered the secondary insult of AIS can exacerbates the inflammatory reaction within the brain [43, 44]. These results all suggest that diabetes magnify the complement cascade response after stroke in central nervous system. Extracellular vesicles (EVs) are double membrane-enclosed vesicles exuded into the extracellular space by astrocyte or other cells, which include exosomes with a size of 30–50 nm and microvesicles ranging in size from 100 to 1000 nm and their membranous nature can protect their cargo from degradation [45, 46]. Take together, complement proteins as one of cargo of EVs, the complement activation in exosomes of astrocyte may reflect the inflammatory reaction of central nervous system better than traditional methods of complement activation. Previous studies have been proven that astrocyte derived exosomes levels of C3b were significantly higher in AD patients than controls and which is also significantly higher at the later stage of AD with documented dementia than preclinical stage of AD and controls [19], which is consisted with the mixed linear regression result of our study that C3b in serum ADEs was negatively associated with MoCA scores at follow-up periods after AIS in T2DM patients. B4Crry as the site targeted inhibitor of C3 activation which can limit the chronic neuroinflammation and neurodegeneration after stroke in mice by inhibit complement cascade reaction of C3 invertase and its downstream products such as C3b [47]. In animal models, the complement system activated by AIS binds to ischemia-induced damage-related molecular patterns via IgM antibodies, while stressed neurons in the ischemic penumbra also exhibit damage-related molecular patterns, leading to the deposition of complement on the surface of stressed neurons in an IgM-dependent manner for microglial cells to phagocytose, which reduces the recovery probability of neurons and ultimately leads to cognitive dysfunction [20, 48]. Interestingly, complement protein Factor B due to over-activation of the complement system showed a lower level than normal subjects or downward trend in several neurodegenerative diseases including AD, Parkinson’s disease, and multiple sclerosis [49, 50], which is consistent with the results of the mixed effect model of our study showed that complement protein Factor B was positively associated with MoCA score during the follow-up periods after AIS in T2DM patients. The reason why there is no significant association between C5b9 in serum ADEs and MoCA score after AIS in T2DM may be that our study is a small-sample single-center study and the enrollment population was limited in patients with T2DM combined with AIS. Our team will explore the relationship between C5b-9 in serum ADEs and PSCI in the general AIS population in the future.

The strengths of the present study are that this is a cohort study and the first study to explore the relationships between complement proteins in serum ADEs and PSCI in T2DM patients. The complement activation in exosomes of astrocyte may reflect the inflammatory reaction of central nervous system better than traditional methods of complement activation. However, there are also some study limitations that should be addressed. First, this was a single-center study with a relatively small sample size, and the follow-up duration after AIS was only 12 months. Larger-scale, multicenter, longer follow-up studies are needed to identify the long-term effect of complement proteins in serum ADEs on PSCI in T2DM patients. Second, our findings were limited to AIS patients with T2DM; therefore, they cannot be generalized to all stroke patients. Third, we only recruited mild-AIS patients who were able to complete neuropsychological assessment and poststroke follow-up, and therefore could not assess the relationship between complement proteins in serum ADEs and PSCI with T2DM in severe AIS patients. Fourth, a total of 32 participants were lost during the follow-ups periods, which may affect the results of our study and we will take into account selection bias of lost follow-ups in future study. If the 32 patients with PSCI and high levels of complement proteins in serum ADEs, the true effect of complement proteins in serum ADEs on PSCI may be greater than the current study shows. The reasons that lead to patient lost follow-ups may include that personal experience and views on PSCI and the complexity of neuropsychological assessment scales and “perceived cost”, especial in terms of time spent [51].

Conclusion

The present study suggested that complement proteins in serum ADEs may be a partially mediating factor for OSA, SVD, depression, and infarction volume on cognitive impairment in the acute stage of stroke in T2DM patients and associated with MoCA scores at follow-up periods after AIS in T2DM patients.

AUTHOR CONTRIBUTIONS

Yaxuan Wu (Data curation; Writing – original draft); Ming Tan (Data curation; the extraction of complement proteins in serum astrocyte derived exosomes); Yanling Gao (Data curation); Geng Na (Data curation); Weibin Zhong (Data curation); Hairong Sun (Project administration); Zhenguang Li (Project administration); Chenxi Wu (Data curation); Xuemei Li (Supervision; conception and design of the study and instructor for data analysis and art); Jin-Biao Zhang (Supervision).

Footnotes

ACKNOWLEDGMENTS

We would like to thank all participants of this study for their important contributions.

FUNDING

The present study was supported by the Key Projects of Discipline Climbing Project of Weihai Municipal Hospital (FH-2021-XC034).

CONFLICT OF INTEREST

The authors have no conflict of interest to report.

DATA AVAILABILITY

To protect the privacy of individuals participating in this study, the data in this article cannot be shared publicly. The data will be shared at the reasonable request of the corresponding author.

The supplementary material is available in the electronic version of this article: .

References

, Cho

, Oh

, Jung

, Lee

, Shin

, Koh

, Cha

, Park

, Bae

, Kang

, Lee

(2013) Cognitive impairment evaluated with vascular cognitive impairment harmonization standards in a multicenter prospective stroke cohort in Korea. Stroke 44, 786–788.

Yan

, Yan

, Niu

, Wang

, Gui

, Zhang

(2017) Dl-3-n-butylphthalide can improve the cognitive function of patients with acute ischemic stroke: A prospective intervention study. Neurol Res 39, 337–343.

Tang

, Chen

, Lu

, Wong

, Mok

, Chu

, Ungvari

, Wong

(2011) Absence of cerebral microbleeds predicts reversion of vascular ‘cognitive impairment no dementia’ in stroke. Int J Stroke 6, 498–505.

, Ding

, Yang

, Bai

, Tu

, Liu

, Wang

, Tao

, Jin

, Wang

, Tang

(2014) The current situation on vascular cognitive impairment after ischemic stroke in Changsha. Arch Gerontol Geriatr 58, 236–247.

Ben Assayag

, Eldor

, Korczyn

, Kliper

, Shenhar-Tsarfaty

, Tene

, Molad

, Shapira

, Berliner

, Volfson

, Shopin

, Strauss

, Hallevi

, Bornstein

, Auriel

(2017) Type 2 diabetes mellitus and impaired renal function are associated with brain alterations and poststroke cognitive decline. Stroke 48, 2368–2374.

Zhu

, Ouyang

, Ma

, Wang

(2022) Obstructive sleep apnea is associated with cognitive impairment in minor ischemic stroke. Sleep Breath 26, 1907–1914.

Williams

, Demeyere

(2021) Association of depression and anxiety with cognitive impairment 6 months after stroke. Neurology 96, e1966–e1974.

Georgakis

, Fang

, Düring

, Wollenweber

, Bode

, Stösser

, Kindlein

, Hermann

, Liman

, Nolte

, Kerti

, Ikenberg

, Bernkopf

, Poppert

, Glanz

, Perosa

, Janowitz

, Wagner

, Neumann

, Speck

, Dobisch

, Düzel

, Gesierich

, Dewenter

, Spottke

, Waegemann

, Görtler

, Wunderlich

, Endres

, Zerr

, Petzold

, Dichgans

(2023) Cerebral small vessel disease burden and cognitive and functional outcomes after stroke: A multicenter prospective cohort study. Alzheimers Dement 19, 1152–1163.

Prodjohardjono

, Vidyanti

, Susianti

, Sudarmanta , Sutarni

, Setyopranoto

(2020) Higher level of acute serum VEGF and larger infarct volume are more frequently associated with post-stroke cognitive impairment. PLoS One 15, e0239370.

10.

, Sun

, Xue

, Leng

, Sun

, Shen

, Liu

, Li

, Shang

, Zhang

(2023) Complement protein levels in serum astrocyte-derived exosomes are associated with cognitive impairment in obstructive sleep apnea. J Clin Sleep Med 19, 727–739.

11.

Xie

, Shen

, He

, Chen

, Han

, Fang

, Zeng

, Chen

, Deng

(2016) IL-1β induces hypomyelination in the periventricular white matter through inhibition of oligodendrocyte progenitor cell maturation via FYN/MEK/ERK signaling pathway in septic neonatal rats. Glia 64, 583–602.

12.

Wang

, Hou

, Li

, Xu

, Ma

, Dong

, Yu

, Tan

(2020) Association of serum complement C1q concentration with severity of neurological impairment and infarct size in patients with acute ischemic stroke. J Stroke Cerebrovasc Dis 29, 105363.

13.

Rost

, Brodtmann

, Pase

, van Veluw

, Biffi

, Duering

, Hinman

, Dichgans

(2022) Post-stroke cognitive impairment and dementia. Circ Res 130, 1252–1271.

14.

Liu

, Chopp

(2016) Astrocytes, therapeutic targets for neuroprotection and neurorestoration in ischemic stroke. Prog Neurobiol 144, 103–120.

15.

Patabendige

, Singh

, Jenkins

, Sen

, Chen

(2021) Astrocyte activation in neurovascular damage and repair following ischaemic stroke. Int J Mol Sci 22, 4280.

16.

Pekny

, Pekna

, Messing

, Steinhäuser

, Lee

, Parpura

, Hol

, Sofroniew

, Verkhratsky

(2016) Astrocytes: A central element in neurological diseases. Acta Neuropathol 131, 323–345.

17.

Weng

, Lai

, Wang

, Zhuang

, Cheng

, Mo

, Liu

, Zhao

, Zhang

, Qiao

(2022) The role of exosomes as mediators of neuroinflammation in the pathogenesis and treatment of Alzheimer’s disease. Front Aging Neurosci 14, 899944.

18.

Chen

, Arumugam

, Cheng

, Lee

, Chigurupati

, Mattson

, Basta

(2018) Combination therapy with low-dose IVIG and a C1-esterase inhibitor ameliorates brain damage and functional deficits in experimental ischemic stroke. Neuromolecular Med 20, 63–72.

19.

Goetzl

, Schwartz

, Abner

, Jicha

, Kapogiannis

(2018) High complement levels in astrocyte-derived exosomes of Alzheimer disease. Ann Neurol 83, 544–552.

20.

Stephan

, Barres

, Stevens

(2012) The complement system: An unexpected role in synaptic pruning during development and disease. Annu Rev Neurosci 35, 369–389.

21.

Koton

, Pike

, Johansen

, Knopman

, Lakshminarayan

, Mosley

, Patole

, Rosamond

, Schneider

ALC

, Sharrett

, Wruck

, Coresh

, Gottesman

(2022) Association of ischemic stroke incidence, severity, and recurrence with dementia in the Atherosclerosis Risk in Communities Cohort Study. JAMA Neurol 79, 271–280.

22.

(1989) Stroke–1989. Recommendations on stroke prevention, diagnosis, and therapy. Report of the WHO Task Force on Stroke and other Cerebrovascular Disorders. Stroke 20, 1407–1431.

23.

, Jia

(2012) The Informant Questionnaire on Cognitive Decline in the elderly individuals in screening mild cognitive impairment with or without functional impairment. J Geriatr Psychiatry Neurol 25, 227–232.

24.

Webb

, Pendlebury

, Li

, Simoni

, Lovett

, Mehta

, Rothwell

(2014) Validation of the Montreal cognitive assessment versus mini-mental state examination against hypertension and hypertensive arteriopathy after transient ischemic attack or minor stroke. Stroke 45, 3337–3342.

25.

Zhang

, Chi

, Wang

, Zhang

, Shen

, Leng

, Sun

, Li

(2021) Altered amyloid-β and tau proteins in neural-derived plasma exosomes of type 2 diabetes patients with orthostatic hypotension. J Alzheimers Dis 82, 261–272.

26.

Chakraborty

, Hegde

, Praharaj

, Prabhu

, Patole

, Shetty

, Mayya

, Acharya

, Hande

, Prabhu

, Upadhya

(2021) Age related prevalence of mild cognitive impairment in type 2 diabetes mellitus patients in the Indian population and association of serum lipids with cognitive dysfunction. Front Endocrinol (Lausanne) 12, 798652.

27.

Alawieh

, Elvington

, Zhu

, Yu

, Kindy

, Atkinson

, Tomlinson

(2015) Modulation of post-stroke degenerative and regenerative processes and subacute protection by site-targeted inhibition of the alternative pathway of complement. J Neuroinflammation 12, 247.

28.

Aaronson

, van Bennekom

, Hofman

, van Bezeij

, van den Aardweg

, Groet

, Kylstra

, Schmand

(2015) Obstructive sleep apnea is related to impaired cognitive and functional status after stroke. Sleep 38, 1431–1437.

29.

Liu

, Ma

, Ouyang

, Zeng

, Zhan

, Lu

, Cui

, Dai

, Luo

, He

, Li

, Zong

, Chen

(2020) The relationship between inflammation and neurocognitive dysfunction in obstructive sleep apnea syndrome. J Neuroinflammation 17, 229.

30.

Dewan

, Nieto

, Somers

(2015) Intermittent hypoxemia and OSA: Implications for comorbidities. Chest 147, 266–274.

31.

Feng

, Tripathi

, Pillai

(2020) Inflammatory pathways in psychiatric disorders: The case of schizophrenia and depression. Curr Behav Neurosci Rep 7, 128–138.

32.

Ishii

, Hattori

, Miyakawa

, Watanabe

, Hidese

, Sasayama

, Ota

, Teraishi

, Hori

, Yoshida

, Nunomura

, Nakagome

, Kunugi

(2018) Increased cerebrospinal fluid complement C5 levels in major depressive disorder and schizophrenia. Biochem Biophys Res Commun 497, 683–688.

33.

Poon

, Ho

, Chiu

, Wong

, Chang

(2015) Sickness: From the focus on cytokines, prostaglandins, and complement factors to the perspectives of neurons. Neurosci Biobehav Rev 57, 30–45.

34.

, Xu

, Lu

, Peng

, Chen

, Bi

, Li

, Chen

, Li

, Jin

, Hu

(2021) Morinda officinalis oligosaccharides alleviate depressive-like behaviors in post-stroke rats via suppressing NLRP3 inflammasome to inhibit hippocampal inflammation. CNS Neurosci Ther 27, 1570–1586.

35.

Chen

, Akinyemi

, Hase

, Firbank

, Ndung’u

, Foster

, Craggs

, Washida

, Okamoto

, Thomas

, Polvikoski

, Allan

, Oakley

, O’Brien

, Horsburgh

, Ihara

, Kalaria

(2016) Frontal white matter hyperintensities, clasmatodendrosis and gliovascular abnormalities in ageing and post-stroke dementia. Brain 139, 242–258.

36.

Ding

, Xu

, Wang

, Li

, Mao

, Yu

, Cui

, Dong

(2019) Predictors of cognitive impairment after stroke: A prospective stroke cohort study. J Alzheimers Dis 71, 1139–1151.

37.

Zhang

, Pan

, Mamtilahun

, Zhu

, Wang

, Venkatesh

, Shi

, Tu

, Jin

, Wang

, Zhang

, Yang

(2020) Microglia exacerbate white matter injury via complement C3/C3aR pathway after hypoperfusion. Theranostics 10, 74–90.

38.

Martinez

Sosa S

, Smith

(2017) Understanding a role for hypoxia in lesion formation and location in the deep and periventricular white matter in small vessel disease and multiple sclerosis. Clin Sci (Lond) 131, 2503–2524.

39.

Puy

, Barbay

, Roussel

, Canaple

, Lamy

, Arnoux

, Leclercq

, Mas

, Tasseel-Ponche

, Constans

, Godefroy

(2018) Neuroimaging determinants of poststroke cognitive performance. Stroke 49, 2666–2673.

40.

Ward

, Valenzuela

, Li

, Dong

, Fagan

, Ergul

(2018) Poststroke cognitive impairment and hippocampal neurovascular remodeling: The impact of diabetes and sex. Am J Physiol Heart Circ Physiol 315, H1402–H1413.

41.

Zhang

, Yin

, Shao

, Yang

, Liu

, Wang

, Diao

, Huang

, Xue

, Ni

, Yang

(2021) High serum complement component C4 as a unique predictor of unfavorable outcomes in diabetic stroke. Metab Brain Dis 36, 2313–2322.

42.

Lin

, Lin

, Li

, Huang

, Dai

(2018) Complement component C3 promotes cerebral ischemia/reperfusion injury mediated by TLR2/NFκB activation in diabetic mice. Neurochem Res 43, 1599–1607.

43.

Jackson

, Dong

, Althomali

, Sayed

, Eldahshan

, Baban

, Johnson

, Filosa

, Fagan

, Ergul

(2020) Delayed administration of angiotensin II type 2 receptor (AT2R) agonist compound 21 prevents the development of post-stroke cognitive impairment in diabetes through the modulation of microglia polarization. Transl Stroke Res 11, 762–775.

44.

, Wang

, Dai

, Xu

, Gao

, Johnson

, Xu

, Leak

, Hu

, Luo

, Chen

(2018) Diabetes mellitus impairs white matter repair and long-term functional deficits after cerebral ischemia. Stroke 49, 2453–2463.

45.

Vidal

(2019) Exosomes: Revisiting their role as “garbage bags”. Traffic 20, 815–828.

46.

Thompson

, Gray

, Heman-Ackah

, Mäger

, Talbot

, Andaloussi

, Wood

, Turner

(2016) Extracellular vesicles in neurodegenerative disease –pathogenesis to biomarkers. Nat Rev Neurol 12, 346–357.

47.

Alawieh

, Langley

, Tomlinson

(2018) Targeted complement inhibition salvages stressed neurons and inhibits neuroinflammation after stroke in mice. Sci Transl Med 10eaao6459.

48.

Narang

, Qiao

, Atkinson

, Zhu

, Yang

, Kulik

, Holers

, Tomlinson

(2017) Natural IgM antibodies that bind neoepitopes exposed as a result of spinal cord injury, drive secondary injury by activating complement. J Neuroinflammation 14, 120.

49.

Finehout

, Franck

, Lee

(2005) Complement protein isoforms in CSF as possible biomarkers for neurodegenerative disease. Dis Markers 21, 93–101.

50.

Krance

, Wu

, Zou

, Mao

, Toufighi

, He

, Pakosh

, Swardfager

(2021) The complement cascade in Alzheimer’s disease: A systematic review and meta-analysis. Mol Psychiatry 26, 5532–5541.

51.

Coley

, Coniasse-Brioude

, Igier

, Fournier

, Poulain

, Andrieu

(2021) Disparities in the participation and adherence of older adults in lifestyle-based multidomain dementia prevention and the motivational role of perceived disease risk and intervention benefits: An observational ancillary study to a randomised controlled trial. {Alzheimers Res Ther 13, 157.