Type 3 diabetes (T3D) accurately reflects that dementia, e.g., Alzheimer’s disease, represents insulin resistance and neurodegeneration in the brain. Similar retinal microvascular changes were observed in Alzheimer’s and chronic stressed individuals. Hence, we aimed to show that chronic stress relates to T3D dementia signs and retinopathy, ultimately comprising a Stress syndrome prototype reflecting risk for T3D and stroke. A chronic stress and stroke risk phenotype (Stressed) score, independent of age, race or gender, was applied to stratify participants (N = 264; aged 44 – 9 years) into high stress risk (Stressed, N = 159) and low stress risk (non-Stressed, N = 105) groups. We determined insulin resistance using the homeostatic model assessment (HOMA-IR), which is interchangeable with T3D, and dementia risk markers (cognitive executive functioning (cognitiveexe-func); telomere length; waist circumference (WC), neuronal glia injury; neuron-specific enolase/NSE, S100B). Retinopathy was determined in the mydriatic eye. The Stressed group had greater incidence of HOMA-IR in the upper quartile (≥5), larger WC, poorer cognitiveexe-func control, shorter telomeres, consistently raised neuronal glia injury, fewer retinal arteries, narrower arteries, wider veins and a larger optic cup/disc ratio (C/D) compared to the non-Stressed group. Furthermore, of the stroke risk markers, arterial narrowing was related to glaucoma risk with a greater C/D, whilst retinal vein widening was related to HOMA-IR, poor cognitiveexe-func control and neuronal glia injury (Adjusted R2 0.30; p ≤ 0.05). These associations were not evident in the non-Stressed group. Logistic regression associations between the Stressed phenotype and four dementia risk markers (cognitiveexe-func, telomere length, NSE and WC) comprised a Stress syndrome prototype (area under the curve 0.80; sensitivity/specificity 85%/58%; p ≤ 0.001). The Stress syndrome prototype reflected risk for HOMA-IR (odds ratio (OR) 7.72) and retinal glia ischemia (OR 1.27) and vein widening (OR 1.03). The Stressed phenotype was associated with neuronal glia injury and retinal ischemia, potentiating glaucoma risk. The detrimental effect of chronic stress exemplified a Stress syndrome prototype reflecting risk for type 3 diabetes, neurodegeneration and ischemic stroke.
The Sympathetic activity and Ambulatory Blood Pressure in Africans (SABPA) prospective study was used as the data source for the present analysis. Methods are well-described elsewhere [26]. Urban-dwelling Black and White teachers with similar socio-economic status and who had access to medical aid benefits were under investigation. A complete dataset was compiled of individuals participating at both baseline and the 3-year follow-up phase, and seasonal changes were avoided (Figure 1). The Sympathetic activity and Ambulatory Blood Pressure in African (SABPA) prospective cohort stratified into Stressed and non-Stressed groups. Exclusion criteria at baseline were pregnancy, lactation, tympanum temperature ≥37.5 °C, the use of psychotropic substances or α- and β-blockers and blood donors or individuals vaccinated within 3 months prior to data collection. A 24-h ambulatory blood pressure and ECG monitoring apparatus, the Cardiotens CE120® (Meditech, Budapest, Hungary), was fitted between 07:00 and 09:00 on working days (Monday–Thursday) at the teachers’ schools (Supplementary Methodology). The 24-h ECG data were used to determine time domain heart rate variability (HRV) (Supplementary Methodology). A standardized 24-h diet commenced and participants resumed their normal school and extra-curricular activities. Thereafter, they were transported to the North-West University for extensive clinical assessments to be performed under well-controlled conditions [26]. Each participant received his/her own private bedroom for an overnight stay at the North-West University guest house facilities. Demographic and General Health questionnaires were completed. Participants took a last beverage before 22:00 (coffee/tea and biscuits) and were asked to fast from 22:00 to 07:00. The next morning at 07:00, anthropometric measures were taken, followed by carotid intima media sonar scanning to determine bifurcation stenosis (Supplementary Methodology). Blood sampling was performed and physical activity measures proceeded before being transported to school. Feedback reports and referral letters to doctors were given to each participant within 7 days. Retinal vessel imaging considering arteriolar and venular calibers (hereafter referred to as arteries and veins) was performed at 3-year follow-up (Figure S1a,b). Food and alcohol intake, smoking and exercise were prohibited one hour prior to measurements. Fundus imaging history and voluntary consent were obtained after participants were introduced to the procedure. A trained registered nurse screened participants for acute anterior angle chamber glaucoma risk using a small light source. Mydriasis was induced in the right eye of 262 (99.2%) individuals by means of a drop containing tropicamide, 1%, and benzalkonium chloride 0.01% (m/v). Fundus imaging was performed with a retinal vessel analyzer with a Zeiss FF450Plus camera and the software VesselMap 2, Version 3.02 (Imedos Systems GmbH, Jena, Germany) (Figure S1a), which automatically determined the retinal artery and vein count. Retinal vessel calibers were measured as monochrome images by manually selecting first-order vessel branches in a measuring zone located between 0.5 and 1.0 optic disc diameters from the margin or the optic disc. After selection of the vessel, software automatically delineated the vessels’ measuring area. The Knudtson formula was used to determine estimates from the 6 largest arteries and veins. As the image scale of each eye was unknown, the values of the retinal arteries and veins were expressed as measuring units (MU). One measuring unit is equivalent to 1 micrometer when the dimensions of the eye being examined correspond to those of the normal Gullstrand eye. Reproducibility was computed for a randomly selected cohort with a correlation coefficient of 0.84. Diastolic ocular perfusion pressure was measured to indicate hypo-perfusion risk. Hypo-perfusion was evaluated after instilling a local anesthetic drop (Novasine Wander 0.4% Novartis) and measuring intra-ocular pressure (IOP) with the Tono-Pen Avia Applanation Tonometer (Reichert 7-0908, ISO 9001, New York, NY, USA). Right-eye diastolic ocular perfusion pressure (mmHg) was calculated (office diastolic blood pressure minus intra-ocular pressure). Clinical observations for diabetic and hypertensive retinopathy were evaluated on color stereo optic disc photographs by an experienced and qualified ophthalmologist blinded to the group. Evaluations included, e.g., determination of the optic cup/disc ratio, arterio-venous nicking (AV nicking) and focal narrowing. AV nicking indicates indentation of retinal veins by stiff (arteriosclerotic) retinal arteries as a sign of chronic hypertension and vascular dysregulation [27]. Focal narrowing of retinal arteries increases with progressive glaucomatous optic neuropathy and may include thinning of the neuroretinal rim area when optic nerve head damage is apparent [28]. Accumulating evidence suggests that cognitive executive functioning control or the Stroop test [29] elicits attentional control mechanisms as an early marker for the transition from healthy cognitive aging to early-stage dementia of the Alzheimer type in the frontotemporal cortex areas [30]. For the current investigation, we used the Stroop test as an early dementia risk marker to indicate cognitive dysfunction [30,31]. The Stroop test (Figure S1c,d) assesses attentional processing of simultaneously occurring sensory information in the context of selective attention, cognitive set shifting and response inhibition in the prefrontal cortex. A template containing five words printed in highly distinguishable colors (“blue”, “green”, “red”, “yellow”) in random order but written in incongruent colors was shown to participants (Figure S1d). The ink color of a given word had to be identified verbally and individuals were encouraged to progress as fast as possible within 1 min and were corrected when wrong answers were given. Prior to the test, they were informed that they will receive a monetary incentive according to their performance on completion of the test, which served as motivation to improve performance. An interference score was calculated that represents the number of correct answers produced during the fixed period of 1 min. A lower score indicates that the individual found it more difficult to inhibit interference. Two trained scientists (registered nurse and MD) were involved in supervision of the Stroop test and scoring of all teachers at baseline. Perception of the stressfulness of the Stroop test was assessed on a 7-point Likert scale (Supplementary Methodology). Waist circumference (WC) as marker of central obesity was measured in triplicate by registered level II anthropometrists (N = 2) and the mean was used for statistical analyses. A non-extensible flexible anthropometric tape was used to measure WC at the midpoint between the lower costal rib and the iliac crest, perpendicular to the long axis of the trunk, and not at the narrowest point, for standardization purposes. The intra- and inter-variability was less than 5%. Mean total energy expenditure over 7 days was derived from an Actiheart accelerometer (GB06/67703; CamNtech Ltd., Upper Pendrill Court, Papworth Everard, Cambridgeshire CB233UY, UK). Participants were in a semi-recumbent position for at least 30 min before blood sampling in both study phases. A registered nurse obtained all fasting blood samples before 09:00 from the antebrachial vein branches of the dominant arm of each participant using a winged infusion set. All blood samples were obtained from never-thawed serum/plasma/citrate samples, handled according to standardized procedures and stored at −80 °C until analyses. The following dementia-related risk markers indicating cardiometabolic perturbations [1,2,3,4,5,6], neurodegeneration [1,2,3,4,5,6,7] and vascular dysregulation [14,15,27] were specified and included in the current investigation (in no particular order): poorer cognitive executive functioning control [30]; neuronal glia injury (increased S100B and NSE) [19]; increased central obesity or waist circumference (WC) [32]; endothelial dysfuntion (increased von Willebrand factor (VWF)) [33]; depressed time domain heart rate variability (HRV) (Supplementary Methodology) [34]; increased insulin resistance/HOMA-IR [8] and inflammation (C-reactive protein/CRP) [35]; and shorter telomeres [21]. Sodium fluoride glucose, serum levels of insulin, triglycerides, total cholesterol:high-density lipoprotein, cotinine (nicotine metabolite as indicator of smoking habits [36], gamma-glutamyl-transferase/GGT (as marker of alcohol consumption [37] and ultra-sensitive C-reactive protein (CRP) were analyzed with an enzyme rated method (Unicel DXC 800-Beckman and Coulter, 4300 N. Harbor Blvd., Fullerton, CA 92835 U.S.A), homogeneous immunoassay (Modular ROCHE Automized systems, Basel, Switzerland) and a particle-enhanced turbidimetric assay (Cobas Integra 400 plus, Roche, Basel, Switzerland), respectively. A dysregulated hypothalamic–pituitary–adrenal cortex axis (HPA axis) contributes to cognitive decline and dysfunctional growth hormones [38], e.g., serum total insulin-like growth factor 1 (IGF-1) and insulin-like growth factor binding protein 3 (IGFBP-3). IGF-1 and IGFBP-3 further protect nerve cells against neurodegenerative processes [39] and were determined using immunoradiometric assays from Immunotech, Beckman Coulter, Brea, California: Catalogue no. A15729 (inter-assay % CV 4.49; intra-assay % CV 2.92) and Catalogue no. DSL-6600 (inter-assay % CV 3.2–9.3; and intra-assay % CV 2.71–7.95), respectively. Whole-blood EDTA glycated hemoglobin (HbA1C) was analyzed with turbidimetric inhibition immunoassays (Cobas Integra 400 Plus, ROCHE Basel, Switzerland). The American Diabetes Association guidelines were used to define pre-diabetes (HbA1C ≥ 5.7%) and diabetes status (HbA1C ≥ 6.5%) [10]. The homeostatic model assessment of insulin resistance (HOMA-IR) was determined using the following formula: fasting glucose × fasting insulin/405. The median (interquartile range) HOMA-IR result was 2.9 (1.8–4.7), and Q4 was at least 4.7. HOMA estimates β-cell function and insulin sensitivity based only on fasting glucose and insulin concentrations and gives a representation of insulin secretion adjusted for insulin sensitivity [40]. Citrate VWF antigen concentration as a marker of endothelial dysfunction, coagulation activation and ischemic stroke risk [41] was measured with a “sandwich” ELISA assay. A polyclonal rabbit anti-VWF antibody and a rabbit anti-VWF-HRP antibody (DAKO, Bloemfontein, Free State, South Africa) were used to form the assay. The 6th International Standard for VWF/FVIII was used to set the standard curve against which the samples were measured. Serum NSE and S100B were analyzed with an electrochemiluminescence immunoassay (e411, ROCHE, Basel, Switzerland) with intra- and inter-assay coefficients of less than 5%. Mitochondrial DNA (mtDNA) variation utilizing the MutPred program in the SABPA study [42] was determined to establish an association between chronic stress and maternal lineage. To determine leukocyte telomere length (LTL), genomic DNA was extracted. Reference DNA samples were prepared, and all isolated DNAs were mixed together in equal proportions, representing the average of all analyzed SABPA patients (N = 255). All experimental DNA samples were assayed in triplicate and PCRs were performed with the CFX96 Touch™ Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA) in a 25-µL volume (Supplementary Methodology) and stored at −20 °C. HIV-positive status (N = 10) was considered in regression analyses (Supplementary Methodology). Statistica version 13.3 (TIBCO Software Inc., Palo Alto, Santa Clara, CA, USA, 2018) was used for data analyses. We proceeded from previous investigations in the SABPA study [14,15] by applying a validated chronic stress and stroke risk phenotype (hereafter Stressed) score, independent of age, race or gender (see Section 6). Participants were stratified into high stress risk (Stressed, N = 159) and low stress risk (non-Stressed, N = 105) groups. The phenotype biomarkers for the current investigation were not included in any statistical models. All variables with non-normal distributions were Box–Cox transformed except for age, cognitiveexe-func, diastolic ocular perfusion pressure, artery count and bifurcation stenosis. Independent t-tests were used to compare clinical characteristics of the Stressed vs. non-Stressed groups whilst comparisons of proportions and prevalence were tested by using chi-square (χ2) tests. Dementia risk markers were compared between Stressed vs. non-Stressed groups using single one-way ANCOVAs, adjusted for age. Mean changes over 3 years in dementia risk markers were calculated using dependent sample t-tests within the Stressed and non-Stressed groups. McNemar’s case–control tests were performed to compare incidence and recovery frequencies of HOMA-IR over 3 years within the Stressed and non-Stressed groups. Pearson correlations were determined between cognitiveexe-func, dementia risk markers (telomere length, NSE, S100B, WC, HOMA-IR, VWF, CRP, standard deviation of normal-to-normal (NN) intervals (SDNN)) and perception of stressfulness of the Stroop test. Multiple linear regression analyses determined associations between retinal vessel calibers, dementia and retinopathy risk markers within the Stressed and non-Stressed groups at baseline and 3-year follow-up. Dependent variables in Models 1 and 2 were retinal artery and venous calibers obtained at 3-year follow-up. Independent variables included baseline age, dementia risk markers (cognitiveexe-func, NSE, S100B, HOMA-IR, CRP) and dyslipidemia as well as retinal risk marker data obtained at follow-up (including diastolic ocular perfusion pressure, optic nerve cup-to-disc ratio, retinal artery count and carotid bifurcation stenosis). When the retinal artery was a dependent variable, the retinal venous diameter was included as a predictor and vice versa. The F value to enter in regression models was fixed at 2.5. A logistic regression analysis was performed, and the chronic stress and stroke risk phenotype (Stressed) was used as a dichotomous dependent variable to develop a Stress syndrome prototype to maximize the clinical utility of stress risk. The prototype development was scientifically grounded and conducted empirically to increase clinical usefulness [43]. Baseline dementia risk markers (cognitiveexe-func, Box–Cox transformed markers, telomere length, NSE, S100B, WC, VWF, CRP and time domain 24-h HRV) were used as continuous predictors. An additional logistic regression analysis was performed to determine the association between Stressed and Box–Cox transformed mitochondrial DNA MutPred load, as a marker of maternal lineage [42]. Only significant Stressed-related dementia risk marker associations were used to develop the novel Stress syndrome prototype score using a receiver operating characteristic (ROC) area under the curve analysis. The statistical significance level was set at p ≤ 0.05 (two-tailed). A final logistic regression analysis computation was used to determine probability of risk for the Stress syndrome by including continuous predictors, i.e., Box–Cox transformed insulin resistance (HOMA-IR), S100B, retinal artery and vein caliber/AV nicking, retinal artery count, optic nerve cup/disc ratio, carotid bifurcation stenosis, diastolic ocular perfusion pressure and dyslipidemia. The statistical significance level was set at p ≤ 0.05 (two-tailed). Statistical significance level was set at p ≤ 0.05 (two-tailed).
DIMA AI Care
