Plasma oxysterols are associated with serum lipids and dementia risk in older women

Abstract INTRODUCTION Apolipoprotein E4 (APOE4) carriers’ tendency toward hypercholesterolemia may contribute to Alzheimer's disease (AD) risk through oxysterols, which traverse the blood‐brain barrier. METHODS Relationships between baseline plasma oxysterols, APOE status, serum lipids, and cognitive impairment risk were examined in 328 postmenopausal women from the Women's Health Initiative Memory Study. Women were followed for 25 years or until incident dementia or cognitive impairment. RESULTS Levels of 24(S)‐hydroxycholesterol (24‐OHC), 27‐hydroxycholesterol (27‐OHC), and 24‐OHC/27‐OHC ratio did not differ by APOE status (p’s > 0.05). Higher 24‐OHC and 27‐OHC were associated with higher total, low density lipoprotein (LDL), non‐high density lipoprotein (HDL), remnant, LDL/HDL, and total/HDL cholesterol and triglycerides (p’s < 0.05). Higher 24‐OHC/27‐OHC was associated with greater dementia risk (hazard ratio = 1.51, 95% confidence interval:1.02‐2.22), which interaction analyses revealed as significant for APOE3 and APOE4+, but not APOE2+ carriers. DISCUSSION Less favorable lipid profiles were associated with higher oxysterol levels. A higher ratio of 24‐OHC/27‐OHC may contribute to dementia risk in APOE3 and APOE4+ carriers.

DISCUSSION: Less favorable lipid profiles were associated with higher oxysterol levels.A higher ratio of 24-OHC/27-OHC may contribute to dementia risk in APOE3 and APOE4+ carriers.

BACKGROUND
The Apolipoprotein E ε4 (APOE4) allele is the main genetic risk factor for Alzheimer's disease (AD). 1,2APOE4 carriers often have unfavorable lipid profiles, increasing their likelihood of hypercholesterolemia. 3,4 Midlife hypercholesterolemia is associated with heightened AD risk, 5 suggesting that APOE's involvement in lipid metabolism may contribute to AD risk.The underlying mechanisms are unclear, however, as blood cholesterol does not permeate the blood-brain barrier. 6e literature suggests that oxysterols, which are oxidized products of cholesterol, may be an intermediary between hypercholesterolemia and AD due to their ability to cross the blood-brain barrier. 6,7][8][9][10] 24-OHC is produced in the brain and serves as a primary route for excess brain cholesterol to exit into the periphery. 11,1224-OHC appears to prevent AD pathology [13][14][15][16] but may be harmful at high concentrations. 17[20][21] In contrast to 24-OHC, 27-OHC promotes AD pathology 9,10,16,22,23 and may even inhibit the protective effects of 24-OHC. 21,24Compared to controls, AD patients have lower 24-OHC and higher 27-OHC levels in the brain, 6,25,26 and higher levels of both in cerebrospinal fluid. 27,28Literature on plasma oxysterols in AD, however, is less consistent, 6 potentially due to changing levels throughout the disease course. 29ven that AD pathology begins years before symptom onset, 30 departure from the normal range of oxysterol levels may be most noticeable before AD diagnosis.Investigations of plasma 24-OHC and 27-OHC preceding diagnosis are therefore needed to clarify their involvement in early disease processes.Additionally, examinations of plasma oxysterols in relation to blood lipids prior to AD incidence are needed to corroborate speculations that oxysterols are a mechanistic link between hypercholesteremia and AD. 6,7,10We previously found that less favorable serum lipid profiles were associated with greater risk for dementia and cognitive impairment in the Women's Health Initiative (WHI) Memory Study (WHIMS 31,32 ). 33In the present study, we sought to determine the potential involvement of oxysterols in lipid-dementia associations by investigating relationships of plasma oxysterols with APOE, serum lipids, and incident dementia and cognitive impairment in WHIMS.

Participants
WHIMS, which evaluated the effects of hormone therapy (HT) on incident dementia and cognitive impairment in postmenopausal women, 31,32 is an ancillary study of the WHI clinical trial of HT. 34 The current sample included WHIMS participants with available plasma oxysterol measurements from baseline blood draw (N = 328), previously assayed by another WHI ancillary study. 35A total of 281 participants had baseline serum lipids, and 264 were genotyped for

APOE. Participants were cognitively unimpaired at baseline based on
Modified Mini-Mental State Examination (3MS) scores. 31,32All participants provided written informed consent, and approval was obtained from Institutional Review Boards of participating sites.

Measures
6][37] Briefly, plasma oxysterol concentrations (24-OHC and 27-OHC; ng/mL) were measured at the University of Texas Southwestern Medical Center, Department of Molecular Genetics (Dallas, TX) using high-performance liquid chromatography-mass spectrometry (HPLC-MS). 36,37Assays were performed using a tertiary Shimadzu LC-20XR HPLC system (Shimadzu Scientific Instruments; Columbia, MD) coupled to an AB Sciex API-5000 triple quadrupole MS equipped with Turbo V ESI source (Foster City, CA). 36The average interassay coefficients of variability (CVs) of 24

Statistical analysis
We classified participants as APOE2+ (ε2/ε2, ε2/ε3), APOE3 (ε3/ε3), or APOE4+ (ε3/ε4, ε4/ε4) carriers for analyses.APOE ε2/ε4 carriers (N = 5) [3][4] Multiple linear regression was performed to investigate associations of plasma oxysterols with APOE (APOE3 as reference) and serum lipids.Regression models were initially adjusted for age, BMI, WHI HT assignment, and cholesterol-lowering medication, followed by further adjustment for education, smoking, alcohol intake, hyperten- diabetes, stroke, total serum cholesterol, serum triglycerides, total dietary cholesterol intake, total calorie intake, total fat intake, total protein intake, and total carbohydrate intake.Power calculations were performed for all analyses due to the relatively small sample size (details and results [Tables A1 and A2] can be found in Appendix A).
An increased risk for dementia and cognitive impairment was previously reported in relation to estrogen and estrogen plus progestin therapy in WHIMS. 31,32We therefore performed supplementary analyses of all associations stratified by HT assignment to assess for possible effect modification by hormone therapy.

Sample characteristics
After a mean follow-up of 9.

Relationships between oxysterols, APOE, and blood lipids
Plasma oxysterols did not significantly differ by APOE status (p's > 0.11), albeit higher in APOE4+ and APOE3 compared to APOE2+ carriers (Table A3 and Figure

Supplementary analysis
Results from supplementary analyses are detailed in Appendix A (Tables A4-A6).Briefly, associations of oxysterols with APOE and serum lipids were largely similar to those in the entire sample, regardless of HT.There were some differences in associations between oxysterols and cognitive impairment based on HT groups, but many models failed to converge and these findings should be interpreted with caution given the limited statistical power due to small sample sizes.

DISCUSSION
In this prospective cohort of postmenopausal women, we report: (1)   higher plasma 24-OHC and 27-OHC in relation to higher lipid levels, and (2) greater dementia and cognitive impairment risk in relation to a higher ratio of 24-OHC/27-OHC in APOE4+ and APOE3 carriers.
The present study is not alone in reporting a lack of variation in oxysterols by APOE status.No APOE-related differences in serum 24-OHC were reported in a sample of 80 AD patients, 43 although another sample of 53 AD patients found lower plasma 24-OHC in APOE4 compared to noncarriers. 29Literature on oxysterol-APOE relationships remains limited, especially in cognitively unimpaired older adults.Our sample, while one of the larger examined to date, is still relatively small for genetic analyses.Investigations in much larger samples will be necessary to understand APOE-related variation in oxysterols with some certainty.
The associations between less favorable lipids and higher 24-OHC and 27-OHC are consistent with some of the literature, 6,9,19,20,21,44 although a recent meta-analysis of case-control studies reported no differences in 24-OHC between MCI (n = 260) or AD (n = 509) patients and controls (n = 438 and n = 428, respectively). 8Higher dementia risk observed in those with a higher 24-OHC/27-OHC ratio is in agreement with the marginally significant findings of increased cognitive impairment risk reported by Hughes et al., but, in contrast, we did not observe a relationship between 24-OHC levels and dementia risk in our exclusively female sample. 45Nonetheless, our study is the first to report significantly greater dementia risk in relation to higher 24-OHC/27-OHC, and specifically in APOE3 and APOE4+ carriers.While this finding could reflect a protective effect of a higher 24-OHC/27-OHC ratio against dementia among APOE2+ carriers, the consistent reports of the deleterious effects of altered oxysterol levels [6][7][8][9][10]16,17,[21][22][23][24] suggest a greater likelihood that APOE ε2 allele may instead be protective against pathological effects of abnormal oxysterol levels. Nevertheles, this interaction between oxysterols and APOE in relation to cognitive impairment should be interpreted with caution given the small number of cases among APOE2+ carriers.We did not observe greater dementia risk for APOE4+ compared to APOE3 carriers in relation to 24-OHC/27-OHC ratio, which also raises questions and further necessitates examination in larger samples to clarify APOE-oxysterol relationships.
Overall, our results align with discussions in the literature regarding alterations in oxysterol levels along the AD trajectory.Hypercholesterolemia likely increases 27-OHC levels, 6,7,12,[19][20][21]44 which is consistent with the positive association between lipids and 27-OHC in this cohort. Hig amounts of 27-OHC may then enter the brain, promote AD pathology, and drive out 24-OHC-resulting in subsequent elevation of 24-OHC/27-OHC levels in the peripheral circulation.6,7,9,10,16,18,20,21,24 WHIMS participants were ≥65 years at the time of oxysterol collection, by which time those who were hypercholesterolemic in midlife would presumably have both a higher ratio of plasma 24-OHC/27-OHC and greater dementia risk-a hypothesis that seems consistent with our findings.If the observed increase in age is in fact a reflection of this proposed sequence of events, our findings may be in agreement with prior reports of 27-OHC as an instigator of AD pathology and an inhibitor of 24-OHC's ability to safeguard against AD pathology.6,7,9,10,16,[21][22][23][24] To confirm whether abnormal oxysterol levels serve as a mechanistic link between hypercholesterolemia and AD will require investigations in larger samples with available midlife lipid levels, subsequent oxysterol levels, and AD incidence.
If our findings can be confirmed in larger samples, this would suggest the possibility of dementia risk modification via oxysterol management.0][21] Vascular risk management including dietary and pharmacological interventions appears to reduce 27-OHC and 24-OHC blood levels, 43,48  Note: Hazard ratios of oxysterol levels were calculated per 1 ng/mL increase in 24-OHC and 27-OHC levels, and per 0.1 unit increase in 24-OHC/27-OHC ratio.
prerandomization plasma sample. 35Although this previous study's selection criteria limited our sample, fracture risk in the current study did not differ by cognitive outcomes (Table 1) and therefore should not have influenced our findings.WHIMS assessed for dementia due to all causes. 31,32While AD was the most common dementia classification, 31,32 the potential inclusion of several vascular dementia cases (the second most common cause of dementia after AD 30,49 ) could have led to underestimation of the associations between oxysterols and cognitive outcomes.Additionally, this sample did not have midlife lipid levels, preventing us from analyzing prior hypercholesterolemia in relation to subsequent oxysterol levels.Nonetheless, WHIMS is a well-screened, prospectively followed sample.To our knowledge, this is the first study to comprehensively examine plasma oxysterols in older adults in relation to APOE, a large lipid panel, and adjudicated incident dementia and cognitive impairment.
Overall, higher 24-OHC and 27-OHC are associated with less favorable lipid profiles, and APOE3 and APOE4+ carriers with a higher 24-OHC/27-OHC ratio may be at greater risk for dementia and cognitive impairment.These findings offer support for the involvement of 24-OHC and 27-OHC in dementia, motivating further research into the clinical significance of oxysterols as modifiable factors contributing to AD risk.Whether AD risk reduction is possible through the maintenance of optimal oxysterol levels warrants further investigation.
APOE3, and 59 (22%) APOE4+ carriers.Baseline characteristics are reported in Table 1.Histograms showing the distribution of plasma oxysterols are displayed in Figure A1 (Appendix A).
a Age, BMI, cholesterol-lowering medication, and HT trial assignment.b TA B L E 4 Interactions between oxysterols and APOE in relation to dementia risk