Sage Journals: Discover world-class research

Abstract

Age-related memory loss shares similar risk factors as cardiometabolic diseases including elevated serum triglycerides (TGs) and low-density lipoprotein cholesterol (LDL-C) and reduced high-density lipoprotein cholesterol (HDL-C). The mechanisms linking these aberrant blood lipids to memory loss are not completely understood but may be partially mediated by reduced integrity of the hippocampus (HC), the primary brain structure for encoding and recalling memories. In this study, we tested the hypothesis that blood lipid markers are independently associated with memory performance and HC viscoelasticity—a noninvasive measure of brain tissue microstructural integrity assessed by high-resolution magnetic resonance elastography (MRE). Twenty-six individuals across the adult lifespan were recruited (14 M/12 F; mean age: 42 ± 15 y; age range: 22–78 y) and serum lipid profiles were related to episodic memory and HC viscoelasticity. All subjects were generally healthy without clinically abnormal blood lipids or memory loss. Episodic memory was negatively associated with the TG/HDL-C ratio. HC viscoelasticity was negatively associated with serum TGs and the TG/HDL-C ratio, independent of age and in the absence of associations with HC volume. These data, although cross-sectional, suggest that subtle differences in blood lipid profiles in healthy adults may contribute to a reduction in memory function and HC tissue integrity.

Keywords

Cognitive impairment/decline MRI hippocampus lipids brain imaging

Introduction

Advancing age is the primary risk factor for Alzheimer’s disease (AD), which is the most common form of dementia and is characterized by neurodegeneration and irreversible memory loss. As the number of older adults in developed societies continues to increase, the number of AD cases is projected to triple by 2050 if preventive measures are not undertaken.¹ Age-related cognitive decline shares many of the same modifiable risk factors as cardiometabolic diseases, especially when they emerge earlier in life, prior to the development of clinically relevant memory loss.² In this regard, elevated blood triglycerides (TGs) and low-density lipoproteins (LDL-C) and reduced high-density lipoprotein cholesterol (HDL-C) are established clinical risk factors and independent predictors of ischemic heart disease^3–9 and have been associated with cognitive impairment and late-life risk for AD.^10–14 Despite their association with cognitive aging, the underlying mechanisms linking aberrant lipid profiles to neurodegeneration and memory loss in humans are not fully understood.

Insight into the mechanisms of cognitive aging can be gained by studying the viscoelastic mechanical properties of the hippocampus (HC), a brain structure essential for memory encoding and performance.¹⁵ In this regard, magnetic resonance elastography (MRE) is a novel and noninvasive imaging modality for measuring brain viscoelastic mechanical properties in humans¹⁶,¹⁷ and reflects the distribution and organization of neurons, axons, and glial cells.^18–20 Indeed, alterations in neuronal microstructure via measures of viscoelastic properties have been demonstrated in neurodegenerative diseases including AD,^21–23 Parkinson’s disease,²⁴,²⁵ multiple sclerosis,²⁶,²⁷ and amyotrophic lateral sclerosis,²⁸ as well under conditions of increased inflammation.¹⁸,²⁹ Changes to brain tissue viscoelastic properties assessed using MRE also occur during normal healthy aging, and properties of the HC specifically have been strongly associated with memory performance^30–32 such that they are potentially more predictive of episodic memory than traditional clinical measures of HC volumes.³³ As such, MRE-based assessments of HC tissue integrity in healthy adults could provide important insight into pre-clinical associations between blood lipids and age-related memory loss that precede HC atrophy. Therefore, in this study we determined if blood lipids are independently associated with HC viscoelastic properties and memory function in healthy individuals across the adult lifespan. We hypothesized that elevated serum TGs, elevated LDL-C, and reduced HDL-C would be independently associated with lower memory performance. We also hypothesized that aberrant serum lipids would be associated with reduced HC integrity—defined as either a higher HC damping ratio (ξ) or lower HC shear stiffness (µ)—and that these associations would occur independent of age, sex, and other cardiometabolic risk factors. We further predicted that these associations would occur in the absence of an association with HC volume.

Materials and methods

This study was reviewed and approved by University of Delaware’s (UD) Institutional Review Board and all experimental procedures and protocols conformed to the standards outlined in the Declaration of Helsinki. The purpose of the study, along with the possible benefits and risks, was explained to all subjects and their written informed consent was obtained prior to participation. All measurements were made in the Neurovascular Aging Laboratory and the Center for Biomedical and Brain Imaging at UD between August 2018 and December 2019.

Study participants and recruitment

Twenty-six healthy adult men and women (14 M/12 F) between the ages of 22 and 78 years were recruited for this study from Newark, DE, and surrounding areas. All participants were non-smokers and were apparently free from chronic clinical diseases and major psychological and neurological disorders as assessed by medical history and baseline blood chemistries. Participants were excluded if they exhibited clinically abnormal blood chemistries such as a GFR <60 ml/min/1.73 m² and values that were ±2.5 × the upper or lower limit of the normal range. Participants taking any statin medications and medications acting on the central nervous system (CNS) were excluded as well. Participants were also excluded if they were unable or unwilling to participate in an MRI scan (e.g. metal implants, pacemakers, claustrophobia). Women of childbearing age were tested for possible pregnancy (hCG Pregnancy Test Kit, McKesson Consult) before entering the MRI. Subjects refrained from all over-the-counter medications for 48 h prior to the experimental visits and refrained from alcohol and vigorous exercise 24 h prior to experimental visits. Height and weight were measured during the initial screening. BMI was calculated as body mass in kilograms over square of height in meters. Physical activity was obtained from the Modifiable Activity Questionnaire and was measured as metabolic equivalents (METs) in hours per week during leisure time activity.³⁴

Blood lipid analysis

A standard blood lipid panel (Quest Diagnostics, USA) was performed at the time of initial screening following a 12 h overnight fast. Measured lipid markers included serum TGs, total blood cholesterol (TC), and HDL-C. LDL-C concentrations were calculated from the Martin–Hopkins equation. Generally, the Friedewald calculation (LDL-C = TC − HDL-C − TG/5) is commonly used to obtain blood LDL-C concentrations in which TG/5 is used to represent very low-density lipoprotein cholesterol (VLDL-C).³⁵ However, Martin et al. derived a more accurate factor that calculates VLDL-C from TGs, ranging from 3.1 to 11.9 depending on individual TG and non-HDL-C concentrations.³⁶ We also calculated an index of atherogenic risk that includes the TG/HDL-C ratio.³⁷,³⁸

Memory function

Episodic memory was assessed using the California Verbal Learning Test (CVLT-III).³⁹ The CVLT-III consists of the examiner reading a pre-recorded list of 16 words (List A) with four items belonging to each of four categories (e.g. vegetables, animals, modes of transportation, and furniture). Participants are asked to recall as many items as possible, in any order, over five trials. Another list of 16 words (List B) is administered after List A for one trial as an interference. After List B, participants are asked to freely recall words in any order from List A and then categorize List A into their four semantic categories (cued recall), without the examiner re-reading the list. Following a 20-min delay, free and cued recall of words from List A is assessed followed by a Yes/No recognition from List A. For this study, raw scores from the delayed recall portion of the test were used, as robust associations between HC structural changes and delayed recall memory exist both in healthy aging as well as in AD patients.^40–42

Magnetic resonance elastography

All participants completed a 50-min scanning session at the UD Center for Biomedical and Brain Imaging on a 3 T Siemens Magnetom Prisma whole-body MRI scanner with a 64-channel head coil (Siemens, Erlangen, Germany). MRE images were acquired using a 3 D multiband, multishot spiral sequence at 1.25 mm³ isotropic resolution (240 × 240 mm² field of view (FOV), 96 slices, repetition time (TR)/echo time (TE) = 3360/70 ms) in order to capture high-resolution displacement data with an acquisition time of 10 min and 45 s.⁴³ The sequence was synchronized with an external 50 Hz vibration frequency delivered to the head by a pneumatic actuator system and soft pillow driver (Resoundant Inc., Rochester, MN, USA) that has already been reported to be acceptable over a wide age range³⁰ as depicted in Figure 1. Motion from the actuator was sampled with flow-compensated motion encoding gradients to generate full vector displacement fields, following iterative image reconstruction, as described previously by our collaborators.⁴⁴

Figure 1.

MRE protocol. (a) Shear waves generated via external vibration at 50 Hz from an active driver to a head pillow-driver; (b) shear waves imaged as tissue displacement in three directions; (c) nonlinear inversion algorithm to determine shear stiffness (µ) and damping ratio (ξ).

Bilateral HC masks were obtained from a T₁-weighted magnetization-prepared rapid gradient echo structural scan at 0.9 mm³ resolution (240 × 240 mm² FOV, 192 slices, TR/TE = 2300/2.32 ms) and segmented using FreeSurfer 6.0. They were then registered from the anatomical T₁ image space to MRE space using the FLIRT tool in FSL (FMRIB Software Library v. 6.0.0).⁴⁵ Bilateral HC volumetric measures were determined from Freesurfer outputs and were normalized using estimated total intracranial volume to correct for participant head size.⁴⁶

MRE data quality was assessed using octahedral shear-strain based signal-to-noise ratio (OSS-SNR) and a cut-off of OSS-SNR >3 was considered as acceptable data for stable inversion.⁴⁷ A nonlinear inversion algorithm, combined with soft prior regularization (SPR), was used to estimate the viscoelastic complex shear modulus (G = G′ + G*) of the HC.⁴⁸,⁴⁹ SPR employs prior anatomical information to reduce regional variability and promote mechanical homogeneity in regions of interest;⁵⁰ SPR weighting of 10⁻¹² was applied in this study. Maps of the complex shear modulus were used to calculate viscoelastic parameters, shear stiffness (µ = 2 ${|| G ||}^{2} / (G^{'} + || G ||$ ) and damping ratio (ξ = G″/2G′), which are commonly reported in brain MRE literature.³⁰,^51–54 Shear stiffness µ is a property of viscoelastic tissue that describes resistance to deformation due to applied loading. Stiffness is related to the speed of shear wave propagation where waves will propagate more quickly through a stiff material and have a longer wavelength, as opposed to a softer material, where waves are slower and have a shorter wavelength.⁵⁵ Damping ratio ξ is a dimensionless quantity that refers to the relative attenuation level in a material. Higher ξ values indicate that the material attenuates rapidly, exhibiting behavior of a more viscous fluid while lower ξ values indicate a more elastic solid. Higher ξ also suggests a less densely connected solid phase, allowing for more viscous and frictional losses, which is an indicator of poorer tissue integrity.²⁰

Statistical analyses

Normality of residuals of each variable was assessed using the Shapiro–Wilk test.⁵⁶ Bivariate correlation analyses (Pearson correlations) were used to identify whether blood lipid markers and potential covariates were associated with any of the dependent variables. The Spearman rank correlation test was performed on variables that were not normally distributed (BMI, TG, and CVLT Delayed Recall scores). Because age is a strong predictor of both HC integrity and memory,⁵⁷ it was included in the bivariate correlation analysis. Sex, habitual physical activity, and BMI are also associated with brain viscoelasticity³¹,⁵³,⁵⁸ and blood lipid profiles;⁵⁹,⁶⁰ therefore, these variables were included as potential covariates. Multiple linear regressions were performed to identify the association between each covariate and independent variable of interest from the bivariate correlation analyses and the primary dependent variables. Any variable that was not normally distributed was log transformed prior to being entered into the regression models. Inferences from simple linear regressions were age-adjusted data based on residuals of the independent and dependent variables; however, for simplicity, figures are presented with raw data. Statistical significance was set at p < 0.05. All analyses were performed with RStudio v.1.2.1355 (RStudio Inc., Boston, MA, USA) and GraphPad Prism 8.0 (GraphPad Software Inc., San Diego, CA, USA).

Results

Participant characteristics

Twenty-six adult men and women between the ages of 22 and 78 years were included in this study. Basic characteristics and clinical laboratory values are presented in Table 1. Participants were generally healthy and free of overt cardiovascular diseases or other chronic conditions and the group mean laboratory values were within normal clinical ranges.

Table 1.

Participant characteristics.

Characteristic	Mean	Range
Sex (M/F)	14/12	N/A
Age (years)	42 ± 15	22–78
Body mass (kg)	76 ± 17	44–107
BMI (kg/m²)	26 ± 5	21.2–38.9
Systolic blood pressure (mmHg)	114 ± 11	90–142
Diastolic blood pressure (mmHg)	69 ± 11	48–93
Heart rate (bpm)	62 ± 9	45–78
Total cholesterol (mg/dl)	191 ± 33	132–256
HDL-C (mg/dl)	64 ± 21	34–115
LDL-C (mg/dl)	105 ± 27	54–175
Triglycerides (mg/dl)	112 ± 60	39–230
TG/HDL-C ratio	2.1 ± 1.6	0.5–5.8

BMI: body mass index; HDL-C: high-density lipoprotein cholesterol; LDL-C: low-density lipoprotein cholesterol; TG: triglycerides.

Data are mean ± SD.

Bivariate correlation between serum lipids and hippocampal structure and function

Bivariate correlations of serum lipid markers and their relation to memory and HC viscoelastic properties are presented in Table 2. HDL-C was significantly associated with delayed recall memory (r_s = 0.48, p = 0.01) and HC stiffness (r = 0.43, p = 0.03), while serum TGs were strongly associated with HC damping ratio (r_s = 0.63, p = 0.0005). TC was also significantly associated with HC stiffness (r = 0.42, p = 0.03); however, we observed no significant relation between LDL-C and HC stiffness (r = 0.25, p = 0.22), suggesting that this association was solely due to the influence of HDL-C. Hence, TC was not included in subsequent multiple linear regressions to avoid overfitting the models. Although not statistically significant, we also observed negative associations between delayed recall memory with both TGs (r_s= −0.37, p = 0.06) and LDL-C (r_s = −0.36, p = 0.08) as well as a negative association between HDL-C and HC damping ratio (r = −0.34, p = 0.09). Collectively, these results support an association between serum lipid profiles and HC function (i.e., memory) and integrity.

Table 2.

Bivariate correlation between dependent variables and potential covariates.

	Independent variables				Covariates
Dependent variables	TC	LDL-C	HDL-C	TG (r_s)	Age	Sex	BMI (r_s)	METs
CVLT Delayed Recall Score (r_s)
r	−0.05	−0.36	0.48*	−0.37	−0.48*	0.16	−0.14	0.26
p value	0.79	0.08	0.01	0.06	0.01	0.43	0.50	0.19
Hippocampal stiffness
r	0.42*	0.25	0.43*	−0.08	0.06	0.27	−0.22	0.21
p value	0.03	0.22	0.03	0.70	0.77	0.18	0.27	0.29
Hippocampal damping ratio
r	−0.03	0.05	−0.34	0.63*	0.39*	−0.15	0.27	0.03
p value	0.88	0.79	0.09	0.0005	0.05	0.45	0.17	0.87

CVLT: California Verbal Learning Test III; BMI: body mass index; METs: leisure physical activity metabolic equivalents hours/week; TC: total cholesterol; HDL-C: high-density lipoprotein cholesterol; LDL-C: low-density lipoprotein cholesterol; TG: triglycerides.

r_s represents variables for which Spearman rank correlation test was run in which the r values are Spearman correlation coefficients; rest of the r values represent Pearson correlation coefficients.

*p < 0.05.

Because serum lipids, HC structure, and memory are potentially influenced by age and other cardiometabolic risk factors or lifestyle behaviors, we next determined whether any of our independent or dependent variables were influenced by other covariates including age, sex, BMI, and physical activity (Table 2). Age was significantly associated with delayed recall memory (r_s = −0.48, p = 0.01) and nearly associated with HC damping ratio (r = 0.39, p = 0.05), therefore we included age as a covariate in all subsequent models. There were no associations between any of the other covariates with our independent or dependent variables.

Independent association between serum lipids and hippocampal structure and function

Next, we performed multiple linear regressions to determine the independent associations between serum lipid markers of interest and HC viscoelastic properties or memory performance while correcting for age and other lipids of interest (Table 3). Because we observed no significant associations between LDL-C and HC viscoelastic properties, we only included TGs and HDL-C in our models. Consistent with our bivariate correlations, we observed a significant association between TGs and HC damping ratio (β = 0.54, p = 0.012) while correcting for age and HDL-C (Table 3). We no longer observed a significant association between HDL-C and HC stiffness after correcting for age and TGs; however, the effect size remained strong and was close to significant (β = 0.43, p = 0.066). Finally, after correcting for blood lipid markers of interest, only age was independently associated with delayed recall memory (β = −0.45, p = 0.02); however, the explained variance (R²= 0.28) was larger than that explained by age alone (R²= 0.23), suggesting at least some contribution of blood lipids to the total variance in cognitive function.

Table 3.

Multiple linear regression models.

	β	p value
CVLT Delayed Recall Score
TG	−0.09	0.655
HDL-C	0.27	0.192
Age	−0.45*	0.020
Model variance (adjusted R²)	0.28*	0.016
Hippocampal stiffness
TG	−0.03	0.903
HDL-C	0.43	0.066
Age	0.12	0.543
Model variance (adjusted R²)	0.09	0.168
Hippocampal damping ratio
TG	0.54*	0.012
HDL-C	−0.03	0.885
Age	0.22	0.212
Model variance (adjusted R²)	0.36*	0.005

CVLT: California Verbal Learning Test III; TG: triglycerides; HDL-C: high-density lipoprotein cholesterol.

Standard regression coefficients (β) represent the association between each independent variable and the dependent variable while holding the other independent variables constant.

*p < 0.05.

Association between TG/HDL-C ratio and hippocampal structure and function

Because serum TGs and HDL-C emerged as the blood lipid markers most associated with HC viscoelastic properties, we next assessed the association between the TG/HDL-C ratio (an established clinical marker of atherogenic risk) and HC viscoelastic properties and memory while correcting for age. Consistent with our findings in the previous analyses, we observed a significant association between the TG/HDL-C ratio and HC damping ratio (Figure 2(a); r = 0.54, p = 0.005) and a close association with HC stiffness, although not significant (Figure 2(b); r = 0.36, p = 0.07). Interestingly, we also observed a significant association between the TG/HDL-C ratio and delayed recall memory, even when corrected for age (Figure 2(c); r = 0.47, p = 0.02). Representative images depicting a lower HC damping ratio in a participant with lower TG/HDL-C ratio characterized by a low serum TG concentration and high HDL-C concentration (Figure 2(d)) compared to a higher HC damping ratio in a subject with higher TG/HDL-C ratio characterized by a high serum TG concentration and low HDL-C concentration (Figure 2(e)) are also presented.

Figure 2.

Serum lipids, memory function and brain mechanical properties, and representative images of hippocampal damping ratio. Linear regression analysis between TG/HDL-C ratio and (a) hippocampal damping ratio; closed downward facing triangle symbol represents individual in (d) and closed upward facing triangle symbol represents individual in (e); (b) hippocampal stiffness and (c) CVLT delayed recall score; r and p values are adjusted for age; dashed lines represent 95% confidence interval; statistical significance set at p < 0.05; N = 26 in all panels. (d) image of participant with a low TG/HDL-C ratio (TG = 79 mg/dl; HDL-C = 76 mg/dl; TG/HDL-C = 1.04) and relatively low hippocampal damping ratio (0.13) compared with (e) image of participant with high TG/HDL-C ratio (TG = 222 mg/dl; HDL-C = 50 mg/dl; TG/HDL-C = 4.44) and relatively high hippocampal damping ratio (0.22).

Serum lipids and HC volumes

Finally, to investigate whether serum lipids were associated with standard volumetric measures of the HC, we examined their association with HC volumes. These associations between HC volume and serum TGs (r = 0.01, p = 0.96), HDL-C (r = 0.10, p = 0.61), and TG/HDL-C ratio (r = 0.02, p = 0.92) were not significant after correcting for age (data not shown).

Discussion

The primary finding of this study is that serum concentrations of TGs and the TG/HDL-C ratio are associated with memory performance and HC viscoelastic properties in healthy, cognitively normal adults, independent of age. Importantly, these associations exist in the absence of any relation with HC volumes suggesting a potential mechanistic role of serum lipids in the loss of HC viscoelastic properties in the absence of HC atrophy or clinically relevant memory loss. Collectively, our results may provide important insight into the early events linking cardiometabolic risk factors to late life memory impairment.

A novel aspect of our study is the use of high-resolution MRE to characterize HC structure and integrity. MRE has recently emerged as an important imaging modality for assessing brain tissue viscoelasticity¹⁶,¹⁷,⁵⁵,⁶¹ and reflects the microstructural composition and organization of neural structures.^18–20 Two important properties can be derived from MRE that provide important insight regarding the structural integrity of the brain. Damping ratio, or the relative viscous-to-elastic behavior of tissue, has been correlated with the density of solid phase connections of soft tissues in animal models¹⁸,¹⁹ and has been proposed to reflect neuronal microstructural tissue organization in vivo.²⁰ Accordingly, a lower HC ξ indicates a more elastic solid likely with densely packed neuronal connections, whereas a higher ξ reflects a more viscous tissue, likely with fewer neuronal connections.²⁰ In contrast, brain tissue stiffness assessed by MRE has been correlated with neuronal density, myelin content, and neural network strength in mouse and bovine animal models, respectively;²⁰,⁶²,⁶³ thus, a higher HC µ has been suggested to reflect an increased number of neurons within a stronger cellular matrix.²⁰ In this regard, normal aging is characterized by an increase in HC ξ and a decrease in HC µ, with a change in either property suggested to represent a loss of brain tissue integrity,²⁰ which in turn is associated with memory impairment⁵¹,⁵⁴ and potentially the transition from mild cognitive impairment to AD.⁶⁴

Our results provide important mechanistic insight into early events linking changes in blood lipids to the loss of HC integrity and function and may help explain why cardiometabolic risk factors during midlife predict memory loss later in life. The most striking observations were positive, independent associations between serum TGs and the TG/HDL-C ratio with HC damping ratio suggesting that even mildly aberrant blood lipids may disrupt existing neuronal connections and organization prior to any clinically detectable HC atrophy. Though not significant after correcting for age and other lipids, we also observed an association between HDL-C and HC stiffness that may suggest an important independent role of HDL-C in maintaining neuronal composition and integrity and is worthy of future investigation.

Although previous studies have demonstrated a similar link between serum lipids and memory function,¹¹,¹³,¹⁴ our study is the first to relate these important cardiometabolic risk factors to markers of HC tissue integrity in otherwise healthy humans. Reductions in brain tissue viscosity (loss modulus G”) and elasticity (storage modulus G’) have been observed in mice following acutely induced neuroinflammation¹⁸ and may mimic the effects of chronic neuroinflammation that occurs in the HC of older adults.⁶⁵ In this regard, TG transport occurs within VLDL particles which can bind to toll-like receptors on vascular endothelial cells leading to the release of pro-inflammatory cytokines and the development of vascular endothelial dysfunction.^66–69 Although VLDL and LDL particles themselves are not known to be present in the CNS, there is some evidence that lower-density lipoproteins may penetrate the blood–brain barrier (BBB) during conditions of increased inflammation.⁷⁰ Thus, elevated serum lipids may contribute to a loss of HC integrity either directly, or through a mechanism involving reduced cerebral perfusion, which is an established early clinical feature of AD.^71–73

TG-rich lipoproteins (VLDL and LDL) exchange their TGs with cholesterol esters from HDL-C via cholesteryl ester-transfer protein;⁶⁸,⁷⁴ thus, higher levels of HDL-C facilitate removal of harmful lipids which may explain why we also observed an association between the TG/HDL ratio and HC damping ratio. A major component of plasma HDL is apolipoprotein-A-I (ApoA-I), which has been shown to be protective against cognitive decline and AD.¹⁴,⁷⁵,⁷⁶ HDL/ApoA-I inhibits expression of pro-inflammatory cytokines by suppressing expression of adhesion molecules such as VCAM-1 and ICAM-1.⁷⁷ Accordingly, the mechanistic role of HDL-C and its associated lipoproteins on HC integrity and function needs to be explored further.

Lifestyle factors, such as consumption of an unhealthy diet, contribute to the increase of several cardiometabolic risk factors and may reduce HC integrity through the modulation of blood lipids. In this regard, dietary intake of added sugars (i.e., all caloric sweeteners added to food during processing) have been shown to increase the production of blood TGs and lower concentrations of HDL-C, in large part via the metabolism of fructose by the liver.⁷⁸,⁷⁹ Several animal studies have linked a high-sugar (along with high-fat) diet to elevated blood TGs,⁸⁰,⁸¹ increased inflammation,^82–84 and reduced concentrations of the neuroprotective protein brain-derived neurotrophic factor.⁸⁵,⁸⁶ Thus, it is possible that chronic consumption of sugar-sweetened food and beverages, particularly those containing fructose, may contribute to age-related memory loss by altering HC integrity; however, this hypothesis has not yet been tested in humans. Future randomized controlled trials examining the effects of a high added sugar diet (e.g., containing foods sweetened with high fructose corn syrup), on HC integrity and memory performance in humans are needed before changes to public health policy or nutritional guidelines can be made.

It is important to note that our study was conducted in generally healthy individuals, most of whom had clinically normal serum concentrations of TGs and HDL-C suggesting that even modest, pre-clinical changes in lipid metabolism may be sufficient to evoke changes to HC structure and function. Importantly, we did not observe a significant association between blood lipids and HC volumes, even after correcting for age, suggesting that elevated TGs or reduced HDL-C can induce a loss of HC microstructural integrity at any age in the absence of clinically significant brain tissue atrophy. These observations from our study could have important therapeutic implications for slowing or reversing the progression of age-related memory loss prior to the development of irreversible neuronal tissue loss; however, it is important to note that the present study was cross-sectional and did not investigate a causal link between serum lipids and markers of HC structure or function. Longitudinal studies are needed to determine whether changes in blood lipids mediate the loss of HC integrity or memory performance with aging.

Although the focus of our study was on healthy individuals, our findings have important implications for cognitive aging and dementia risk. Future studies should examine whether these associations are impacted by apolipoprotein-E (ApoE) genotype, which not only plays a significant role in lipid metabolism, but is involved in brain amyloid-beta (Aβ) transport and BBB integrity.^87–89 Individuals with one or more copy of the E4 variant of ApoE are at greater risk for AD, while the E2 variant has been shown to be neuroprotective and is associated with increased HC volume and improved cognitive performance.⁹⁰,⁹¹ Thus, it will be important to characterize the influence of ApoE genotype on the association between serum lipids and HC integrity in future studies to further elucidate the mechanisms underlying age-related memory loss.

In summary, this is the first study to explore the relation between serum lipid markers and HC viscoelastic properties in healthy adults. We found that elevated serum TGs are associated with a reduction in HC integrity while HDL-C may help maintain HC integrity. Collectively, our findings suggest that the TG/HDL-C ratio is inversely associated with memory performance, even after correcting for age, and precedes any loss of HC volume which underscores the importance of cardiometabolic risk factors on memory function before any clinical changes of cognitive impairment can occur. Future research should explore how diet and other lifestyle factors that are known to modulate serum lipids affect HC integrity to delay or reverse cognitive aging.

Footnotes

Authors’ contributions

FS, CLJ, and CRM conceived and designed the study; FS, PLD, TMD, and JCH performed the study procedures; FS, PLD, LVH, and MLC analyzed the data; FS, PLD, and LVH prepared the figures; FS, PLD, LVH, TMD, JCH, MLC, CLJ, and CRM interpreted the results; FS and CRM drafted the manuscript; all authors critically revised and approved the final manuscript.

Acknowledgments

We thank our research nurse, Wendy Nichols, for assisting with blood draws and our undergraduate research volunteer, Elizabeth Habash, for help with data entry.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the following grants from National Institutes of Health K01AG054731, R01AG058853, P20GM113125, and P20GM103653.

ORCID iDs

Faria Sanjana

Christopher R Martens

References

Hebert

Weuve

Scherr

, et al. Alzheimer disease in the United States (2010–2050) estimated using the 2010 census. Neurology 2013; 80: 1778–1783.

Santos

Snyder

W-C

, et al. Pathophysiologic relationship between Alzheimer’s disease, cerebrovascular disease, and cardiovascular risk: a review and synthesis. Alzheimer’s Dement Diagn Assess Dis Monit 2017; 7: 69–87.

Brown

Kinch

Doyle

JT.

Serum triglycerides in health and in ischemic heart disease. N Engl J Med 1965; 273: 947–952.

Jeppesen

Hein

Suadicani

, et al. Triglyceride concentration and ischemic heart disease: an eight-year follow-up in the Copenhagen Male Study. Circulation 1998; 97: 1029–1036.

Kannel

Castelli

Gordon

, et al. Serum cholesterol, lipoproteins, and the risk of coronary heart disease. The Framingham study. Ann Intern Med 1971; 74: 1–12.

Nordestgaard

Benn

Schnohr

, et al. Nonfasting triglycerides and risk of myocardial infarction, ischemic heart disease, and death in men and women. J Am Med Assoc 2007; 298: 299–308.

Castelli

Doyle

Gordon

, et al. HDL cholesterol and other lipids in coronary heart disease. The cooperative lipoprotein phenotyping study. Circulation 1977; 55: 767.

Lerner

Kannel

WB.

Patterns of coronary heart disease morbidity and mortality in the sexes: a 26-year follow-up of the Framingham population. Am Heart J 1986; 111: 383–390.

Gordon

Castelli

Hjortland

, et al. High density lipoprotein as a protective factor against coronary heart disease. The Framingham study. Am J Med 1977; 62: 707–714.

10.

Nägga

Gustavsson

A-M

Stomrud

, et al. Increased midlife triglycerides predict brain β-amyloid and tau pathology 20 years later. Neurology 2018; 90: e73–e81.

11.

Farr

Yamada

Butterfield

, et al. Obesity and hypertriglyceridemia produce cognitive impairment. Endocrinology 2008; 149: 2628–2636.

12.

Burgess

McIsaac

Naus

, et al. Elevated plasma triglyceride levels precede amyloid deposition in Alzheimer’s disease mouse models with abundant Aβ in plasma. Neurobiol Dis 2006; 24: 114–127.

13.

Parthasarathy

Frazier

Bettcher

, et al. Triglycerides are negatively correlated with cognitive function in nondemented aging adults. Neuropsychology 2017; 31: 682–688.

14.

Atzmon

Gabriely

Greiner

, et al. Plasma HDL levels highly correlate with cognitive function in exceptional longevity. Journals Gerontol Ser A 2002; 57: M712–M715.

15.

Eichenbaum

Yonelinas

Ranganath

The medial temporal lobe and recognition memory. Annu Rev Neurosci 2007; 30: 123–152.

16.

Muthupillai

Lomas

Rossman

, et al. Magnetic resonance elastography by direct visualization of propagating acoustic strain waves. Science 1995; 269: 1854–1857.

17.

Hiscox

Johnson

Barnhill

, et al. Magnetic resonance elastography (MRE) of the human brain: technique, findings and clinical applications. Phys Med Biol 2016; 61: R401–R437.

18.

Riek

Millward

Hamann

, et al. Magnetic resonance elastography reveals altered brain viscoelasticity in experimental autoimmune encephalomyelitis. NeuroImage Clin 2012; 1: 81–90.

19.

Schregel

Wuerfel Nee Tysiak

Garteiser

, et al. Demyelination reduces brain parenchymal stiffness quantified in vivo by magnetic resonance elastography. Proc Natl Acad Sci 2012; 109: 6650–6655.

20.

Sack

Jöhrens

Würfel

, et al. Structure-sensitive elastography: on the viscoelastic powerlaw behavior of in vivo human tissue in health and disease. Soft Matter 2013; 9: 5672.

21.

Murphy

Huston

Jack

, et al. Decreased brain stiffness in Alzheimer’s disease determined by magnetic resonance elastography. J Magn Reson Imaging 2011; 34: 494–498.

22.

Murphy

Jones

Jack

, et al. Regional brain stiffness changes across the Alzheimer’s disease spectrum. NeuroImage Clin 2016; 10: 283–290.

23.

Hiscox

Johnson

McGarry

MDJ

, et al. Mechanical property alterations across the cerebral cortex due to Alzheimer’s disease. Brain Commun 2019; 2: fcz049.

24.

Lipp

Trbojevic

Paul

, et al. Cerebral magnetic resonance elastography in supranuclear palsy and idiopathic Parkinson’s disease. NeuroImage Clin 2013; 3: 381–387

25.

Lipp

Skowronek

Fehlner

, et al. Progressive supranuclear palsy and idiopathic Parkinson’s disease are associated with local reduction of in vivo brain viscoelasticity. Eur Radiol 2018; 28: 3347–3354.

26.

Streitberger

Sack

Krefting

, et al. Brain viscoelasticity alteration in chronic-progressive multiple sclerosis. PLoS One 2012; 7: e29888.

27.

Wuerfel

Paul

Beierbach

, et al. MR-elastography reveals degradation of tissue integrity in multiple sclerosis. Neuroimage 2010; 49: 2520–2525.

28.

Romano

Guo

Prokscha

, et al. In vivo waveguide elastography: effects of neurodegeneration in patients with amyotrophic lateral sclerosis. Magn Reson Med 2014; 72: 1755–1761.

29.

Millward

Guo

Berndt

, et al. Tissue structure and inflammatory processes shape viscoelastic properties of the mouse brain. NMR Biomed 2015; 28: 831–839.

30.

Hiscox

Johnson

McGarry

MDJ

, et al. High-resolution magnetic resonance elastography reveals differences in subcortical gray matter viscoelasticity between young and healthy older adults. Neurobiol Aging 2018; 65: 158–167.

31.

Arani

Murphy

Glaser

, et al. Measuring the effects of aging and sex on regional brain stiffness with MR elastography in healthy older adults. Neuroimage 2015; 111: 59–64.

32.

Sack

Streitberger

K-J

Krefting

, et al. The influence of physiological aging and atrophy on brain viscoelastic properties in humans. PLoS One 2011; 6: e23451–e23451.

33.

Van Petten

Relationship between hippocampal volume and memory ability in healthy individuals across the lifespan: review and meta-analysis. Neuropsychologia 2004; 42: 1394–1413.

34.

Kriska

Knowler

LaPorte

, et al. Development of questionnaire to examine relationship of physical activity and diabetes in Pima Indians. Diabetes Care 1990; 13: 401–411.

35.

Friedewald

Levy

Fredrickson

DS.

Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972; 18: 499–502.

36.

Martin

Blaha

Elshazly

, et al. Comparison of a novel method vs the Friedewald equation for estimating low-density lipoprotein cholesterol levels from the standard lipid profile. JAMA 2013; 310: 2061–2068.

37.

Da Luz

Favarato

Faria-Neto

, et al. High ratio of triglycerides to HDL-cholesterol predicts extensive coronary disease. Clinics 2008; 63: 427–432.

38.

Gaziano

Hennekens

O’Donnell

, et al. Fasting triglycerides, high-density lipoprotein, and risk of myocardial infarction. Circulation 1997; 96: 2520–2525.

39.

Delis

Kaplan

Kramer

, et al. California verbal learning test (CVLT). ECPA, 2017.

40.

Golomb

Kluger

de Leon

, et al. Hippocampal formation size in normal human aging: a correlate of delayed secondary memory performance. Learn Mem 1994; 1: 45–54.

41.

Mortimer

Gosche

Riley

, et al. Delayed recall, hippocampal volume and Alzheimer neuropathology. Neurology 2004; 62: 428–432.

42.

Foster

Meikle

Goodson

, et al. The hippocampus and delayed recall: bigger is not necessarily better? Memory 1999; 7: 715–733.

43.

Johnson

Holtrop

Anderson

, et al. Brain MR elastography with multiband excitation and nonlinear motion-induced phase error correction. In: Proceedings of the 24th annual meeting of the International Society for Magnetic Resonance in Medicine, Singapore, 7–13 May 2016, p.1951.

44.

Johnson

Holtrop

McGarry

MDJ

, et al. 3D multislab, multishot acquisition for fast, whole-brain MR elastography with high signal-to-noise efficiency. Magn Reson Med 2014; 71: 477–485.

45.

Jenkinson

Beckmann

Behrens

TEJ

, et al. FSL. Neuroimage 2012; 62: 782–790.

46.

Buckner

Head

Parker

, et al. A unified approach for morphometric and functional data analysis in young, old, and demented adults using automated atlas-based head size normalization: reliability and validation against manual measurement of total intracranial volume. Neuroimage 2004; 23: 724–738.

47.

McGarry

MDJ

Van Houten

EEW

Perrĩez

, et al. An octahedral shear strain-based measure of SNR for 3D MR elastography. Phys Med Biol 2011; 56: N153–N164.

48.

Van Houten

EEW

Miga

Weaver

, et al. Three-dimensional subzone-based reconstruction algorithm for MR elastography. Magn Reson Med 2001; 45: 827–837.

49.

McGarry

MDJ

Van Houten

EEW

Johnson

, et al. Multiresolution MR elastography using nonlinear inversion. Med Phys 2012; 39: 6388–6396.

50.

McGarry

MDJ

Johnson

Sutton

, et al. Including spatial information in nonlinear inversion MR elastography using soft prior regularization. IEEE Trans Med Imaging 2013; 32: 1901–1909.

51.

Hiscox

Johnson

McGarry

MDJ

, et al. Hippocampal viscoelasticity and episodic memory performance in healthy older adults examined with magnetic resonance elastography. Brain Imaging Behav 2020; 14: 175–185.

52.

Johnson

Schwarb

McGarry

MDJ

, et al. Viscoelasticity of subcortical gray matter structures. Hum Brain Mapp 2016; 37: 4221–4233.

53.

Schwarb

Johnson

Daugherty

, et al. Aerobic fitness, hippocampal viscoelasticity, and relational memory performance. Neuroimage 2017; 153: 179–188.

54.

Schwarb

Johnson

McGarry

MDJ

, et al. Medial temporal lobe viscoelasticity and relational memory performance. Neuroimage 2016; 132: 534–541.

55.

Murphy

Huston

Ehman

RL.

MR elastography of the brain and its application in neurological diseases. Neuroimage 2019; 187: 176–183.

56.

Shapiro

Wilk

MB.

An analysis of variance test for normality (complete samples). Biometrika 1965; 52: 591–611.

57.

Raz

Lindenberger

Rodrigue

, et al. Regional brain changes in aging healthy adults: general trends, individual differences and modifiers. Cereb Cortex 2005; 15: 1676–1689.

58.

Hetzer

Hirsch

Braun

, et al. Viscoelasticity of striatal brain areas reflects variations in body mass index of lean to overweight male adults. Brain Imaging Behav 2019: 1–11.

59.

Sternfeld

Sidney

Jacobs

, et al. Seven-year changes in physical fitness, physical activity, and lipid profile in the CARDIA study. Ann Epidemiol 1999; 9: 25–33.

60.

Denke

Sempos

Grundy

SM.

Excess body weight: an under-recognized contributor to dyslipidemia in white American women. Arch Intern Med 1994; 154: 401.

61.

Johnson

Schwarb

Horecka

, et al. Double dissociation of structure-function relationships in memory and fluid intelligence observed with magnetic resonance elastography. Neuroimage 2018; 171: 99–106.

62.

Weickenmeier

de Rooij

Budday

, et al. Brain stiffness increases with myelin content. Acta Biomater 2016; 42: 265–272.

63.

Freimann

Müller

Streitberger

, et al. MR elastography in a murine stroke model reveals correlation of macroscopic viscoelastic properties of the brain with neuronal density. NMR Biomed 2013; 26: 1534–1539.

64.

Farias

Mungas

Reed

, et al. Progression of mild cognitive impairment to dementia in clinic- vs community-based cohorts. Arch Neurol 2009; 66.

65.

Barrientos

Kitt

Watkins

, et al. Neuroinflammation in the normal aging hippocampus. Neuroscience 2015; 309: 84–99.

66.

Rosenson

Davidson

Hirsh

, et al. Genetics and causality of triglyceride-rich lipoproteins in atherosclerotic cardiovascular disease. J Am Coll Cardiol 2014; 64: 2525–2540.

67.

Toth

PP.

Triglyceride-rich lipoproteins as a causal factor for cardiovascular disease. Vasc Health Risk Manag 2016; 12: 171–183.

68.

Welty

FK.

How do elevated triglycerides and low HDL-cholesterol affect inflammation and atherothrombosis?

Curr Cardiol Rep 2013; 15: 400.

69.

De Nardo

Labzin

Kono

, et al. High-density lipoprotein mediates anti-inflammatory reprogramming of macrophages via the transcriptional regulator ATF3. Nat Immunol 2014; 15: 152–160.

70.

Loving

Bruce

KD.

Lipid and lipoprotein metabolism in microglia. Front Physiol 2020; 11.

71.

Johnson

Jones

Holman

, et al. Preclinical prediction of Alzheimer’s disease using SPECT. Neurology 1998; 50: 1563–1571.

72.

Hanyu

Sato

Hirao

, et al. The progression of cognitive deterioration and regional cerebral blood flow patterns in Alzheimer’s disease: a longitudinal SPECT study. J Neurol Sci 2010; 290: 96--101.

73.

Jagust

Eberling

Reed

, et al. Clinical studies of cerebral blood flow in Alzheimer’s disease. Ann NY Acad Sci 1997; 826: 254–262.

74.

Hottman

Chernick

Cheng

, et al. HDL and cognition in neurodegenerative disorders. Neurobiol Dis 2014; 72PA: 22–36.

75.

Merched

Xia

Visvikis

, et al. Decreased high-density lipoprotein cholesterol and serum apolipoprotein AI concentrations are highly correlated with the severity of Alzheimer’s disease. Neurobiol Aging 2000; 21: 27–30.

76.

Kuriyama

Takahashi

Yamano

, et al. Low levels of serum apolipoprotein AI and AII in senile dementia. Psychiatry Clin Neurosci 1994; 48: 589–593.

77.

Stukas

Robert

Wellington

CL.

High-density lipoproteins and cerebrovascular integrity in Alzheimer’s disease. Cell Metab 2014; 19: 574–591.

78.

Mensink

Zock

Kester

ADM

, et al. Effects of dietary fatty acids and carbohydrates on the ratio of serum total to HDL cholesterol and on serum lipids and apolipoproteins: a meta-analysis of 60 controlled trials. Am J Clin Nutr 2003; 77: 1146–1155.

79.

Hellerstein

MK.

Carbohydrate-induced hypertriglyceridemia: modifying factors and implications for cardiovascular risk. Curr Opin Lipidol 2002; 13: 33–40.

80.

Davidson

Kanoski

Chan

, et al. Hippocampal lesions impair retention of discriminative responding based on energy state cues. Behav Neurosci 2010; 124: 97–105.

81.

Stranahan

Norman

Lee

, et al. Diet-induced insulin resistance impairs hippocampal synaptic plasticity and cognition in middle-aged rats. Hippocampus 2008; 18: 1085–1088.

82.

Hsu

Konanur

Taing

, et al. Effects of sucrose and high fructose corn syrup consumption on spatial memory function and hippocampal neuroinflammation in adolescent rats. Hippocampus 2015; 25: 227–239.

83.

Pistell

Morrison

Gupta

, et al. Cognitive impairment following high fat diet consumption is associated with brain inflammation. J Neuroimmunol 2010; 219: 25–32.

84.

White

Pistell

Purpera

, et al. Effects of high fat diet on Morris maze performance, oxidative stress, and inflammation in rats: contributions of maternal diet. Neurobiol Dis 2009; 35: 3–13.

85.

Molteni

Ying

, et al. A saturated-fat diet aggravates the outcome of traumatic brain injury on hippocampal plasticity and cognitive function by reducing brain-derived neurotrophic factor. Neuroscience 2003; 119: 365–375.

86.

Kanoski

Meisel

Mullins

, et al. The effects of energy-rich diets on discrimination reversal learning and on BDNF in the hippocampus and prefrontal cortex of the rat. Behav Brain Res 2007; 182: 57–66.

87.

Deane

Sagare

Hamm

, et al. apoE isoform-specific disruption of amyloid β peptide clearance from mouse brain. J Clin Invest 2008; 118: 4002–4013.

88.

Verghese

Castellano

Garai

, et al. ApoE influences amyloid-β (Aβ) clearance despite minimal apoE/Aβ association in physiological conditions. Proc Natl Acad Sci U S A 2013; 110: E1807–E1816.

89.

Hawkes

Sullivan

Hands

, et al. Disruption of arterial perivascular drainage of amyloid-β from the brains of mice expressing the human APOE ε4 allele. PLoS One 2012; 7: e41636.

90.

Zhao

ApoE2 and Alzheimer’s disease: time to take a closer look. Neural Regen Res 2016; 11: 412–413.

91.

Castellano

Kim

Stewart

, et al. Human apoE isoforms differentially regulate brain amyloid-β peptide clearance. Sci Transl Med 2011; 3: 89ra57.

Blood lipid markers are associated with hippocampal viscoelastic properties and memory in humans

Abstract

Keywords

Introduction

Materials and methods

Study participants and recruitment

Blood lipid analysis

Memory function

Magnetic resonance elastography

Statistical analyses

Results

Participant characteristics

Bivariate correlation between serum lipids and hippocampal structure and function

Independent association between serum lipids and hippocampal structure and function

Association between TG/HDL-C ratio and hippocampal structure and function

Serum lipids and HC volumes

Discussion

Footnotes

Authors’ contributions

Acknowledgments

Declaration of conflicting interests

Funding

ORCID iDs

References