Abstract
Background:
Serum uric acid (SUA) affects the reaction of oxidative stress and free radicals in the neurodegenerative processes. However, whether SUA impacts Alzheimer’s disease (AD) pathology remains unclear.
Objective:
We aimed to explore whether high SUA levels can aggravate the neurobiological changes of AD in preclinical AD.
Methods:
We analyzed cognitively intact participants (n = 839, age 62.16 years) who received SUA and cerebrospinal fluid (CSF) biomarkers (amyloid-β [Aβ], total tau [t-Tau], and phosphorylated tau [p-Tau]) measurements from the Chinese Alzheimer’s Biomarker and LifestylE (CABLE) database using multivariable-adjusted linear models.
Results:
Levels of SUA in the preclinical AD elevated compared with the healthy controls (p = 0.007) and subjects with amyloid pathology had higher concentration of SUA than controls (p = 0.017). Roughly, equivalent levels of SUA displayed among cognitively intact individuals with or without tau pathology and neurodegeneration. CSF Aβ1 - 42 (p = 0.019) and Aβ1 - 42/Aβ1 - 40 (p = 0.027) were decreased and CSF p-Tau/Aβ1 - 42 (p = 0.009) and t-Tau/Aβ1 - 42 (p = 0.043) were increased with the highest (> 75th percentile) SUA when compared to lowest SUA, implying a high burden of cerebral amyloidosis in individuals with high SUA. Sensitivity analyses using the usual threshold to define hyperuricemia and precluding drug effects yielded robust associations. Nevertheless, the quadratic model did not show any U-shaped relationships between them.
Conclusion:
SUA may aggravate brain amyloid deposition in preclinical AD, which corroborated the detrimental role of SUA.
INTRODUCTION
Uric acid (UA), a major natural antioxidant, accounts for approximately 60%proportion of the antioxidant capacity in humans [1, 2]. A growing body of evidence has further advanced the beneficial roles of UA against pathological processes of neurodegenerative diseases by reducing oxidative stress and free radicals [3, 4]. Hyperuricemia as an initial factor of gout, however, links with multiple vascular risk factors and diseases, which may predispose individuals to dementia [5]. Hence, there may not preclude a U-shaped association of UA and dementia, including Alzheimer’s disease (AD) and vascular dementia. Increasing research has investigated the association between UA and AD [6–9] or cognitive impairment [10–12] and yielded inconsistent findings. The population-based longitudinal studies found a decreased risk of developing dementia with higher serum UA (SUA) levels [11] and revealed similar risk estimates for AD and vascular dementia [7]. While a 12-year follow-up cohort recruiting healthy elders from the community (Three-City Dijon cohort) stated an elevated risk of dementia in individuals with high SUA [9], there still lacks evidence concerning the association of SUA with cognitive decline or with predictive biomarkers of AD to determine the role of UA in early and later neurodegenerative progression.
AD is the most common cause of dementia and throws major challenges for global health and social security [13]. Extracellular amyloid-β (Aβ) aggregation and intracellular neurofibrillary tangles constitute two hallmarks of AD [14, 15]. Recent studies have confirmed that aggregated Aβ as the first emerging pathology of AD can trigger or drive tau pathology, thereby leading to synaptic dysfunction and neurodegeneration [14, 16]. These pathologies are initiated a decade or more before the clinical symptoms [17]. In neuropathological series, a considerable proportion of cognitively intact elders has AD pathology [18]. PET studies reported aggregated amyloid and tau pathology in one-third healthy elderly from control samples [19–21]. Biomarkers [(Aβ1 - 42, total-tau (t-Tau), phosphorylated-tau (p-Tau)] in cerebrospinal fluid (CSF) reflect the earliest pathological changes [22–25] and facilitate the determination of preclinical AD in accordance with the National Institute on Aging–Alzheimer’s Association (NIA-AA) criteria [26]. Clarifying the roles of risk factors in early neurobiological alterations would advance perceptions of underlying mechanisms, clinical trial recruitment, and promote existing preventative strategies targeting the early stage of AD (i.e., preclinical AD) [26].
Therefore, we assumed that high SUA levels could aggravate the neurobiological changes of AD. We aimed to investigate 1) whether the changes of SUA levels linked with the preclinical AD and influenced CSF indicators of AD pathology in a large sample of cognitively intact older adults; 2) whether a nonlinear relationship existed between SUA and CSF AD biomarkers; 3) whether the associations differed by gender or apolipoprotein E ɛ4 (APOE ɛ4) status. The primary findings of our study would provide paramount clues about the mechanism of SUA in the long-term neurodegenerative process, thereby reinforcing the role of SUA as a potentially modifiable factor of AD.
MATERIALS AND METHODS
CABLE database and study participants
Chinese Alzheimer’s Biomarker and LifestylE (CABLE) is an ongoing large-scale cohort study mainly focusing on the risk factors and biomarkers of AD in the northern Chinese Han population, thereby providing evidence for early diagnosis and disease prevention. The enrolled individuals visited the Qingdao Municipal Hospital, Shandong Province, China. All participants were Han Chinese aged between 40 to 90 years. Subjects were included consecutively and consisted of cognitively intact older adults and individuals with MCI or AD. The exclusion criteria for this study were 1) central nervous system infection, epilepsy, multiple sclerosis, or other major neurological disorders; 2) major psychological disorders; 3) severe systemic diseases (such as malignant tumors); 4) and family history of genetic disease [27]. If the subject has any one of the exclusion criteria, it will not be included. Demographic information, medical history, and risk factors were collected by a self-reported questionnaire and an electronic medical record system. The enrolled participants received clinical and neuropsychological assessments, blood and CSF specimen collection, and MRI examinations to determine their cognitive status. Preclinical AD was determined according to the NIA-AA criteria in 2011 among asymptomatic individuals with evidence of AD pathology [26]. General cognition was assessed using the China Modified Mini-Mental State Examination (CM-MMSE) and Montreal Cognitive Assessment (MoCA) [28]. The CM-MMSE is a fully structured screening instrument evaluating questions about memory, orientation to time and place, registration, attention and calculation, recall, language and visual construction. The score of this scale ranges from 0 to 30, with a higher score representing better cognitive performance.
No statistical method was used to determine sample size. A total of 885 cognitively intact older adults were selected initially for the present study from the CABLE database who had measurements of SUA and CSF Aβ or tau proteins. Participants who received urate-lowering treatments were precluded owing to the potential confounding effect of gout status on the risk of AD (n = 12) [9, 29]. After the removal of individuals with incomplete covariate data consisting of age, gender, education, and APOE ɛ4 alleles, 870 participants were included in subsequent association analyses. Quality control additionally excluded outliers sitting outside three standard deviations and CSF biomarker data with an inter-batch/intra-assay coefficient of variation (CV) values greater than 15%. Thus, the whole 839 study participants eventually consisted of those with SUA measurements and available CSF biomarkers data on Aβ1 - 42 (n = 839), Aβ1 - 40 (n = 797), p-Tau (n = 832), and t-Tau proteins (n = 804). No randomization was performed to allocate subjects in the study.
Standard protocol approvals and patient consents
Written informed consent was obtained from all participants or their caregivers. CABLE study was in accordance with the Helsinki Declaration of 1975 and approved by the Institutional Review Board of Qingdao Municipal Hospital. The study was not pre-registered.
CSF sampling and biomarker analyses
CSF specimens were collected in 15 ml polypropylene tubes following the standard operational process of lumbar puncture and centrifuged at 2000×g for 10 min within 2 h after collection. These specimens were snap-frozen in an enzyme-free EP tube at –80°C until assay. CSF concentrations of Aβ1 - 42, Aβ1 - 40, p-Tau, and t-Tau were measured with the ELISA kit (INNOTEST β-AMYLOID(1-42), INNOTEST β-AMYLOID(1-40), INNOTEST PHOSPHO-TAU (181P), and INNOTEST hTAU Ag; Fujirebio, Ghent, Belgium) by trained technicians who were blinded for clinical information. All sample were analyzed in duplicate. Average intra-assay variation of duplicate measurements was well below 15%(11.96%for Aβ1 - 42, 8.15%for Aβ1 - 40, 11.12%for p-Tau, and 13.89%for t-Tau) and average inter-assay CVs were below 5%(4.64%for Aβ1 - 42, 4.22%for Aβ1 - 40, 2.59%for p-Tau, and 4.64%for t-Tau).
Serum uric acid determination
SUA was determined by uricase-peroxidase coupled assays [30] using Uric acid Test Kit (Ningbo Ruiyuan Biotechnology Co., Ltd, China) after a 12-h overnight fast at the laboratory of the Department of Clinical Chemistry at Qingdao Municipal Hospital in China with the use of standard routine clinical laboratory procedures.
Apolipoprotein E ɛ4 genotyping
Genomic DNA was isolated from EDTA blood using QIAamp®DNA Blood Mini Kit (Qiagen, Hilden, Germany). Genotypes were determined by restriction endonuclease digestion of polymerase chain reaction–amplified genomic DNA. One or two APOE ɛ4 alleles classified individuals as APOE ɛ4 carriers, whereas no APOE ɛ4 allele classified individuals as non-carriers.
Statistical analysis
Epidemiological evidence reported a consistently higher SUA levels in males than in females over time [31]. Study participants in the CABLE study did present gender differences in SUA levels (mean SUA [SD]: 376.75 [87.06] in males versus 326.33 [75.15] in females, p < 0.0001). Gender-specific cutoffs of SUA quantiles were thus applied for subsequent analyses to avoid the potential confounding effect of gender: quartile 1 (Q1): < 312μmol/L for males, < 271μmol/L for females; quartile 2 (Q2): 312–370 for males, 271–325 for females; quartile 3 (Q3): 370–433 for males, 325–371 for females; quartile 4 (Q4): > 433 for males, > 371 for females.
Evidence from neuroimaging [19, 32] and neuropathological [33–36] studies has consistently approved approximately one-third prevalence of AD pathology in cognitively unimpaired individuals with similar observation in Asian populations [35, 36]. Hence, the cutoffs to determine abnormal CSF AD biomarkers were < 16.46 pg/ml (lower one-third) for Aβ1–42, > 38.47 pg/ml (upper one-third) for p-Tau, and > 176.46 pg/ml (upper one-third) for t-Tau. Participants were divided into healthy controls (stage 0) and preclinical AD (stage 1 and stage 2) on the basis of AD continuum scheme [26, 38].
Characteristics of the study population by SUA level were described using the Mann-Whitney U test for continuous variables and the Chi-square test for categorical variables. In terms of skewed CSF biomarkers data (Shapiro-Wilk test < 0.05), the Box-Cox transformations were performed to address the distribution of non-normality prior to statistical analyses via “car” package of R software. We detected the association of SUA (categorical and continuous data) with CSF biomarkers of AD pathology (Aβ1 - 42, p-Tau, and t-Tau) and their ratios (Aβ1 - 42/Aβ1 - 40, p-Tau/Aβ1 - 42, and t-Tau/Aβ1 - 42) using linear regression models with age (continuous), gender (male or female), education (continuous), APOE ɛ4 carrier status (carriers or non-carriers), and CM-MMSE (continuous) as covariates. No multi-collinearity existed in each model after assessments of tolerance, variance inflation factor (VIF), and Pearson’s correlation coefficients. Sensitivity analyses were then conducted in three ways to evaluate the robustness of results: 1) using two models additionally adjusted for body mass index (BMI) (continuous), regular alcohol consumption (yes or no), regular smoking (yes or no), diabetes mellitus (yes or no), hypertension (yes or no), history of stroke (yes or no), history of cardiovascular disease (yes or no), total cholesterol (continuous), triglyceride (continuous), and estimated glomerular filtration rate (eGFR), as well as aspirin (yes or no), diuretics (yes or no), and non-steroidal anti-inflammatory drugs (NSAIDs) (yes or no); 2) repeating the primary results using the usual definition of hyperuricemia; 3) analyzing the linear associations in participants not taking NSAIDs. To explore whether several associations differed by gender and APOE ɛ4 status, interaction terms (SUA * sex and SUA * APOE ɛ4 status) were further performed. In case of any potential interaction (p < 0.05), subgroup analysis was further performed. To investigate whether there were U-shaped associations between SUA level and CSF AD biomarkers, non-linear regression based on the quadratic model (y =αx2 +βx + c) was conducted with age, gender, educational level, and APOE ɛ4 carrier status as covariates. The significantly larger or smaller coefficients of the quadratic term (α) indicated the non-linear association.
All analyses involved the use of R (version 3.6.0) software program, and a two-tailed p value < 0.05 was considered statistically significant.
RESULTS
Study participants
The study population involved 839 cognitively intact older adults from the CABLE database (mean [SD] age 62.16 [10.93] years, males 57.69%) (Table 1). There were no differences in age, gender, educational level, CM-MMSE score, and APOE ɛ4 status between high SUA group (SUA level > 75th percentile distribution) and controls. Whereas high SUA level was associated with the use of aspirin or diuretics, low eGFR, history of cardiovascular disease and related risk factors (alcohol consumption, high BMI, hypertension, and high triglyceride and total cholesterol levels), as well as CSF biomarkers of brain amyloidosis (high CSF Aβ1 - 42 and CSF Aβ1 - 42/Aβ1 - 40 and low CSF p-Tau/Aβ1 - 42 and CSF t-Tau/Aβ1 - 42).
Characteristics of study participants in CABLE database
*SUA level >gender-specific 75th percentile of the distribution (433μmol/L for males; 371μmol/L for females). 1Mann-Whitney U test for continuous variables; χ2 test for categorical variables. APOE, apolipoprotein E; Aβ, amyloid-β; BMI, body mass index; CABLE, Chinese Alzheimer’s Biomarker and LifestylE; CM-MMSE, China modified Mini-Mental State Examination; CSF, cerebrospinal fluid; eGFR, estimated glomerular filtration rate; NSAIDs, non-steroidal anti-inflammatory drugs; p-Tau, phosphorylated tau; IQR, Interquartile range; SUA, serum uric acid; t-Tau, total tau.
Differences in SUA levels between different biomarker classifications
A representative population of preclinical AD (stage 1 and stage 2, n = 267) with abnormal amyloid markers or with the combination of amyloid and injury markers was analyzed in this study in accordance with the definition of NIA-AA scheme [26]. To assess the associations of changes in SUA levels with preclinical AD, we compare the differences between preclinical AD individuals and healthy controls after removal of participants with isolated tauopathy (i.e., p-Tau and t-Tau pathology). The results of ANCOVA controlling for age, gender, educational levels, and APOE ɛ4 carrier status showed that preclinical AD individuals had significantly higher SUA compared to healthy controls (*p = 0.007) (Fig. 1A). To elucidate the further relations of SUA levels with the β-amyloidosis and the downstream process (i.e., pathological p-Tau and neurodegeneration), we leveraged the biomarkers classification framework to compare the differences in SUA levels among three biomarker groups. The A + subgroup had significantly increased SUA concentrations compared to the A- subgroup (*p = 0.017) (Fig. 1B) while levels of SUA were roughly equivalent in both T (p = 0.442) and N (p = 0.791) subgroups (Fig. 1C, D). It can be inferred that elevated SUA levels might contribute to the brain Aβ-deposition but not the aggregated p-Tau or t-Tau in downstream tau pathology and neurodegeneration.
Levels of SUA in the CSF biomarker classifications. Scatter plots depict the levels of SUA for each biomarker profile and diagnostic groups. p-values were assessed by the ANCOVA controlling age, gender, education, and APOE ɛ4 status. β-amyloidosis (A) was operationally determined by CSF assay for Aβ1 - 42. Tau pathology (T) was defined by CSF assays for p-Tau. Neurodegeneration (N) was defined by CSF assays for t-Tau. The cutoff values to define abnormal CSF core biomarkers were < 116.46 pg/ml for Aβ1 - 42 (A+), > 38.47 pg/ml for P-tau (T+), and > 176.46 pg/ml for T-tau (N+). A, β-amyloidosis; APOE, apolipoprotein E; Aβ, amyloid-β; CM-MMSE, China modified Mini-Mental State Examination; CSF, cerebrospinal fluid; N, neurodegeneration; p-Tau, phosphorylated tau; SUA, serum uric acid; T, pathologic tau; t-Tau, total tau.
High SUA associated with CSF indicators of amyloidosis
Three ratios to CSF Aβ1 - 42, including CSF Aβ1 - 42/Aβ1 - 40, CSF p-Tau/Aβ1 - 42, and CSF t-Tau/Aβ1 - 42, have been extensively expressed as more sensitive predictors of cerebral amyloidosis and cognitive decline than the individual biomarkers alone [39–42]. Multivariable-adjusted models showed that CSF indicators of amyloidosis including CSF Aβ1 - 42 (β= –5.1×10–4, *p = 0.019) and CSF Aβ1 - 42/Aβ1 - 40 (β= –0.372, *p = 0.027) were decreased, and CSF p-Tau/Aβ1 - 42 (β= 0.098, *p = 0.009) and CSF t-Tau/Aβ1 - 42 (β= 0.088, *p = 0.043) were increased with the highest quartile of SUA level (> 433μmol/L for males, > 371μmol/L for females) as compared with the lowest quartile (< 312μmol/L for males, < 271 for females) (Table 2). The relationships were robustness after adjustment for risk factors (model 2) and drugs affecting SUA metabolism (model 3). Further, associations of SUA with CSF Aβ1 - 42 and three ratios to Aβ1 - 42 in CSF persisted when we used the usual threshold to define hyperuricemia [43, 44], indicating a high burden of potential amyloidosis in individuals with hyperuricemia (Table 3).
The association of SUA and CSF biomarkers of AD pathology in three multi-adjusted models
Model 1: adjusted for age, gender, educational level, APOE ɛ4 status, and CM-MMSE. Model 2: model 1 + BMI, alcohol consumption, smoking, diabetes mellitus, hypertension, history of stroke, history of cardiovascular disease, total cholesterol, triglyceride, and eGFR. Model 3: model 2 + NSAIDs, aspirin, or diuretics. Gender-specific cutoffs for SUA: Q1: < 312 in males, < 271 for females; Q2:312–370 for males, 271–325 for females; Q3:370–433 for males, 325–371 for females; Q4: > 433 for males, > 371 for females. AD, Alzheimer’s disease; APOE, apolipoprotein E; Aβ, amyloid-β; BMI, body mass index; CM-MMSE, China modified Mini-Mental State Examination; CSF, cerebrospinal fluid; eGFR, estimated glomerular filtration rate; NSAIDs, non-steroidal anti-inflammatory drugs; p-Tau, phosphorylated tau; Q, quartile; SUA, serum uric acid; t-Tau, total tau.
Sensitivity analyses using the usual hyperuricemia threshold
*Hyperuricemia: ≥417μmol/L for males, ≥357μmol/L for females. Analyses adjusted for age, gender, educational level, APOE ɛ4 status, and CM-MMSE. APOE, apolipoprotein E; Aβ, amyloid-β; CM-MMSE, China modified Mini-Mental State Examination; CSF, cerebrospinal fluid; p-Tau, phosphorylated tau; SUA, serum uric acid; t-Tau, total tau.
SUA linearly associated with CSF indicators of amyloidosis
SUA as a continuous variable was linearly associated with CSF indicators of amyloidosis. Af-ter adjusting for age, gender, educational level, APOE ɛ4 status, and CM-MMSE, higher SUA levels were associated with lower levels of CSF Aβ1 - 42 (β= –2.2×10–6; *p = 0.017) and CSF Aβ1 - 42/Aβ1 - 40 (β= –1.6×10–3; *p = 0.033), as well as higher levels of CSF p-Tau/Aβ1 - 42 (β= 4.4×10–4; *p = 0.007) and CSF t-tau/Aβ1 - 42 (β= 4.0×10–4; *p = 0.033), but not associated with CSF p-Tau and CSF t-Tau levels (Table 4, Fig. 2). Sensitivity analyses additionally adjusting cardiovascular risk factors (model 2) and drugs affecting SUA levels (model 3) revealed similar associations with CSF indicators of brain amyloidosis (Table 4). When repeating linear analyses in participants not taking NSAIDs, the associations with CSF indicators of amyloidosis were maintained (Supplementary Table 1).
The linear relationships between SUA and CSF biomarkers
Model 1: adjusted for age, gender, educational level, APOE ɛ4 status, and CM-MMSE. Model 2: model 1+ BMI, alcohol consumption, smoking, diabetes mellitus, hypertension, history of stroke, history of cardiovascular disease, total cholesterol, triglyceride, and eGFR. Model 3: model 2+ NSAIDs, aspirin, or diuretics. APOE, apolipoprotein E; Aβ, amyloid-β; BMI, body mass index; CM-MMSE, China modified Mini-Mental State Examination; CSF, cerebrospinal fluid; eGFR, estimated glomerular filtration rate; NSAIDs, non-steroidal anti-inflammatory drugs; p-Tau, phosphorylated tau; SUA, serum uric acid; t-Tau, total tau.
Linear associations between SUA and CSF biomarkers in total participants. SUA was negatively correlated with CSF Aβ1 - 42 (A) and Aβ1 - 42/Aβ1 - 40 (B) and was positively correlated with p-Tau/Aβ1 - 42 (C) and t-Tau/Aβ1 - 42 (D). All the above analyses were adjusted for age, gender, educational level, APOE ɛ4 status, and CM-MMSE. APOE, apolipoprotein E; Aβ, amyloid-β; CM-MMSE, China modified Mini-Mental State Examination; CSF, cerebrospinal fluid; p-Tau, phosphorylated tau; SUA, serum uric acid; t-Tau, total tau.
Evidence has indicated that SUA has possible U-shaped links with neurodegenerative diseases risk including cognitive impairment and dementia [45]. However, we did not observe any non-linear relationships between SUA level and CSF biomarkers of AD pathology (Supplementary Table 2).
Associations of SUA with CSF AD biomarkers in different subgroups
We also explored whether SUA in the associations with CSF AD biomarkers differed by APOE ɛ4 status or gender. Interaction analyses did not find significant strata differences regarding APOE ɛ4 status or gender (Supplementary Table 3), which thus suggested that the associations of SUA with CSF biomarkers of AD pathology were independent of the effects of APOE ɛ4 status or gender.
DISCUSSION
In the present study, the well-established CSF and clinical datum enabled us to examine the levels of SUA in the associations with AD pathophysiology (including amyloid and tau pathology and neurodegeneration) in humans deriving from a large sample population within the early stage of AD. The levels of SUA elevated in preclinical stage of AD and in subjects with aggregated amyloid pathology. In line with these observations, increased burden of cerebral amyloidosis reflected by CSF indicators (including CSF Aβ1 - 42, CSF Aβ1 - 42/Aβ1 - 40, CSF p-Tau/Aβ1 - 42, and CSF t-Tau/Aβ1 - 42) with elevated SUA levels allowed us to infer the potentially detrimental effects of SUA affecting amyloid metabolism in cognitively intact older adults. Sensitivity analyses yielded robust results, indicating the associations barely driven by intricate confounders. However, we did not find any non-linear relationships between them and any association of SUA with CSF tau proteins.
Mechanism evidence supported the primary findings of our study. An experimental study in vitro uncovered that UA could aggravate the detrimental effects of Aβ deposition on neuronal loss, neu-ronal cytoskeletal lesions, and neuronal cell interconnections [46]. The differentiated SHSY5Y neuroblastoma cells incubated with Aβ were used to establish an AD model in vitro. Higher UA concentrations significantly reduced cell viability and potentiated the pro-apoptotic effect of Aβ in neuronal cells [46]. As an initiating factor in the pathogenesis of AD, Aβ can trigger a cascade of neurodegenerative events inducing cognitive impairment across the AD spectrum [16, 47]. Studies have proved that the expression of integral membrane protein 2B (ITM2B) can not only inhibit amyloid precursor proteolysis and reduce the production of Aβ [48] but also inhibit glucose transporter 9 (GLUT9)-mediated urate uptake and reduce SUA concentration, thereby indirectly supporting the link between CSF Aβ and urate [49]. Recent cohorts revealed an involvement of SUA in occurrence of Lewy body disorder and an involvement of CSF UA in cognitive decline of Lewy body disorders possibly through the Aβ pathway [50]. There is currently a lack of studies that directly explore the relationship between SUA and tau pathology, and the role of UA in tauopathies has not yet been fully evaluated. Tommaso et al. observed that patients with different tauopathies had similar but significantly lower SUA levels compared to healthy controls, which indicated that low concentrations of SUA might represent a common risk factor for diverse tauopathies [51]. However, Tommaso’s study did not prove the protective effects of higher urate on tauopathies. In addition, a previous study of patients with primary tauopathies reported normal content of UA [52]. Due to the moderate sample sizes of the above studies, they could not rule out small effect sizes. And these studies are retrospective and did not evaluate many other potential factors affecting the SUA concentration. All these contributed to the inconsistent results. In conclusion, these studies do not support the hypothesis that UA is associated with other tauopathies in a similar manner as it is to Parkinson’s disease. Overall, available evidence implied that SUA influenced neurodegeneration via promoting Aβ pathology more than tau. However, further considerable experimental explorations in vitro by establishing AD models are still warranted to confirm the related molecular mechanism.
Indeed, our findings on the association between SUA and CSF biomarkers of AD pathology in the preclinical stage complemented the evidence of previous observational studies showing that raised SUA might cause cognitive decline and functional brain changes [9, 53–55]. Besides, a 12-year follow-up cohort of healthy elderly in the community (Three-City Dijon cohort) consistently showed that individuals with high SUA had higher risks of dementia [9]. Still, the association between SUA and dementia was under debated [56]. The observation of reduced SUA in AD patients indicated UA might have a potential protective role against aging process [11, 58]. Further, there was an association of low SUA with AD-related cerebral hypometabolism [51, 59]. Additionally, some longitudinal studies conducted in the Prospective Population Study of Women (PPSW) in Sweden and the Rotterdam Study also reported that risk of AD was lower among those with a higher SUA concentration [7, 11]. Interestingly, a longitudinal study noted a protective effect on cognition in MCI and dementia groups. However, this study also revealed, higher SUA was associated with faster cognitive decline in healthy subjects [60]. In the present study, we have observed that high SUA is associated with high Aβ pathology in cognitively intact older adults, which indicates SUA may aggravate brain amyloid deposition in preclinical AD. As for longitudinal studies, consistent with our findings, the Three-City Dijon cohort [9] reported that individuals with high SUA had a higher risk of dementia, although the Rotterdam Study [11] and the PPSW study [7] found that people with higher SUA had a lower risk of dementia. The explanation for the inconsistent results of the PPSW study and our study is that the former only included younger (47.4±6.2 years) females with relatively normal UA concentration at baseline. The Rotterdam study showed that only after correcting for cardiovascular risk factors, higher SUA was associated with a decreased risk of dementia. But this study cannot rule out the selective attrition of participants with relatively high levels of UA and concurrent worse cognitive function. However, our sensitivity analyses after correcting for cardiovascular risk factors yielded robust conclusions, which implied that these associations were less likely to be biased by confounders. Moreover, previous studies mainly explored the effects of UA on AD risk or cognitive function, while our study directly explored the association between SUA and the early pathology of AD. Therefore, our result reflects more refined biological metabolism in the body. Of course, the other probable explanation for these inconsistent findings is the diverse population backgrounds, because UA may exert different biological properties depending on its chemical microenvironment and its concentration in biological fluids [46, 61]. Urate radicals are generated to propagate the pro-oxidant state when UA is oxidized by peroxynitrites, thereby breaking the balance between pro-oxidation and anti-oxidation. In vitro and cellular evidence has underlined its pro-oxidant action, such as the aggravating oxidation of already oxidized low-density lipoprotein, and this dual role may be partly determined by the presence of transition metals [62]. In addition, Hershfield et al. found that urate maybe not a major factor controlling oxidative stress in vivo, which partly rejected the hypothesis that UA, as the main antioxidant, might reduce oxidative stress and protect against the detrimental effects of free radicals in the brain [63]. Furthermore, what most previous studies overlooked was that increased cardiovascular risk related to hyperuricemia might exceed the antioxidant effect of SUA. Together, the existing information implies that UA has complex biochemical effects, and the pro-oxidant property of UA may explain its detrimental association with neurobiological alterations in the development of AD.
This study mainly found strong associations between UA and CSF biomarkers of AD pathology and the non-significant U-shaped relationship. Amyloid and tau pathological proteins have accumulated silently before the clinical symptoms become evident [64]. High SUA might increase the risk of preclinical AD by inducing cerebral Aβ metabolism disturbance in the present study. Chronic hyperuricemia can directly cause endothelial dysfunction, vascular damage, and then cognitive impairment [65, 66]. Besides, multiple cardiovascular risk factors related to high SUA levels [67–69], such as hypertension and metabolic syndrome, are confirmed to induce a decline of chronic cerebral perfusion. Notably, cerebral perfusion dysfunction is associated with decreased clearance and increased deposition of amyloid plaques [66, 70].
Several limitations need to be acknowledged. Firstly, the cross-sectional datum limits us to infer the causality between SUA and CSF biomarkers, prospective cohorts are thus warranted to tract the neurobiological alterations of AD pathology with the changes of SUA. Next, our study population is limited to cognitively intact middle-aged and older adults, so this finding cannot be generalized to individuals with symptoms of MCI or dementia. Besides, compared with most neuroimaging/neuropathological studies reporting one-third of cognitively normal older adults had AD pathology, the subjects (mean age 62.16) included in the CABLE study were relatively young. However, in our study, the pattern of results was similar when biomarker abnormality was defined as having CSF Aβ1 - 42 levels in the lower quarter of the distribution of participants or having tau or p-tau levels in the upper quarter of the distribution (Supplementary Table 4), suggesting robustness to cut-point variations. Of course, in future, repeating this finding in more representative cohorts and neuropathological/neuroimaging studies would further supplement these identified associations. Further, the single measurement of SUA level and menopause affecting the SUA metabolism in females might lead to heterogeneous data. While the analyses in our study using SUA as a gender-specific categorical variable or a continuous variable both generated consistent findings and avoided any gender bias. Future research also should involve multiple measurements of SUA levels over time. Notably, we aimed to detect the association of CSF AD biomarkers with SUA levels and not with gout or hyperuricemia. Results delivered herein are thus not generalizable to hyperuricemia or gout research. Strengths, however, reinforce the significance of the present study. Large-sample and comprehensive data from the CABLE database ensured the credibility of the uncovered relationships. Besides, the investigations on cognitively intact population complemented evidence for the potential influence of SUA on neurobiological alteration of AD pathology and indicated the necessity of early manage plasma risk markers targeting the preclinical stage of AD. The significant associations deduced in the present study were less driven by confounding risk factors. Although NSAIDs could modulate systemic inflammation and oxidative stress, the findings of sensitivity analyses precluding participants with NSAIDs and drug treatments related to UA metabolism further validated the robustness and low heterogeneity of our study. Of course, large prospective studies are still needed to verify the long-term effects of uric acid on pathological biomarkers.
In summary, the findings corroborated the SUA as the potential risk factor of AD and demonstrated strong associations with CSF AD biomarkers in preclinical stage of AD, indicating high SUA might induce the disruption of brain Aβ metabolism, the-reby increasing the risk for developing AD. Besides, there was no evidence supporting the non-linear association. Large-scale and prospective studies considering time-varying effects of SUA are required to validate longitudinal alteration of AD pathology with increased SUA levels.
Footnotes
ACKNOWLEDGMENTS
The authors thank all participants of the present study as well as all members of staff of the CABLE study for their role in data collection. This study was supported by grants from the National Natural Science Foundation of China (91849126, 81571245, and 81771148), the National Key R&D Program of China (2018YFC1314702), Qingdao Applied Basic Research Project (18-2-2-43-jch), Shanghai Municipal Science and Technology Major Project (No.2018SHZDZX03) and ZJLab.
