Associations of plasma SMOC1 and soluble IL6RA levels with the progression from mild cognitive impairment to dementia

Abstract

Despite the central role attributed to neuroinflammation in the etiology and pathobiology of Alzheimer's disease (AD), the direct link between levels of inflammatory mediators in blood and cerebrospinal fluid (CSF) compartments, as well as their potential implications for AD diagnosis and progression, remains inconclusive. Moreover, there is debate on whether inflammation has a protective or detrimental effect on disease onset and progression. Indeed, distinct immunological mechanisms may govern protective and damaging effects at early and late stages, respectively. This study aims to (i) identify inflammatory mediators demonstrating robust correlations between peripheral and central nervous system (CNS) compartments by means of plasma and CSF analysis, respectively, and (ii) assess their potential significance in the context of AD and disease progression from mild cognitive impairment (MCI) to dementia. To achieve this, we have examined the inflammatory profile of a well-defined subcohort comprising 485 individuals from the Ace Alzheimer Center Barcelona (ACE). Employing a hierarchical clustering approach, we thoroughly evaluated the intercompartmental correlations of 63 distinct inflammation mediators, quantified in paired CSF and plasma samples, using advanced SOMAscan technology. Of the array of mediators investigated, only six mediators (CRP, IL1RAP, ILRL1, IL6RA, PDGFRB, and YKL-40) exhibited robust correlations between the central and peripheral compartments (proximity scores <400). To strengthen the validity of our findings, these identified mediators were subsequently validated in a second subcohort of individuals from ACE (n = 873). The observed plasma correlations across the entire cohort consistently have a Spearman rho value above 0.51 (n = 1,360, p < 1.77E-93). Of the high CSF-plasma correlated proteins, only soluble IL6RA (sIL6RA) displayed a statistically significant association with the conversion from MCI to dementia. This association remained robust even after applying a stringent Bonferroni correction (Cox proportional hazard ratio [HR] = 1.936 per standard deviation; p = 0.0018). This association retained its significance when accounting for various factors, including CSF amyloid (Aβ42) and Thr181-phosphorylated tau (p-tau) levels, age, sex, baseline Mini-Mental State Examination (MMSE) score, and potential sampling biases identified through principal component analysis (PCA) modeling. Furthermore, our study confirmed the association of both plasma and CSF levels of SPARC-related modular calcium-binding protein 1 (SMOC1) with amyloid and tau accumulation, indicating their role as early surrogate biomarkers for AD pathology. Despite the lack of a statistically significant correlation between SMOC1 levels in CSF and plasma, both acted as independent biomarkers of disease progression (HR > 1.3, p < 0.002). In conclusion, our study unveils that sIL6RA and SMOC1 are associated with MCI progression. The absence of correlations among inflammatory mediators between the central and peripheral compartments appears to be a common pattern, with only a few intriguing exceptions.

Keywords: Alzheimer's disease; Dementia; Disease progression; Inflammation; Mild cognitive impairment; SMOC1; sIL6RA.

Fig. 1

Global overview of pairwise comparisons involving Euclidean proximities calculated during the NNC analysis. We conducted this analysis on a subset of 485 individuals, carefully selected for having all protein measurements registered in both plasma and CSF, thus eliminating the need for any data imputation. NNC dendrogram generated by SPSS permitted the identification of the paired plasma-CSF protein levels with proximity scores below 400. The blue square highlights the plasma-CSF comparisons that appear considerably more distant. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 2

Scatterplots of proteins demonstrating highly significant correlations between plasma and CSF compartments (n = 1369).

Fig. 3

Plasma sIL6RA levels and their relationships with amyloid, p-tau, MCI to AD conversion, and sex. Horizontal lines presented in panels a and b show the lack of significant associations of plasma sIL6RA levels with CSF amyloid levels or CSF p-tau levels.