Dementia is a clinical syndrome triggered by many differing pathologies, mainly including Alzheimer’s disease dementia (AD), vascular dementia (VD) and mixed AD/VD, dementia with Lewy bodies (DLB) and frontotemporal dementia (FTD) (Kokkinou et al., 2021). In 2030, it was estimated that there will be about 75.6 million people with dementia and the numbers are projected to rise to 131.5 million by 2050 (Ijaopo, 2017). Due to the incurable character of the disorder, most patients in aging often suffer various complications, high frequency of readmission, poor self-care ability, heavy psychological and physical burden, which may take up extensive social and economic resources. Therefore, the prevention of dementia relying on early diagnosis and distinction is a priority. Great effort has been made over the past decade to identify early and accurate disease-specific biomarkers as optimized tools for the differential diagnosis of dementia subtypes (Llorens et al., 2016; Maul et al., 2020). In 2020, Bai et al. demonstrated new potential proteins and molecular networks in brain tissue, cerebrospinal fluid during the progression of AD through proteomics based on mass spectrometry (MS) and multilayer omics methods, and further identified them in mouse models (Bai et al., 2020). However, to date, there are no valid options for the earliest possible and reliable diagnosis, but so far, these candidate biomarkers have only been established by AD, mostly requiring invasive and expensive procedures (Maul et al., 2020; Kokkinou et al., 2021). The clinical diagnosis of VD still depends on imaging criteria and there is no specific biochemical marker.

The Label-free proteomics quantitative techniques based on Liquid chromatography tandem-mass spectrometry (LC–MS/MS) are simple and practical, and play an important role in clinical research, biomarker discovery and personalized medicine recently (Krey et al., 2014; Zhao et al., 2020). Unprecedented proteomic coverage, such as in brain tissue, cerebrospinal fluid, and serum, has yielded new putative AD biomarkers (Bai et al., 2021). In addition, urine is an excellent biospecimen utilized as a desirable source of protein-biomarker discovery, because it can be collected recurrently by non-invasive techniques and reflect pathophysiologic state, especially in the early stages of disease (Barratt and Topham, 2007; Harpole et al., 2016). In addition, urine biomarkers could be used not only to diagnose but also to monitor the pathological changes of disease therapeutics (Zhang et al., 2018). The goals of this study, we sought to use LC–MS/MS label-free methods to characterize the urine proteome in VD, AD and normal controls (NC) subjects to identify disease-associated alterations in proteins and biological pathways. Furthermore, we aimed to identify urine proteins that could distinguish VD from both NC and AD by applying feature analysis to our urine proteomic profiles.

VD and AD are the leading causes of dementia in the elderly, together accounting for more than 70 to 75% of cases (Jia et al., 2020). The etiology of AD comprise multiple genetic and environmental factors, which is characterized by amyloid plaques and neurofibrillary tangles, (Knopman et al., 2021) while VD is defined as loss of cognitive function due to hypoperfusion, ischemic or hemorrhagic encephalopathy, which attributable to cerebrovascular or cardiovascular pathologies (Iadecola, 2013). However, due to frequent overlap with AD and the lack of specificity and sensitivity of clinical diagnostic criteria currently, it is difficult to distinguish VD from AD in clinical practice (Karantzoulis and Galvin, 2011). Thus, it is important to explore ancillary biological markers as a complement to clinical assessments for the differential diagnosis.

In this study, the urine proteomic results showed quantitative changes in patients with VD compared to patients with NC and AD groups; among 4,699 identified urine proteins, 939 and 1,147 proteins displayed quantitative changes unique to VD vs. NC and AD, respectively, including 484 overlapped common DEPs. Then, 10 unique proteins named in KEGG database (including PLOD3, SDCBP, SRC, GPRC5B, TSG101/STP22/VPS23, THY1/CD90, PLCD, CDH16, NARS/asnS, AGRN) were confirmed by a ROC curve method. SRC-AGRN was the strongest network node according to the combined score according to the PPI network with tightly connected clusters generated by the STRING database (Figure 6). SRC-AGRN was the top one network node according to the combined score (Figure 6; Supplementary Table 5).

Firstly, Collagen is the major component of extracellular matrix. Collagen cross-link and deposition depend on lysyl hydroxylation.Lysyl hydroxylase is composed of lysyl hydroxylase 1(LH1), LH2a, LH2b and LH3, (Scietti et al., 2018) among which LH3 encoded by human PLOD3 has a variety of enzymatic activities and catalyzes lysine hydroxylation and subsequent glycosylation of collagen (Scietti et al., 2018) PLOD3 gene is localized on chromosome 7Q36 and plays an important role in the biosynthesis of collagen IV and IV (Valtavaara et al., 1998). Abnormal PLOD expression and loss-of-function mutations have been implicated in many collagen-related diseases such as bone fragility with arterial rupture and deafness, and contractures or LH3 deficiency (Salo et al., 2008; Vahidnezhad et al., 2019). Notably, some recent studies have demonstrated that LH3 is essential for the intermolecular cross-linking and stabilization of collagen IV. LH3 deficiency affects the assembly and secretion of collagen IV and vascular basement membrane (BM) integrity (Li et al., 2019). In addition, LH3 overexpression can reduce the susceptibility of hypertensive ICH, and its regulation may be a feasible method for the prevention of intracranial hemorrhage (ICH) in the future (Li et al., 2019). In our study, the high expression of LH3 in VD may indicate an enhanced vasoprotective effect, which is distinct from AD and NC.

Another important protein marker is SRC, a member of the SRC family of non-receptor protein tyrosine kinases (PTKs), many of which are essential for growth, proliferation, and differentiation (Kalia et al., 2004). Five Src family kinases (SFKs) are expressed throughout the CNS in humans (including Src, Fyn, Yes, Lck and Lyn) and participate in many cellular functions (Ali and Salter, 2001). Specifically Src, as a key regulator of the NMDA (N-methyl-D-aspartate) subtype of glutamate receptors (NMDARs) and NMDAR-dependent synaptic plasticity and excitotoxicity may be crucial for learning and memory, pain, epilepsy and neurodegeneration (Yu and Salter, 1999; Salter and Kalia, 2004). Excitotoxicity mediated by NMDAR is associated with neuronal death in many pathological conditions, including CNS trauma, ischemia and neurodegeneration. This suggests that Src-mediated pathways and tyrosine phosphorylation of NMDAR may contribute to the pathophysiology of ischemic neuronal death in VD patients (Takagi et al., 1997, 1999; Cheung et al., 2003).

As shown in our result, AGRN-SRC was the top point of convergence according to the PPI network. Agrin (AGRN) is a kind of heparan sulfate proteoglycan rich in cerebrovascular basement membrane. It is a secreted neuroprotein, which is essential for synapse formation and stability (Bogdanik and Burgess, 2011). It is expressed in different tissues and non-neuronal cells, such as brain, lung, heart, kidney etc., and has been found to be regulated in various disease conditions, such as diabetes, cardiovascular, pulmonary, renal, muscle wasting, nerve and brain damage, and neurodegenerative diseases, and subsequently explored as a biomarker (Donahue et al., 1999; Rauch et al., 2011; Drey et al., 2015; Rauch et al., 2018). It has been reported that AGRN is important for maintaining the composition of the blood–brain barrier (BBB) in mouse models of AD and that changes in AGRN expression (especially vascular-associated AGRIN) affect a-amyloid (AA) homeostasis (Barber and Lieth, 1997; Berzin et al., 2000). Similarly, the effects of altered Agrn expression in the brain of VD patients affecting cerebrovascular and BBB may also be part of its pathological mechanism. However, how SRC-AGRIN interaction act in VD remains to be investigated. Perhaps advances in genetics and proteomics will provide answers to unanswered questions in this exciting field in the future.

In addition, we found a very interesting question: why are the majority of dysregulated proteins in VD upregulated proteins?

Nevertheless, this study also has certain limitations. First, the samples included were too limited to adjust the effects of some other factors, such as age, sex, blood pressure levels, potential drug effects etc. Second, VD is often difficult to distinguish from AD and mixed dementia because of the limited sensitivity and specificity of clinical criteria. Second, in urine proteomic studies, proteomic variability cannot be avoided, as has been demonstrated in many previous studies, (Nagaraj and Mann, 2011) due to technical variation or many physiological factors such as daily, gender, aging, diet, exercise, and intra - and inter-individual variation. Third, it is also a limitation that there is no cross-verification using any other biological specimen/analytical technique. Therefore, given the limitations of the research, our team is currently collecting more samples of patients and is preparing to further verify the specificity of each possible differently expressed protein using targeted proteomics and ELISA/Western Blot technique. We will expand the sample size, increase stratified analysis to adjust complicated influencing factors, standardize sampling procedures and develop more stringent p-values to attenuate this variability by continuing in future studies. In addition, we will also conduct in-depth analyses of VD proteomes and post-translational modifications (PTMs) to identify high-confidence biomarker candidates for VD by integrating multiple proteomes including cortex, cerebrospinal fluid, and serum in the future, as in previous AD studies (Wang et al., 2020).

This study demonstrated that we successfully screened ten upregulated urinary protein biomarkers of VD different from AD and NC using HPLC-MS/MS analysis, which is an effective and specific method. In addition, VD and AD share some urinary proteins, suggesting that they may share some similar pathophysiological mechanisms. SRC-AGRN interaction may act an essential role in VD pathological mechanism, which merit further attention.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found at: mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD022189.

The studies involving human participants were reviewed and approved by this study was approved by the Ethics Committee of the Second Xiangya Hospital of Central South University, and all participants provided written informed consent following the Declaration of Helsinki and the independent ethics committee or institutional review board. The patients/participants provided their written informed consent to participate in this study.

YZ designed the research and determined the structure of the paper. RC selected the references and contributed to the writing. YY collected the clinical data. WX and BZ helped to analyze experiment results. LZ contributed to the revision and finalization of the article. All authors contributed to the article and approved the submitted version.

This study was supported by the National Science and Technology Fundamental Resources Investigation Program of China (2018FY100900), the National Natural Science Foundation of China (81571151 and 81641039), and the Hainan Provincial Natural Science Foundation of China (822QN499).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.