Progress in understanding primary glomerular disease: insights from urinary proteomics and in-depth analyses of potential biomarkers based on bioinformatics

Abstract

Chronic kidney disease (CKD) has become a global public health challenge. While primary glomerular disease (PGD) is one of the leading causes of CKD, the specific pathogenesis of PGD is still unclear. Accurate diagnosis relies largely on invasive renal biopsy, which carries risks of bleeding, pain, infection and kidney vein thrombosis. Problems with the biopsy procedure include lack of glomeruli in the tissue obtained, and the sampling site not being reflective of the overall lesion in the kidney. Repeated renal biopsies to monitor disease progression cannot be performed because of the significant risks of bleeding and kidney vein thrombosis. On the other hand, urine collection, a noninvasive method, can be performed repeatedly, and urinary proteins can reflect pathological changes in the urinary system. Advancements in proteomics technologies, especially mass spectrometry, have facilitated the identification of candidate biomarkers in different pathological types of PGD. Such biomarkers not only provide insights into the pathogenesis of PGD but also are important for diagnosis, monitoring treatment, and prognosis. In this review, we summarize the findings from studies that have used urinary proteomics, among other omics screens, to identify potential biomarkers for different types of PGD. Moreover, we performed an in-depth bioinformatic analysis to gain a deeper understanding of the biological processes and protein–protein interaction networks in which these candidate biomarkers may participate. This review, including a description of an integrated analysis method, is intended to provide insights into the pathogenesis, noninvasive diagnosis, and personalized treatment efforts of PGD and other associated diseases.

Keywords:

• Primary glomerular disease
• urine
• proteomics
• biomarker
• diagnosis

Introduction

In the past few decades, chronic kidney disease (CKD) has become a global public health challenge [1]. The prevalence of CKD is high, between 11 and 13%, and glomerular disease (GD), especially primary glomerular disease (PGD), is one of the most common causes of CKD [2]. Patients with PGD may develop end-stage renal disease (ESRD) [3]. Determining the pathological type of PGD together with understanding the pathogenesis of the disease is therefore pertinent and conducive to treatment. However, the invasive method of renal biopsy remains the gold standard for determining the pathological type of PGD, and there is a lack of painless diagnostic methods for PGD. Renal biopsy is challenging for patients with severely impaired renal function. Risks associated with renal biopsy include bleeding, pain, infection and kidney vein thrombosis. Problems with biopsy include lack of glomeruli in the tissue obtained, and the sampling site not being reflective of the overall lesion in the kidney. Repeated renal biopsies to monitor disease progression cannot be performed because of the significant risks of bleeding and kidney vein thrombosis. Thus there is great interest in identifying biomarkers for the diagnosis of PGD.

PGD is an immune-mediated inflammatory disease, and the immune complex can damage the glomeruli, leading to clinical symptoms [4]. The application of proteomics technology using mass spectrometry (MS) has helped to identify a number of potential biomarkers in urine and has recently played an important role in studying the pathogenesis of several diseases as well as achieving diagnosis and prognosis without the use of biopsies [5].

In this review, we summarize articles on urinary proteomics involving PGD and provided an overview of the developments in this area. Additionally, differentially expressed proteins (DEPs) were selected for in-depth analysis using a bioinformatic approach.

Advantages of urinary proteomics in the study of glomerular disease

Urine is an easily accessible biological fluid with the advantage of noninvasive collection [6]. When there is an abnormality in the urinary system or a kidney-related disease, the levels of several proteins in the urine change. Thus, urine is an ideal source for identifying biomarkers related to kidney disease [7].

Proteomics is the study of all proteins expressed by specific cells, tissues, organs, or organisms, and it can reveal changes in protein levels at a given time. Various functional proteins are not only important active substances with distinct physiological roles but also can be directly related to pathological states. Urinary proteomics describes all the proteins in the urine, the levels of which may change in disorders of the kidney and following kidney damage. Increasing attention has been given to urinary proteomics studies, especially in the following fields: identifying biomarkers for noninvasive diagnosis, studying the function of DEPs, and investigating interactions between target proteins to understand the mechanisms of pathogenesis.

Many emerging technologies have been developed based on core proteomics methods, and proteomics technologies based on MS have played crucial roles in identifying biomarkers in urine for the diagnosis of GD, and in expanding our knowledge regarding its origin.

The roadmap for the study of biomarkers in the urinary proteome

When investigating biomarkers in urine, the first step is to identify individuals with the disease of interest along with appropriate controls. Then, their urine is collected. The participants should be divided into discovery and validation groups. Urine samples should be processed by a suitable method [8]. The urine collected from individuals in the discovery group may be analyzed using proteomics technologies to identify DEPs. The most commonly-used proteomic technique for screening biomarkers is MS. Chu’s team identified three DEPs using MS between patients with type 2 diabetes and HCs [9]. Fang et al. collected urine samples from four patients with IgAN, four with Henoch-Schönlein purpura nephritis (HSPN), and four HCs [10], and applied LC-MS/MS to identify DEPs between the disease groups and HCs. While 276 DEPs between IgA nephropathy (IgAN) patients and HCs, and 125 DEPs between HSPN patients and HCs were found, they observed that urinary α-1B-glycoprotein and afamin were considerably increased in children with IgAN and HSPN compared with HCs [10]. The expression of these candidate biomarkers was assessed in a validation set of 15 patients with IgAN, 15 patients with HSPN, and 10 HCs. Then they performed bioinformatic analysis using Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment, and protein–protein interaction (PPI) analyses of the DEPs [10]. Through GO annotation enriched analysis, they found that most of the DEPs in the two diseases related to cell adhesion and biological regulation and that the DEPs were distributed mainly in the extracellular matrix and intrinsic component of the plasma membrane. KEGG pathway annotation analysis showed that the DEPs of IgAN and HSPN were involved in cell adhesion and the complement and coagulation cascades. The PPI analysis showed the complex interactions between these DEPs.

The protein concentration in urine is very low under physiological conditions (urine protein excretion <150 mg/24 h), approximately 1/1000th the plasma protein content. Thus, the presence of serum protein in urine could be indicative of glomeruli damage or another pathological state of the urinary system; consequently, changes in the levels of several target proteins in the urine may help to characterize the underlying molecular pathways of certain diseases.

Despite the importance of MS-based urine proteomics in the screening of biomarkers and the study of the pathological mechanisms of kidney diseases, urine proteomics studies also have a number of drawbacks. Proteins of low abundance are not easily detected, and urine contains high concentrations of urea and salts, which can interfere with MS results. Therefore, the removal of high-concentration proteins such as albumin, as well as salts, from urine is necessary for the identification of low-abundance proteins in MS studies. It is important to note that 70–75% of urine protein is produced by the urinary system, and 25–30% comes from plasma filtration, thereby making urine a suitable source for the study for biomarkers of urinary diseases. In this regard, urinary proteomics has a key role in further delineating the pathogenesis of PGD.

The common procedures for studies of urinary proteomics are shown in Figure 1.

Figure 1. Common research steps on urinary proteomics. The first stage is to identify candidate biomarkers via proteomics analysis. When differentially expressed proteins are screened, candidate biomarkers in the validation group can be targeted. After identifying biomarkers, functional analysis of specific proteins can be carried out.

Urinary proteomics in healthy people

In 1996, Wilkins first put forward the concept of proteomics as the study of all the protein components expressed by the genome in specific cells or tissues [11]. In the same year, Marshall et al. for the first time separated proteins from human urine using two-dimensional electrophoresis, which helped in creating a two-dimensional gel map of the human urine proteome [12]. The following year, Heine et al. identified 34 highly expressed proteins via coupled microbore high-performance liquid chromatography and electrospray ionization coupled with tandem MS [13]. Since then, proteomics research has developed rapidly.

The combination of two different separation techniques, acetone precipitation and ultracentrifugation, contributed to the identification of more proteins, and a relatively simple urinary proteome map of healthy people was constructed [14]. Adachi et al. analyzed urinary proteins using liquid chromatography combined with linear ion trap-Fourier transform and a linear ion trap-orbitrap MS and drew up the first comprehensive protein expression map of healthy human urine [15]. Zhang et al. observed the differential expression of urinary proteins between healthy Li and Han ethnic groups in a study using LC-MS/MS [16]. A total of 25 urinary proteins were found to differ significantly, 16 of which had been previously reported to be biomarkers of many diseases, e.g. Ig kappa chain V-III region (which is increased in the urine of type 2 diabetic patients [17]) and neutrophil gelatinase-associated lipocalin (reported to be a urinary biomarker of patients with acute kidney injury and urinary tract infection [18,19]). The authors confirmed that protein variability was based on ethnic rather than random differences [16]; thus, ethnic group differences may be an influencing factor in the study of urinary proteomics and should be considered when urinary samples are used for biomarker discovery. Using a new enrichment method (lectin affinity chromatography), Marimuthu et al. identified 671 new proteins [20]. Moreover, several proteins that were previously thought to be present only in plasma were detected in urine. Thus, there are numerous connections between blood and urinary proteins among healthy people.

Because different techniques were used in these studies, the direct comparison between results is difficult, which has in turn limited the development of proteomics to a certain extent [21]. Nevertheless, progress made in research on urinary proteomics of healthy people has concomitantly allowed research on kidney diseases to flourish.

Progress of urinary proteomics in studying primary glomerular disease

Urinary proteomics and IgA nephropathy

IgA nephropathy (IgAN) is the most common chronic proliferative glomerulonephritis worldwide. Approximately one-third of patients progress to ESRD, and the associated mortality is steadily increasing [22,23]. In Europe, IgAN accounts for 19–51% of the total renal biopsy patients with GD [24,25]. The disease is related to galactose-deficient IgA1 (Gd-IgA1) [26,27]. While the pathogenesis of IgAN remains unclear, the theory of “four blows” is current the common consensus [28–30]. This theory suggests that the development of the disease involves four main steps: hit 1. production of galactose-deficient lgA1; hit 2. production of anti-galactose-deficient lgA1 autoantibodies; hit 3. formation and deposition of immune complexes; and hit 4. deposition of immune complexes in the mesangial area and activation of the complement system, leading to glomerular damage (Figure 2(A)). The main pathological feature of IgAN is the granular or massive deposition of immune complexes, mainly IgA, in the glomerular mesangial region [31–34] (Figure 3(A)). Interestingly, glycosylation of autoantibodies also occurs in other autoimmune diseases such as rheumatoid arthritis [35]. Therefore, there may be a relationship between IgAN and rheumatoid arthritis. Although potentially promising, this hypothesis warrants further examination.

Figure 2. The current widely accepted pathogenesis and pathological characteristics of several glomerular diseases. (a) The popular theory of “four blows” on the pathogenesis of IgA: hit 1, production of galactose-deficient lgA1; hit 2, production of anti-galactose-deficient lgA1 autoantibodies; hit 3, immune complex formation and deposition; hit 4, deposition of immune complexes in the mesangial area, activating the complement system and causing glomerular damage. (b) Depiction of immune complexes deposited under the epithelial cells of GBM, resulting in its thickening and podocyte injury, which could damage the barrier of glomerular filtration and cause further proteinuria. (c) Depiction of glomerular changes in patients with MCD and FSGS. The glomerular lesion of the former is mild, while that of the latter shows local sclerosis. GBM: glomerular basement membrane.

Figure 3. Pathological features of different glomerular diseases. (a) Immunofluorescence microscopy in a patient with IgA shows IgA deposition along the mesangial area (original magnification, 200×). (b) Glomerulus from a patient with idiopathic membranous nephropathy shows the thickened GBM stained with PASM (arrow) under a light microscope. (c) HE staining shows mild glomerular lesions and mild mesangial hyperplasia (arrow) from a patient with MCD. (d) The glomeruli with segmental sclerosis (arrow) are shown via PAS staining in a patient with FSGS. These images were provided by the Department of Nephrology of Shengjing Hospital. HE: hematoxylin-eosin; PAS: Schiff periodic acid; PASM: periodic acid-silver metheramine.

Preliminary studies on the urinary proteome of patients with IgAN were performed to detect DEPs that would differentiate IgAN from other nephropathies. A previous study established the urinary polypeptide maps of patients with IgAN using two-dimensional electrophoresis, which distinguished these individuals from HCs and patients with membranous nephropathy (MN), minimal change disease (MCD), focal segmental glomerulosclerosis (FSGS), as well as diabetic nephropathy [36]. Another team also used DEPs to distinguish IgAN from HCs and other kidney diseases. [37]. It is worth mentioning, the overall specificity of this panel was as high as 82.3%. Subsequently, with more data being obtained, together with advances in recent technology, several researcher groups screened for markers related to the severity of IgAN and established relevant models [38,39]. The well-known IgAN 237 classifier, which contains 237 different urinary peptides of progressive and nonprogressive IgAN and was developed through CE-MS, was found to indicate the severity of IgAN accurately [40].

One team examined the urinary proteome of Uygur-ethnic patients with IgAN in Xinjiang, and screened out four candidate biomarkers (adiponectin, antithrombin-III, intercellular adhesion molecule 1, and metalloproteinase inhibitor 1) [41]. However, another study that included urine samples from children revealed different biomarkers through CE-MS/MS, while also reporting that the contents of α-1B-glycoprotein and afamin in the urine were higher in children with IgAN than in healthy children [10]. The variability of the results might be caused by variances in the chosen sample sources and techniques used as per the particular purpose of each study.

We have summarized urinary proteomic studies on IgAN in Table 1.

Table 1. Summary of research progress on urinary proteomics of glomerular disease.

Nephropathy type	Method	Sample: training set/test set	Signature	Clinical utility	Comment	Year	Reference
IgAN	CE-ESI-TOF-MS, MALDI-TOF-MS	IgAN-45, MN-13, HC-57, MCD-16, FSGS-10, DN-23	22 Proteins/peptides that could discriminate between HC and either MN or IgAN	Diagnostic value and evaluating the prognosis of IgAN	Pattern could discriminate IgAN from HCs and MN with a sensitivity of 100 and 77%, and specificity of 90 and 100%, respectively.	2005	[36]
	CE-MS	IgAN-45, HC-207, DC-253/IgAN-10, HC-12, DC-27	Panel of 25 peptides	Diagnosis	25-Peptide panel could separate IgAN from HC and patients with other types of renal diseases, with an overall specificity as high as 82.3%.	2007	[37]
	MALDI-TOF MS	IgAN (55, comprising 23 patients with severe pathologic presentation and 32 patients with mild ones), HC-14	Diagnosis model of mild pathological injury and severe pathological injury (model 1); diagnosis model of mild pathological injury and HC (model 2); diagnosis model of severe pathological injury and HC (model 3).	Diagnosis and prognosis	Models could distinguish between IgA severe pathological damage, mild pathological damage, and healthy people. Sensitivity: 90.48% (model 1), 93.55% (model 2), 100% (model 3). Specificity: 96.77% (model 1), 85.71% (model 2), 92.86% (model 3).	2012	[38]
	nLC-MS/MS, GeLC-MS/MS	IgAN (13)	Afamin; leucine-rich α-2-glycoprotein, ceruloplasmin, α-1-microglobulin, hemopexin, apolipoprotein A-I, complement C3, vitamin D-binding protein, apolipoprotein A-IV, beta-2-microglobulin, retinol-binding protein 4	Assessing the prognosis and pathological severity of IgAN	Classification model could be built which demonstrated 60% sensitivity and 87% specificity in comparison with Lee’s classification and 80% sensitivity and 100% specificity in comparison with the Oxford classification.	2013	[39]
	iTRAQ, LC‑MS/MS	IgAN (30), HC (30)	Afamin↑, serum albumin↑, apolipoprotein A‑I↑, zinc‑α‑2‑glycoprotein↑, α‑2‑macroglobulin↑, carbonic anhydrase 1↑, ceruloplasmin↑, complement C3↑, complement C4A↑, fibronectin↓, prothrombin↑, vitamin D‑binding protein↑, haptoglobin↑, insulin‑like growth factor‑binding protein 7↓, proactivator polypeptide↑, α‑1‑antitrypsin↑, α‑1‑antichymotrypsin↑, antithrombin‑III↑	Diagnosis and aid in understanding the pathogenesis of IgAN	Number of potential biomarkers for patients with IgAN identified	2014	[42]
	Nano-scale LC-MS	IgAN-13, HC-8	CD44↓, GP2↓, epidermal growth factor (EGF)↓, CLM9↓, protocadherin-1 (PCDH1)↑, uteroglobin (UTER)↑, dipeptidyl peptidase 4 (DPP4)↑, NHL repeat-containing protein 3 (NHLC3)↑	Identified novel biomarkers for diagnosis and further delineated the pathogenesis pathways of IgAN	Diagnostic model created with accuracy of 97%.	2015	[43]
	iTRAQ-MS	IgAN-4, HC-4/ IgAN-30, HC-10, MCD-10, MN-10	Complement C9↑, Ig kappa chain C region↑, Keratin- type I cytoskeletal 10↓, Keratin- type II cytoskeletal 1 ↓, Keratin- type II cytoskeletal 5↓	Diagnosis	It was speculated that certain immunological indicators in urine may be related to glomerular lesions.	2017	[44]
	2D-LC-MS/MS and iTRAQ	IgAN-5, HC-5/ IgAN-7, HC-7	Adiponectin↑, antithrombin-III↑, intercellular adhesion molecule 1↑, metalloproteinase inhibitor 1↑	Diagnosis	Four markers were found in the urine of patients with IgAN, with AUCs of 0.701, 0.840, 0.840 and 0.639, respectively.	2018	[41]
	CE-MS	IgAN progressors-47, IgAN non-progressors-47/IgAN progressors-23, IgAN non-progressors-23	Collagen I ↓ (in progressors), α-1 antitrypsin↑ (in progressors)	Evaluated the prognosis	IgAN237 classifier that contains 237 differential urine peptides of progressive IgAN and nonprogressive IgAN showed a value for the prediction of IgAN progression with the AUC of 0.89.	2021	[40]
	LC-MS/MS	IgAN-4, HC-4/IgAN-15, HC-10	α-1B-glycoprotein↑, afamin↑	Diagnosis	276 Proteins identified with potential use for achieving noninvasive diagnosis in children with IgAN.	2021	[10]
MN	MALDI-TOF MS, HPLC-MS/MS	MN-6, FSGS-9, MCD-3, IgAN-9, HC-7/MN-7, FSGS-8, MPGN-3, MCD-3, IgAN-9, MPGN-3, HC-7	UMOD↓, A1AT↑	Diagnosis of different types of GKD	Workflow of urinary peptide profiling described that could distinguish GD noninvasively with a sensitivity of 91.7% and specificity of 85.7%.	2012	[45]
	LC-MS/MS, iTRAQ	MN-5, FSGS-5, HC-5/MN-2, FSGS-2, HC-2	Lysosome membrane protein-2 (LIMP-2)↑	Diagnosis and insight gain into the pathogenesis of PMN	Value of urinary microvesicles in identifying candidate biomarkers was demonstrated	2015	[46]
	LC-MS/MS	NS-16 (MN-4, MCD-4, FSGS-4), HC-4/ MCD-13, MN-26, FSGS-5, IgAN-9, HC-8	SERPINA7, CD44	Diagnosis of three main causes of NS.	Protein panel developed that could discriminate among MN, MCD, and FSGS. The logistic regression models with three proteins (C9, CD14 and SERPINA1) could discriminate MN from MCD and FSGS with the AUC of 0.852 and 0.817 respectively.	2017	[47]
	TMT1 and TMT2+ nano LC-MS/MS	MN positive for anti-PLA2R antibody-23, MN negative for anti-PLA2R antibody-22, HC-23/MN positive for anti-PLA2R antibody-9, MN negative for anti-PLA2R antibody-9, HC-9	α-1-Antitrypsin (A1AT)↑, afamin (AFM)↑	Contributed in further elucidating the pathogenetic mechanisms of PMN	Abnormal expression of A1AT and AFM in the urine of PMN patients could help to elucidate its pathogenetic mechanisms.	2018	[48]
	ELISA	PMN-134, FSGS-72, MCD-18	C3a↑, C5b-9↑	Understanding the pathogenesis of PMN	Findings helped in promoting understanding of the role of complement activation in the pathogenesis of PMN	2019	[49]
	MALDI-TOFMS, DIA-MS	PMN-22, HC-20, DC-15 (MCD-5, FSGS-5, IgAN-5)/FSGS-12, IgAN-26, MCN −8, IMN-21, HC-8	Calpain↑, matrix metalloproteinase 14↑, and cathepsin S↑ (in kidney tissues of patients with PMN)	Diagnosis	Role of proteases in PMN through urinary peptidomics was systematically explored.	2021	[50]
	CE-MS	DN-11, MCD-14, MN-23, HC-10	The combination of urinary afamin and complement C3 urine/plasma ratio	Diagnosis (distinguishing between MN and DN)	The combination of urinary afamin and complement C3 urine/plasma ratio could help distinguish between MN and DN with an AUC of 0.9842.	2021	[51]
MCD/FSGS	MALDI-TOF MS, HPLC-MS/MS	MN-6, FSGS-9, MCD-3, IgAN-9, HC-7/MN-7, FSGS-8, MPGN-3, MCD-3, IgAN-9, MPGN-3, HC-7	UMOD↓ (MCD or FSGS vs HC), A1AT↑ (MCD or FSGS vs HC)	Disease diagnosis and prognosis	Workflow described in which urinary peptide profiling was combined with histological findings. This could distinguish different types of GD noninvasively, with high sensitivity (91.7%) and specificity (85.7%).	2012	[45]
	MALDI-TOF MS, HPLC-MS/MS	MCD- 11, FSGS- 11, HC-16/ MCD- 11, FSGS- 11	Fragment of uromodulin (FSGS vs. MCD)↑, a fragment of α-1-antitrypsin (FSGS vs. MCD)↓	Distinguishing patients with FSGS from those with MCD with similar pathological features	Prediction model generated to classify patients with MCD and FSGS (correctly classified 81.8% of patients with MCD and 72.7% of those with FSGS)	2014	[52]
	Nano-flow LC-MS/MS	FSGS-11 (with different prognosis features)	Patients with poor prognosis: ribonuclease (RNAS2)↑, haptoglobin (HPT)↓, CD59 glycoprotein (CD59)↑, Prostaglandin H2 D-isomerase (PTGDS)↑, Beta-2-microglobulin (B2MG)↑, α-1-microgolbulin (AMBP)↑	Monitoring prognosis	A predictive model with 100% accuracy was constructed.	2014	[53]
	Nano-LCMS/MS	FSGS-11, IgAN-6, HC-8	CD59↑, CD44↑, IBP7↓, Robo4↑, DPEP1↓	Diagnosis	A panel of candidate biomarkers can be applied to the diagnosis of FSGS noninvasively, while the complete absence of DPEP1 may represent a novel marker of FSGS	2014	[54]
	LC-MS	MCD-4, MN-4, FSGS-4, HC-4/MCD-13, MN-26, FSGS-5, IgAN-9, HC-8	C9(MCD vs.MN) ↑, CD14(MCD vs.MN)↑, SERPINA1(MCDvs.MN) ↑	Diagnosis	The logistic regression model with three proteins (C9, CD14 and SERPINA1) could help discriminate FSGS from MCD, with an AUC of 1.	2017	[47]
	2D-DIGE-MS	MCD-10, FSGS −11/MCD-14, FSGS-14	A1AT, transferrin histatin-3 39S ribosomal protein L17↓ (FSGS vs MCD), calretinin↑ (FSGS vs MCD)	Diagnosis	Panel of proteins differentiated FSGS from MCD.	2017	[55]
	CE-MS	FSGS-79, MCD-25/FSGS-31, MCD-10	Uromodulin↑, keratin ↑, apolipoprotein C-IV↑, β-2-microglobulin↓, clusterin↓, complement C3↓ (MCD vs other types of CKD) Collagens↑, A1AT, clusterin↓, uromodulin ↓, polymeric immunoglobulin receptor↓, Golgi-associated olfactory↑, signaling regulator↓ (FSGS vs other types of CKD)	Diagnosis	SVM-based classifiers could differentiate different types of CKD in a noninvasive manner (AUC ranged from 0.77 to 0.95).	2017	[56]
	2D-LC-MS/MS	NS-5/FSGS −30, MCD-30, HC-30	Ubiquitin-60S ribosomal protein L40 (UBA52) (FSGS vs MCD) ↑	Diagnosis	Potential novel urinary protein biomarkers for NS identified.	2017	[57]
	CE-MS	DN-11, MCD-14, MN-23, HC-10	Urinary retinol- binding protein 4, SH3 domain-binding glutamic acid-rich-like protein 3	Diagnosis	The combination of urinary retinol-binding protein 4 and SH3 domain-binding glutamic acid-rich-like protein 3 could help distinguish between MCD and DN, with an AUC of 0.9740.	2021	[51]

2D-DIGE-MS: two-dimensional differential gel electrophoresis coupled with mass spectrometry; 2D-LC-MS/MS: two-dimensional liquid chromatography coupled to tandem mass spectrometry; AUC: area under the receiver operating characteristic curve; CE-ESI-TOF-MS: capillary electrophoresis coupled on-line to an electrospray ionization-time-of-flight mass spectrometry; CE-MS: capillary electrophoresis coupled mass spectrometry; CE-MS: capillary electrophoresis-mass spectrometry; DC: disease control; DIA-MS: data-independent acquisition MS; DN: diabetic nephropathy; ELISA: enzyme linked immunosorbent assay; FSGS: focal segmental glomerulosclerosis; GeLC-MS/MS: gel electrophoresis followed by liquid chromatography tandem mass spectrometry; HC: healthy control; HPLC-MS/MS: high-performance liquid chromatography coupled to a tandem mass spectrometer; IgAN: IgA nephropathy; iTRAQ: isobaric tags for relative and absolute quantitation; LC-MS/MS: liquid chromatography with tandem mass spectrometry; MALDI-TOF-MS: Matrix-assisted laser desorption ionization time-of-flight mass spectrometry; MCD: minimal change disease; MN: membranous nephropathy; MPGN: membranoproliferative glomerulonephritis; nano LC-MS/MS: nano-scale liquid chromatography coupled to tandem mass spectrometry; nLC-MS/MS: nano-scale liquid chromatography-mass spectrometry; NS: nephrotic syndrome; PMN: primary membranous nephropathy; TMT: tandem mass tag; ↑: upregulated protein; ↓: downregulated protein.

As these collective findings highlight, potential IgAN biomarkers are related mainly to podocytes, immune complexes, and complement components. Podocytes, together with basement membrane and capillary endothelial cells, constitute the glomerular hemofiltration barrier and are one of the main cell types that maintain normal structure and function of the glomerular hemofiltration membrane. When podocytes are damaged, symptoms such as proteinuria occur. IgAN is an autoimmune disease. Galactose-deficient IgA1 exposes its hinge region epitopes, resulting in the production of IgA antibodies against abnormal IgA1, along with the formation and deposition of immune complexes in the mesangial region. IgA deposited in the mesangial region can activate the complement system through the bypass pathway and mannose-binding lectin pathway, which promotes inflammation and further pathological injury. Therefore, the level of complement components in urine may be closely related to the disease.

Urinary proteomics and membranous nephropathy

Membranous nephropathy (MN) is the main pathological type of adult nephrotic syndrome. It can occur in all age groups, although children are less likely than adults to develop the disease [58–62]. Approximately, 30% of MN patients progress to ESRD [63]. The main pathological feature of MN is the appearance of autoantibodies and activation of the complement system, which together form a membrane attack complex that targets the glomerular podocytes and causes podocyte damage and proteinuria. Immune complexes are deposited under epithelial cells of the glomerular basement membrane, leading to its thickening [64,65] (Figures 2(B) and 3(B)). The M-type phospholipase A2 receptor (PLA2R) is the main antigen of primary MN [66,67]. The detection of serum anti-PLA2R antibody levels has been widely used as a noninvasive diagnostic method in clinical settings [68,69].

Several studies have been published on the urinary proteomics of MN. In a study using MALDI-TOF-MS/MS to compare the urinary proteome of passive Heymann nephritis in mice with the proteome prior to modeling, the content of α-1-antitrypsin (A1AT) increased in the urine of the nephritic mice [70], Several proteomic studies have also found A1AT to be abnormally expressed in the urine of patients with kidney diseases [40,42,45,47,48,71]. A1AT is the most abundant serum serine protease inhibitor, whose main physiological role is to inhibit the activity of specific serine proteases [48,72]. A1AT overexpression can lead to inhibition of neutrophil elastase, accelerating the accumulation of mesangial matrix [72]. As MN is a disease related to glomerular podocyte injury and A1AT is found in the cytoplasm of sclerotic glomeruli podocytes, further studies that examine the relationship between A1AT and podocyte dysfunction may help in deepening our understanding of this disease [73]. Lin et al. analyzed polypeptides in the urine of patients with MN using LC-MS/MS and screened for DEPs [50]. They predicted the proteases that were related to the DEPs through relevant database investigations and provided the impetus for identifying biomarkers of other kidney diseases using a similar method. Another study found that levels of C3a and C5a in the urine of patients with MN increased significantly compared to controls, indicating that the complement system was strongly activated [49]. Renal biopsy has demonstrated the deposition of complement complexes in patients with MN. The products of complement activation are expected to serve as sensitive biomarkers for disease diagnosis, and the mechanism related to complement activation may also provide additional insight for the treatment of idiopathic membranous nephropathy.

Choi et al. performed LC-MS/MS on urine samples from 16 patients with MCD, MN, and FSGS and HC (4 individuals each) to identify candidate biomarkers for the differential diagnosis of these diseases [47]. After validating several proteins by enzyme-linked immunosorbent assay in a validation set of 51 patients, they reported that SERPINA7 and CD44 were specific to MN. The urinary proteomic studies performed by different teams related to MN along with the relevant contents are summarized in Table 1.

Urinary proteomics and minimal change disease and focal segmental glomerulosclerosis

Minimal change disease (MCD) and focal segmental glomerulosclerosis (FSGS) are diseases related to podocyte injury and are characterized by excessive proteinuria [74–77]. Of the two, MCD often occurs in children and adolescents [78,79]. Glomeruli do not change drastically in patients with MCD, but extensive fusion of the podocyte foot process can be observed (Figures 2(C) and 3(C)). Patients with FSGS often present with hematuria and hypertension, while podocytes are also damaged. Segmental hyaline sclerosis can be detected in a number of glomeruli during light microscopy (Figure 3(D)).

Podocyte injuries in MCD and FSGS are similar, a finding that has prompted several researchers to regard MCD and FSGS as the same disease [80]. Under this assumption, MCD was thought to represent early podocyte injury and FSGS, a late form [81]. Despite their similar characteristics, there are many differences between the two pathologies. Patients with MCD often have better prognosis, whereas patients with FSGS are more likely to develop renal insufficiency. Glucocorticoid therapy is more effective against MCD than against FSGS [1,75,82]. In addition, urinary CD80 levels are increased in patients with MCD, but not in those with FSGS [83]. C1q level is high in FSGS patients, but mannose-binding lectin is rarely observed [84]. The foot process width in patients with FSGS is higher than that seen in patients with MCD, which may suggest more severe podocyte pathological changes [85,86].

Several classic omics studies have identified biomarkers of MCD and FSGS. Williams et al. were the first to undertake molecular analysis at the genomic level in glomeruli and tubulointerstitial regions with histological characteristics and to compare the differences in small RNA maps between patients with FSGS and HCs [87]. They observed a large number of small RNAs that were differentially expressed between the two groups [87]. Nafar et al. analyzed urine samples of six patients with IgAN, eleven individuals with FSGS and eight healthy people using nano-scale liquid chromatography tandem mass spectrometry and reported significant changes in the differential expression of CD59, CD44, IBP7, Robo4 and DPEP1 in people with FSGS [54]. Another study utilizing this technique found differences in the urinary protein landscape of patients with steroid-resistant and steroid-sensitive FSGS [88]; in particular, they observed that apolipoprotein A-1 (APOA1) expression was markedly elevated in the urine samples of the latter patients compared to the former [88].

Pérez et al. compared the urinary proteome between patients with MCD and FSGS using two-dimensional differential gel electrophoresis coupled with MS, and reported that levels of A1AT, transferrin, histatin-3 and 39S ribosomal protein L17 increased, while levels of calretinin decreased, in MCD patients relative to those in FSGS patients [55]. Urinary fibrinogen could also aid in distinguishing between the two diseases [89]. Choi’s study found that CD14, C9, and SERPINA1 were specific to MCD. [47]. Certain candidate biomarkers specific to FSGS include calmodulin-like mucin 26, RNase A family 1, DIS3-like nucleic acid exonuclease 1, and Golgi-associated olfactory signaling regulators [47,56].

These DEPs may be involved in the development of MCD and FSGS and could potentially be used for the noninvasive differential diagnosis and monitoring of treatment of the two pathologies.

The abnormal expression of these proteins in urine may reflect mass cell death and release of cell contents during the separation of podocytes from glomeruli. These results may also suggest specific roles for immunity, inflammation, and apoptosis in the development of FSGS. Cell proliferation, differentiation, and death may be related to the development of MCD [90]. Other potential biomarkers include CD44, MXRA8, and APOA1. The CD44 glycoprotein reflects the activation of parietal epithelial cells that trigger glomerulosclerosis, while MXRA8 is involved in fibrosis and disease progression. Finally, APOA1 indicates that oxidative stress is related to hyperlipidemia, which is one of the pathogenic factors in the development and progression of FSGS. Thus, it has become evident that uroproteomics could be instrumental for the development of noninvasive differential diagnostic methods for MCD and FSGS as well as for the expansion of our understanding of each disease’s pathogenesis.

The collective results of urinary proteomics studies for MCD and FSGS are presented in Table 1.

Bioinformatic analysis

In the context of gaining additional insight related to candidate biomarkers of PGD, their predicted protein interactions and their molecular functions, we performed an in-depth bioinformatic analysis of 42 DEPs obtained from the data of key urinary proteomics studies for biomarker screening by means of comparing the urine protein profiles between PGD and HC via MS discussed in this review. The analyzed DEPs were: AFAM, A2GL, A1M, HEMO, APOA1, CO3, APOA4, B2MG, RET4, ZA2G, A2MG, CAH1, CERU, CO4A, FINC, THRB, VTDB, HPT, IBP-7, A1AT, AACT, ANT3, CD44, CLM-9, PCDH1, UTER, DPP4, NHLC3, CO9, IGKC, K1C10, K2C1, K2C5, ADIPO, ICAM1, TIMP1, A1BG, UROM, SCRB2, THBG, CD59, and SH3L3 (Table 2).

Table 2. Summary of research on urinary protein biomarkers of glomerular diseases.

Biomarker	DEP abbreviation	Gene name	Nephropathy				Molecular function	Biological process	Pathways
			IgAN	MN	MCD	FSGS
Afamin ↑	AFAM	AFM	[10,39,42]	[48,51]			Fatty acid binding; vitamin E binding	Protein stabilization; protein transport within extracellular region
Leucine-rich α-2-glycoprotein	A2GL	LRG1	[39]				Type I or type II transforming growth factor beta receptor binding	Brown fat cell differentiation; keratinocyte migration; leukocyte adhesion to vascular endothelial cell
α-1-microglobulin	A1M	AMBP	[39]				Serine-type endopeptidase inhibitor activity; Extracellular matrix structural constituent	Regulation of protein phosphorylation; regulation of protein activation cascade; regulation of heterotypic cell-cell adhesion; regulation of blood coagulation; acute inflammatory response
Hemopexin	HEMO	HPX	[39]				Hemetransmembrane transporter activity	Cellular iron ion homeostasis
Apolipoprotein A-I ↑	APOA1	APOA1	[39,42]				High-density lipoprotein particle receptor binding; lipase inhibitor activity; low-density lipoprotein particle binding	Cholesterol metabolism; lipid metabolism; lipid transport	Cholesterol metabolism; vitamin digestion and absorption; fat digestion and absorption
Complement C3 ↑	CO3	C3	[39]	[49,51]			Complement component C3b binding; complement receptor activity	Complement alternate pathway; Complement pathway; Fatty acid metabolism	Complement and coagulation cascades
Apolipoprotein A-IV	APOA4	APOA4	[39]				High-density lipoprotein particle receptor binding; Lipoprotein lipase activator activity; lipase binding	Lipid transport; Transport	Cholesterol metabolism; fat digestion and absorption
Beta-2-microglobulin	B2MG	B2M	[39]				Beta-2-microglobulin binding; CD8 receptor binding; T cell receptor binding	Amyloid fibril formation	Allograft rejection
Retinol-binding protein 4	RET4	RBP4	[39]		[51]		Retinol transmembrane transporter activity		Reactome pathways: Retinoid cycle disease events
Zinc-α-2-glycoprotein ↑	ZA2G	AZGP1	[42]				Protein transmembrane transporter activity	Cell adhesion	Reactome Pathways
α‑2‑macroglobulin ↑	A2MG	A2M	[42]				Serine-type endopeptidase activity; protease binding	Regulation of cell adhesion molecule production; regulation of heterotypic cell-cell adhesion; blood coagulation	African trypanosomiasis; cholesterol metabolism
Carbonic anhydrase 1 ↑	CAH1	CA1	[42]				Arylesterase activity; carbonate dehydratase activity; zinc ion binding	One-carbon metabolic process	Erythrocytes take up oxygen and release carbon dioxide; erythrocytes take up carbon dioxide and release oxygen
Ceruloplasmin ↑	CERU	CP	[39,42]				Chaperone binding; oxidoreductase activity; copper transmembrane transporter activity	Iron ion transport; heme oxidation; copper ion export	Porphyrin and chlorophyll metabolism
Complement C4-A ↑,	CO4A	C4A	[42]				Complement component c4b binding; complement binding	Regulation of complement activation; regulation of apoptotic cell clearance;	Complement and coagulation cascades
Fibronectin ↓	FINC	FN1	[42]				Fibronectin binding; vascular endothelial growth factor receptor 2 binding; extracellular matrix binding	Cell-cell adhesion; cell-matrix adhesion; negative regulation of lipoprotein metabolic process	ECM-receptor interaction
Prothrombin ↑	THRB	F2	[42]				Thrombin-activated receptor activity	Regulation of protein activation cascade	Complement and coagulation cascades
Vitamin D-binding protein ↑	VTDB	GC	[39,42]				D3 vitamins binding; vitamin binding; tetrapyrrole binding	Vitamin D metabolic process; vitamin transport; regulation of blood coagulation	Reactome pathways: vitamin D (calciferol) metabolism
Haptoglobin ↑	HPT	HP	[42]				High-density lipoprotein particle receptor binding; oxygen carrier activity; high-density lipoprotein particle binding; cholesterol transfer activity	Negative regulation of very-low-density lipoprotein particle remodeling; positive regulation of cholesterol esterification	African trypanosomiasis; cholesterol metabolism
Insulin-like growth factor-binding protein 7 ↓	IBP-7	IGFBP7	[42]				Insulin-like growth factor receptor binding; insulin-like growth factor binding; insulin receptor binding	Positive regulation of glycogen biosynthetic process	Aldosterone-regulated sodium reabsorption; IL-17 signaling pathway
α-1-antitrypsin ↑	A1AT	SERPINA1	[40,42] (in progressors)	[45]	[45,47,48]	[45,47]; (FSGS vs. MCD) ↓[52]	Serine-type endopeptidase activity; protease binding	Antibacterial peptide production; negative regulation of retrograde protein transport, ER to cytosol	Protein processing in endoplasmic reticulum
α-1-antichymotrypsin↑	AACT	SERPINA3	[42]				Serine-type endopeptidase activity	Antibacterial peptide production	Renin-angiotensin system
Antithrombin-III ↑	ANT3	SERPINC1, AT3	[41,42]				Extracellular matrix structural constituent	Regulation of protein activation cascade; blood coagulation; positive regulation of heterotypic cell-cell adhesion	Complement and coagulation cascades
CD44 ↓	CD44	CD44	[43]	[47]			Collagen binding; hyaluronic acid binding	Negative regulation of DNA damage response	Bladder cancer; ECM-receptor interaction; endometrial cancer; adherens junction
CLM-9 ↓	CLM-9	CLM9	[43]				Transmembrane signaling receptor activity	Immune system process	Immunoregulatory interactions between a lymphoid and a non-lymphoid cell
Protocadherin-1 ↑	PCDH1	PCDH1	[43]				Calcium ion binding	Cell adhesion; cell-cell signaling
Uteroglobin ↑	UTER	SCGB1A1	[43]					Respiratory gaseous exchange by respiratory system; negative regulation of T cell proliferation	Reactome pathways: Surfactant metabolism
Dipeptidyl peptidase 4 ↑	DPP4	DPP4	[43]				Glucagon receptor binding; virus receptor activity; protease binding; signaling receptor binding	Positive regulation of gap junction assembly; calcium-independent cell-matrix adhesion; regulation of cell-cell adhesion mediated by integrin; membrane raft organization	Primary immunodeficiency; protein digestion and absorption; proteoglycans in cancer
NHL repeat-containing protein 3 ↑	NHLC3	NHLRC3	[43]				Ubiquitin protein ligase activity	proteasome-mediated ubiquitin-dependent protein catabolic process
Complement C9 ↑	CO9	C9	[44]		[47]	[47]	Complement binding	Regulation of activation of membrane attack complex; complement activation; positive regulation of heterotypic cell-cell adhesion	Complement and coagulation cascades
Ig kappa chain C region ↑	IGKC	IGKC	[44]
Keratin-type I cytoskeletal 10 ↓,	K1C10	KRT10	[44]				Structural constituent of skin epidermis; structural constituent of cytoskeleton	Keratinocyte migration; peptide cross-linking	Reactome pathways: formation of the cornified envelope
Keratin-type II cytoskeletal 1 ↓	K2C1	KRT1	[44]				Structural constituent of cytoskeleton	Peptide cross-linking; protein heterotetramerization; blood coagulation	Reactome pathways: Intrinsic pathway of fibrin clot formation
Keratin-type II cytoskeletal 5 (K5)↓	K2C5	KRT5	[44]				Structural constituent of skin epidermis; structural constituent of cytoskeleton	Keratinocyte migration; intermediate filament cytoskeleton organization	Bladder cancer; endometrial cancer; adherens junction
Adiponectin ↑	ADIPO	ADIPOQ	[41]				Adiponectin binding	Response to metformin; negative regulation of receptor biosynthetic process	Longevity regulating pathway; adipocytokine signaling pathway; AMPK signaling pathway
Intercellular adhesion molecule 1 ↑	ICAM1	ICAM1	[41]				ICAM-3 receptor activity; complement component C3b binding	Positive regulation of prostaglandin-e synthase activity	Viral myocarditis; cell adhesion molecules
Metalloproteinase inhibitor 1 ↑	TIMP1	TIMP1	[41]				Metalloendopeptidase activity; serine-type endopeptidase activity	Iron coordination entity transport; extracellular matrix disassembly	Bladder cancer; IL-17 signaling pathway
α-1B-Glycoprotein ↑	A1BG	A1BG	[10]				Insulin receptor substrate binding; transmembrane receptor protein tyrosine kinase adaptor activity	Insulin-like growth factor receptor signaling pathway	Renal cell carcinoma; chronic myeloid leukemia
Uromodulin ↓	UROM	UMOD		[45]	[45]	[45]	Extracellular matrix structural constituent; IgG binding	Positive regulation of blood microparticle formation; regulation of plasma lipoprotein particle levels
Lysosome membrane protein-2 ↑	SCRB2	SCARB2		[46]			K48-linked polyubiquitin modification-dependent protein binding; protein tag	Regulation of ubiquitin-specific protease activity	Protein processing in endoplasmic reticulum
Thyroxine-binding globulin	THBG	SERPINA7		[47]			Serine-type endopeptidase inhibitor activity	Negative regulation of endopeptidase activity	Thyroid hormone synthesis
CD59↑	CD59	CD59				[54]	Complement binding	Blood coagulate; negative regulation of activation of membrane attack complex; regulation of complement activation	Complement and coagulation cascades; systemic lupus erythematosus
SH3 domain-binding glutamic acid-rich-like protein 3	SH3L3	SH3BGRL3			[51]				TFAP2 (AP-2) family regulates transcription of growth factors and their receptors

↑: upregulated protein; ↓; downregulated protein.

To reveal the possible physiological roles of DEPs, GO and KEGG enrichment analyses were carried out using Metascape (https://metascape.org). The result of the GO enrichment analysis is shown in Figure 4. The molecular functions, biological processes, cellular components, and pathways in which the identified DEPs are predicted to be involved are presented in Table 2. Interactions between DEPs were analyzed via the STRING database (https://cn.string-db.org), while Cytoscape (version 3.9.1, https://cytoscape.org) was used to construct and visualize the resulting PPI network (Figure 5). By utilizing these approaches, we found that DEPs were related to complement activation, immune response, cell adhesion, and cytoskeleton dynamics.

Figure 4. GO and KEGG enrichment analysis. Bar graph of enriched terms across input gene lists (colored according to p-values). (a) Enrichment analysis results of GO molecular functions. (b) Enrichment analysis results of GO biological process. (c) Enrichment analysis results of GO cellular components. (d) Enrichment analysis results of KEGG pathways. Adjusted p-value <0.05 was considered significant.

Figure 5. PPI network diagram of candidate biomarkers. Excluding four proteins without segment connections, the PPI network diagram includes 37 DEPs. At the outermost part of the concentric circle, the BC value that represents the centrality of a protein increases counterclockwise from SERPINA7, while decreasing clockwise from HP for the middle circle. The PPI enrichment p-value was lower than 1.0e⁻¹⁶ in STRING. The number of edges on a protein is proportional to its the connectivity with other proteins in the network. The BC was selected to represent the importance of differentially expressed proteins. The higher BC values correlate with stronger centrality of these proteins and consequently with more protein cross interactions through the relevant mediations. The size of the circle and the font size reflect the BC value, with bigger circles and fonts corresponding to higher betweenness values. BC: betweenness centrality

The complement system plays a vital role in many diseases and can be activated in three ways. Its activation can also trigger an attack against the body’s own cells, consequently causing autoimmune diseases [91,92]. It is also associated with kidney pathologies, especially MN. Cell adhesion is a dynamic process that plays an essential role in cell proliferation, differentiation, and migration, while perturbations in cell adhesion often lead to severe pathological alterations [93]. Abnormal cell adhesion may cause kidney disease such as IgAN, which is associated with mesangial cell and cell adhesion matrix proliferation [94,95]. Further delineating the relationship between IgAN and cell adhesion may help toward promoting noninvasive methods for diagnosing this disease. Cytoskeleton. the protein fiber grid system in eukaryotic cells, is a dynamic scaffolding network that is essential in maintaining basic cell morphology. Cytoskeleton proteins have been found to be abnormally expressed in patients with MCN and FSGS [96], while podocyte cytoskeleton injury can cause proteinuria [97]. Another study reported that rearrangements of the cytoskeleton could lead to functional abnormality of podocyte foot processes and symptoms of albuminuria, which trigger podocyte-related diseases [98]. Additional studies on cytoskeleton dynamics may help to characterize the pathogenesis of MCD and FSGS in greater detail. Further research in these three areas may lead to new discoveries about PGD.

We believe that the occurrence of PGD may be due to complement attacks on glomerular cells, which disrupt cytoskeleton integrity and further lead to abnormal expression of certain proteins and other macromolecules through a complex intracellular protein interaction network, culminating in the collapse of the glomerular filtration barrier and the development of albuminuria, among other pathological symptoms. However, the specific pathogenesis of PGD warrants more investigation.

Performing in-depth studies of these proteins may in turn lead to advancements toward the pain-free diagnosis of kidney diseases and facilitate research in the design of drugs for the personalized treatment and assessment of prognosis of patients, as well as improve our current knowledge regarding the pathogenesis of PGD.

Conclusions and prospects

In this review, we have discussed studies on urinary proteomics of primary glomerular diseases and we have selected potential biomarkers of PGD to analyze via bioinformatics to obtain a more informed overview of their molecular functions, biological processes, and protein–protein crosstalk interactions, to gain a deeper insight into the pathogenesis of PGD. Based on the results gleaned from our analysis, PGD appears to be related to complement activation, immune response, cell adhesion, and cytoskeleton dynamics.

PGD is an autoimmune disease with unknown pathogenesis, and there is no effective pain-free diagnosis method that can replace renal biopsy in clinical practice. Compared with blood and tissue fluid, urine as the fluid for analysis offers the advantages of noninvasiveness, sustainable collection, and repeated detection. As urine is generated and excreted by the urinary system, it may also be used for detecting urinary system-related abnormalities. In recent years, research in the fields of noninvasive diagnosis, treatment, prognostic evaluation and in-depth understanding of the pathogenesis of PGD has progressed significantly, owing to advances in urinary proteomics. Nevertheless, as results are largely dependent on the choice of proteomic techniques, these methods should be carefully selected according to the purpose of the study. The combination of urine, blood, and kidney tissues, together with high-throughput omics studies, may reveal key findings relevant to the study of PGD. In this regard, urine proteomics and bioinformatics may provide a sound basis for the development of novel biomarkers to diagnose disease and assess prognosis, and to delineate the molecular mechanisms of the disease.

Although there is a number of reviews on urine proteomics of PGD, few articles have integrated the results of different studies to assess their biological impact on the mechanism of PGD through bioinformatic analysis. In this context, protein network analysis may be more informative than protein analysis alone owing to the complexity of the biological processes in which proteins participate. Our present review discusses candidate biomarkers from several studies while also analyzing these data in a descriptive bioinformatic process, thereby proposing an integrated method for future proteomics research.

However, our analysis has some limitations. We assumed that all the proteins analyzed are involved in at least some relevant biological processes during the occurrence and development of PGD and are related to its pathogenesis. Owing to the complex network of protein interactions in an organism, the level of some proteins might be influenced by other proteins, and the change in their levels may not have an impact on PGD.

In conclusion, numerous studies discuss urine proteomics in PGD and several potential biomarkers have been identified. Here, our bioinformatic analysis of candidate biomarkers showed that they tend to be associated with processes such as complement response, cell adhesion, and cytoskeleton dynamics. The network of protein interactions is complex, and our bioinformatic analysis of these potential biomarkers may facilitate research efforts for the further elucidation of the pathogenesis of PGD. Our summary of urinary proteomics studies of PGD is helpful in promoting diagnostic alternatives, in the form of biomarkers, for PGD. We believe that urinary proteomics combined with bioinformatic approaches will not only help in the noninvasive diagnosis of PGD but also provide a framework for precision therapy and a deeper understanding of the pathogenesis for this disease.

Abbreviations
CKD	=	chronic kidney disease
DEPs	=	differentially expressed proteins
ESRD	=	end-stage renal disease
FSGS	=	focal segmental glomerulosclerosis
GD	=	glomerular disease
GO	=	Gene Ontology
HCs	=	healthy controls
IgAN	=	IgA nephropathy
KEGG	=	Kyoto Encyclopedia of Genes and Genomes
MCD	=	minimal change disease
MN	=	membranous nephropathy
MS	=	mass spectrometry
PGD	=	primary glomerular disease
PPI	=	protein–protein interaction

Acknowledgments

We thank Figdraw (https://www.figdraw.com/) and Untitled (https://app.biorender.com/) for providing us with a drawing platform.

Disclosure statement

The research was conducted in the absence of any commercial or financial relationships that might be construed as potential conflicts of interest. No potential conflict of interest was reported by the author(s).

Additional information

Funding

Our study was supported by “345 Talent Project” of Shengjing Hospital of China Medical University [2022146], Liaoning Province Central Government’s special project to guide local scientific and technological development [2019JH6/1], and National Key Research and Development Program of China [2018YFE0207300].