MMPs, a large family of endopeptidases, could affect CKD progression through reshaping and degrading cellular ECM components and regulating non-ECM molecules that play an important role in the development of CKD, including adhesion molecules, cytokines, and growth factors, under the strict regulatory mechanisms (Lindsey et al., 2016). Overall, It had been found that 11 members (MMP-1, MMP-2, MMP-3, MMP-7, MMP-9, MMP-13, MMP-14, MMP-24, MMP-25, MMP-27, and MMP-28) are expressed in the kidney.

Among them, MMP-9 plays the most representative role in the progression of renal fibrosis. The expression of MMP-9 is low in a normal physiological state but active in persistent renal injury. In the experimental model of renal fibrosis, the expression of MMP-9 was found in mesangial cells, glomerular cells, epithelial cells and endothelial cells, fibroblasts, macrophages, and neutrophils (Tan and Liu, 2012). Furthermore, inhibition of MMP-9 could reduce the infiltration of neutrophils and other inflammatory cells and reduce the degree of renal fibrosis (Wang H et al., 2019). MMP-9 was also abnormally expressed in CKD patients. Urinary MMP-9 levels were significantly higher in children with CKD than those in age-matched healthy controls (Musiał et al., 2015) and were increased in patients with focal segmental glomerulosclerosis (Korzeniecka-Kozerska et al., 2013). In addition, a cohort study involving 251 adults showed that the increase in plasma MMP-9 could predict the progression of CKD after 8.5 years with a risk ratio of 4.7 (Hsu et al., 2013). Above all, current evidence supported the potential of MMP-9 as a biomarker of CKD progression.

In addition to MMP-9, MMP-2 also shows a high correlation with CKD progression. Nadkarni GN et al. found that the increase of MMP-2 per unit of urine led to a decrease in eGFR (0.1 ml/min/m²) during the 38-month follow-up (AUC 0.74) (Nadkarni et al., 2016). Such results were also confirmed by another 8-year follow-up study on patients after coronary angiography: patients with higher levels of MMP-2 are 2.5 times more likely to have a decline in eGFR (Hsu et al., 2013). Moreover, a systematic review showed that MMP-2 has the potential to identify patients at risk of renal fibrosis that leads to worse renal outcomes (Mansour et al., 2017). However, more evidence is still needed to determine the role of MMP-2 in evaluating the progression of renal fibrosis in CKD.

MMP-7 is a secreted zinc- and calcium-dependent endopeptidase and a transcriptional target of canonical Wnt/β-catenin signaling. It had been demonstrated that Wnt/β-catenin could be activated in kidney disease and the level of urinary MMP-7 might be used as a noninvasive surrogate biomarker (Liu Z et al., 2020) and a therapeutic target for renal fibrosis (Wozniak et al., 2021). Zhou et al. (2017) conducted a cross-sectional study under measuring urinary MMP-7 levels in a cohort of 102 patients with CKD compared with normal subjects. It indicated that urinary MMP-7 levels, elevated in kidney disorder patients, closely correlated with renal fibrosis scores. Meanwhile, the knockout of MMP-7 could ameliorate fibrotic lesions and matrix gene expression in mice. The expression of E-cadherin protein could be preserved by the genetic ablation of MMP-7 and substantially reduced the expression of total and dephosphorylated β-catenin. Li et al. (2021) reviewed that MMP-7 could activate the Wnt/β-catenin signaling pathway after renal injury and urinary MMP-7 may be a noninvasive biomarker of profibrotic signaling in the kidney. In contrast, they discovered that MMP-7 exerts protective effects on the kidney as an adaptive response in cisplatin administration or folic acid-induced AKI animal models. Therefore, further study should be performed to explore the role of MMP-7 as a therapeutic target of renal fibrosis.

MCP-1 belongs to the chemokine family and has a strong chemotactic effect on monocytes (Das et al., 2021). This unique cytokine is mainly produced by chemokine in renal intrinsic cells and macrophages in response to hypercholesterolemia and other arterial injuries (Gregg et al., 2017; Tam and Ong, 2020). MCP-1 plays an important role in developing renal inflammation and fibrosis under the stimulation of oxidative stress, cytokines, or growth factors.

It had been reported that MCP-1 was deeply implicated in albuminuria and kidney dysfunction. The upregulation of MCP-1 was accompanied by the activation of IκB/NF-κB signaling in CKD patients with macroalbuminuria; the similarity result had also been found that the activation of NF-κB was accompanied by significant upregulation of MCP-1 in CKD rats (Feng et al., 2019). The researchers found in animal models that blocking MCP-1 receptors might suppress inflammation and alleviate glomerulonephritis (Khalili et al., 2020). In addition, the potential of MCP-1 as a biomarker of renal fibrosis has also been confirmed in clinical trials. In a cohort study of 58 patients with IgA nephropathy, urinary MCP-1 levels provided significant additional non-invasive information for a better predictive performance of the severity of interstitial fibrosis beyond traditional markers (Segarra-Medrano et al., 2017). Recently, a probabilistic sampling longitudinal cohort study involving 3,257 participants was found that the MCP-1 of CKD patients was higher than that of non-CKD participants, and the level of MCP-1 was negatively correlated with eGFR. Furthermore, this study also proved that a high level of MCP-1 was an independent risk factor for CKD death (Gregg et al., 2018). Nadkarni GN et al. found that the ratio of urinary MCP-1 to creatinine was associated with a sustained decline of more than 40% in eGFR and showed better risk predictions than other traditional indicators (Tam and Ong, 2020). Besides, a case-control study also found that the level of MCP-1 was positively correlated with the probability of CKD (Zhang et al., 2018). Another sensitivity analysis involving 171 matched cases and controls suggested that the ratio of urinary MCP-1 to creatinine was correlated with the incidence of CKD stages (Musiał and Zwolińska, 2020). It indicated that MCP-1 is of high predictive value in evaluating the progression and prognosis of renal fibrosis in CKD.

DKK-3, a member of the glycoprotein family (DKK1-4) (Gröne et al., 2017), is a recently discovered stress-induced tubular-derived renal biomarker for interstitial fibrosis (Federico et al., 2016; Fang et al., 2020; Seibert et al., 2021). It mainly promotes renal tubular epithelial cells producing fibrogenic molecules secreted by stress stimulation (Schunk et al., 2021a). Expression of DKK-3 in the developing kidney would cease after kidney maturation and would reignite when the kidney tissue is damaged (Seibert et al., 2021). Lipphardt et al. (2019) have found that DKK3 could significantly reduce the area percentage, total length, and bifurcation number of capillary structures in renal microvascular endothelial cells. Furthermore, it indicated that the nephropathy induced by adriamycin could lead to the upregulation of the DKK-3 level in renal tubular. As a result, DKK-3 could induce EMT and impair angiogenic competence, giving rise to renal fibrosis. In addition, DKK-3 could stimulate the expression of TGF-β, which is recognized as an important factor leading to renal fibrosis (Karamariti et al., 2018).

In the animal model of CKD, DKK-3 could regulate the signal transduction of the Wnt/β-catenin signal pathway and induce renal tubulointerstitial fibrosis (Schunk et al., 2021b). Another cohort study showed that increased urinary DKK-3 concentration in patients with primary glomerular disease and primary interstitial disease was significantly associated with higher levels of tubulointerstitial fibrosis, and the significant eGFR decline within 6 months is closely related to the level of urinary DKK-3 (Zewinger et al., 2018). The results suggested that DKK-3 might be a sensitive predictor in the early progress of CKD.

5-MTP, converted by TPH-131, is an endogenous tryptophan metabolite that could reduce the production of inflammatory cytokines and thus ameliorate inflammation and tissue damage by inhibiting NF-κB activation and the consequent inhibition of COX-2 transcriptional activation.

In recent research, Chen D. Q et al. (2019) identified 5-MTP for the first time as the most promising biomarker metabolite for detecting early CKD. An untargeted metabolomics study of 2,155 participants, including patients with stage 1–5 CKD and healthy controls, was performed, and the level of 5-MTP decreased with the progression of CKD that strongly correlates with clinical markers of kidney disease. Furthermore, the biological effects of 5-MTP were performed using UUO mice and HK-2 and HMC cells. It indicated that 5-MTP attenuated the expression of the pro-inflammatory factor NF-κB p65 and its target gene products MCP-1 and COX-2 while increasing the expression of the anti-inflammatory and antioxidant transcription factor Nrf 2 and its target gene products HO-1 and NQO-1. In addition, 5-MTP treatment could significantly attenuate the upregulation of the pro-fibrotic proteins collagen I, fibronectin, wave protein and α-SMA and reverse the downregulation of E-cadherin and Thy1 in vivo and in vitro settings. The present study demonstrated that 5-MTP could separate patients with early stage CKD from healthy controls and serve as potential biomarkers of the early stage of CKD.

A recent study have discovered that aryl-containing metabolites might have an important effect on chronic kidney disease progression. Most importantly, two polycyclic aromatic hydrocarbon metabolites, 1-aminopyrene (1-AP) and 1-hydroxypyrene (1-HP), showed strong positive and negative correlation with serum creatinine and creatinine clearance, respectively. Miao et al. (2020), Miao et al. (2021b) analyzed 5,406 urine and serum samples from patients with stage 1–5 CKD using metabolomics; 1-AP and 1-HP were identified and validated using longitudinal and drug intervention cohorts and 5/6 nephrectomized and adenine-induced rats. Furthermore, the increased levels of 1-AP and 1-HP in serum and kidney tissues correlated with decreased renal function in two rat models. Upregulated mRNA expression of aryl hydrocarbon receptor and its target genes, including CYP1A1, CYP1A2, and CYP1B1, were observed in patients and rats with progressive CKD. It indicated that 1-AP and 1-HP were demonstrated to mediate renal fibrosis through activation of the aryl hydrocarbon receptor signaling pathway and the endogenous 1-AP and 1-HP are novel mediators of CKD progression.

Multiple evidence have shown that the commonly inflammatory indicators, such as IL-1, IL-6, C-reactive protein (CRP), and tumor necrosis factor (TNF)-like weak inducer of apoptosis (TWEAK), could effectively judge the progression and prognosis of CKD. A study showed that patients with chronic kidney disease with lower eGFR levels had higher levels of plasma IL-1 β, IL-1RA, IL-6, TNF-α, hypersensitive CRP, and fibrinogen and that the inflammation score was inverted with GFR estimates and urinary albumin/creatinine ratio (UACR) (Gupta et al., 2012). A recent cohort study involving 3430 patients with chronic renal insufficiency also found that increased levels of TNF-α and decreased serum albumin in patients with chronic renal disease were associated with rapid loss of renal function (Amdur et al., 2016). Studies also have shown that TEWAK worked as an important factor in inflammation-related renal injury and accelerates the progression of CKD (Sanz et al., 2014).

The inflammasome is a large polyprotein complex induced by lipopolysaccharide (LPS). It had been a hot spot in the field of nephrology recently and was first reported as an innate immune signaling pathway in 2002 (Martinon et al., 2002).

Intracellular NOD-like receptors (NLRs) play a key role in forming and activating inflammatory bodies. They could form the inflammatory complex and release proinflammatory cytokines IL-1 β and IL-18 (Akchurin and Kaskel, 2015; Hutton et al., 2016). NLRP3 inflammasome is closely related to apoptosis, autophagy, and proinflammatory cytokines regardless of traditional microbial innate immune stimulation; therefore, it is of extensive significance for a variety of renal diseases (Chang et al., 2014).

The role of NLRP3 in the formation of the inflammatory state of CKD had been confirmed in animal experiments. In the model of unilateral ureteral obstruction, the renal tubular injury, inflammation, and fibrosis were less severe in NLRP3-knockout mice (Vilaysane et al., 2010). In addition, a bidirectional Mendelian randomization (MR) analysis showed that elevated NLRP3 inflammasome had a profound impact on the pathogenesis and severity of CKD (Jia et al., 2019). A clinical cohort study also showed that the expression of NLRP3 mRNA in CKD patients was significantly upregulated compared with that of the control group (El-Deeb et al., 2019). However, although NLRP3 has the potential to be used as a biomarker to evaluate the progression of CKD, future research is needed to provide solid evidence to evaluate its diagnostic performance.

Tryptophan metabolites from gut microbiota have been demonstrated to participate in renal fibrosis as aryl hydrocarbon receptor ligands. Indole derivatives are derived from tryptophan metabolism pathways modulated by gut microbiota directly or indirectly. IS is a metabolite of indole produced by tryptophan through the intestinal flora (including E. coli) (Evenepoel et al., 2009). IS could activate the renin-angiotensin-aldosterone (RAS) system (Sun et al., 2012), induce cell senescence and apoptosis (Han et al., 2018), and promote EMT (Kim et al., 2012), thus accelerating the progression of fibrosis, renal dysfunction, and CKD.

A variety of trials had confirmed the value of IS in predicting CKD. A study showed that at the earliest stage of CKD, the change of IS is proportional to the severity of the disease (Kim et al., 2020). Serum IS levels in CKD patients were directly related to dialysis events; therefore, IS might potentially predict the progression of advanced CKD (Wu et al., 2011). Moreover, a recent cohort study involving 100 CKD patients and 30 patients with normal renal function showed that the level of IS reflected the degree of renal impairment and peaked in the late stage of CKD (Wu et al., 2020). IS could be used as a novel biomarker for a thorough evaluation of CKD progression and future research is needed to provide solid evidence to evaluate its diagnostic function.

PCS is produced through the transformation of cresol by intestinal epithelial cells. PCS induces renal injury and fibrosis by inhibiting klotho gene expression, activating the RAS/TGF-β pathway (Sun et al., 2012), inducing EMT, and causing NADPH oxidase-driven ROS (Watanabe et al., 2013). A prospective observational trial involving a cohort of 268 patients with specific CKD stages showed that serum total PCS was independently associated with renal progression (Barrios et al., 2015). Another cohort study of 4,439 individuals with different eGFR also demonstrated that PCS was negatively correlated with eGFR and PCS might gradually increase along the early process of CKD (Chen L et al., 2019). Furthermore, several studies have also shown that PCS levels in patients with CKD are associated with poor prognosis, including death (Shafi et al., 2015; Lu et al., 2016). Since the change of PCS might be sensitive to the slight process of early CKD, it could be a valuable predictor for CKD progression.

TMAO, one of the major uremic toxins, is byproduct of bacterial metabolism of phosphatidylcholine, choline, or L-carnitine. It had been confirmed to be elevated in patients with CKD and closely associated with a decreased renal function (Tang et al., 2015). Chen et al. investigated that both plasma and urine TMAO were significantly elevated in 180 CKD patients compared to those of 120 age-matched healthy controls patients by determining their fasting plasma and urine samples by UPLC-HDMS-metabolomics and quantitative real-time RT-PCR techniques. The concentration of TMAO in CKD patients plasma had been shown to directly correlate with plasma concentration of urea and creatinine, indicating the close association of TMAO with the degree of renal insufficiency (Chen et al., 2017).

TMAO might induce oxidative stress by inhibiting the expression of the oxidative stress inhibitor SIRT1, increasing H₂O₂, and reducing SOD activity (Ke et al., 2018). TMAO could affect systemic inflammatory response (Missailidis et al., 2016) by regulating inflammatory inducers (IL-6, TNF-α, and so on), promoting the phosphorylated NF-κB to enter the nucleus (Seldin et al., 2016), and activating the expression of inflammasomes (Boini et al., 2017). In clinical practice, a cohort study of 80 patients at all stages of CKD observed significant increases in urinary and plasma TMAO levels in the CKD group and significantly higher levels in stage 4 CKD (Wu et al., 2020). Consistent results have been reported in an experiment that TMAO level was negatively correlated with renal function decline and returned to normal after kidney transplantation (Missailidis et al., 2016). Furthermore, TMAO was an independent predictor of mortality in patients with CKD.

Microflora refers to all the coexisting microorganisms living in the host, which are mainly composed of bacteria, viruses, archaea, fungi, and unicellular eukaryotes (Evenepoel et al., 2017; Castillo-Rodriguez et al., 2018; Cosola et al., 2018). It has been estimated that there are at least 35,000 different microorganisms in the human body (Frank et al., 2007). Under normal circumstances, the intestinal microflora is mainly composed of five phyla, among which Firmicutes and Bacteroidetes account for the largest proportion in the colon, followed by Actinobacteria, Verrucomicrobia, and Proteobacteria (Vaziri, 2012).

The microflora is responsible for metabolizing dietary components and improving digestion by expressing several enzymes (Plata et al., 2019). The microflora, with its metabolic activity, could affect the health of the host by interfering with the physiology, nutrition, metabolism, and immune system (Le Chatelier et al., 2013)^; (Human Microbiome Project Consortium, 2012). Researchers have investigated that obesity, type 2 diabetes (T2DM), and cardiovascular diseases (CVD) might be associated with complex interactions between gut microflora and human hosts (Gurung et al., 2020; Zhou et al., 2020).

Emerging evidence also suggested that intestinal microflora plays an important role in the occurrence and development of CKD (Vanholder and Glorieux, 2015; Nallu et al., 2017). Therefore, the changes of intestinal microflora might provide a possibility for measuring the progression of CKD. Barrios et al. found that the changes in Ruminococcaceae, Christensenellaceae, and Lachnospiraceae families were significantly correlated with eGFR at the early stage of CKD (Barrios et al., 2015). Furthermore, Wu and his colleagues found that accessory Escherichia coli (AUC 0.78; 95% CI 0.7–0.87) and pseudobutyric acid bacillus (AUC 0.76; 95% CI 0.67–0.84) could effectively distinguish CKD from the health control compared to urinary protein/creatinine ratio and sterins Collinsella stercoris has shown excellent performance in identifying healthy people and early CKD patients (Wu et al., 2020). This indicated that microflora could be a valuable tool to evaluate the progress of CKD, especially in the early stage.

As a beta-galactoside binding protein, Gal-3 is ubiquitous in cells and it could be secreted into the extracellular space by epithelial cells, endothelial cells, and macrophages (Wang and Guo, 2016) in the kidney and heart. It had been indicated that Gal-3 could play an important role in initiating cardiac fibrosis and ventricular remodeling (Gleissner et al., 2017). Gal-3 could modulate kidneys pro-inflammatory effects, regulate growth, differentiation, proliferation of the cells and moreover mediate aldosterone-induced fibrosis of the heart and blood vessels (Calvier et al., 2013; Madrigal-Matute et al., 2014; Vergaro et al., 2016). Furthermore, researchers have found that Gal-3 is also highly expressed in injured kidneys and is involved in the progression of renal fibrosis (Nikolic-Paterson et al., 20112014). O’Seaghdha CM et al. found that higher plasma Gal-3 levels were associated with a rapid decline in eGFR based on an analysis of kidneys from 2,450 patients who were followed for an average of 10 years (O’Seaghdha et al., 2013). Considering its important role in the progression of both cardiac and renal fibrosis, Gal-3 has the potential to predict the risk and prognosis of CVD in patients with CKD.

A postmortem analysis of a four-dimensional study of 1,168 diabetic HD patients showed that circulating Gal-3 levels were negatively correlated with renal function and were four to five times above the reference range; meanwhile, Gal-3 was associated with cardiovascular events (Drechsler et al., 2015). Studies with 5,226 patients have confirmed that the level of Gal-3 increased by 1% and the risk of all-cause death increased by 37.9% in CKD patients (Zhang et al., 2019). In summary, Gal-3 could be an effective biomarker for predicting cardiovascular events in CKD.

As a member of the IL-1 receptor family, ST-2 contains two forms as soluble ST-2 (sST-2) and transmembrane ST-2 (ST-2L). Elevated sST-2 might be highly associated with adverse outcomes in patients with CVD, acute and chronic heart failure, or even death (Dieplinger et al., 2014; Savic-Radojevic et al., 2017). It has been approved by FDA as a new type of biomarker for clinical use to evaluate risk stratification and treatment guidance for patients with acute and chronic heart failure (Aimo et al., 2017; Dalal et al., 2018).

In recent years, emerging evidence suggested its potential role as a biomarker for predicting cardiovascular events in CKD. A study confirmed that heart failure patients with elevated sST-2 and eGFR<60 ml/min/1.73 m² had a significantly increased risk of death (Bayes-Genis et al., 2013). A head-to-head comparison study with Gal-3 in patients with chronic heart failure also demonstrated that sST-2 performed better than Gal-3 and could predict 5-year risk of cardiovascular death (Bayes-Genis et al., 2014). The significant correlation between plasma sST-2 levels and the progression of CKD to ESRD was also confirmed in a recent cohort study involving 219 patients who participated in the German Chronic Kidney Disease (GCKD) study (Mirna et al., 2020). All the evidence demonstrated the importance of sST-2 in evaluating the cardiovascular risk of CKD.

As a member of the TGF-β cytokine family, GDF-15 is widespread in mammalian tissues, including the prostate, intestinal mucosa, and kidney (Unsicker et al., 2013). It has been found to be involved in cancer, obesity, and cardiovascular and kidney diseases (Nair et al., 2017). Previous studies suggested that GDF-15 might take parts in tissue inflammation, oxidative stress, and injured cardiomyocyte repair and show anti-apoptosis and anti-hypertrophy effects (Kempf et al., 2006).

Several clinical studies have found that higher levels of GDF-15 are associated with poor prognosis, including CVD, heart failure, and death in patients with CKD (Daniels et al., 2011; Breit et al., 2012; Lindman et al., 2015; Bansal et al., 2019). Meanwhile, GDF-15 is associated with CKD events and rapid decline in renal function. Adding GDF-15 to clinical covariates could also improve the prediction of CKD events (Ho et al., 2013). These studies confirmed that the increase of GDF-15 may be a physiological signal of early heart failure, and GDF-15 is helpful in evaluating the risk of CVD in patients with CKD.

CypA, a cytoplasmic protein, has isomerase activity and could catalyze the trans-to-cis change of the peptide bond on the proline residue (Hoffmann and Schiene-Fischer, 2014). CypA could act as inflammatory mediators under the stimulation of oxidative stress, inflammation, and hypoxia and participate in the process of inflammation and apoptosis by affecting multiple processes of transcriptional signal transduction (Sherry et al., 1992; Kim et al., 2004; Suzuki et al., 2006). In the kidney, CypA exists mainly in proximal tubules and renal tubular injury has been proved to be the leading cause of proteinuria and hyperfiltration of renal tissue in diabetic nephropathy (Demeule et al., 2000; Zeni et al., 2017). Studies found that plasma monocytes in patients with diabetes secrete a large amount of CypA in response to hyperglycemia, suggesting that CypA may be a potential secretory marker for T2DM (Ramachandran et al., 2012). Therefore, it could be detected in urine or plasma of patients with diabetes and may increase in patients with DN.

In 2015, a cross-sectional study found the connection between the concentration of urinary CypA and the progression of renal function with good discriminatory power (sensitivity of 90.0%, specificity of 72.7%, AUCs = 0.85) for diagnosing stage 2 DN; therefore, the authors proposed that urinary CypA could act as a new biomarker for early DN (Tsai et al., 2015). A cross-sectional and longitudinal study showed that baseline plasma CypA was positively correlated with changes in glomerular filtration rate for patients with T2DM regardless of the cutoff level or persistent level and could serve as indicators of renal disease progression in T2DM patients (Chiu et al., 2018). In addition, another cross-sectional study of 137 patients with T2DM also verified that CypA has high accuracy in the early diagnosis of DN: AUCs of CypA were 0.914 and 0.937, respectively, for the prediction of incipient and overt DN (Abdel Ghafar et al., 2020).

Periostin is an original member of the stromal cell protein family (Frangogiannis, 2012). When binding to αVβ3-and α Vβ5-integrins, periostin could activate a scaffold protein complex with adaptor proteins and α-parvin (Wallace, 2019). In normal tissue, ILK-PINCH-parvin (IPP) activates EMT and stimulates the cellular signal pathway involved in tissue repair, thus protecting the integrity of the extracellular matrix (Cobo et al., 2016).

EMT is one of the key mechanisms involved in DN development (Loeffler and Wolf, 2015); during EMT, increasing mesenchymal biomarkers like periostin is very characteristic (Conway et al., 2014). Previous studies have shown that periostin could increase the expression of TGF-β, one of the most important fibrotic factors that could directly promote EMT and stimulate ECM synthesis, and therefore induce extracellular matrix deposition (Gordon et al., 2012). Interestingly, renal fibrosis and inflammatory mediators, including TGF-β, angiotensin II, PDGF-B, and IL-4 and IL-13, could upregulate the expression of periostin (Olsan et al., 2018). The study also showed that the administration of TGF-β to adrenal epithelial cells could increase the expression of periostin and stimulate the transformation of renal epithelial cells (Mael-Ainin et al., 2014). Therefore, it is reasonable to believe that the positive feedback regulation between periostin and TGF-β may play a key role in EMT, and abnormal overexpression of periostin would lead to glomerulonephritis and interstitial fibrosis and cause persistent damage to the kidney in DN.

Recent studies have found that periostin could be an effective tissue biomarker for predicting renal damage in patients with DN. Periostin is not expressed in healthy adult kidneys (Satirapoj et al., 2015); however, renal tubular epithelial cells would secrete a high level of periostin to the renal tubule interstitial in response to different severity of renal injury (Sen et al., 2011; Satirapoj et al., 2012). Both experimental and clinical studies showed that periostin might act as a key biomarker of DN. Increased expression of periostin was positively related to the severity of renal fibrosis in the mouse model of bilateral kidney and unilateral ureteral obstruction (Um et al., 2017). In addition, a significant increase in periostin was also found in the kidneys of streptozotocin-induced diabetic mice after nephrectomy (Satirapoj et al., 2012). An observational study involving 19 healthy controls and 71 DN patients found a strong positive correlation between urinary albumin-creatinine ratio (UACR) and urinary periostin levels, and the AUC of periostin for the diagnosis of established microalbuminuria was 0.833, indicating that periostin could be considered as reliable biomarkers in the diagnosis of DN (El-Dawla et al., 2019). The study also found that the level of urinary periostin was significantly higher than the normoalbuminuria group, with the AUCs of periostin being 0.954 and 0.997, respectively, for the prediction of DN (Abdel Ghafar et al., 2020). In addition, it has been found that the level of urinary periostin was gradually increased for healthy controls and T2DM patients with normal albuminuria, microalbuminuria, and massive albuminuria; therefore, the increased level of urine periostin could be detected in T2DM patients earlier than significant albuminuria (Satirapoj et al., 2015). Therefore, periostin could be regarded as a valuable urinary biomarker for predicting early DN in patients with T2DM.

MicroRNAs are a group of non-coding RNA composed of 19–24 nucleotides (Zhou et al., 2014). miRNAs could cleave mRNA or suppress the translation by interacting with the complementary sequence in the 3′-untranslated region of its mRNA target, followed by regulating gene expression (He and Hannon, 2004). They contributed to the physiological processes of cell proliferation, differentiation, and death (Sayed and Abdellatif, 2011). The specific expression of miRNA has been found in the development of many diseases, such as cancer and chronic lymphoblastic leukemia (Khan et al., 2019; Javandoost et al., 2020). Recent studies have found that the abnormal expression of miRNAs is one of the important mechanisms of the occurrence and development of DN (Ishii et al., 2020), and it may become a potential biomarker for predicting DN due to its specificity and stability in body fluids.

Mir-192 is one of the earliest miRNAs that could regulate DN-induced pathological pathways. It was highly expressed in renal tissue and played a key role in renal development and differentiation (Abdelsalam et al., 2019). Later experiments showed that mir-192 could promote the fibrosis of glomeruli and tubules (Mu et al., 2013). Putta et al. demonstrated that TGF-β1 increased mir-192 levels in cultured mesangial cells and the glomeruli of diabetic mice, while decreasing renal mir-192 resulted in decreased renal fibrosis and improved proteinuria (Putta et al., 2012). It was also observed that in the early stage of renal injury, mouse mesangial cells treated with TGF-β1 showed upregulation of mir-192 and collagen α-2(I), and the mechanism may be related to mir-192 downregulating the expression of Smad-interacting protein 1 (SIP1) and δEF1 (also called ZEB1) to control TGF-β1-induced type 1 collagen 2α (COL1A2) and participate in the development of matrix accumulation (Kato et al., 2007).

On the contrary, some opposite evidence indicated that low mir-192 expression could be detected in tubular cells cultured with high TGF-β1 and glucose, which is similar to the renal biopsy samples obtained from patients with DN under decreased eGFR and tubulointerstitial fibrosis (Kato et al., 2007; Wang et al., 2010). It was also confirmed in cohort clinical trials that mir-192 levels in patients with large amounts of albuminuria were significantly lower than those in patients with normal albuminuria (Ma et al., 2016). Interestingly, a study observed the expression level of mir-192 in EVs of patients with different albuminuria levels (n = 80). The results showed that mir-192 was positively correlated with albuminuria. The level of mir-192 in EVs of the microalbuminuria group was significantly higher than that of the normal albuminuria group and control group, and mir-192 could distinguish the patients with normal albuminuria from those with microalbuminuria (AUC = 0.802, p < 0.001). It is proved that the significant increase of EVmiRNAs expression has potential value in diagnosing early DN. The different results may be caused by the multiple effects of miR-192 in the kidney, and its anti-fibrotic and pro-fibrotic effects are obviously dependent on the different regulatory roles of cells and in different stages of the kidney disease (Jia et al., 2016). However, more evidence in vivo and in vitro is needed to determine the significance of mir-192 for DN early prediction.

It has been proved that the translation of superoxide dismutase and p21-activated kinase could be inhibited by mir-377 in blood, which could induce fibronectin accumulation in DN (Wang et al., 2008). A study that examined the levels of mir-377 in the blood showed that the expression of mir-377 increased in T2DM patients and the level of mir-377 was significantly related to the severity of albuminuria. Furthermore, the AUC of mir-377 was 0.851 for diabetic patients and healthy subjects, while the AUC of normal albuminuria and abnormal albuminuria was 0.711 (Al-Kafaji and Al-Muhtaresh, 2018). Therefore, blood mir-377 might also be a promising biomarker for DN.

In addition, the decrease of mir-16 and mir-451-5p in the kidney may downregulate the expression of IL-6 and MMP-9 and therefore contribute to the pathogenesis of DN (Li et al., 2014; Xu et al., 2014). Another experimental study showed that the level of mir-451-5p in the urinary exosomal (UE) increased gradually in diabetes induced by streptozotocin, while the rise of the miR-16 level was delayed comparatively (Xu et al., 2014). Therefore, the UE mir-451-5p might be a more sensitive biomarker of DN. Moreover, miR-200 families, including mir-200a, mir-200b and mir200c, and mir-429 and miR-141, are considered to play an important role in maintaining epithelial differentiation and have an anti-fibrosis effect (Bracken et al., 2015). However, more evidence is still needed to verify their value as early predictive biomarkers of DN.

PTH, a single-chain hormone, is mainly produced in the parathyroid gland and is considered to be the most classical biomarker for estimating bone turnover by KDIGO. As a direct regulator of bone formation, it could play an essential role in vitamin D and phosphate metabolism (Ardawi et al., 2012). PTH could be metabolized into PTH fragments, which could be measured by different PTH assays (Smit et al., 2019).

A lot of research has confirmed the effectiveness of PTH as a biomarker of CKD-MBD. Monier et al. found that PTH-(1–84)/C-PTH fragment ratio could be a predictor for bone turnover (Monier-Faugere et al., 2001). A recent multicenter cross-sectional retrospective diagnostic trial involving 492 dialysis patients demonstrated the value of serum biomarker intact PTH (iPTH) could distinguish between high bone formation rate/bone surface (BFR/BS) and non-high BFR/BS (Sprague et al., 2015). In addition to iPTH, the different biological effects of PTH fragments provide a new target for accurate clinical diagnosis and treatment of CKD-MBD patients, but its diagnostic value in CKD-MBD needs more in-depth research (Chen et al., 2018). However, the oxidation in the CKD process, the low reactivity of parathyroid hormone, and the accumulation of inhibitory C-terminal fragments at the target tissue level affect the accuracy of PTH in evaluating CKD-MBD to a certain extent (Katherine et al., 2010; Evenepoel et al., 2016). Nevertheless, accurate determination of PTH is an important part of clinical management for patients with CKD-MBD.

Activin A is a multifunctional cytokine, which is the most abundant transforming TGF-β family protein found in the bone matrix. SakaiR et al. found that the expression of activin A was related to bone resorption firstly (Sakai et al., 2000).

Previous studies have shown that activin A could stimulate skeletal growth and inhibit activin signals (Peng et al., 2018). Furthermore, it has been shown to enhance the activity of osteoclasts, and the inhibitory effect of ligand trap RAP011 on activin type IIA could inhibit osteoclast formation and bone remodeling in CKD diabetic mice in vitro (Sugatani et al., 2017). Another study demonstrated that using natural antagonists of activin A inhibin in mice could lead to an increase in bone mineral density (BMD) (Perrien et al., 2007). A recent cross-sectional study involving 104 patients with CKD found that activin A could reflect histomorphological parameters (BFR/BS, Acf, ObS/BS, and OcS/BS) of bone turnover similar to those of PTH and FGF-23. Interestingly, blood activin A levels in patients with CKD increased earlier than those in PTH and FGF-23, suggesting the role of activin A in evaluating early CKD-MBD development (Lima et al., 2019).

TRAP5b is an enzyme mainly derived from osteoclasts (Guañabens et al., 2019). It has been found that osteoclast-like cells produce TRAP5b in cell lines differentiated by exposure to an osteoblast-derived ligand receptor activator for NF-κB ligand (RANKL) (Lv et al., 2015). TRAP5b could affect the function of phosphate by separating it from the protein. Furthermore, it is the only available useful biomarker reflecting bone resorption in CKD patients regardless of CKD progression, hemodialysis, or peritoneal dialysis (Shidara et al., 2008; Vervloet and Brandenburg, 2017). However, the bone histomorphometric evidence for its reliability as a biomarker of CKD-MBD is still insufficient.

In addition to the biomarkers mentioned above, N-terminal propeptide of procollagen-1, bone-specific alkaline phosphatase, and C-terminal crosslaps of collagen-1 might also be of certain value in CKD-MBD (Vervloet and Brandenburg, 2017).