Podocytes are the highly specialized cells whose foot processes cover the basement membrane of the glomerulus and comprise the filtration slit diaphragms, therefore regulating blood filtration (Pavenstadt et al., 2003). Most human chronic kidney diseases exist with podocyte injury or podocyte loss (Mundel and Shankland, 2002; Nagata, 2016). The loss of podocytes and the inability to renew a damaged glomerulus with functional podocytes will ultimately result in glomerulosclerosis or scarring of the glomerulus (Barisoni et al., 1999). Diabetes and other systemic disease states can lead to podocyte injury and loss, which in turn results in ESRD (Kriz, 2002; Welsh et al., 2010).

Podocytes express cyclin A, cyclin B1, and cyclin D1 and CDK inhibitors (such as p21, p27, and p57). In the early stage of kidney development, Ki-67, which is a marker of the proliferated cells, was highly expressed in immature podocytes, whereas cyclin D1 and CKIs were dramatically downregulated; in the capillary loop stage, CKIs and cyclin D1 were intensely increased, whereas Ki-67, cyclin A, and cyclin B1 were not detectable (Nagata et al., 1998; Barisoni et al., 2000b). The expression changes of cyclins, CDKs, and CKIs were associated with podocytes exiting the cell cycle and differentiate into mature podocytes expressing the podocyte markers, such as WT-1 or podocalyxin (Nagata et al., 1998; Barisoni et al., 2000b). Under normal physiological conditions, mature podocytes are arrested in G0 quiescent phase and express high levels of CDK inhibitors. The constitutive and intense production of CKIs is necessary to maintain the function of the differentiated quiescent podocytes (Nagata et al., 1998). The high level of CKIs probably leads to mature podocytes lacking the ability to renew during adult life.

Severe injuries induce cell death and promote the proliferation of survival cells so as to compensate for the cell loss. As postmitotic and quiescent cells, podocytes do not readily proliferate after injuries; however, in some diseased situations, such as collapsing focal segmental glomerulosclerosis (FSGS), the podocytes were stained positive for proliferating cellular markers and some podocytes even existed as binuclear (Barisoni et al., 1999). Cyclins, CDKs, and CKI expression were also changed in collapsing FSGS and human immunodeficiency virus-associated nephropathy (HIVAN); in these diseased situations, p27, p57, and cyclin D disappeared in podocytes, whereas the p21, cyclin A, and Ki-67 were highly expressed (Barisoni et al., 2000a; Barisoni et al., 2000b; Shankland et al., 2000). These podocytes bypassed cell cycle restriction points and entered the cell cycle, but they were unable to complete cell cycle and finally causing podocytes loss via podocyte mitosis (mitotic catastrophe) (Liapis et al., 2013). In the setting of adriamycin-induced podocyte injury, the presence of p21 has a protective effect on the podocytes in this model of toxic podocytopathy (Marshall et al., 2010). In other glomerular diseases, such as membranous nephropathy, immune-mediated injury led to cyclin A and Cdk2 upregulation in podocytes, mitosis entry, and DNA synthesis. Although these podocytes entered mitosis, they were unable to successfully complete it, and podocytes manifested as multinucleated and absence of cytoplasmic division (cytokinesis).

Diabetic nephropathy is characterized by podocyte hypertrophy. In various experimental models of diabetic nephropathy, such as Zucker diabetic rats and db/db mice, both models of type II diabetes, or type I model, induced by streptozotocin administration, the increasing expression of p27 and p21 was identified (Kuan et al., 1998; Hoshi et al., 2002; Baba et al., 2005). Although diabetic p21 or p27 knockout mice were protected from glomerular hypertrophy and the development of progressive renal failure, the specific mechanism of podocyte hypertrophy and its role in renal fibrosis remain unknown (Kuan et al., 1998; Awazu et al., 2003). As the terminal consequence of podocyte injury, glomerulosclerosis is characterized by segmental obliteration of glomerular capillaries with the extracellular matrix and has been believed to be a process to the complete sclerosis without regression (D’Agati, 2012). As a typical feature of kidney disease, proteinuria is induced by the podocyte injury because slit membrane molecules, the actin cytoskeleton, and cell adhesion molecules have formed a tight network so as to maintain filtration barrier function, and defection of these components leads to proteinuria (D'Agati et al., 2011). Persistent or severe podocyte injuries lead to cell detachment, which is probably caused by mitotic catastrophe (Kriz and Lehir, 2005). As normal function of podocytes requires specific arrangement of the cellular actin cytoskeleton, this may lead to podocytes unable to further form the actin contractile ring required by cytokinesis, resulting in mature podocytes that are unable to complete cytokinesis. Binucleated podocytes are frequently seen in the urine, suggesting podocyte loss caused by mitotic catastrophe is involved in podocyte detachment (Liapis et al., 2013). Moreover, podocyte loss in adriamycin-induced nephropathy was alleviated through administration of an inhibitor of p53-dependent cell cycle arrest, MDM-2, further strengthening the hypothesis (Mulay et al., 2013).

In summary, under normal physiological conditions, mammalian podocytes are arrested in G0 quiescent phase and express high levels of CDK inhibitors. Injuries can induce podocyte death, whereas the remaining podocytes are unable to undergo regenerative proliferation to compensate for the loss of podocytes. Although podocytes can enter the cell cycle and can even undergo nuclear division in a variety of glomerular diseases, they were unable to complete normal cell division. The expression of CKIs, such as p21, p27, and p57, could lead to podocyte G1/S phase arrest, causing the abundant podocyte hypertrophy seen in progressive renal failure (Figure 3). When mature podocytes are forced to override cell cycle restriction point, they fulfill an aberrant mitosis followed by detachment and death through mitotic catastrophe (Figure 3). Such podocytes appeared multinucleated with aberrant mitotic spindles or micronuclei and were often found in several human and experimental glomerular diseases, such as HIVAN, FSGS, minimal change disease, immunoglobulin A (IgA) nephropathy, or adriamycin-induced nephropathy (Liapis et al., 2013; Mulay et al., 2013); all of these diseases exist with different degree of renal fibrosis.

Under normal physiological conditions, mammalian mature TECs proliferate at a very low rate, which could be proved by PCNA and Ki-67 immunostaining (Nadasdy et al., 1994; Witzgall et al., 1994). Through this low rate proliferation, kidney can remedy the loss of TECs into the urine, which is few under normal conditions, probably one TEC per human nephron daily (Prescott, 1966). However, the rate of dividing cells remarkably increases after AKI so as to remedy TEC loss (Humphreys et al., 2008). If the injury is mild, the surviving TECs could cover the exposed basal membrane and restore cell number by proliferation (Prescott, 1966; Witzgall et al., 1994; Lin et al., 2005). In addition to proliferation, surviving TECs can also differentiate and express the embryologic markers such as vimentin (Witzgall et al., 1994; Lin et al., 2005), and then redifferentiate into specialized TECs resulting in the recovery of the nephron (Humphreys et al., 2011).

However, when the damage is more severe or repeated, the repair process can be maladaptive, which will lead to incomplete structural and functional recovery of kidney tissue with persistent inflammation, activation, and proliferation of myofibroblasts, vascular rarefaction, increased production of interstitial matrix, and finally resulting in the progression of fibrosis (Grgic et al., 2012). Upon severe injuries, some TECs will arrest in the G2/M phase of cell cycle and mediate renal fibrosis by secreting profibrotic cytokines such as CTGF and TGF-β1 (Yang et al., 2010; Cosentino et al., 2013; Tang et al., 2013; Zhao et al., 2020) (Figure 4). Drug intervention causing an increasing TECs G2/M phase arrest after AKI could aggravate kidney fibrosis, whereas interventions which reduce TECs G2/M phase arrest result in less renal fibrosis (Yang et al., 2010; Cosentino et al., 2013; Gasparitsch et al., 2013; Tang et al., 2013; Zhao et al., 2020). Hence, TECs G2/M phase arrest properly is a novel histologic biomarker of renal fibrosis.

Many factors could influence TECs G2/M phase arrest and therefore affect renal fibrosis outcome. During the process of renal fibrosis, TECs undergo a partial EMT program; during this process, TECs still keep associated with their basement membrane but can express cellular markers of both epithelial and mesenchymal cells (Lovisa et al., 2015). During fibrotic injury, the partial EMT program led to a TEC G2/M phase arrest of the cell cycle; inhibition of partial EMT program can alleviate TEC G2/M phase arrest and attenuate interstitial fibrosis (Lovisa et al., 2015). Specific knockout of Atg5 gene in mouse TECs can destroy TEC autophagy and aggravate the TEC G2/M phase arrest, leading to aggravation of renal fibrosis upon kidney injury (Li et al., 2016). Specific knockout Numb in mouse TECs can alleviate TEC G2/M phase arrest and renal fibrosis induced by unilateral ureteral obstruction or unilateral ischemic renal injury (Zhu et al., 2016). Conventional knockout of Cyclin G1 can alleviate TEC G2/M phase arrest and renal fibrosis induced by severe kidney injury in mice (Canaud et al., 2019). Inhibition phosphorylation of 4E-BP1, a downstream effector molecule of mTORC1 pathway, can alleviate the TECs G2/M phase arrest and renal fibrosis (Sun et al., 2019). Therefore, TECs G2/M phase arrest is a common characteristic of renal fibrosis induced by various injures; however, the specific mechanism causing TECs G2/M phase arrest is still unclear.

Cell cycle regulators probably involve in the process of TECs G2/M phase arrest. Injuries can induce DNA damage of TECs, causing the activation of ATM/ATR (Kishi et al., 2019), which can further activate their downstream target genes Chk1/Chk2 and p53 (Reinhardt and Yaffe, 2009). Chk1/Chk2 inhibits Cdc25, the activator of Cdk1/cyclin B, causing G2/M phase arrest by downregulating the Cdk1/cyclin B kinase activity (Figure 4). p53 could activate its downstream target gene p21, causing G2/M phase arrest by downregulating the Cdk1/cyclin B kinase activity (Vogelstein et al., 2000) (Figure 4). Roscovitine, an inhibitor of Cdks, have been found to have the anti-fibrosis function (Steinman et al., 2012). In high glucose cultured HK-2 cells, a human proximal renal tubular epithelial cell line, roscovitine, can successfully reduce α-SMA expression and increase E-cadherin expression, suggesting that it can inhibit the EMT process of TECs (Wang et al., 2019). Further studies showed that roscovitine inhibited TECs EMT by inhibiting the upregulation of TGF-β1/p38MAPK pathway in HK-2 cells cultured with high glucose (Wang et al., 2019). In diabetic mice, administration of roscovitine can remarkably alleviate renal functional and histological injuries through inhibiting the expression of collagen, α-SMA, and TGF-β1 (Wang et al., 2019).

Consequently, although the TECs do not abundantly transdifferentiate into myofibroblasts, the G2/M phase arrested TECs could mediate renal fibrosis through paracrine pathway that is reinforced by a state of senescence characterized by the production of profibrotic cellular factors (Canaud et al., 2019) (Figure 4). In addition, p21 overexpression could induce the senescence of TECs which is involved in early stage of diabetic nephropathy in streptozotocin-induced diabetes 1 model (Kitada et al., 2014). In contrast to protective effect of p21 in AKI, the continued p21 activation may result in renal fibrosis, as the p21 knockout mice did not develop chronic kidney failure after 5/6 nephrectomy (Megyesi et al., 1999). It has been suggested that deletion of p21 allows hyperplastic compensatory proliferation of residual kidney tissue and prevents maladaptive hypertrophy (Megyesi et al., 1999).

Mesangial cells (MCs) offer structural support for the glomerular tuft partially through the secretion and maintenance of the extracellular matrix. There is less MC proliferation in the adult mammalian healthy kidney, probably because under normal conditions MCs are either not exposed to mitogens or protected from them by inhibitory factors. Mature MCs remain in the G0 quiescent state by upregulation of the cell cycle inhibitor p27 (Combs et al., 1998). The initial of MC division could be stimulated by mitogen of injuries accompanied by a decrease in the expression of p27 (Figure 5).

In response to all kinds of injury stimulation, the quiescent MCs were stimulated to proliferation, resulting in an increasing number of MCs and persistent cellular matrix accumulation, and finally causing glomerulosclerosis. These characteristics could be found in diseases such as IgA nephropathy, lupus nephritis, membranoproliferative glomerulonephritis, and diabetic nephropathy. In the model of mesangial proliferative glomerulonephritis, Thy1 nephritis, the expression of cyclin D, cyclin E, cyclin A, CDK2, and CDK4 were increased during the phase of marked mesangial proliferation (Schöcklmann et al., 1999).

MC proliferation and the resulting matrix formation are the major characteristics of glomerular injury and fibrosis (Johnson, 1994; Shankland et al., 1996). Roscovitine, which can block the activity of CDK2, has been studied in experimental glomerulonephritis (Pippin et al., 1997b). It can decrease MC proliferation and the resulting matrix production (such as collagen type IV, laminin, and fibronectin), leading to alleviated renal fibrosis and improved renal function, suggesting that inhibiting MC overproliferation may be a useful therapeutic targeting for renal fibrosis (Pippin et al., 1997a). In the Thy1 model, accompanied by the onset of MC proliferation, the expression of p27 strikingly decreased. If experimental nephritis is induced in p27 knockout mice, the MC proliferation initial earlier and the proliferative response is bigger, accompanied by prominent extracellular matrix (ECM) accumulation (Marshall and Shankland, 2006). T-type calcium channels play an essential role in MC proliferation by targeting the G1/S checkpoint. Blocking of these channels by pharmacological drugs could inhibit MC proliferation by arresting them in G1 phase and alleviates glomerular damage in Thy1 model (Cove-Smith et al., 2013).

Statins, the cornerstone hormone drugs to manage dyslipidemia, have been found to have lipid-independent benefits against renal injury and fibrosis (Kostapanos et al., 2009; Chen et al., 2019). For example, statins can inhibit mesangial expansion, and the resulting extracellular matrix accumulation in the glomeruli of diabetic animal kidneys, and therefore attenuate renal fibrosis (KIM et al., 2000; Fujii et al., 2007). Moreover, in vitro studies had found that statins can inhibit the proliferation of cultured MC, which focal or diffuse proliferation is a typical characteristic of glomerular pathology (O'Donnell et al., 1993; Terada et al., 1998; Danesh et al., 2002). Lovastatin can dose-dependently inhibit DNA replication and proliferation of rat MCs, which can be reversible by added mevalonate (O'Donnell et al., 1993). Further study showed that the effect of lovastatin to inhibit MC proliferation through upregulation of a CDK inhibitor, p27Kip1, protein levels, as knockdown of p27Kip1 showed strikingly decreasing lovastatin-induced cell cycle arrest (Terada et al., 1998). Another study found that the proliferation of MCs induced by high glucose was accompanied by the decrease in p21 protein expression and the increase in CDK4 and CDK2 kinase activities. Simvastatin can increase p21 protein expression and downregulate CDK4 and CDK2 kinase activities (Danesh et al., 2002). These studies suggest that statins can inhibit renal injury and renal fibrosis independently of their cholesterol-lowering effect, such as inhibition of MC proliferation.

Therefore, MC overproliferation after injuries can contribute to renal fibrosis through persistent accumulation of ECM. Under normal physiological conditions, mammalian mature MCs are arrested in the G0 phase of cell cycle through upregulation of the CDK inhibitor p27 (Figure 5). However, upon injuries, the quiescent MCs are stimulated to overproliferation causing persistent cellular matrix accumulation, and finally causing renal fibrosis (Figure 5).

Renal fibrosis is characterized by deposition of extracellular matrix in the potential space between tubules and peritubular capillaries. It is generally believed that myofibroblasts are the primary extracellular matrix-producing cells that produce a fair amount of interstitial matrix components, such as fibronectin and type I and type III collagens. Considering this, one of the key problems in the field is to study the origin of these matrix-producing myofibroblasts (Grande and Lopez-Novoa, 2009; Meran and Steadman, 2011; Schrimpf and Duffield, 2011).

It has been supposed that myofibroblasts have at least five different sources in mammalian fibrotic kidney, including activation of interstitial fibroblasts, differentiation of pericytes, translation of tubular epithelial cells and endothelial cells and recruitment of circulating fibrocytes (Barnes and Gorin, 2011). It has been believed that matrix-producing myofibroblasts mostly derive from resident fibroblasts through activation after kidney injury (Hewitson, 2009). Although this perception has recently been challenged, it is generally accurate (Strutz and Zeisberg, 2006; Grande and Lopez-Novoa, 2009). Recently, Kuppe et al. have found that distinct subpopulations of pericytes and fibroblasts were the main sources of myofibroblasts during human kidney fibrosis through the single-cell RNA sequencing technology (Kuppe et al., 2021). Moreover, they also showed that NKD2 may be a myofibroblast-specific target in human kidney fibrosis, as overexpression of NKD2 in human fibroblast cell line promoted the expression of ECM molecules, whereas knockout of NKD2 markedly downregulated the expression of ECM molecules (Kuppe et al., 2021). Blocking fibroblast to myofibroblast transformation can effectively inhibit renal fibrosis (Gerarduzzi et al., 2017; Li N. et al., 2020), suggesting that fibroblast is the major source of myofibroblast.

Under normal physiological conditions, renal fibroblasts are located in the interstitial space between the capillaries and the tubular epithelia and take shape a network in the whole renal parenchyma, so as to stabilize tissue structure (Kaissling and Le Hir, 2008). These cells are stellate shaped and contain abundant rough endoplasmic reticulum, collagen-secreting granules and actin filaments. They involve multiple cell processes, which keep them in contact with the tubular and capillary basement membranes (Kaissling and Le Hir, 2008). Under normal physiological conditions, renal fibroblasts stay in quiescent G0 phase of cell cycle and express CD73 (also known as ecto-5′-nucleotidase) in their plasma membrane and produce erythropoietin (Kaissling and Le Hir, 2008; Paliege et al., 2010). PDGFRβ and FSP1 were also expressed in fibroblasts (Floege et al., 2008; Grigorian et al., 2008; Boye and Mælandsmo, 2010; Boor and Floege, 2011). Fibroblasts control interstitial matrix physiological homeostasis by producing a few ECM components in normal conditions. However, after injury, fibroblasts are activated and acquiring the ability to proliferate and translating to myofibroblast that expressing α-SMA, generating a huge amount of ECM components. Myofibroblasts also retain FSP1 and PDGFRβ expression and express vimentin de novo.

Activated fibroblasts possess two typical characteristics, that is, proliferation and myofibroblastic activation. The latter manifests as the expression of α-SMA and the production of extracellular matrix. Fibroblasts and myofibroblasts are overproliferated under the stimulus of cytokines, which results in the increasing number of myofibroblasts and accumulation of ECM in injured kidney. Growth factors such as PDGF, TGF-β, FGF2, and CTGF are well-known mitogens promoting fibroblast overproliferation (Strutz et al., 2000; Böttinger, 2007; Phanish et al., 2010; Ostendorf et al., 2012) (Figure 6). Besides these well-known cytokines, tissue-type plasminogen activator can also promote fibroblast overproliferation and myofibroblastic activation by recruitment of β1 integrin (Hu et al., 2007; Hu et al., 2008; Hao et al., 2010; Lin et al., 2010).

Therefore, fibroblasts are the main source of matrix-producing myofibroblasts. Under normal physiological condition, fibroblasts stay in quiescent state of G0 phase. Upon injuries, fibroblasts are activated and overproliferated, translating to myofibroblasts, causing renal fibrosis by consistent accumulation of ECM (Figure 6). However, the specific mechanism as to how the fibroblast is activated and overproliferated upon injuries remains unclear, and other types of renal cells, such as TECs, play an essential role during this process (Li X. et al., 2020). For example, tubule-derived exosomes can promote renal fibrosis through promoting fibroblast activation and proliferation (Liu et al., 2020).