Single-cell RNA sequencing and spatial transcriptomics in the discovery of kidney disease potential biomarkers: a narrative review

Abstract

The kidney maintains the structural and functional integrity of the human body through its sophisticated internal architecture. Specialized cell types are organized into distinct compartments, each serving specific functions. Recent advancements in sequencing technologies, such as single-cell RNA sequencing (scRNA-seq) and spatial transcriptomics (ST), have significantly enhanced our understanding of the cellular-level pathophysiological mechanisms underlying disease onset and progression, leading to the discovery of numerous biomarkers. This review summarizes the latest applications and advancements in scRNA-seq and its integration with ST in the discovery of potential biomarkers for kidney diseases, offering new perspectives and directions for future research and clinical applications.

Introduction

The kidney, a complex organ composed of diverse parenchymal and immune cell types, maintains homeostasis and responds to injury through dynamic intercellular networks [1, 2]. Understanding the specific roles of certain cell types and their intercellular interactions is crucial for elucidating the mechanisms underlying kidney diseases. Single-cell RNA sequencing (scRNA-seq), introduced in 2009, has since become an invaluable tool for characterizing transcriptomic profiles across diverse cell types [3, 4] and for identifying key genes and pathways in various fields, including nephrology (Fig. 1) [5–9]. However, scRNA-seq often lacks crucial spatial context and provides limited information on intercellular communication. Spatial transcriptomics (ST), an emerging technology, addresses this limitation by preserving both mRNA expression patterns and spatial localization within intact tissue samples. However, ST relies on intact tissue samples and therefore cannot be directly applied to non-tissue samples, such as blood or urine [10]. The integration of scRNA-seq and ST has proven particularly powerful for defining cell-type-specific transcriptomes, mapping spatial cell localization, and elucidating dynamic transitions among specific cell states.

Fig. 1.

Schematic of using single-cell RNA sequencing to identify different cell types and analyze intercellular communication (Created with BioGDP.com) [97]. Single cells were isolated from biological samples such as blood, urine, and kidney tissues of humans or animal models. The single-cell RNA sequencing (scRNA-seq) technique can identify distinct cell clusters using cell-type-specific biomarkers. The functional molecules involved in cell-cell communication are predicted based on the CellChat database. The communication between immune cells and kidney parenchymal cells is the current focus of study in kidney diseases. Mesangial cells, podocytes, proximal tubular epithelial cells, and endothelial cells constitute the majority of kidney parenchymal cells. PBMCs: peripheral blood mononuclear cells

Furthermore, the application of scRNA-seq and ST to renal biopsies holds great promise for identifying potential biomarkers for kidney diseases, although the clinical implementations of these technologies remain limited. In basic research, this combination is becoming increasingly popular. It is anticipated that differentially expressed genes, proteins, and specific cell types identified by these technologies may help in clinical diagnosis, prognostic assessment, and therapeutic target selection. Although urine and blood samples are not suitable for ST, they can be analyzed by scRNA-seq. Genes identified in urinary sediments or peripheral blood mononuclear cells may provide complementary information for potential biomarker discovery, as they can reflect molecular and cellular alterations occurring in kidney diseases.

This review summarizes recent advances in scRNA-seq and its integration with ST, emphasizing differentially expressed genes as well as intercellular interactions across various kidney diseases, though these interactions are mostly predicted through computational techniques and require further experimental validation. (Fig. 2) A comprehensive literature search was conducted in electronic databases in English over the past 10 years by inputting the keywords “single-cell RNA sequencing” or “spatial transcriptomics” combined with “kidney”. We excluded studies related to kidney tumors and kidney transplantation, as well as studies lacking validation of identified genes. Furthermore, single-cell RNA sequencing is sometimes performed via single-nucleus RNA sequencing in kidney research, and the outcomes of these two approaches may differ slightly. We refer to both sequencing techniques as “scRNA-seq” to make the text more readable and concise. These insights offer novel perspectives for understanding disease mechanisms and identifying potential targets for diagnosis and therapy. Of note, all the molecules remain at the discovery stage, with few or no clinical studies available. To critically evaluate the translational potential of these molecules, we categorized them into two groups based on the strength of the evidence, which also reflects their translational status. Group 1: molecules that have been tested in human studies, with at least one independent cohort. Group 2: molecules that have been revealed merely at the experimental level. Of the 50 molecules, 23 fall into Group 1, and 27 into Group 2. Molecules in Group 1 are closer to clinical application, but none have yet entered routine clinical practice. Specific details, including study size and reproducibility, are presented in Tables 1, 2 and 3.

Fig. 2.

Cellular landscape of various kidney diseases revealed by scRNA-seq and ST. scRNA-seq and ST have revealed a diverse array of cell types in blood, urine, and renal tissue samples across various kidney diseases. In PBMCs of blood, immune cells are detected in PMN (H), IgAN (H), and LN (H). In urine sediments, researchers primarily focus on immune cells detected in PMN (H) and LN (H). Additionally, TEC in urine sediments also represents a potential diagnostic and therapeutic target in AKI (H). In renal tissue samples, analysis of a broad spectrum of kidney diseases reveals the following cell types: parietal epithelial cell (PEC) in AKI (M); podocyte (POD) in IgAN (H), PMN (H), DKD (H, M), and LN (M); mesangial cell (MC) in IgAN (H), LN (H), and DKD (M); endothelial cell (EC) in IgAN (H), FSGS (H), DKD (M), LN (M), AKI (M), and AKI-CKD (M); tubular epithelial cell (TEC) in DKD (H, M), LN (H), AKI (M), AKI-CKD (M), and CKD (M); and immune cells in PMN (H), LN (H, M), DKD (M), ANCA-GN (H), AKI (M), AKI-CKD (M), and CKD (M). H, human; M, mouse; scRNA-seq, single-cell RNA sequencing; ST, spatial transcriptomics; PBMCs, peripheral blood mononuclear cells; PMN, primary membranous nephropathy; IgAN, IgA nephropathy; LN, lupus nephritis; TEC, tubular epithelial cell; DKD, diabetic kidney disease; MC, mesangial cell; POD, podocyte; AKI, acute kidney injury; PEC, parietal epithelial cell; EC, endothelial cell; FSGS, focal segmental glomerulosclerosis; CKD, chronic kidney disease; ANCA-GN, anti-neutrophil cytoplasmic antibody-associated glomerulonephritis

Table 1.

scRNA-seq and ST in human primary glomerular diseases

Disease	Specimens	Cells	Potential Biomarkers	Expression	Underlying Mechanism	Sample Size (Patients/Controls)	Independent Validation	Translational Status	Ref
IgA nephropathy	Kidney biopsies	Mesangial cells	JCHAIN	Up	IgA recognition, transport, and deposition	13/6	No	Group 2	[13]
	Peripheral blood mononuclear cells	Macrophages	CCR2	Up	Interactions between podocytes and macrophages	3/2	Yes	Group 2	[14]
		Follicular helper T cells, B cells	TNFSF14-TNFRSF14	Up	Communication between Tfh and B cells	30/29	No	Group 2	[15]
	Kidney biopsies	Endothelial cells	TXNIP	Up	Link to oxidative stress and inflammatory response	4/1	No	Group 2	[17]
		Endothelial cells	SPARCL1 CD74	Up	Cell adhesion, migration, and proliferation	4/1	No	Group 2	[17]
		Mesangial cells	SPARC, ROCK2	Up	Associate with the extracellular matrix	4/1	No	Group 2	[17]
		Podocytes	PTGDS	Down	Podocyte injury	3/2	No	Group 2	[14]
Focal segmental glomerulosclerosis	Kidney biopsies	Endothelial cells	A2M	Up	Glomerular selective filtration function	10/24	Yes	Group 1	[19]
Primary membranous nephropathy	Peripheral blood mononuclear cells and urine sediments	B cells and plasma cells	APRIL	Up	Immune dysregulation	5/3	Yes	Group 2	[23]
	Kidney biopsies	Podocytes	BMP2	Up	glomerular basement membrane thickening and podocyte dysfunction	11/7	No	Group 2	[25]
		Tubular cells	MMP7	-	Differentially expressed in massive and non- massive proteinuria groups	6/2	No	Group 2	[26]

Group 1 includes molecules that have been tested in human studies, with at least one independent cohort. Group 2 includes molecules that have been revealed merely at the experimental level. Molecules in Group 1 are the closest to clinical application, but none have yet entered routine clinical practice. Group 2 remains at preclinical or early exploratory stages. scRNA-seq, single-cell RNA sequencing; ST, spatial transcriptomics; JCHAIN, joining chain of multimeric IgA and IgM; SPARC, secreted protein acidic and cysteine; ROCK2, Rho associated coiled-coil containing protein kinase 2; PTGDS, prostaglandin D2 synthase; CCR2, C-C motif chemokine receptor 2; TNFSF, tumor necrosis factor superfamily; A2M, α-2-macroglobulin; MMP7, Matrix metalloprotease 7; BMP2, Bone morphogenetic protein 2; APRIL, a proliferation-inducing ligand

Table 2.

scRNA-seq and ST in secondary glomerular diseases

Disease	Species	Specimens	Cells	Potential Biomarkers	Expression	Mechanism	Sample Size (Patients or model/ Controls or sham)	Independent Validation	Translational Status	Ref
Diabetic kidney disease	Human	Kidney biopsies	Tubular epithelial cells	MMP7	Up	Link to interstitial fibrosis and the risk of declining kidney function	23/10	Yes	Group 1	[28]
			Podocytes	TRAIL	Up	TRAIL/DR5 pathway-PANoptosis	3/3	Yes	Group 2	[29]
	Mouse	Kidney tissues	Podocytes	RIPK3	Up	Podocyte loss	2/2	Yes	Group 1	[30]
			Tubular epithelial cells	SPP1	Up	Promoted pathogenic cross-talk among tubular cells	3/3	Yes	Group 1	[32]
			Mesangial cells	SEMA3C	Up	Communication between endothelial cells and mesangial cells	3/3	No	Group 1	[32]
Lupus nephritis	Human	Kidney biopsies	CD163 + dendritic cells	TNF, IL1B, CCL17, 22	Up	T cell migration and tissue damage	40/6	Yes	Group 1	[36]
			Tubular epithelial cells	IFN	Up	Response to treatment	21/3	Yes	Group 2	[37]
			Monocytes	APOE	Up	Antigen presentation and IFN secretion	2/2	Yes	Group 2	[40]
	Mouse	Kidney tissues	Monocytes	CCR2	Up	Monocyte differentiation	5/5	Yes	Group 1	[41]
			Tissue-resident macrophages	TNFSF13B	Up	B cell survival, proliferation, and plasma cell niche formation	5/5	Yes	Group 2	[43]
			Podocytes	PIK3α	Up	Podocyte proliferation, hypertrophy, and crescent formation	28/27	Yes	Group 1	[44]

TRAIL, TNF-related apoptosis-inducing ligand; MMP7, matrix metalloprotease 7; RIPK3, Receptor-interacting protein kinase 3; SPP1, secreted phosphoprotein 1; SEMA3C, semaphorin 3 C; RIPK3, receptor-interacting protein kinase 3; IFN, Type I interferon; TNF, tumor necrosis factor; CCR2, C-C motif chemokine receptor 2; APOE, apoprotein E; TNFSF13B, TNF superfamily member 13b; PIK3α, phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha

Table 3.

scRNA-seq and ST in kidney diseases with tubular injury in various kidney injury mouse models

Disease	Models	Cells	Potential Biomarkers	Expression	Underlying Mechanism	Sample Size (model/ sham)	Independent Validation	Translational Status	Ref
Acute kidney injury	Ischemia- reperfusion	Mesenchymal stem cells	miR-26a-5p	Up	Suppress the recruitment of immune cells to the profibrotic tubular epithelial cells	4/4	Yes	Group 2	[52]
		Regulatory T cells	AREG	Up	T cell proliferation and differentiation	3/3	Yes	Group 2	[53]
	Sepsis	Endothelial cells	IL-6	Up	Regulated by F4/80hi macrophages	6/6	Yes	Group 2	[55]
	Ischemia- reperfusion	Macrophages	S100A8/ A9	Up	Relate to tissue injury	6/6	Yes	Group 1	[61]
		Proximal tubular epithelial cells	Slc6a19	Up	Isoleucine transport and absorption	3/3	No	Group 1	[53]
		Parietal epithelial cells	WT1	Up	Parietal epithelial cell differentiation	3/3	Yes	Group 2	[63]
		Distal tubular cells	Nrp1	Up	Exacerbate AKI	7/7	Yes	Group 1	[64]
	Maleic Acid	Tubular cells	ALDH2	Up	ALDH2 lactylation impedes mitophagy	6/6	Yes	Group 1	[66]
	Drug-induced	Tubular epithelial cells	ATF4	Up	Promotes pyroptosis	6/6	Yes	Group 1	[67]
	High-dose folic acid	Proximal tubular cells	Lrig1	Up	Repair proximal tubular injury	6/6	No	Group 2	[68]
	Ischemia- reperfusion	Endothelial cells	VE-PTP	Up	Deactivate angiopoietin-Tie2 signaling	8/7	No	Group 2	[69]
Acute Kidney Injury to Chronic Kidney Disease	Ischemia- reperfusion	Macrophages	THBS1	Up	Cycling M2 macrophage proliferation	5/5	Yes	Group 1	[72]
		Endothelial cells	MCT4	Up	Monocyte-endothelial interactions and pro-inflammatory endothelial cell activity	1/1	No	Group 2	[74]
		Macrophages	Mincle	Up	TNF production	3/3	Yes	Group 2	[75]
		Tubular epithelial cells	YAP1	Up	Tubular epithelial cell polyploidization	6/6	Yes	Group 1	[76]
Chronic Kidney Disease and Renal Fibrosis	Aristolochic acid	Tubular epithelial cells	TβRII	Down	Mitochondrial dysfunction and Th1 immune response	2/2	Yes	Group 2	[84]
	Folic acid	Proximal tubular cells	ESRRA	Down	Regulate PTC-specific genes	2/6	Yes	Group 2	[85]
	Transgenic mouse	Distinct cell types	Rtn3	Down	Disrupt cell-cell contact	3/3	Yes	Group 1	[87]
	Unilateral ureteric obstruction	Proximal tubular cells	CXCL1	Up	Recruit CXCR2 + basophils	2/6	Yes	Group 1	[78]
		Tubular epithelial cells	Cx43	Up	ATP release	4–7/ group	Yes	Group 1	[88]
		Fibroblasts	TRPC6	Up	Relate to interactions between fibroblasts and endothelial cells	3/3	Yes	Group 1	[89]
		Myofibroblasts	WWP2	Up	Metabolic reprogramming and the profibrotic response	3/3	Yes	Group 1	[91]
		Macrophages	Src	Up	Macrophage-myofibroblast transition	4/4	Yes	Group 2	[93]
		Fibroblasts	lncRNA Gas5	Up	Positively correlated with the severity of kidney fibrosis	3/3	Yes	Group 2	[94]

Slc6a19, solute carrier family 6 member 19; WT1, Wilms’ tumor 1; Nrp1, Neuropilin-1; VE-PTP, vascular endothelial protein tyrosine phosphatase; AREG, amphiregulin; S100A8, S100 calcium binding protein A8; ALDH2, Aldehyde dehydrogenase 2; ATF4, Activating transcription factor 4; Lrig1, Leucine-rich repeats and immunoglobulin-like domains protein 1; MCT4, monocarboxylate transporter 4; YAP1, yes1 associated transcriptional regulator; THBS1, thrombospondin-1; Mincle, macrophage-inducible C-type lectin; RTN3, Reticulon 3; Cx43, connexin 43; CXCL, C-X-C motif chemokine ligand; TRPC6, Transient receptor potential canonical 6; WWP2, WW domain containing E3 ubiquitin protein ligase 2; Src, a proto-oncogene tyrosine protein kinase; lncRNA Gas5, Long noncoding RNA growth arrest-specific 5; ESRRA, Estrogen-related receptor alpha; TβRII, TGF-β type II receptor

Primary glomerular diseases

Primary glomerular diseases are among the leading causes of end-stage kidney disease in young adults and represent the third most common cause of kidney failure worldwide, following diabetes and hypertension. The most prevalent subtypes of primary glomerular diseases include IgA nephropathy, focal segmental glomerulosclerosis, and primary membranous nephropathy (Table 1) [11].

IgA nephropathy

IgA nephropathy, characterized by the mesangial deposition of IgA-containing immune complexes, leads to inflammation and eventually to glomerulonephritis. It is recognized as a prototypical autoimmune disease, with immunological, genetic, environmental, and nutritional factors contributing to its pathogenesis [12].

Applied to human kidney biopsies, scRNA-seq identified a diverse array of cell-type-specific molecular alterations associated with disease pathogenesis, which could serve as biomarkers. From an immunological perspective, the joining chain of multimeric IgA and IgM (JCHAIN) showed higher expression in mesangial cells, which contributed to IgA recognition, transport, and deposition, highlighting its roles in disease development [13]. In pediatric IgA nephropathy presenting with nephrotic syndrome, Chen et al. observed an increased abundance of intermediate monocytes, which exhibited elevated expression of C-C motif chemokine receptor 2 (CCR2). The researchers identified the CCL2 (C-C motif chemokine ligand 2)/CCR2 pathway modulating the interactions between podocytes and macrophages, though this finding is based on computational prediction and requires experimental validation [14]. CCR2 also played a crucial role in lupus nephritis, as discussed below. In addition to tissues, scRNA-seq studies on the patients’ peripheral blood mononuclear cells also predicted significant immune cell interactions. Specifically, B cells exhibited extensive interactions with infiltrating CD4⁺ T cells, especially follicular helper T cells, and this crosstalk was mediated by tumor necrosis factor superfamily ligand 14 (TNFSF14) and its receptor TNFRSF14. TNFSF family members also contributed to the progression of lupus nephritis [15]. Furthermore, comprehensive single-cell immune profiling revealed distinct expression patterns: NK cells were negatively correlated with disease progression, whereas B cells and monocytes were positively correlated with disease severity [16]. Collectively, immune signatures detected in the peripheral blood mononuclear cells might reflect intrarenal inflammatory status and serve as non-invasive biomarkers.

Regarding metabolic dysregulation, endothelial cells showed increased expression of TXNIP, which was associated with oxidative stress and inflammatory responses. In terms of tissue remodeling and fibrosis, endothelial cells showed increased expression of SPARCL1 and CD74, which regulated cell adhesion, migration, and proliferation. In addition, mesangial cells displayed distinct transcriptional profiles according to disease severity. For example, compared with patients in the microproteinuria group, those with overt proteinuria exhibited increased expression of extracellular matrix-associated glycoprotein secreted protein acidic and rich in cysteine (SPARC) and Rho-associated coiled-coil containing protein kinase 2 (ROCK2), which were associated with extracellular matrix production and renal fibrosis [17]. For cell injury, the downregulated prostaglandin D2 synthase (PTGDS) in podocytes was related to podocyte injury, suggesting its potential as a clinical biomarker [14]. Additionally, increasing evidence reveals the roles of intercellular communication, particularly mesangial-tubular cell interactions. However, most of the cellular interactions described above are computational predictions and await further experimental validation.

Focal segmental glomerulosclerosis

Focal segmental glomerulosclerosis is a major component of nephrotic syndrome in both adults and children. It is characterized by a common glomerular lesion, primarily induced by various insults targeting or originating from podocytes [18].

Using scRNA-seq analysis of human renal biopsies, Menon et al. generated an extensive adult kidney single-cell transcriptomic atlas. They observed two distinct focal segmental glomerulosclerosis patient groups with significantly divergent intrarenal alpha-2-macroglobulin (A2M) gene expression in endothelial cells. From the perspective of glomerular filtration function, A2M contributed to the ability of the glomerular capillary wall to prevent macromolecules from entering the urinary space. Higher A2M expression was associated with lower proteinuria remission rates, indicating that A2M upregulation might serve as a biomarker of worse prognosis [19].

Beyond tissue-based profiling, scRNA-seq analysis of urinary sediments also provided valuable insights into disease progression. Latt et al. identified several distinct cell types, indicating that urinary sediments can reflect the status of kidney injury and suggesting a promising future for disease diagnosis [20].

Primary membranous nephropathy

Primary membranous nephropathy is an autoimmune glomerulopathy characterized by increased protein excretion and thickened glomerular basement membrane. It is the most common type of nephrotic syndrome in individuals without diabetes [21]. The advent of scRNA-seq has enabled the exploration of this disease at an unprecedented level of detail.

From an immune perspective, Feng et al. performed single-cell B cell receptor sequencing and scRNA-seq on the kidney and circulating CD19⁺ cells from a patient negative for seven common podocyte autoantibodies. Their study revealed critical roles of memory B cells and naive B cells, particularly CD38⁺ naive B cells, in regulating the circulating immunological microenvironment and autoantibody production [22]. Beyond human kidney biopsies, peripheral blood mononuclear cells and urinary sediments were also useful for exploring immunological dysregulation. Specifically, APRIL (a proliferation-inducing ligand from B cells) was a promising regulator of B-cell and plasma cell activity and might also be involved in cell-cell contacts [23]. Furthermore, the origins and functions of urinary macrophages also varied between patients in complete remission and non-responders to primary membranous nephropathy therapies [24].

Regarding cell injury, in human biopsies analyzed by scRNA-seq, podocytes showed upregulation of bone morphogenetic protein 2 (BMP2), which was associated with glomerular basement membrane thickening and podocyte dysfunction, probably through the phosphorylated SMAD1 (pSMAD1)/collagen type IV (COL4) pathway [25]. In the context of kidney dysfunction, an unbiased analysis of kidney biopsies by Xu et al. revealed profound cellular diversity and novel disease-associated pathways. According to this study, patients with massive proteinuria exhibited distinct differentially expressed genes compared with those in the non-massive proteinuria group, such as matrix metallopeptidase 7 (MMP7) in all tubular cells, suppressor of cytokine signaling 3 (SOCS3) in many intrinsic cells, and KLF transcription factor 6 (KLF6) in nearly all cell types. These molecules might be related to kidney lesions and could be used in clinical diagnosis and classification after experimental and clinical validation. Moreover, MMP7 played a significant role in diabetic kidney disease, which will be discussed below [26].

Secondary glomerular diseases

In recent years, secondary glomerular diseases, such as diabetic kidney disease, lupus nephritis, and anti-neutrophil cytoplasmic antibody (ANCA)-associated glomerulonephritis, have also seen widespread applications of scRNA-seq (Table 2).

Diabetic kidney disease

Diabetic kidney disease, a major microvascular complication of diabetes, is a leading cause of end-stage kidney disease. The primary pathological features of diabetic kidney disease include interstitial fibrosis, renal tubule atrophy, glomerular basement membrane thickening, and mesangial cell dilation [27].

In studies using scRNA-seq on human kidney biopsies, from a fibrotic perspective, researchers found that MMP7, primarily generated from proximal tubules, connecting tubules, and principal cells, was highly correlated with interstitial fibrosis. Furthermore, plasma MMP7 levels also correlated with the risk of declining kidney function, indicating that it was a viable biomarker for both diagnosis and therapy [28]. Regarding cell injury, TNF-related apoptosis-inducing ligand (TRAIL) was discovered to be differentially expressed in kidney biopsies. Functionally, TRAIL interacted with death receptor 5 (DR5) and induced PANoptosis (pyroptosis, apoptosis, and necroptosis), which further contributed to podocyte injury and accelerated diabetic kidney disease progression. Experiments conducted both in vitro and in vivo confirmed this outcome. Thus, TRAIL might represent a tissue-based biomarker candidate for clinical application [29].

Animal models are frequently used in basic research, though these findings often require cautious extrapolation to humans. scRNA-seq has been substantially applied to mouse kidney tissues, yielding significant findings. For cell injury, receptor-interacting protein kinase 3 (RIPK3) in podocytes was a pivotal factor associated with renal ultrastructure alterations, podocyte loss, and albuminuria. Through activating PGAM5-Drp1 signaling, RIPK3 contributed to mitochondrial fission and facilitated albuminuria [30]. Podocyte injury could further exacerbate disruption of intercellular communication among podocytes, endothelial cells, and mesangial cells, thereby significantly contributing to the progression of kidney disease [31]. Regarding fibrosis, a study found that urine SPP1 (secreted phosphoprotein 1)/creatinine levels correlated with tubular damage, and SPP1 could promote pathogenic crosstalk across multiple types of tubular cells. Besides, SPP1 was also expressed in macrophages and contributed to ANCA-associated glomerulonephritis, as discussed later. Concurrently, endothelial cells and mesangial cells communicated through semaphorin 3 C (SEMA3C). There was also a positive correlation between urine SEMA3C/Cr level and glomerular damage, indicating the diagnostic potential of SEMA3C [32]. Notably, these inferred cell-cell interactions are mostly computational predictions and require experimental validation.

Additionally, the integration of scRNA-seq with ST on kidney tissues has also proven valuable. In rat models, this approach dissected the functional diversity of the fibrotic niches and complex interactions among cell subtypes. These studies demonstrated that persistent hyperglycemia in diabetic kidney disease activated macrophages. Fibrotic factors were then released and promoted renal interstitial fibrosis by inducing the trans-differentiation of glomerular mesangial cells into myofibroblasts [33].

Lupus nephritis

Lupus nephritis, a severe manifestation of systemic lupus erythematosus, affects 60–80% of systemic lupus erythematosus patients, leading to various degrees of renal dysfunction throughout the disease course [34].

Der et al. were the first to demonstrate that scRNA-seq was a feasible and informative technique applicable to renal biopsies from human lupus nephritis [35]. From an immunological standpoint, a CD163⁺ dendritic cell subset was identified, and its number had the potential to indicate therapeutic response after immunosuppressive and glucocorticoid induction therapy. CD163⁺ dendritic cells were implicated in a lupus nephritis-specific inflammatory network: injured proximal tubular epithelial cells expressed elevated levels of pro-inflammatory cytokines and chemokines, thereby recruiting blood-derived CD163⁺ dendritic cells to the kidney. Infiltrating CD163⁺ dendritic cells exhibited significant expression of TNF, IL1B, CCL17, and CCL22, confirming their roles as potent producers of inflammatory mediators that contributed to T cell migration and tissue damage [36]. A further study demonstrated that type I interferon (IFN) response signatures in tubular cells could distinguish patients with lupus nephritis from healthy controls. A high IFN response profile in tubular cells was associated with treatment failure. Interestingly, the same high IFN response was also observed in keratinocytes from skin biopsies [37]. Beyond tissue biopsies, scRNA-seq analysis of patient urinary sediments and peripheral blood mononuclear cells offered a complementary perspective. Compared with peripheral blood mononuclear cells, urinary sediments were possibly a more suitable surrogate for kidney biopsies due to the strong correlation in immune cell gene expression profiles between kidney biopsies and urinary sediments [38]. A further study conducted by Fava et al. supported this point. They analyzed urinary proteins at various diagnostic stages and identified hundreds of potential biomarkers as well as their associated molecular pathways [39]. Furthermore, by integrating scRNA-seq with ST on human kidney samples, researchers identified lymphangiogenesis as a potential route for lupus nephritis-specific monocyte/macrophage trafficking into and out of glomerular lesions and mapped lupus nephritis-specific monocyte/macrophage subclusters across ST layers. A novel, functionally impaired monocyte subset (APOE⁺ Mono) was characterized by its reduced antigen presentation and IFN secretion [40].

Concurrently, scRNA-seq analysis of lupus nephritis mouse models yielded significant insights that may inform human studies. At the immune level, researchers identified that CCR2 promoted the conversion of monocytes into immature macrophages. These immature macrophages might appear in urinary sediments from patients with lupus nephritis, and their abundance was correlated with clinical disease scores. Therefore, CCR2-mediated transition of monocytes into immature macrophages might play a crucial role in disease progression, and CCR2 might be a biomarker for future clinical application [41]. Besides, the spatiotemporal distribution patterns of human umbilical cord-derived mesenchymal stem cells were correlated with the dynamic deterioration of renal function. The C-X-C motif chemokine ligand 10 (CXCL10)-C-X-C motif chemokine receptor 3 (CXCR3) axis was found to be a crucial regulator of mesenchymal stem cells that migrated to the kidneys affected by lupus nephritis [42]. In addition, scRNA-seq along with ST in mouse models characterized kidney macrophage activation, heterogeneity, signaling, and the distinct functions of tissue-resident and monocyte-derived renal macrophages. Notably, tissue-resident macrophages showed higher expression of TNF superfamily member 13b (TNFSF13B), which encoded BAFF, a cytokine that supported B-cell survival, proliferation, and plasma cell niche formation. Selective inhibition of BAFF might offer therapeutic potential for lupus nephritis [43]. Regarding both cell injury and immune responses, phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha (PIK3α) played vital roles in podocyte proliferation, hypertrophy, and crescent formation, and it also contributed to immune cell activity. In the mouse model, B- and T-lymphocyte populations were impacted by the PIK3α inhibitor alpelisib treatment, which improved kidney function by reducing the generation of pro-inflammatory cytokines [44].

ANCA-associated glomerulonephritis

ANCA-associated vasculitis is a rare systemic disorder characterized by relapsing inflammation and severe injury of small- and medium-sized blood vessels. This vascular injury can jeopardize organ function and lead to mortality, commonly resulting in significant multi-organ involvement. Renal involvement, specifically ANCA-associated glomerulonephritis, often manifests as rapidly progressive crescentic glomerulonephritis, leading to a rapid decline in renal function [45].

One feature of ANCA-associated glomerulonephritis is renal macrophage infiltration. Researchers created a single-cell immune atlas based on kidney biopsies from patients. They discovered four subsets of macrophages, including SPP1⁺ lipid-associated macrophages and classical monocyte-derived macrophages, which might be essential for the disease progression [46]. Furthermore, integrated scRNA-seq and ST analysis of kidney biopsies from patients identified T-helper-1 cells and T-helper-17 cells as key mediators of immune-mediated renal injury. Based on this discovery, researchers administered ustekinumab, an add-on medication that specifically targets T-helper-1 and T-helper-17 cells, to four ANCA-associated glomerulonephritis patients and observed encouraging results (a small uncontrolled case series). These observations were preliminary, but they suggested that further clinical evaluation of ustekinumab in larger trials might be warranted [47]. Consequently, the integration of scRNA-seq and ST might facilitate the identification of potential pathophysiological drivers of ANCA-associated glomerulonephritis and the development of individualized treatment strategies.

Tubular injury

In our exploration of tubular injury, we primarily focus on acute kidney injury, the progression from acute kidney injury to chronic kidney disease, and chronic kidney disease and renal fibrosis, all of which have been extensively investigated using scRNA-seq and ST (Table 3).

Acute kidney injury

Acute kidney injury is a severe clinical condition marked by a sudden decline in renal function and is associated with increased morbidity and mortality. During the early stage of acute kidney injury, some epithelial cells undergo cell death, whereas surviving cells undergo dedifferentiation, thereby triggering inflammatory responses [48].

Using scRNA-seq analysis of human renal biopsies, researchers revealed a distinct cell-specific transcriptomic atlas, dysregulated signaling pathways, and computationally predicted intercellular communication mechanisms involved in the progression of acute kidney injury [49]. Elevated T cell immunoreceptor with Ig and immunoreceptor tyrosine-based inhibitory motif domains (TIGIT) was found to modulate T-cell activity as well as inflammatory and metabolic gene expression. TIGIT knockout in mouse models was shown to attenuate kidney damage, suggesting that TIGIT might warrant further investigation as a therapeutic target in humans [50]. Furthermore, Klocke et al. conducted scRNA-seq analysis on human urinary sediments and discovered that the transcriptomes of tubular epithelial cells might reflect kidney disease status, consistent with findings from both human samples and mouse models of acute kidney injury. This finding indicated that released tubular epithelial cells might serve as valuable targets for studying kidney inflammation and regeneration following injury [51].

Significant scRNA-seq advancements in murine models of ischemia-reperfusion-induced acute kidney injury unveiled critical mechanisms and potential therapeutic targets. In the context of immune regulation, mesenchymal stem cells were proposed as a potential cell-based therapy for tubular epithelial cell damage, based on observations that they might reduce canonical injury markers such as KIM-1. Researchers also discovered that miR-26a-5p generated from mesenchymal stem cells might modulate the therapeutic efficacy by inhibiting the recruitment of profibrotic immune cell subpopulations to profibrotic tubular epithelial cells. These findings raised the possibility that optimizing MSC delivery strategies could be explored for targeted therapy in acute kidney injury [52]. Furthermore, Sabapathy et al. investigated the roles of endogenous IL-33/ST2 (suppression of tumorigenicity 2) signaling in regulatory T cells. They discovered that reparative factors, such as amphiregulin (AREG), which promoted proliferation and differentiation, were enriched in ST2⁺ regulatory T cells. In an ST2- and AREG-dependent manner, regulatory T cells protected kidney organoids against hypoxia, suggesting that regulatory T cells might represent therapeutic targets for resolving kidney injury [53]. Besides, in murine models of glycerol-induced rhabdomyolysis-associated acute kidney injury, Rao et al. uncovered a novel macrophage cluster characterized by a senescence-associated gene signature, suggesting that senolytic treatment might exert protective effects [54]. In sepsis-associated acute kidney injury, F4/80^hi macrophages regulated the disease progression by suppressing interleukin 6 (IL-6) expression in endothelial cells. Mice with reduced numbers of F4/80^hi macrophages showed more severe injury, and anti-IL-6-targeted treatment could alleviate kidney injury. These findings also pointed to an immunoregulatory axis between macrophages and endothelial cells, though it required experimental validation [55]. In a cisplatin-induced acute kidney injury model, researchers characterized renal lymphatic endothelial cell subclusters and further identified key immunomodulatory mechanisms, particularly the interactions between lymphatic endothelial cells and T cells [56]. Additionally, studies examining peritoneal immune cells demonstrated a novel function of CD38 ligation, which stimulated renal stromal cells to produce IL-6 through nicotinamide phosphoribosyl transferase (NAMPT) signaling from CD38⁺ macrophages. This mechanism was consistent across humans and animal models. Blockade of the NAMPT signaling pathway could reduce kidney injury and reverse the increase in IL-6 levels, suggesting that NAMPT blockade might be further explored as a therapeutic approach [57]. The integration of scRNA-seq and ST substantially advanced our understanding of the pathogenesis of acute kidney injury. Utilizing this integrative approach in mouse models of ischemia-reperfusion-induced acute kidney injury, Cheung et al. identified, described, and spatially localized seven distinct kidney-resident macrophage subpopulations, each with specific transcriptomic signatures mapped to niches from cortex to papilla. Following injury, both the spatial distributions and transcriptomic signatures of these macrophage subpopulations were disrupted; defining these alterations might inform more targeted interventions [58]. In a separate study, renal double-negative T cells, with the absence of CD4 and CD8 markers, exhibited suppressive effects on CD4⁺ T cells in vitro, thereby protecting the kidney from acute kidney injury. Fc epsilon receptor Ig (Fcer1g), a gene for double-negative T cells, might function as a biomarker for acute kidney injury [59]. Regarding macrophages, researchers found distinct markers for each subset of kidney-infiltrating macrophages and investigated their spatiotemporal distribution, providing an innovative perspective on the immunological landscape throughout the progression of acute kidney injury. These macrophage subsets warrant further investigation as potential predictive markers in humans [60]. Another study highlighted the complementary roles of tissue-resident macrophages and monocyte-derived kidney macrophages in regulating tissue inflammation and promoting repair. It also identified S100A8/A9⁺ (S100 calcium-binding protein A8/A9⁺) macrophages that infiltrated the kidney and were associated with tissue injury [61].

On the metabolic side, through scRNA-seq in ischemia-reperfusion-induced acute kidney injury, it was found that proximal tubular epithelial cells with high expression of solute carrier family 5 member 2 (Slc5a2) concurrently exhibited higher expression of the neutral amino acid transporter gene Slc6a19. As acute kidney injury progressed, Slc6a19 levels decreased, resulting in impaired plasma isoleucine metabolism. Based on statistical analysis of the association between plasma isoleucine levels and cardiac surgery-associated acute kidney injury, researchers proposed plasma isoleucine as a biomarker of acute kidney injury and suggested Slc6a19 might be further investigated as a therapeutic target [62].

Regarding tissue remodeling and fibrosis, using scRNA-seq in ischemia-reperfusion-induced acute kidney injury models, researchers found that WT1⁺ (Wilms’ tumor 1) proximal tubular epithelial cells could derive from WT1⁺ glomerular parietal epithelial cells. Further experiments indicated that WT1 was necessary for tubular regeneration and that WT1⁺ parietal epithelial cells demonstrated stem cell-like characteristics. These findings suggested that WT1 might be involved in renal repair after injury [63]. In distal tubular cells, upregulated neuropilin-1 (Nrp1) was observed. By activating Smad3 (SMAD family member 3), Nrp1-positive distal tubular cells exacerbated kidney injury by secreting collagen and interacting with myofibroblasts. Further experiments showed that Nrp1 knockout could reduce kidney injury and delay the development of fibrosis [64]. Besides, some pro-fibrotic tubular epithelial cells might promote inflammation and fibrosis by recruiting inflammatory monocytes through various ligand-receptor pairs, suggesting that intercellular communication between renal tubular epithelial cells and immune cells might play a significant role in acute kidney injury. These ligand-receptor interactions, however, are computationally inferred and await experimental validation [52].

From a cell injury perspective, using a mouse model of ischemia-reperfusion-induced acute kidney injury, a comprehensive single-cell transcriptomic profiling study identified several injury markers and novel genes, including AHNAK nucleoprotein (Ahnak), SH3 domain-binding glutamate-rich protein like 3 (Sh3bgrl3), and collagen type XVIII alpha 1 chain (Col18a1), as molecules of interest for further therapeutic investigation [65]. In maleic acid-induced acute kidney injury, researchers discovered that aldehyde dehydrogenase 2 (ALDH2) lactylation might facilitate the degradation of prohibitin 2 (PHB2) and impede mitophagy, thereby exacerbating tubular cell injury. Targeting sirtuin 3 (SIRT3), which reduced ALDH2 lactylation levels, might reverse this process both in vitro and in vivo, indicating that ALDH2 might be a therapeutic target [66]. In drug-induced acute kidney injury, activating transcription factor 4 (ATF4) was found to be upregulated in tubular epithelial cells. ATF4 facilitated the phosphorylation of signal transducer and activator of transcription 1 (STAT1) and its association with guanylate binding protein 2 (GBP2), thereby promoting pyroptosis and exacerbating injury. Inhibition of ATF4 could block the STAT1-GBP2 pathway, thereby reducing inflammation and injury. Of note, ATF4 expression was found to be positively correlated with renal dysfunction in both mouse models and humans, suggesting that ATF4 could be explored as a predictive biomarker [67]. Specifically, in high-dose folic acid-induced acute kidney injury models, Lee et al. identified the cell lineages responsible for proximal tubular damage. Their investigation focused on leucine-rich repeats and immunoglobulin-like domains protein 1 (Lrig1)-expressing cells, which were known for their beneficial roles in epithelial repair. The study characterized the specific transcriptional signatures of these Lrig1⁺ cells and traced their progeny, ultimately confirming their capacity to maintain tubular integrity and repair proximal tubular injury [68]. Besides, in endothelial cells, elevated levels of vascular endothelial protein tyrosine phosphatase (VE-PTP) could suppress the angiopoietin-Tie2 pathway and promote kidney injury. Genetic reduction of VE-PTP provided additional renoprotection. However, VE-PTP mutations might be linked to angiosarcoma, and its inhibitor exhibited adverse impacts while treating ocular diseases; therefore, further studies were required before its broader application [69].

Acute kidney injury to chronic kidney disease

Given the lack of effective therapies to promote kidney repair, a more comprehensive understanding of the mechanisms driving the transition from acute kidney injury to chronic kidney disease is essential [70].

Using scRNA-seq analysis of renal tissues from ischemia-reperfusion mouse models, researchers obtained numerous findings. At the immune level, researchers discovered that as acute kidney injury progressed to chronic kidney disease, T cells accumulated and could be classified into nine clusters. Among them, CD8⁺ T cells dramatically increased through self-proliferation and recruitment mediated by the macrophage-derived CXCL16-CXCR6 pathway. It was also discovered that CD8⁺ T cells played a major role in peritubular capillary rarefaction. The removal of CD8⁺ T cells significantly reduced proximal tubular cell loss and renal fibrosis, suggesting that they could be explored as therapeutic targets [71]. The integration of scRNA-seq and ST further illuminated the critical roles of macrophages. In mouse ischemia-reperfusion models, a “cycling M2” macrophage subset was identified and characterized by enhanced proliferative capacity and profibrotic activity regulated by thrombospondin-1 (THBS1) signaling. THBS1 was a key regulator that orchestrated interactions between macrophages and fibroblasts; targeted inhibition of THBS1 markedly reduced the cycling M2 macrophages, thereby mitigating fibrotic progression [72]. Besides, a new macrophage subset called extracellular matrix-remodeling-associated macrophages was discovered. This subset was derived from monocytes, adopted a pro-inflammatory macrophage phenotype, and interacted with fibroblasts. As a result, this novel macrophage subset was thought to be a prospective therapeutic target for preventing the progression of chronic kidney disease. Spatiotemporal profiling of macrophages showed that tubules co-localized with macrophages during the early stage after acute kidney injury, whereas during late chronic phases, fibroblasts were spatially proximal to injured tubules [73].

From a fibrotic perspective, using scRNA-seq in the same ischemia-reperfusion model, Tiwari et al. uncovered that ischemia simultaneously inactivated prolyl hydroxylase domains 1, 2, and 3 (PHD1, PHD2, and PHD3), resulting in enhanced activation of hypoxia-inducible factor (HIF) and exacerbation of capillary dropout, inflammation, and fibrosis. Endothelial hypoxia and glycolysis were associated with overexpression of SLC16A3, which encoded the lactate exporter monocarboxylate transporter 4 (MCT4). Moreover, the MCT4 inhibitor syrosingopine might promote kidney repair by reducing monocyte-endothelial interactions and suppressing pro-inflammatory endothelial cell activity, thereby offering a potential strategy to prevent the progression of acute kidney injury to chronic kidney disease [74]. Complementing these findings, combining scRNA-seq with ST, Wang et al. tracked the dynamic expression of Mincle (macrophage-inducible C-type lectin), a transmembrane pattern recognition receptor predominantly expressed in innate immune cells such as neutrophils, dendritic cells, and monocytes/macrophages. They revealed that sustained Mincle-high neutrophils and macrophages contributed to unresolved inflammation and fibrosis through the increased production of tumor necrosis factor (TNF) [75]. Regarding cell injury, De Chiara et al. reported that although Yes1-associated protein 1 (YAP1)-driven polyploidization of tubular epithelial cells played a rapid compensatory role to improve renal function and prevent early fatality, it concurrently drove tubular epithelial cell senescence and progressive interstitial fibrosis. Blockage of YAP1-driven polyploidization might prevent progression to chronic kidney disease. Therefore, YAP1 could be a novel therapeutic target for slowing the transition from acute kidney injury to chronic kidney disease [76].

Chronic kidney disease and renal fibrosis

Chronic kidney disease poses a significant threat to global health. A hallmark of chronic kidney disease is interstitial fibrosis, which is characterized by the accumulation of pathological fibrillar extracellular matrix in the interstitial space surrounding renal tubules and peritubular capillaries. Excessive production of extracellular matrix by myofibroblasts leads to fibrotic remodeling of the renal parenchyma, thereby impairing tubular function and ultimately causing decreased kidney volume and perfusion [77].

scRNA-seq is instrumental in delineating the immune landscape of renal fibrosis. Several studies identified key roles for basophils in human kidney biopsies and revealed distinct monocyte subpopulations in blood, notably a marked increase in HLA-DR^hi (major histocompatibility complex, class II, DR) intermediate monocytes, which exhibited unique chemokine receptor and adhesion molecule expression patterns [78, 79]. Furthermore, scRNA-seq analysis of human peripheral blood mononuclear cells suggested that an imbalance between T-helper-17 cells and regulatory T cells significantly contributed to kidney disease progression, with macrophage inhibitory factor (MIF) signaling mediating interactions between T-helper-17 and regulatory T cells and facilitating the trans-differentiation of regulatory T cells into T-helper-17 cells [80]. From a fibrotic perspective, the integration of scRNA-seq and ST further elucidated key cellular contributors to fibrosis. Specifically, fibroblasts and pericytes were identified as the primary sources of scar-forming myofibroblasts [81]. Complementing this, Abedini et al. found that the fibrotic stroma contained various cell types, including fibroblasts, myofibroblasts, immune cells, and endothelial cells, highlighting intricate cellular interactions within this microenvironment [82]. Most of the inferred intercellular interactions described above were based on transcriptomic data and await experimental validation.

In mouse models, from an immune perspective, researchers found that Zc3h12c (zinc finger CCCH-type containing 12 C) suppressed macrophage activation toward a pro-inflammatory phenotype, modulated macrophage survival, migration, and phagocytosis, and further exacerbated kidney injury, as demonstrated using a novel Tnfrsf11a^Cre-Zc3h12c^flox/flox mouse model, most likely by altering the alternative splicing of STAT1 pre-mRNA [83]. Furthermore, using scRNA-seq and ST to investigate the impact of TGF-β signaling on proximal tubular mitochondrial homeostasis and tubulo-interstitial interactions, Kayhan et al. examined TGF-β type II receptor (TβRII) expression in proximal tubular cells from mice exposed to aristolochic acid. Their results indicated that intact TGF-β signaling in proximal tubular cells supported an adaptive response to chronic kidney disease, while TβRII deletion exacerbated mitochondrial dysfunction and a T-helper-1 cell immune response [84].

Metabolically, various experimental models provided critical insights into the cellular and metabolic adaptations during the development of chronic kidney disease and renal fibrosis, leading to the identification of potential biomarkers. Specifically, in folic acid-induced models, cell-type-specific investigations revealed that fatty acid oxidation and oxidative phosphorylation significantly impacted the development of proximal tubular cells, with estrogen-related receptor alpha (ESRRA) highlighted for its protective role by regulating proximal tubular cell-specific genes [85]. Using scRNA-seq combined with ST to explore proximal tubular epithelial cells, researchers suggested that nicotinamide adenine dinucleotide (NAD⁺) could stimulate the PGC-1α/PPARα (Peroxisome proliferator-activated receptor-α)/FAO (Fatty acid β-oxidation) axis. This axis appeared to restore impaired metabolism and further protect the kidney in chronic kidney disease caused by Alport syndrome. Nicotinamide riboside, a precursor of NAD⁺, was shown to improve cellular metabolism and stop the progression of chronic kidney disease [86].

In the context of fibrosis, a study using reticulon 3 (RTN3)-mutant mice showed that loss of RTN3 caused mitochondrial dysfunction, increased reactive oxygen species levels, enhanced endothelial-to-mesenchymal transition, and impaired cell-cell communication, all of which contributed to renal fibrosis [87]. Specifically, unilateral ureteric obstruction models are also pivotal for studying chronic kidney disease and the progression of renal fibrosis. Doke et al., for instance, identified a novel proximal tubular cell subset with a profibrotic and pro-inflammatory phenotype that secreted CXCL1, which recruited CXCR2⁺ basophils and appeared to play a critical role in kidney fibrosis, suggesting an interaction between proximal tubular cells and basophils as a key contributor to renal fibrosis [78]. Subsequent studies showed that connexin 43 (Cx43) regulated ATP release from tubular epithelial cells through associated channels, thereby activating fibroblasts and inducing pyroptosis in neighboring macrophages, suggesting potential crosstalk between tubular epithelial cells and macrophages [88]. Fibroblasts significantly influenced the development of fibrosis. Transient receptor potential canonical 6 (TRPC6) was an important molecule in fibroblasts and might be involved in interactions between fibroblasts and endothelial cells. Renal fibrosis was attenuated when the TRPC6 inhibitor SH045 activated a Prnp (prion protein PrP) transcription factor regulatory network. Based on these preclinical findings, targeting TRPC6 warrants further investigation as a potential therapeutic strategy [89]. Myofibroblasts, which were crucial contributors to fibrosis, had diverse origins, with the majority originating from resident mesenchymal cells such as pericytes and fibroblasts [90]. Following that, Chen et al. investigated the roles of WW domain-containing E3 ubiquitin protein ligase 2 (WWP2) in myofibroblasts and reported that WWP2 modulated metabolic reprogramming and profibrotic responses, partly through peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC-1α), thus facilitating renal fibrosis. WWP2 deficiency was shown to decrease the proliferation of myofibroblasts and inhibit profibrotic activation, which lessened the severity of renal fibrosis [91]. Additionally, kidney-Gli1 (glioma-associated oncogene homolog-1) cells were shown to faithfully replicate myofibroblast activation in vitro, which might facilitate basic research [92]. Src, a non-receptor tyrosine kinase, was a major regulator of the TGF-β1-mediated macrophage-myofibroblast transition, suggesting that Src could be explored as a therapeutic target for renal fibrosis [93]. Besides, long noncoding RNA growth arrest-specific 5 (lncRNA Gas5) expressed in fibroblasts had a positive correlation with the severity of kidney fibrosis, which also regulated the immune microenvironment. Depletion of lncRNA Gas5 altered the expression of inflammatory mediators and activated the peroxisome proliferator-activated receptor signaling pathway. Targeting lncRNA Gas5 might therefore have therapeutic potential, though this is based on preclinical models [94]. Additionally, using scRNA-seq along with ST, researchers established a proinflammatory, profibrotic, and tenascin C (TNC)-enriched microenvironment that promoted renal inflammation and fibrosis and enabled macrophage activation. This study also predicted interactions between macrophages and fibroblasts, suggesting that investigating the fibrotic niche might be important and that TNC might have translational potential pending further validation [95].

In addition, regarding cell injury, comprehensive datasets revealed numerous genes associated with endo-lysosomal function and proximal tubular cell activity. DAB adaptor protein 2 (DAB2), a connector protein in the TGF-β (transforming growth factor β) pathway, emerged as a key regulatory node. Reduction of DAB2 expression in mice conferred protection against chronic kidney disease, suggesting that DAB2 might be further investigated as a therapeutic target [96].

Limitations

Currently, most biomarker research is still exploratory. A few molecules have been validated in clinically relevant studies, but the limitations are clearly evident. Small sample sizes, limited patient numbers, and a lack of independent replication mean that none of these findings have been translated into clinical practice.

Additional technical limitations of scRNA-seq and ST exist. Both methods have strict requirements for sample quality, and sometimes obtaining sufficient high-quality kidney biopsies from patients can be challenging. This leads to sampling bias and practical difficulties. Batch effects are another concern as they may compromise reproducibility. Additionally, low capture efficiency means loss of information and higher costs. Furthermore, RNA levels inside cells are usually low, so amplification is generally needed before sequencing, and this step can introduce amplification bias. Although methods like SMART-Seq have been developed to reduce this, it remains a concern. Because of these factors, most sequencing results are not entirely stable, reproducibility is difficult to ensure, and findings from different studies are difficult to compare. This makes cross-study integration challenging and undermines the strength of sequencing data as evidence. Therefore, there are still many technical hurdles to overcome before these two emerging technologies can be broadly used in clinical practice.

Though animal models are widely used in kidney disease research, the findings cannot be directly applied to humans. This is especially the case for sequencing technologies like scRNA-seq and ST, which are based on transcriptome changes, aspects in which differences between humans and animals are more likely to manifest. Therefore, results from animal models need to be taken with caution and backed up by rigorous human studies. Besides, while scRNA-seq and ST provide new insights into how different cell types are arranged spatially and how they communicate, such information currently relies heavily on computational predictions. There is no robust, comprehensive experimental validation yet. Future work will need multidimensional experiments to confirm these predicted cell-cell interactions.

Finally, this review has its own limitations. We only included papers from PubMed published in the last ten years, and we only covered common kidney diseases, not renal tumors or kidney transplantation. We also focused mainly on finding potential biomarkers without exploring the pathways or underlying mechanisms. Future studies will need to explore these areas to provide a more comprehensive understanding.

Conclusion

scRNA-seq, along with the integration of scRNA-seq and ST, provides a more comprehensive and unbiased approach for tissue- and single-cell-level analysis. Recent advancements in these techniques have primarily addressed three key areas: defining spatial cell locations, cellular states, and novel cell types; elucidating cell-cell interactions; and discovering new potential biomarkers for disease diagnosis and treatment. These emerging tools have substantially advanced our understanding of nephrology.

However, ST, which indicates the precise definition of spatial cell localization, has not yet been widely adopted, potentially due to the relative novelty of this technology. Moreover, the snapshot nature of fixed-time point analysis presents challenges in capturing dynamic cellular processes. Furthermore, the underlying mechanisms of cell-cell communication remain incompletely understood. Lastly, the majority of existing studies still focus on the potential utility of the discovered biomarkers, while lacking further discussion of the clinical significance. Furthermore, using scRNA-seq and ST together provides richer information, but it also introduces extra technical bias and higher costs, making practical adoption difficult. Another issue is that the cell-cell communications described are mostly computational predictions rather than experimentally confirmed. Looking forward, several priorities should be considered for the field. One is to improve the sequencing methods to reduce bias and enhance reproducibility. Another is to validate the potential biomarkers through dedicated experiments and clinical studies, with greater attention to their clinical relevance. Better integration of the two technologies and broader application across different research domains are also needed. Multi‑omics approaches in particular hold considerable promise for improving the diagnosis and treatment of kidney diseases and could help advance nephrology research more broadly. In terms of timing, most biomarkers discussed here are still exploratory. Achieving clinical applicability will likely take years, potentially a decade or longer. Independent replication and large‑scale human studies are still needed. Currently, these molecules should be viewed as hypotheses requiring further investigation, rather than being ready for clinical application.

Abbreviations

scRNA-seq

Single-cell RNA sequencing

ST

Spatial transcriptomics

JCHAIN

Joining chain of multimeric IgA and IgM

SPARC

Secreted protein acidic and cysteine

ROCK2

Rho-associated coiled-coil containing protein kinase 2

TXNIP

Oxidative stress and inflammatory response

CCL2

C-C motif chemokine ligand 2

CCR2

C-C motif chemokine receptor 2

PTGDS

Prostaglandin D2 synthase

TNFSF

Tumor necrosis factor superfamily

TNFRSF

Tumor necrosis factor superfamily receptor

A2M

Alpha-2-macroglobulin

MMP7

Matrix metalloprotease 7

SOCS3

Suppressor of cytokine signaling 3

KLF6

KLF transcription factor 6

BMP2

Bone morphogenetic protein 2

pSMAD1

Phosphorylated SMAD1

COL4

Collagen type IV

APRIL

A proliferation-inducing ligand from B cell

ANCA

Anti-neutrophil cytoplasmic antibody

TRAIL

TNF-related apoptosis-inducing ligand

DR5

Death receptor 5

PANoptosis

Pyroptosis, apoptosis, and necroptosis

RIPK3

Receptor-interacting protein kinase 3

SPP1

Secreted phosphoprotein 1

SEMA3C

Semaphorin 3 C

IFN

Type I interferon

CXCL10

C-X-C motif chemokine ligand 10

CXCR3

C-X-C motif chemokine receptor 3

APOE

Apoprotein E

TNFSF13B

TNF superfamily member 13b

PIK3α

Phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha

TIGIT

T cell immunoreceptor with Ig and immunoreceptor tyrosine-based inhibitory motif domains

Ahnak

AHNAK nucleoprotein

Sh3bgrl3

SH3 domain binding glutamate-rich protein like 3

Col18a1

Collagen type XVIII alpha 1 chain

Slc5a2

Solute carrier family 5 member 2

WT1

Wilms’ tumor 1

Nrp1

Neuropilin-1

Smad3

SMAD family member 3

VE-PTP

Vascular endothelial protein tyrosine phosphatase

ST2

Suppression of tumorigenicity 2

AREG

Amphiregulin

S100A8

S100 calcium-binding protein A8

ALDH2

Aldehyde dehydrogenase 2

PHB2

Prohibitin 2

SIRT3

Sirtuin 3

ATF4

Activating transcription factor 4

STAT1

Signal transducer and activator of transcription 1

GBP2

Guanylate-binding protein 2

Lrg1

Leucine-rich α-2-glycoprotein 1

IL-6

Interleukin 6

NAMPT

Nicotinamide phosphoribosyl transferase

Fcer1g

Fc Epsilon Receptor Ig

PHD1

Prolyl hydroxylase domains 1

HIF

Hypoxia-inducible factor

MCT4

Monocarboxylate transporter 4

YAP1

Yes1 associated protein

THBS1

Thrombospondin-1

Mincle

Macrophage-inducible C-type lectin

TNF

Tumor necrosis factor

HLA-DR

Major histocompatibility complex, class II, DR

MIF

Macrophage inhibitory factor

DAB2

DAB adaptor protein 2

TGF-β

Transforming growth factor β

RTN3

Reticulon 3

Cx43

Connexin 43

TRPC6

Transient receptor potential canonical 6

Prnp

Prion protein PrP

WWP2

WW domain-containing E3 ubiquitin protein ligase 2

PGC-1α

Peroxisome proliferator-activated receptor gamma coactivator 1-alpha

Gli1

Glioma-associated oncogene homolog-1

lncRNA Gas5

Long noncoding RNA growth arrest-specific 5

ESRRA

Estrogen-related receptor alpha

Zc3h12c

Zinc finger CCCH-type containing 12 C

TβRII

TGF-β type II receptor

ACE2

Angiotensin-converting enzyme 2

NRBF2

Nuclear receptor binding factor 2

NAD+

Nicotinamide adenine dinucleotide

PPAR-α

Peroxisome proliferator-activated receptor-α

FAO

Fatty acid β-oxidation

TNC

Tenascin C

Author contributions

Y.G. wrote the original manuscript. D.Y. and Y.Z. helped with figure preparation and designing. J.L. and H.Z. supervised and revised the manuscript. All authors read and approved the final copy of the manuscript.

Funding

This work was supported by Chinese Nature Science Foundation 82170740(H.Z.), 82100743 (J.L.), Science and Technology Plan Joint Program (Key R&D Program Project) of Liaoning Province 2025JH2/101800408 (H.Z), Science and Technology Plan Joint Program (General Project of Natural Science Foundation) of Liaoning Province 2024-MSLH-584 (J.L.), Outstanding Scientific Fund of Shengjing Hospital 202206 (H.Z.).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Declaration of AI use

During the writing of this manuscript, the authors used DeepSeek exclusively for language polishing. No AI was used to generate scientific content, interpret data, or draw conclusions. The authors take full responsibility for the final content.