Single-cell RNA sequencing in diabetic kidney disease: a literature review

Abstract

Diabetic kidney disease (DKD) is the leading cause of end-stage renal disease (ESRD), and its pathogenesis has not been clarified. Current research suggests that DKD involves multiple cell types and extra-renal factors, and it is particularly important to clarify the pathogenesis and identify new therapeutic targets. Single-cell RNA sequencing (scRNA-seq) technology is high-throughput sequencing of the transcriptomes of individual cells at the single-cell level, which is an effective technology for exploring the development of diseases by comparing genetic information, reflecting the differences in genetic information between cells, and identifying different cell subpopulations. Accumulating evidence supports the role of scRNA-seq in revealing the pathogenesis of diabetes and strengthening our understanding of the molecular mechanisms of DKD. We reviewed the scRNA-seq data this time. Then, we analyzed and discussed the applications of scRNA-seq technology in DKD research, including annotation of cell types, identification of novel cell types (or subtypes), identification of intercellular communication, analysis of cell differentiation trajectories, gene expression detection, and analysis of gene regulatory networks, and lastly, we explored the future perspectives of scRNA-seq technology in DKD research.

Keywords:

• scRNA-seq
• diabetic kidney disease
• gene expression
• cell communication
• trajectory analysis

1. Introduction

Diabetic Kidney Disease (DKD) carries a high risk of death, it affects about 30–40% of people with diabetes mellitus in the United States and significantly increases mortality rates [1,2]. Although great progress has been made in the development of clinical therapies for DKD [3,4], its progression cannot be controlled. An in-depth understanding of the pathogenesis of DKD is of great significance in the prevention of the disease as well as in the search for new treatments. In the last few years, several studies have reported that The pathobiology of DKD involves many different types of cells in the kidney, and cellular interactions and their dysfunction [5–10], such as renal tubular hypertrophy, podocyte dysfunction, mesangial expansion and mesangiolysis, endothelial cells, etc. However, the current diagnosis, treatment, and experimental research on renal diseases heavily rely on morphological cell identification, molecular mechanisms, cellular communication, and known biomarkers [11–15]. To fully understand the complex mechanism of kidney organization, function, and disease, it is necessary to gain a deeper understanding of the diverse cell types within and beyond the heterogeneous kidney milieu, signal regulation, and cellular communication.

With interest and technological advances, numerous reports on the systems biology of human DKD have been published [16–19], including changes in gene expression [19–21], epigenomics, proteomics and metabolomics analyses [16,22,23], it provides a very meaningful reference value for our subsequent studies, but has the limitation that a few cell type signals may be missing. However, single-cell RNA sequencing (scRNA-seq) is capable of capturing complete cell type signals, enabling the creation of comprehensive cell atlases that reveal the identities and clarify the characteristics and status of various cell types in health and disease so that new scientific and technological support can be provided for the study of kidney health and disease [24,25]. Here, we briefly introduce existing scRNA-seq technologies, outline their recent developments, and discuss relevant applications of scRNA-seq in DKD as well as future perspectives.

2. Developments and breakthroughs in scRNA-seq

In 2009, Tang et al. used scRNA-seq for the first time to analyze individual mouse blastomeres [26]. They then combined this modified single-cell cDNA amplification method with Applied Biosystems’ next-generation sequencing technology, the SOLiD^™ system, to establish a single-cell whole transcriptome assay [27]. Subsequently, many single-cell sequencing technologies have been developed and validated and are now commonly utilized for mRNA and protein analyses [28,29]. Initially, 10–100 cells were included in the scRNA-seq study [30,31]. Through continuous improvement and the adoption of new technologies, individual programs can perform transcriptional analysis of tens of thousands of single cells [32,33].

With the recent advancements in DNA barcoding and combinatorial indexing strategies, it is expected that millions of cells or nuclei can be sequenced at one time, enabling ultra-high-throughput scRNA-seq sequencing of samples in a variety of disease and tissue contexts [34]. At the same time, the sensitivity, accuracy, and efficiency of scRNA-seq are improving because of ongoing technological developments in single-cell isolation methods as well as scRNA-seq protocols, the technology of scRNA-seq for mapping the whole transcriptome expression of individual cells has made great development, Its applications include explore cell heterogeneity [35,36], cell type identification [24,36,37], analyze cell–cell communication [36,38], gene-regulatory network identification [36], allelic expression patterns [36]. In addition, scRNA-seq explores some of the changes that occur during biological development [39], depicting cell differentiation trajectories in pseudo-time [40]. In particular, advances in technology and computational analysis methods have facilitated the development of scRNA-seq techniques, including valve-based, nanopore, enabling highly sensitive, accurate, and high-throughput transcriptomic analysis of individual cells [41]. The integration of single-cell transcriptomic data has important implications. It helps to address medical challenges that cannot be solved by RNA sequencing alone and supports precision medicine for diseases. The scRNA-seq technology has brought new breakthroughs in the biological and medical fields, and its wide application has greatly facilitated the research progress in related fields.

3. Applications of scRNA-seq on DKD

3.1. scRNA-seq with DKD kidney cells and applications

3.1.1. Clustering to annotate cell types

Over the past decades, numerous outstanding studies have been conducted to investigate mammalian kidney development and nephrogenesis, as well as to characterize the adult kidney [42,43]. However, the remarkable developmental complexity has posed a significant challenge for scientists across various fields in determining the exact number of renal cell types. Such as inter-individual heterogeneity, including the timing of kidney development, number of nephrons, cell numbers per segment, and anatomical differences, had a major impact on determining the number of renal cell types [44]. However, there is no consensus on the number and range of renal cell types in humans or rodents.

ScRNA-seq technology is capable of labeling and identifying cell types through clustering, and understanding the development of these different cell types in the kidney is critical to understanding kidney homeostasis, disease, and regeneration. Mature mammalian kidneys reportedly include 18–26 renal cell types [45–47]. These cells mainly consist of epithelial, endothelial, interstitial, and immune cells. While the source of these numbers is challenging to trace, possible reasons include the unclear definition of the distinct cell type, incomplete knowledge of cell-specific markers, technical limitations, a certain degree of variability in healthy subjects, and the difficulty of defining cell identities and distinguishing cell types [44]. Additionally, the types of kidney cells are different in different disease models. For instance, Wu et al. [48] validated snRNA-seq on fibrotic kidneys from mice after surgery for unilateral ureteral obstruction, they identified two novel populations of proximal renal tubules, with a strong value-added gene signature and a strong cellular motility transcriptional signature, respectively, and expressed a number of damage markers, including Havcr1 and Vcam1.

The application of scRNA-seq in DKD remains in its infancy, but recent correlative studies have provided many insights into the expression of genes that characterize normal as well as mutant renal cells. Table 1 summarizes the diverse numbers of clusters/cell types identified in the various scRNA and single-nucleus RNA sequencing studies performed on DKD to date. In addition, the number of endothelial cells, immune cells, interstitial cell types may be even larger. However, there was a significant variation in the results across experiments, especially, the results are quite different between Chung et al. [49] and Fu et al. [50], which both used isolated glomerular cells and both differed from the results of Karaiskos et al. [51]. This may be related to the differences in animal types, isolation methods, cell numbers, etc., and the subtypes of these cells in DKD were also identified.

Table 1. Donor information and detected cell clusters from scRNA-seq studies of DKD.

Publication	Donor and tissue source	Number of cells	Number of clusters	Distribution of clusters (%)
Wilson et al. [15]	Diabetic patients and controls; kidney cortex	Total of 68,458 cells	20	Most abundant cell type was proximal convoluted tubule
Wilson et al. [67]	Diabetic patients and nondiabetic controls; kidney cortex	Total of 23,980 cells (13,860 control, 10,120 diabetic)	15	A total of 347 leukocytes (control and diabetic) were identified and consisted of 49% T cells, 21% B cells, 23% monocytes, and 7% plasma cells. Diabetics had an approximate 7- to 8-fold increase in leukocytes versus controls
Wu et al. [53]	db/db mice; whole kidneys	Total of 58,259 cells	13	77.6% proximal tubule cells, 9.4% immune cells (Mac, DC, B and T cells), 13.0% other cells
Fu et al. [54]	Diabetic OVE26 mice and WT; whole kidneys (CD45-positive myeloid cells)	Total of ∼17,000 CD45+ cells (∼10,500 WT and ∼6,500 OVE26)	11	Small increase in resident macrophage and infiltrating macrophage populations in OVE26 kidneys versus WT
Wu et al. [25]	db/db mice; whole kidneys	Total of 946,660 cells	18	Including the rare cell types such as macular densa cells and renin + cells from juxtaglomerular apparatus
Wu et al. [52]	Type 2 diabetic db/db mice; whole kidneys	Total of 83,585 cells	10	94% tubular cell types, 0.6% lymphoid cells types, (5%)myeloid cell types, (0.4%) endothelial cell type
Lu et al. [68]	Diabetic patients and control (from GSE131882 dataset)	–	10	–
Chung et al. [49]	BTBR ob/ob mice (type 2 diabetes model) and control; isolated glomerular cells	0.5 × 10^6 cells/mouse	7	Mesangial cells (41.5% in ob/ob mice, 11.9% in control mice), podocytes (4.4% in ob/ob mice, 6.5% in control mice)
Fu et al. [50]	Streptozotocin-induced diabetic eNOS–deficient (eNOS−/−) mice and control eNOS−/− mice; isolated glomerular cells	644 cells (326 control, 318 diabetic)	5	Glomerular endothelial cells (65% diabetic, 49% control), mesangial cells (12% diabetic, 32% control), podocytes (5% diabetic, 15% control), immune cells (15% diabetic, 1.5% control), tubular cells (2% diabetic, 2% control)

DC: dendritic cell; DKD: diabetic kidney disease; eNOS: endothelial nitric oxide synthase; Mac: macrophage; scRNA-seq: single-cell RNA sequencing; WT: wild type.

In 2022, Wu et al. [52] reported performing scRNA-seq analyses in the kidneys of db/db mice, an animal model for type 2 diabetes and diabetic kidney disease, and identified five tubular cell subtypes, two lymphoid cell subtypes, and two myeloid cell subtypes. In the same year, Wu et al. [53] divided epithelial tubular cells into three subpopulations in the kidneys of db/db mice using scRNA-seq analyses, which is similar to a previous experiment [45]. Wu et al. [25] identified six endothelial subtypes, seven distinct subtypes of thick ascending limb cells, and three fibroblast subtypes. Additionally, to examine the kidney macrophage subtypes, Fu et al. [54] identified 11 subclusters of mononuclear phagocytes in the kidneys of diabetic mice (four dendritic cell and seven macrophage populations) and showed an increase in the proportions of interferon IFN^hi, Trem2^hi, and Mrc1^hi Mac subpopulations in diabetic kidneys.

3.1.2. Novel cell type (or subtype) identification

ScRNA-seq technology can help us identify new cell types, reveal disease mechanisms, and explore cellular communication in the human kidney from multiple perspectives [45,55,56]. For instance, scRNA-seq identified PT-3 cells (a subtype of proximal tubular cells) and their marker genes SLC36A2 and PEPD, but these cells are susceptible to infection by the sudden acute respiratory syndrome coronavirus 2 [57]. Researcher identified two novel PT subtypes in a mouse model of UUO using seRNA-seq. The gene expression of these two novel PT cells differed significantly. One is a proliferative subtype and the other is a dedifferentiated subtype [48].

However, there has been limited research on this topic in DKD. During research on phenotypic changes in macrophages in DKD [54], a certain number of macrophage subpopulation 14 (M14) overlapping with the expression profile of the Mrc1hi Macs gene was observed, whereas classical M2 macrophage markers (e.g., Cd163, Fcna, Retnla) expression appears to be higher. From this we surmise that it may be a novel subtype of mannose receptor C-type 1-high expressing macrophages (Mrc1^hi Macs). In addition, the proportion of the Mrc1hi Macs subpopulation was significantly elevated in diabetic kidneys, and both pro- and anti-inflammatory pathways of macrophages may be simultaneously regulated in early DKD. Wilson et al. [15] identified a new proximal tubular cell cluster (PT_PROM1 cluster) by mononuclear RNA and ATAC sequencing, which represents a previously identified CD133 + VCAM1- cell cluster, which partly supports the possibility of the existence of multiple proximal tubular states involving cellular damage or inflammation [58]. In 2022, Wu et al. [52] revealed a new subcluster of cells called ‘proliferative PT’ through clustering analysis, and this cell population has both strong proliferation genes and aPT marker genes and has increased in diabetic kidneys, but this cluster had distinct markers from the other subclusters [48,52,59]. Proliferative PT cells are indicators of tubular cell injury and repair.

3.1.3. Cell–cell communication identification

Cell-cell interactions between renal parenchymal cells, resident immune cells and infiltrating immune cells are critical in the development of DKD [60–62], scRNA-seq technology has greatly enhanced our understanding of renal disease pathogenesis and renal cell-cell interactions [45,49,63,64]. Regarding the cell–cell communication of renal parenchymal cells, some scholars [50] conducted scRNA-seq analysis in diabetic and control mice and revealed that communication between endothelial cells and renal tubular cells, which are in close proximity to each other and capable of generating cross-talk, may further influence the direction of DKD progression, which is consistent with many other studies [7,65,66]. In the same year, Wilson et al. [67] explored alterations in intercellular signaling in diabetic patients using scRNA-Seq, they found that CCN1 and SLIT3 in glomerular thylakoid cells may act as communication mediators between podocytes and endothelial cells and thus promote each other’s expression, both matrix protein CCN1 and secretory protein SLIT3 can perform tissue repair or regulate cell migration through interaction with podocytes and endothelial cells.

Regarding the cell–cell communication of immune cells, Wei et al. [60] used a single nucleus transcriptome dataset to investigate human DKD and discovered that glomerular endothelial cells and podocytes play a crucial role in glomerular and tubular cell cross-talk. They also identified VEGFA and FGF1 as the key molecule involved in this process. Lu et al. [68] showed that immune cells had close interactions with other cells via receptor–ligand interactions through analysis of scRNA-seq data from human DKD specimens. Furthermore, it was found that the collecting duct intercalated cell, collecting duct principal cell, descending loop of Henle, and PTCs play important roles in the communication pathways of the kidney, in contrast, mac cells and T cells participate in MHC I and MHC II communications, and 13 renal cell types were found to highly express markers, including proximal epithelial tubular cells (Slc5a2, Slc5a12, Spp2), endothelial cells (Flt1, Ptprb, Eng), macrophages (Lyz2, Ncf2, Alox5ap), and immune cells (Ccl5, Rgs1) [53]. These cell types play specific roles in intercellular information exchange.

3.1.4. Cell differentiation trajectories analysis

Pseudo-time analytical calculations can be used to reconstruct the dynamic processes that cells undergo and to organize the cells, it has also helped reveal alterations in the differentiation state of renal parenchymal cells or immune cells in diseased kidneys [59,69,70]. In experimental studies of DKD, although the single-cell transcriptome of isolated glomerular cells from both control and diabetic mice, Fu et al. [50] conducted trajectory analyses of endothelial and mesangial cells from control and diabetic patients, revealing a distinct separation of cells between the two groups and presents a pattern of changes in gene expression along pseudotimes. Meanwhile, pseudo-temporal reconstruction analysis of macrophages from WT and OVE26 kidneys showed changes consistent with a continuum of activation and differentiation states, with a tendency for gene expression to shift toward an undifferentiated phenotype but with an increase in the M1-like inflammatory phenotype over time [54]. Lineage tracing using scRNA-seq analysis has the potential to reveal the exact mechanisms of renal cell transformation and to tailor the expression of key signaling pathways and regulatory factors at specific time points to improve cell differentiation protocols, as well as to elucidate the origin of the cells and potentially provide greater insight into kidney development.

The development of DKD drives cell type-specific metabolic alterations, injury to proximal tubular epithelial cells triggers different cell death programs and cytokine release and activates immune cells and fibroblasts, with macrophages involved in metabolic adjustments, and synergistic involvement of multiple cell types such as epithelial, endothelial, and inflammatory cells. The discovery of cellular markers has also provided new perspectives for monitoring the therapeutic response of DKD patients. In addition to the relevant markers expressed by various cell types that we have previously described, studies have reported additional biomarkers that mark tubular cell injury (e.g., KIM-1, MCP-1) and biomarkers that mark tubular cell dysfunction (e.g., α1M, UMOD) [71]. Exploring these cell types and their biomarkers facilitates a better understanding of the molecular mechanisms of DKD and offers the possibility of exploring effective cell-based therapeutic strategies.

3.2. scRNA-seq with DKD kidney gene and applications

3.2.1. Gene expression detection

Gene expression profiles as the basis for other analyses, such as identification of cell types, regulatory networks, and cell–cell communication, are important for scRNA-seq. ScRNA-seq studies provide a comprehensive dynamic gene expression profile of human kidney and other organ tissues at the single-cell level [45,72–76], and by means of gene expression detection, ScRNA-seq technology helps us to identify the biomarker genes of DKD, especially in immune cells, which brings a very meaningful potential value to our clinical diagnosis and treatment of DKD.

Previous studies have used ScRNA-seq to determine the resident immunophenotypes in renal tissues of lupus nephritis patients [56]. In addition, Wilson et al. [67] discovered that infiltrating monocytes expressed IFN gamma (IFNGR1 and IFNGR2) downstream signaling genes, such as human leukocyte antigen class II genes (HLA-DRB1, HLA-DRB5, HLA-DQA1) and tumor necrosis factor receptor superfamily member 1B (TNFRSF1B), which have been implicated as biomarkers for DKD [77,78]. Lu et al. [68] used ScRNA-seq to reveal EIF4B, RICTOR, and PRKCB, which will be marker genes for diabetic nephropathy. Hirohama et al. analyzed gene expression levels by bulk RNA and scRNA-seq, and used renal tissue MMP7 and blood MMP7 as biomarkers of renal fibrosis and decreased renal function [79], Ye et al. also determined that urinary PART1 and PLA2R1 can be used together for early diagnosis of DKD [80]. Meanwhile, Zhang et al. [81] selected 17 hub genes present in the immune cells by analysis from an ScRNA-seq dataset, among these genes, BTG2, CDH2, GADD45B, and CLDN4 has the potential to be a new target for the study of DKD, these findings may provide important insights into the etiology of DKD and potential molecular events. For gene expression detection, scRNA-seq has been used in both human specimens and animal experiments. Tsai et al. used scRNA-seq to study early renal changes in DKD mice and reported results consistent with human subjects, identifying central genes involved in pathophysiological changes in the early DKD microenvironment [82]. Li et al. also reported that heritability enrichment from human and mouse scRNA-seq datasets showed very good agreement between species [83]. It also confirms the reliability of using scRNA-seq techniques to characterize renal function in both human and animal models.

The scRNA-seq technique also helps us to detect differentially expressed genes, which can lead to the discovery of relevant pathogenic mechanisms of DKD. For example, Wilson et al. [15] studied and analyzed the transcriptional profiles of patients with advanced DKD, and 9632 differentially expressed genes were identified, of which approximately half of these differentially expressed genes were upregulated, while the remaining genes were downregulated. In the same year, Lu et al. [68] screened the top 2,000 highly variable genes with large cell-to-cell differences, such as PLA2R1, SLC26A4, PTPRQ, IL1RL1, VCAM1, SERPINE1, REN, SELE, HIST2H2AA3, and TM4SF1, from scRNA-seq data of human DKD specimens and controls. Furthermore, Wu et al. [53] first analyzed the hub genes expressed differentially in the single-cell resolution transcriptome map of the kidneys of db/db mice. In proximal tubule cells (PTCs) of S1, S2, and S3 subpopulations, the values of the difference between the ideal number of differentially expressed genes and the reduced number of differentially expressed genes were 4, 3, and 13. Recently, several studies have also pointed out the increased expression of dual specificity phosphatase 1 (Dusp1) and G0s2 (G0/G1 switch 2) are involved in the pathogenesis of DKD [81,84].

3.2.2. Gene-Regulatory network analysis

Gene regulatory network analysis can help us better understand the regulatory pathways in the pathogenesis of DKD, such as metabolic processes, stress responses, and apoptotic signaling pathways. In 2020, Zhang et al. [81] validated differentially expressed genes using the DKD patient scRNA-Seq dataset, in which pathways enriched for up-regulation of DEG include extracellular matrix (ECM) receptor interactions, focal adhesion, human papillomavirus infection, malaria, and cell adhesion molecules, down-regulated DEGs were mainly enriched in ascorbate and aldate metabolism, arginine and proline metabolism, mineral uptake and longevity regulatory pathways, and multi-species signaling pathways. Wu et al. [53] used the hub genes expressed differentially in the single-cell resolution transcriptome map of the kidneys in db/db mice to find that biological process pathways obtained from the enrichment of upregulated genes were mainly concentrated in the response to oxygen-containing compounds, steroid hormones, organonitrogen compounds, and metabolism-related pathways, such as organic substance metabolic processes, metabolic processes, and small molecule metabolic processes. The differences in the above results may be related to the different disease models, cell numbers, and analytical methods.

Furthermore, a number of studies have utilized scRNA-Seq to investigate the regulation of pathways in DKD kidney cells. First, for immune cells: Fu et al. [54] identified DEGs in mononuclear phagocytes in kidneys of diabetic OVE26 mice with type 1 diabetes at different periods of time: In the OVE26 mice of 3-month old, its regulatory pathway mainly involves protein translation, oxidative phosphorylation, and myeloid differentiation, With the progression of DKD, macrophage accumulation becomes more evident, particularly in advanced disease at 7 months, these macrophages mainly consisted of inflammatory, stress response and apoptotic signaling pathways.

Meanwhile, Lu et al. [68] utilized scRNA-seq data from the GSE131882 dataset to compare immune cell marker genes in diabetic kidney specimens to controls. They identified 83 upregulated and 56 downregulated genes that are significantly involved in various processes, including cellular protein metabolism, protein modification, multicellular organism development, protein metabolism, and regulation of RNA metabolism. Second, for renal parenchymal cells: Wilson et al. [67] identified DEGs in different cell type of diabetic patient by scRNA-Seq, and found that their cell-associated components, regulated genes and important pathways for ion homeostasis were increased with Na+/K+-ATPase subunits decreased in diabetic thick ascending limbs. Chung et al. [49] identified all cell types in glomeruli utilizing scRNA-seq, they found that the cell proliferation pathway was induced in mesangial cells, whereas the apoptotic pathway was induced in podocytes. Fu et al. [50] found that angiogenesis and migration pathways in endothelial cells of diabetic mice are deregulated, consistents with previous bulk RNA sequencing data on sorted glomerular endothelial cells [85], in mesangial cells of diabetic mice, genes involved in translation regulation, gene expression, and protein stabilization are highly enriched.

3.2.3. Allelic expression pattern analysis

Genetic regulatory mechanisms are valuable for studying the development of diabetic nephropathy. SALSA is an analytical algorithm capable of exploring allele-specific expression using direct genotyping with snRNA-seq and snATAC-seq [58]. Wilson et al. [15] analyzed the renal cortex of normal and diabetic patients by snRNA-seq and snATAC-seq, added a fixed effect of diabetes and the presence of an interaction term between target gene expression and diabetes to the SALSA base model and set a p-value to judge the predictive ability of ATAC peaks for allele-specific effects and promote the heritability of eGFR. They further used 4476 significant peak gene combinations for gene ontology enrichment, and the most enriched pathways were associated with multiple HLA genes, raising the possibility that genetic background influences renal function through allele-specific chromatin accessibility (ASCA).

4. Future outlook for DKD studies

The use of scRNA-seq technologies has facilitated our in-depth research in the areas of renal cell heterogeneity, DKD pathogenesis and cellular cross-talk, and accelerated the pace of implementing precision medicine for DKD. In particular, scRNA-seq has unique advantages in analyzing the heterogeneity of cell populations and exploring the functions of rare cell types. However, the number of cells required for rigorous analysis of complex cell populations containing rare cell types is high, and the number of DKD cells varies considerably across study groups and tissue types. Some renal cells are prone to loss or incomplete breakdown under abnormal conditions, and the introduction of confounding factors may negatively affect the results of the study and bias the results in favor of cell types that are more likely to be released [86]. In a recent study utilizing scRNA-seq with murine kidney samples, it was found that endothelial and mesangial cells were under-represented [48]. Therefore, in future studies, we aim to obtain an effective protocol that is efficient, standardized, and unbiased toward certain cell subtypes.

In the future, we still need to further grasp the dynamic gene expression heterogeneity at multiple time points. In the current study, most scRNA-seq data samples were collected at a single time point, leading to one reason for variation in the results across different research groups in DKD. In particular, because the development of DKD is a dynamic process [87,88], it is difficult to accurately reveal the cellular and molecular mechanisms of pathological changes in diseases using the results at a single time point, and there may even be deviations.

Moreover, burgeoning evidence indicates that long noncoding RNAs (lncRNAs) are essential for diverse biological processes [89,90], and recent studies have confirmed the important role of lncRNAs in DKD [91,92], but there are fewer studies related to scRNA-seq including non-coding RNAs. Due to the low abundance of lncRNAs, some lncRNA sequences are only partially assembled, and full-length sequences is not yet available in current gene annotations. Recent advances in single-cell technology and related bioinformatics tools, as well as efforts to improve cell-type-specific annotation of lncRNAs, will help to overcome the limitations of current scRNA-seq analyses for such RNAs [86,93]. In addition, the main hypothesized principles of scRNA-seq involve the regulation of cellular functions and activities, including signaling, epigenetic, transcriptional and metabolic pathways. Katzenelenbogen et al. [94] established INs-seq, an integrated technology for massive parallel recording of scRNA-seq and intracellular protein activity, as a broadly applicable technology for elucidating integrated transcriptional and intracellular maps. Therefore, in the future, the combination of proteome and metabolome with scRNA-seq will better reveal the mechanism of disease occurrence, development, and targeted intervention.

5. Conclusion

ScRNA-seq technologies have many advantages but also some limitations, such as cell isolation and individualized RNA capture, they remain the most challenging [95], technical, biological, and computational challenges [96], and are unable to reliably detect low-abundance transcripts [97]. Smaller numbers of transcripts can lead to high levels of computational noise, which can interfere with data analysis and may mask underlying biological variation [98]. Sequencing-based genealogical tracing methods are sensitive to the choice of experimental platform and may also influence conclusions [99]. In addition, certain renal cell types may not be amenable to the processing steps of the popular scRNA-seq technique, which hinders the use of this single-cell system [50]. Therefore, it is necessary to improve the sensitivity of scRNA-seq to detect low-copy transcripts in individual cells as well as to achieve higher spatial resolution to fully understand more regulatory processes. It is worth mentioning that the studies in this review have potential biases or limitations in terms of methodological design, sample size, and data analysis methods, such as inconsistencies in the cell types identified between certain studies and therefore a lack of reproducibility, and the identification of transcriptomic features that are not identical from batch to batch or from sample to sample, although this seems to be considered plausible, mainly due to the fact that most human single-cell studies have small sample sizes, constrained by the cost of the technology and the limited availability of healthy human samples.

Many studies have common features of scRNA-seq and highlight the gleaned insights and opportunities available to enhance our understanding of the molecular basis of different diseases through the study of individual cells, and although they may be subject to differences in causal gene assignments, different species, and technical and analytical variability, most of the results are still largely consistent. Numerous insights from scRNA-seq studies have also informed our understanding of the pathogenesis of DKD, as well as informing the development of new therapeutic strategies. Renal fibrosis is a key process in the pathophysiology of DKD, and previous studies using scRNA-seq have identified several hub genes directly related to renal fibrosis, including SPP1 and VCAM-1 [81]. Cai et al. also found that epigenetic regulation of SPP1 is also important in DKD and targets histone markers, which may provide a new way to protect the kidneys from the harmful effects of glucose [100]. Recently, scRNA-seq methods that can be used to fix and freeze samples have also been proposed, and combining these techniques with several other single-cell measurements and multi-omics analyses to create new insights into the complex pathophysiology of glomerular disease will greatly benefit the study of highly heterogeneous clinical samples of DKD [101].

As an emerging cellular sequencing technology, scRNA-seq has great potential for development in the field of transcriptomics, and has unique advantages in exploring human development, health and disease in depth, providing researchers with extensive technical support. We have reviewed studies demonstrating that scRNA-seq technology has provided us with many powerful aids in the in-depth study of the pathogenesis of DKD, including identifying novel cell types and subtypes to offer enormous potential for de novo discovery, revealing gene expression profiles, gene-regulatory networks, and allelic expression to help us gain insight into the signaling pathways and molecular mechanisms of DKD, deciphering cell–cell communication, cell differentiation to further explore cellular mechanisms, and potential therapeutic targeting of DKD, and the identification of new relevant biomarkers will benefit us in diagnosing and preventing progressive DKD more accurately, thus improving the efficiency of diagnosis and treatment. However, because scRNA-seq is still at the beginning stage for the study of DKD, there are still areas that need to be improved. The underlying pathogenesis of DKD is very complex and there is no unanimous conclusion yet, and we still need to invest a lot of work in elucidating the complex mechanisms of DKD.

Author contributions

YW, WT, and JY designed the work of review. WT and JC wrote the paper, JY and YW revised the manuscript. KX, XL, QH, and MH reviewed the literature available on this topic. All authors approved the paper for publication.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by Scientific and Technological Research Program of Chongqing Municipal Education Commission [Grant No. KJQN202200428].