Approach to Studies on Podocyte Lesions Mediated by Hyperglycemia: A Systematic Review

Abstract

Podocyte injury is a central event in the pathogenesis of diabetic nephropathy (DN). We conducted a systematic review across four major databases, identifying 7769 records and including 130 studies that met predefined eligibility criteria. Methodological quality was assessed with Joanna Briggs Institute tools, yielding a mean score of 81.3%, indicating overall moderate-to-high rigor despite design-contingent limitations. Publication activity was sparse until 2018 but increased markedly thereafter, with more than 80% of studies published between 2019 and 2025. Temporal analyses confirmed a strong positive trend (p = 0.86, p < 0.0001), reflecting the rapid expansion of this field. Study designs evolved from early human-only descriptions to integrated multi-model approaches combining human tissue, animal experiments, and in vitro systems, thus balancing clinical relevance with mechanistic exploration. Geographically, Asia emerged as the leading contributor, complemented by increasing multinational collaborations. Mechanistic synthesis highlighted five reproducible pillars of podocyte injury: slit-diaphragm and adhesion failure, mTOR-autophagy-ER stress disequilibrium, mitochondrial and lipid-driven oxidative injury, immune, complement, and inflammasome activation, and epigenetic and transcriptomic reprogramming. Collectively, these findings underscore a convergent mechanistic cascade driving podocyte dysfunction, while also providing a framework for therapeutic interventions aimed at restoring barrier integrity, metabolic balance, and immune regulation in DN.

Keywords: diabetic neuropathy; physiopathology; podocytes.

Figure 1

Flowchart illustrating the selection process of eligible studies. Generated using the PRISMA2020 tool [145].

Figure 2

Annual output of studies on podocyte injury in diabetes mellitus (2002–2025). Y-axis: number of studies per year; X-axis: publication year. Black circles mark yearly counts, linked by a blue line for visual continuity. The solid black line is the linear regression fit (Y = 0.8060 × Year − 1617; slope 0.8060, 95% CI 0.4861–1.126; R² = 0.6089; F_1,18 = 28.02; p < 0.0001). Dashed parallel lines depict the 95% confidence bands around the fitted line. A strong monotonic trend is also supported by Spearman’s p = 0.87 (95% CI, 0.6814–0.9479; p < 0.0001). Counts for 2025 reflect indexing through July only.

Figure 3

Study models and their temporal distribution (2002–2025). (A) Horizontal bar chart summarizing the proportion and count of study models across all eligible records (N = 130). Bars are shown in green; labels at bar ends denote counts. Categories (n, %): Human (37, 28.5%), In vitro + Animal (28, 21.5%), Animal (27, 20.8%), Human + Animal (10, 7.7%), Human + In vitro + Animal (10, 7.7%), Human + In vitro (8, 6.2%), In vitro (7, 5.4%), In silico (2, 1.5%), Unclassified (1, 0.8%). (B) Stacked column chart showing, for each publication year, the number of studies stratified by model. The X-axis is the year (2002–2025) and the Y-axis is the number of studies. Colors encode the study models as indicated in the legend (Human, Animal, Human + Animal, Human + In vitro, Human + In vitro + Animal, In silico, In vitro, In vitro + Animal, Unclassified). The 2025 bar reflects records indexed through July only.

Figure 4

Global distribution of study designs across continents and time. (A) Heatmap—Continent × study model. Matrix of absolute frequencies with rows (continents) and columns (model types); darker shades indicate a stronger association (i.e., higher frequency) between continent and model. Clustering is disabled and values are unscaled. (B) Alluvial—Model → Continent → Year. Sankey-style flows in which band width is proportional to the number of studies linking model type to continent and publication year; strata are labeled, and color encodes model category. Abbreviations: A = Animal; H = Human; H+A = Human + Animal; H+IV = Human + In vitro; IV = In vitro; IV+A = In vitro + Animal; IS = In silico; UnC = Unclassified; AF = Africa; AS = Asia; EU = Europe; NA = North America; SA = South America; OC = Oceania.

Figure 5

Mechanistic vocabulary of podocyte injury in diabetes. Word cloud summarizing the top 100 single-word terms extracted from the mechanistic sections of all eligible studies. Font size is proportional to term frequency; color is purely esthetic (no quantitative meaning) and word placement is random. Singular/plural or closely related variants appear as separate tokens when present in the source text. The emergent lexicon maps onto the five mechanistic domains summarized in Table 2, slit-diaphragm/adhesion failure, mTOR–autophagy/ER stress, mitochondrial–lipid injury, immune/complement/inflammasome activation, and epigenetic–transcriptomic regulation—providing a compact qualitative overview of the corpus.

Figure 6

Integrated architecture of diabetic podocytopathy. (A) Systems map (drivers → hubs → targets → outcomes). Directed graph summarizing hypothesized causal flow from upstream drivers (light blue: chronic hyperglycemia, hemodynamic overload, dyslipidemia/lipotoxicity, AGEs/metabolic memory, microvascular stress, sterile/viral inflammation) to core signaling hubs (light green: mTORC1, ER stress/UPR, autophagy/TFEB, TXNIP, EGFR → p70S6K/RPS6, ERK/VEGF, TGF-β/EMT/Notch, ROCK, HDAC4 → calcineurin, complement C3a/C3aR, NLRP3/pyroptosis/TRAIL-DR5, epigenetic/RNA programs) and then to organellar/structural targets (light amber: mitochondria—FAO ↓/fission ↑; lysosome/autophagic flux; ER; actin/α-actinin-4 cytoskeleton; slit diaphragm, nephrin/podocin; adhesion—Cx43 and α3β1-integrin), culminating in tissue-level outcomes (light rose: foot-process effacement, podocyte loss, proteinuria, mesangial expansion, GBM thickening/sclerosis). Arrow direction encodes putative influence; edges are drawn only where supported in the corpus (see Table 2). (B) Five-pillar interaction network. Circular network collapsing panel A into five interacting modules,) slit-diaphragm/adhesion, mTOR–autophagy/ER stress, mitochondrial lipid injury, immune/complement/inflammasome, epigenetic transcriptomic control. Edge thickness is proportional to the curated crosstalk weight (relative evidence), highlighting dense bidirectional coupling among pillars. (C) Cell-level schematic under hyperglycemia. Representative podocyte blueprint depicting hyperglycemia-induced programs (e.g., TXNIP; Wnt/β-catenin and miR-27a–PPARγ–FOXO1 axes; complement C3a/C3aR and IL-1β/IL-1R1 signaling) converging on slit-diaphragm depletion/mistrafficking (nephrin/podocin), adhesion defects (Cx43, α3β1-integrin), cytoskeletal instability, impaired autophagy/lysosome, and mitochondrial dysfunction (FAO ↓, fission ↑), which together yield effacement, detachment, podocyte loss and proteinuria. Abbreviations: AGE, advanced glycation end-product; AMPK, AMP-activated protein kinase; APA, alternative polyadenylation; C3aR, complement C3a receptor; DR5, death receptor 5; EMT, epithelial–mesenchymal transition; ER, endoplasmic reticulum; FAO, fatty-acid oxidation; GBM, glomerular basement membrane; HDAC, histone deacetylase; IL-1R1, interleukin-1 receptor type 1; lncRNA, long non-coding RNA; m6A, N6-methyladenosine; mTORC1, mechanistic target of rapamycin complex 1; NLRP3, NLR family pyrin domain containing 3; TFEB, transcription factor EB; TXNIP, thioredoxin-interacting protein; VEGF, vascular endothelial growth factor.

Figure 6

Integrated architecture of diabetic podocytopathy. (A) Systems map (drivers → hubs → targets → outcomes). Directed graph summarizing hypothesized causal flow from upstream drivers (light blue: chronic hyperglycemia, hemodynamic overload, dyslipidemia/lipotoxicity, AGEs/metabolic memory, microvascular stress, sterile/viral inflammation) to core signaling hubs (light green: mTORC1, ER stress/UPR, autophagy/TFEB, TXNIP, EGFR → p70S6K/RPS6, ERK/VEGF, TGF-β/EMT/Notch, ROCK, HDAC4 → calcineurin, complement C3a/C3aR, NLRP3/pyroptosis/TRAIL-DR5, epigenetic/RNA programs) and then to organellar/structural targets (light amber: mitochondria—FAO ↓/fission ↑; lysosome/autophagic flux; ER; actin/α-actinin-4 cytoskeleton; slit diaphragm, nephrin/podocin; adhesion—Cx43 and α3β1-integrin), culminating in tissue-level outcomes (light rose: foot-process effacement, podocyte loss, proteinuria, mesangial expansion, GBM thickening/sclerosis). Arrow direction encodes putative influence; edges are drawn only where supported in the corpus (see Table 2). (B) Five-pillar interaction network. Circular network collapsing panel A into five interacting modules,) slit-diaphragm/adhesion, mTOR–autophagy/ER stress, mitochondrial lipid injury, immune/complement/inflammasome, epigenetic transcriptomic control. Edge thickness is proportional to the curated crosstalk weight (relative evidence), highlighting dense bidirectional coupling among pillars. (C) Cell-level schematic under hyperglycemia. Representative podocyte blueprint depicting hyperglycemia-induced programs (e.g., TXNIP; Wnt/β-catenin and miR-27a–PPARγ–FOXO1 axes; complement C3a/C3aR and IL-1β/IL-1R1 signaling) converging on slit-diaphragm depletion/mistrafficking (nephrin/podocin), adhesion defects (Cx43, α3β1-integrin), cytoskeletal instability, impaired autophagy/lysosome, and mitochondrial dysfunction (FAO ↓, fission ↑), which together yield effacement, detachment, podocyte loss and proteinuria. Abbreviations: AGE, advanced glycation end-product; AMPK, AMP-activated protein kinase; APA, alternative polyadenylation; C3aR, complement C3a receptor; DR5, death receptor 5; EMT, epithelial–mesenchymal transition; ER, endoplasmic reticulum; FAO, fatty-acid oxidation; GBM, glomerular basement membrane; HDAC, histone deacetylase; IL-1R1, interleukin-1 receptor type 1; lncRNA, long non-coding RNA; m6A, N6-methyladenosine; mTORC1, mechanistic target of rapamycin complex 1; NLRP3, NLR family pyrin domain containing 3; TFEB, transcription factor EB; TXNIP, thioredoxin-interacting protein; VEGF, vascular endothelial growth factor.