GATA6 marks fibrosis-associated stromal cells rather than repair macrophages in kidney disease

Abstract

The mechanisms underlying kidney repair remain incompletely understood. GATA6-expressing large cavity macrophages are reported to facilitate inflammation resolution and repair in several organs following injury, but their role in kidney disease is unknown. GATA6 expression was minimal in normal mouse and human kidney. Across three experimental kidney disease models (ischemia–reperfusion injury, unilateral ureteric obstruction and nephrotoxic serum glomerulonephritis), GATA6 expression increased progressively in the renal interstitium with chronic, but not acute injury. Immunophenotyping demonstrated that GATA6 + cells mainly expressed stromal markers (RGS4, PDGFRβ, αSMA) as opposed to leukocyte markers (CD45, F4/80, MHC-II). Single-cell RNA sequencing confirmed enrichment of Gata6 expression in kidney stromal cell populations. GATA6 + cells expressed genes associated with stromal cell identity, repair, inflammation, angiogenesis and collagen biosynthesis. In human kidney biopsies from patients with hypertensive nephrosclerosis and IgA nephropathy, interstitial GATA6 expression was increased and correlated with the degree of fibrosis. As in mice, human GATA6 + cells were largely CD45 and MHC-II negative but expressed stromal markers, consistent with non-immune interstitial cells rather than macrophages. Together, these findings demonstrate that GATA6 is predominantly expressed by kidney stromal cells with limited evidence of a major role for GATA6 + large cavity macrophages in kidney injury and repair.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1038/s41598-026-61940-6.

Introduction

Acute and chronic kidney diseases cause significant morbidity, mortality, and are associated with high healthcare costs. There are many different causes of chronic kidney disease (CKD) including vascular (e.g., hypertension), autoimmune, obstructive, drug/toxins and diabetes. CKD arises not only because of persistent injury from underlying diseases, but also from maladaptive resolution and repair mechanisms that lead to sustained activation of immune cells and increased production of profibrotic factors¹. Acute kidney injury (AKI) is also a significant risk factor for CKD². AKI is the result of many different insults including microbial infection/sepsis, nephrotoxic drugs, ischemia/hypovolemia, and urinary tract obstruction³. The recovery from AKI is variable and depends largely on the type and severity of the underlying insult. The capacity for the kidney to minimize injury and enact repair from the various causes of CKD and AKI is critical to recovery and long-term patient outcomes.

While preventing kidney injury and nephron loss remains key to decreasing the burden of AKI, CKD and end stage kidney disease, therapies that enhance renal repair to preserve nephron number following injury are also needed. Unfortunately, the understanding of kidney repair mechanisms is incomplete. Kidney repair is a complex biological process that is activated during injury to maintain cellular integrity and enhance cellular regeneration to preserve kidney structure and function. Most cell types in the kidney participate in repair including tubular epithelium, endothelium and leukocytes, especially macrophages⁴. Anti-inflammatory macrophages resolve inflammation, clear debris and express pro-angiogenic and reparative genes that limit injury and the subsequent loss of renal parenchyma^5,6. Kidney interstitial and perivascular stromal cells, such as pericytes and myofibroblasts also participate in kidney repair⁵. Meflin + and α-smooth muscle actin (SMA) + stromal cells promote tubular epithelial cell proliferation and support vascular integrity by secreting factors such as platelet-derived growth factor (PDGF)-β and retinoic acid^7–9. Anti-inflammatory macrophages and stromal cells however, also contribute to renal fibrosis, and thus kidney outcomes following injury represent a balance between pro-inflammatory, repair and pro-fibrotic pathways⁵. Despite the reported phenotypes of stromal cells and macrophages in kidney repair, markers that identify purely reparative cellular populations or functions have not been identified. Thus, progress to develop therapies to specifically target or enhance kidney repair has been limited.

GATA6 is a zinc-finger transcription factor that is expressed during embryonic development and in most endoderm and mesoderm-derived tissues^10,11. Beyond development, the function of GATA6 has not been studied extensively but suggested to regulate the cell cycle and differentiation in a number of cell types particularly vascular smooth muscle, intestinal epithelium and cancer^10,12–14. More recently, studies have suggested specific anti-inflammatory and reparative roles for a population of GATA6-expressing large cavity macrophages that exist in peritoneal, pleural and pericardial spaces^15–17. In mouse models of heart, liver, peritoneal and intestinal disease, GATA6 + macrophages co-expressing markers such as F4/80 + , CD102 + and CD11b + were reported to home to sites of tissue injury where they limited inflammation, cleared apoptotic bodies, promoted revascularization, and prevented fibrosis to promote tissue repair^18–21. In the pleural cavity, GATA6 + macrophages were essential in resisting infection with the nematode L. sigmodontis²². On the other hand, GATA6 expressed in non-immune cells is suggested to drive pathology in disease models including cancer, vascular injury and arteriosclerosis^13,23,24. Little is known about GATA6, its expression and function in the kidney. Interestingly, retinoic acid, which is a metabolite of vitamin A and produced by retinal dehydrogenase-2 (RALDH2) induces GATA6 expression¹⁵. RALDH2 upregulation in tubular epithelial cells, pericytes and macrophages was associated with improved kidney function following ischemia reperfusion injury in mice⁷, indirectly suggesting that GATA6 may be a downstream effector or marker in the renal repair pathway.

Given the limited knowledge of renal repair in the kidney, this study was undertaken to investigate GATA6 expression and the potential involvement of GATA6-expressing repair macrophages in mouse and human kidney injury and disease. In a variety of immune and non-immune kidney diseases, this study shows that GATA6 is expressed primarily in interstitial stromal cells but largely absent in renal leukocytes suggesting a limited role for GATA6-expressing large cavity macrophages in kidney injury and repair.

Results

GATA6 is upregulated in chronic but not acute kidney injury

To determine GATA6 expression in the context of experimental kidney injury in mice, experiments were first performed in a model of unilateral renal ischemia reperfusion injury (IRI) in wild type and Gata6^H2B-Venus mice that express the Venus reporter under the control of the mouse Gata6 promoter. At 24 and 72 h post-IRI, cells were isolated from injured kidneys and analyzed by flow cytometry. Very few GATA6 + (Venus +) cells were detectable in the kidney (Fig. 1a, b) and similar to contralateral kidney controls and wild type mice over 72 h post-IRI (Fig. 1a, b). The small number of GATA6 + cells detected in the kidney were CD45, CD11b and F4/80 negative. In contrast, peritoneal cells isolated from Gata6^H2B-Venus mice were strongly positive with more than 75% of cells expressing the Venus reporter as well as CD11b and F4/80 (Fig. 1b). Thus, GATA6 is not highly expressed in the uninjured kidney or during acute phase of IRI. Next GATA6 expression was assessed in a more chronic disease model of AKI-to-CKD transition at 7 and 14 days following unilateral IRI that results in tubular injury, interstitial fibrosis and renal inflammation marked by the infiltration of CD45 + and CD206 + cells (Suppl. Figure 1). Unlike AKI at 24 and 72 h, GATA6 + cells and protein expression increased over 14 days following IRI (Fig. 1c-e). To determine the cellular source of GATA6 expression during AKI-to-CKD transition, Gata6 mRNA expression was determined in scRNA-seq datasets derived from control and injured kidneys in mice over 14 days post-unilateral IRI. First, to enhance the ability to detect Gata6 expression macrophages, a leukocyte enriched dataset was probed²⁵. Surprisingly, very little to no Gata6 expression was observed in several leukocyte populations including macrophages (Fig. 2a-c). Gata6 expression was concentrated mainly in the stromal cell compartment, with some enrichment in a sub-population of endothelial cells. Cells enriched in Gata6 mRNA expressed mesenchymal genes such as Pdgfrb, Col1a2, Cald1, and Gpc6 but not leukocyte genes such as Adgre, Itgam, Ptprc or the MHCII genes H2-Ab1 and Cd74 corroborating the flow cytometry findings (Fig. 2d). Gata6 expression was verified in an independent, publicly available single nucleus RNA-seq dataset of total kidney cells at 14 days post IRI²⁶ (Fig. 2e-g). In this dataset that contains a higher number of parenchymal and stromal populations, Gata6 mRNA expression was also found primarily in fibroblasts and a subpopulation of endothelial cells similar to the findings in the leukocyte-enriched scRNA-seq data. Collectively these data suggest that GATA6 is stromal cell, but not leukocyte expressed during chronic experimental kidney injury.

Fig. 1.

GATA6 expression in acute kidney injury. Wild type and Gata6^H2B-Venus reporter mice underwent unilateral renal ischemia reperfusion injury (IRI) and kidneys analyzed at days 1 (D1) and 3 (D3) post injury. (A) Flow cytometry for CD45 + and GATA6 + (Venus +) cells from total live cells, isolated from Gata6^H2B-Venus reporter mice. Uninjured wild type C57Bl/6 (WT, Venus-) and D1 contralateral kidneys were used as controls. (B) Representative flow cytometry zebra plots of GATA6 + (Venus +) CD11b + F4/80 + cells isolated injured kidneys at day 1 and day 3 post-IRI in Gata6^H2B-Venus reporter mice. Peritoneal lavage cells are used as a positive control. (C) Flow cytometry of total live GATA6 + (Venus +) cells isolated from the kidneys of Gata6^H2B-Venus reporter mice at baseline, 7 and 14 days post-IRI. (D) Quantification of GATA6 + cells in the kidneys of Gata6^H2B-Venus reporter mice at baseline, 7 and 14 days post-IRI. Each data point represents the total number of GATA6 + cells from one mouse (mean ± SD, n = 6–9). Statistical analysis was performed using one-way ANOVA followed by Bonferroni’s multiple comparisons test, **p < 0.01. (E) Immunoblotting for GATA6 expression in whole kidney lysates from C57Bl/6 mice at baseline, 7 and 14 days post-IRI. Peritoneal lavage cells (PC) were used as a positive control. Lower MW bands represent the 2 different GATA6 isoforms and higher MW band likely represents post-translational modification.

Fig. 2.

Gata6 mRNA expression in the mouse kidney. (A–D) Single cell RNA-sequencing in leukocyte-enriched cell populations isolated from injured kidneys over 14 days post-ischemia reperfusion injury (IRI) in C57Bl/6 mice²⁵. (A) Uniform Manifold Approximation and Projection (UMAP) displaying annotated cell populations in pooled samples of normal and IRI-injured mouse kidneys at days 1, 7 and 14 (n = 3 mice per condition). Basophil/mast cell (Baso/MC), Ccr7 + dendritic cell (Ccr7 + DC), distal convoluted tubule (DTC), endothelial cell (EC), innate lymphoid cell (ILC2), natural killer cell (NK cell), neutrophil (Neu), proximal convoluted tubule (PCT), proximal straight tubule (PST), podocyte (Podo), ascending and descending loop of Henle (aLOH and dLOH), conventional dendritic cell (cDC), plasmacytoid dendritic cell (pDC). (B) Gata6 expression in cell populations. Populations annotated as endothelial cells (5) and stroma (14) are demarcated. (C) Dot plot identifying the degree of Gata6 mRNA expression in kidney cells. (D) Dot plot of Gata6 mRNA expression and associated genes in stromal cells compared to macrophages. (E–G) Single nucleus RNA-sequencing in total cell populations isolated from injured kidneys at 14 days post-IRI in C57Bl/6 mice²⁶. (E) UMAP displaying annotated cell populations. Connecting tubule (CNT), cortical thick ascending loop of Henle (CTAL), distal convoluted tubule (DCT), descending—thin ascending loop of Henle (DTL-ATL), endothelial cell (EC), fibroblast (Fib), intercalated A (ICA), intercalated B (ICB), macula densa (MD), medullary thick ascending loop of Henle (MTAL), macrophage (MØ), injured proximal tubule (NewPT), principle cell (PC), parietal epithelial cell (PEC), proximal tubule segment (PTS), pericyte (Per), podocyte (Pod), urothelium (Uro). (F) Gata6 expression in cell populations. Populations annotated as fibroblasts (9), urothelium (26) and endothelial cells (8) are demarcated. (G) Dot plot identifying the degree of Gata6 mRNA expression in kidney cells.

GATA6 is expressed in interstitial stromal cells in experimental chronic kidney injury

To validate the transcriptomic findings, GATA6 protein expression was assessed using immunofluorescence microscopy in the course of chronic kidney injury in mice undergoing AKI-to-CKD transition following IRI. Renal IRI induced a significant increase in interstitial GATA6 expression over 14 days post IRI (Fig. 3a). The increase in GATA6 + cells occurred in both male and female mice, with lower numbers in female mice that are known to have a less severe phenotype in IRI-induced AKI-to-CKD model (Suppl. Figure 2)²⁷. Although IRI also triggered substantial interstitial renal inflammation, semi-quantitative analysis revealed less than 10% of all GATA6 + cells co-expressed CD45 (4.3% ± 2.4), F4/80 (7.5% ± 4.1) or MHCII (5.3% ± 1.9) (Fig. 3b) consistent with the flow cytometry and scRNA-seq data. To ascertain if GATA6 was upregulated in other experimental models of chronic kidney injury, immunofluorescence microscopy was performed in the kidneys of mice at 7 days following unilateral ureteric obstruction (UUO) that results in significant renal inflammation and fibrosis (Suppl. Figure 1, 3a). Similar to the AKI-to-CKD model, GATA6 + cells also increased in the kidney interstitium of mice at 7 days following UUO. Once again, immunofluorescence microscopy demonstrated that the majority of GATA6 + cells were CD45 and F4/80 negative (Suppl. Figure 3b). Finally, to determine if the type of chronic injury (immune vs. non-immune) influenced the expression pattern of GATA6 in the kidney, immunofluorescence microscopy was also conducted in the kidneys from mice with experimental anti-GBM glomerulonephritis induced by nephrotoxic serum (NTS)²⁸. As with UUO and AKI-to-CKD models, at 10 days following immunization with NTS, GATA6 expression increased in interstitial cells of the kidney that were largely CD45 and F4/80 negative (Suppl. Figure 4). Taken together, these data show that during immune and non-immune experimental chronic kidney injury, GATA6 expression increases primarily in interstitial cells that do not co-express leukocyte or macrophage markers.

Fig. 3.

GATA6 expression and leukocytes in AKI-to-CKD transition. (A) Immunofluorescence microscopy probing for GATA6 and leukocyte markers CD45, MHC-II, and F4/80 in C57Bl/6 mouse kidneys at 14 days post ischemia reperfusion injury (IRI). Uninjured kidneys are used as controls. Hatched box represents zoomed image in adjacent right panel. Images representative of staining performed in at least 3 mice. Scale bars = 25 µm. (B) Percentage of GATA6 + cells expressing CD45, F4/80 or MHC-II at 14 days post IRI. Each datapoint represents mean percentage GATA6 + cells/FOV per biological replicate. Mean ± SD; **** p < 0.0001; n = 6, 3 male and 3 female; unpaired t-test.

GATA6 cells are interstitial stromal cells in chronic kidney injury

The previous experiments identified GATA6 expression likely within stromal cells located in the renal interstitium. During chronic injury, numerous stromal cell types infiltrate or proliferate the kidney including pericytes, myofibroblasts, fibroblasts and vascular smooth muscle cells^5,29. To further characterize the GATA6 + cell phenotype, immunofluorescence microscopy was performed probing for stromal cell markers in control and injured kidneys 14 days following IRI (Fig. 4). GATA6 + cells uniformly expressed PDGFRβ and RGS4, general markers of mesenchymal and vascular smooth muscle cells respectively^8,30 and partially co-expressed αSMA and NG2 in a mainly peritubular pattern consistent with myofibroblasts⁵. GATA6 + expression was also observed occasionally in cells that expressed the endothelial marker CD31 within the disordered interstitium of injured kidneys (Fig. 4). Staining for renin, which has been previously reported in interstitial Norn cells to co-express with GATA6³¹, revealed only a very small number of positive cells in this model of kidney injury with few expressing GATA6 (not shown). Collectively, these observations show that during chronic kidney injury, GATA6 is a non-leukocyte/macrophage marker expressed mainly within interstitial stromal cells without appearing to be restricted to a specific stromal cell sub-population.

Fig. 4.

GATA6 expression and stromal cells in AKI-to-CKD transition. Immunofluorescence microscopy probing for GATA6 and stromal cell markers α-smooth muscle actin (αSMA), platelet-derived growth factor receptor β (PDGFRβ), nerve/glial-antigen 2 (NG2), regulator of G-protein signaling 4 (RGS4) and CD31 in C57Bl/6 mouse kidneys at 14 days post ischemia reperfusion injury (IRI). Uninjured kidneys are used as controls. Hatched box represents zoomed image in adjacent right panel. Images representative of staining performed in at least 3 mice. Scale bars = 25 µm.

GATA6 cells express genes associated with repair and fibrosis

To further interrogate the phenotype of GATA6 expressing cells, nCounter mRNA analysis was performed on cells isolated from the kidneys of Gata6^H2B-Venus reporter mice. At 14 days post-IRI, CD45 + and GATA6-Venus + cells were sorted from injured kidneys using flow cytometry and total RNA isolated. mRNA expression was analyzed using the nCounter platform and Immunology and Fibrosis gene panels. Hierarchal clustering and differential gene expression analysis revealed important phenotypic differences between GATA6 + and CD45 + cells (Fig. 5a, b). Most notably, GATA6 + cells expressed multiple genes consistent with stromal/mesenchymal cells and associated with wound healing, vascular homeostasis and resolution of inflammation including Lilrb4, Csf1, Arg1, Pdgfrb, Hif1a, Il33, Timp1 and Tgfb3. GATA6 + cells also expressed genes associated with fibrosis such as Serping1, Spp1, Ccn2 and numerous collagens including Col1a2, Col3a1, Col1A3, and Col6a3. In contrast and as expected, CD45 + cells expressed a greater number of inflammatory and leukocyte genes such as Ltb, Cxcr2, Ccl5, Cd2, Cd3d, Il2rb and MHCII genes H2-Ob and H2-DMb2 (Fig. 5b). These data are aligned with the microscopy and flow cytometry findings and support the premise that GATA6 is mostly expressed in kidney interstitial stromal cells and not leukocytes or macrophages.

Fig. 5.

Gene expression in GATA6 and CD45 cells. Male Gata6^H2B-Venus reporter mice underwent unilateral ischemia reperfusion injury (IRI) and kidneys isolated at day 14. Kidney cells were isolated and sorted into CD45 + and GATA6 + (Venus +) cell populations. Cells from 3 mice per group were pooled and mRNA expression was analyzed using the Fibrosis and Immunology gene panels on the nCounter platform. (A) Heat maps and hierarchal clustering comparing gene expression in CD45 + and GATA6 + cell populations. (B) Top 20 upregulated and downregulated genes in GATA6 cells compared to CD45 leukocytes.

GATA6 is expressed in human chronic kidney disease

To determine the pattern of GATA6 expression in humans, immunofluorescence microscopy was performed on non-diseased human kidney specimens derived from patients undergoing nephrectomy for surgically indicated purposes. As in mice, very little to no GATA6 expression was present in the normal human kidney (Fig. 6a, 7a). Next, GATA6 expression was determined in biopsy specimens from patients with IgA nephropathy and hypertensive/vascular nephrosclerosis representing examples of immune and non-immune causes of CKD respectively. Seven biopsies for each condition were randomly chosen from the Biobank for the Molecular Classification of Kidney Disease at the University of Calgary³². For hypertensive/vascular nephrosclerosis biopsies with mild, moderate or severe interstitial fibrosis and tubular atrophy were selected. For IgA nephropathy biopsies graded as T0, T1 and T2 by the Oxford classification were chosen (Suppl. Figure 5). As with the experimental kidney disease models in mice, GATA6 + cells were increased in in the biopsies from patients with both IgA nephropathy and hypertensive/vascular nephrosclerosis (Fig. 6b-d). GATA6 + cells were present mainly in tubulointerstitial areas of the kidney with less than 10% of GATA6 + cells co-expressing CD45 (6.6% ± 4.0) or MHC-II (3.7% ± 1.8) (Fig. 6b, c, e). The majority of GATA6 + cells co-expressed stromal markers PDGFRβ, RGS4 and/or αSMA and correlated directly with severity of tubulointerstitial disease and fibrosis (Fig. 7). Therefore, as with the mouse data, these observations show that in both human hypertensive/vascular nephrosclerosis and IgA nephropathy, GATA6 is upregulated primarily in interstitial stromal cells and not leukocytes or macrophages.

Fig. 6.

GATA6 expression and leukocytes in the human kidney. Immunofluorescence microscopy probing for GATA6 and leukocyte markers CD45 and MHC-II in (A) non-diseased human kidney (nephrectomy), (B) kidney biopsies from patients with non-diabetic hypertensive/vascular nephrosclerosis (CKD) and (C) IgA nephropathy. Representative images of Masson trichrome and hematoxylin & eosin staining for the corresponding biopsy/tissue sections are also shown. Images representative of staining performed in specimens from 4–7 patients/group. Scale bars = 25 µm. (D) Quantification of GATA6 + cells in normal nephrectomy, hypertensive/vascular nephrosclerosis (CKD), and IgA-nephropathy samples. Each datapoint represents the mean number of GATA6 + cells/field of view in each sample. Mean ± SD; * p < 0.04; n = 4 nephrectomy, n = 7 CKD, n = 7 IgA nephropathy; unpaired t-test. (E) Percentage of GATA6 + cells expressing CD45 or MHC-II in kidney biopsies. Each datapoint represents mean percentage GATA6 + cells/FOV per biological replicate. Mean ± SD; **** p < 0.0001; n = 14, 7 hypertensive/vascular nephrosclerosis, 7 IgA-nephropathy; unpaired t-test.

Fig. 7.

GATA6 expression and stromal cells in the human kidney. Immunofluorescence microscopy probing for GATA6 and stromal cell markers α-smooth muscle actin (αSMA), platelet-derived growth factor receptor β (PDGFRβ), and regulator of G-protein signaling 4 (RGS4) in (A) non-diseased human kidney (nephrectomy, hatched box represents zoomed image in adjacent right panel), (B) kidney biopsies from patients with non-diabetic hypertensive/vascular nephrosclerosis (CKD) and (C) IgA nephropathy. Images representative of staining performed in specimens from 4–7 patients/group. Scale bars = 25 µm. (D) Correlation between area of tubulointerstitial fibrosis (Masson trichrome) and GATA6 + cells in hypertensive/vascular nephrosclerosis (CKD) and IgA-nephropathy. Twelve samples that had a representative Masson trichrome stain were pooled and included in the analysis (Spearman’s rank correlation test).

Discussion

In this paper we characterize expression of the transcription factor GATA6 in the kidney. We found GATA6 expression in several examples of murine experimental kidney injury as well as human immune and non-immune-mediated CKD. GATA6 was primarily a marker of interstitial stromal cells that also expressed genes associated with repair and fibrosis. In contrast, little GATA6 expression was evident in renal leukocyte or macrophage populations suggesting a limited contribution of GATA6 + large cavity macrophages in kidney injury and repair.

GATA6 is a transcription factor that is expressed in numerous cell types but recently identified in a population of large cavity macrophages that mediate tissue repair^15–17. In the heart, pleura, liver and peritoneum, GATA6 + macrophages were reported to limit fibrosis and promote cellular regeneration and homeostasis in models of acute non-microbial injury^17–20,22. Given the limited knowledge, and yet important implications to renal repair, we sought to determine GATA6 expression in the kidney. Our data show that GATA6 was largely absent in renal leukocyte/macrophage populations and expressed primarily in interstitial stromal cells, increasing with the severity and chronicity of kidney injury. The lack of evidence for GATA6 + large cavity macrophages following AKI and other models of kidney injury in mice is not surprising given the fact that the kidney exists in the retroperitoneum and separated from the peritoneal cavity by a layer of parietal mesothelium as well as its own fibrous capsule. GATA6 was found to be broadly expressed in interstitial stromal cells during mouse and human kidney disease without restriction to a specific sub-population. The high rate of RGS4 and PDGFRβ co-expression, as well as partial co-labelling with αSMA and NG2 suggests that GATA6 is likely expressed in stromal cells that include pericytes, vascular smooth muscle cells, myofibroblasts and fibroblasts⁵. This premise is consistent with studies reporting GATA6 expression in kidney vascular smooth muscle cells, erythropoietin-expressing interstitial Norn cells and pericytes^23,31. Some GATA6 expression was also identified in a subpopulation of endothelial cells by scRNA-seq as well as in CD31 + cells by immunofluorescence microscopy. The significance of this observation is uncertain but may represent cells undergoing endothelial-mesenchymal transition³³ and remains consistent with the assertion that GATA6 is localized primarily to cells in the kidney that are phenotypically mesenchymal.

GATA6 was expressed in populations of kidney stromal cells that not surprisingly co-expressed genes involved in repair and fibrosis, but its direct impact on the transcriptional landscape in these cells was not evaluated. A study by Sulahian et al. that probed the GATA6 gene regulatory landscape in gastric cancer cell lines suggested that GATA6 primarily regulated genes involved in cell proliferation and survival¹³. Since GATA6 was also found to be expressed in a variety of immune and non-immune experimental and human kidney diseases, it is likely not a disease specific transcription factor but upregulated in interstitial stromal cells in the course of a general response to chronic injury that may include wound healing and fibrosis.

The primary limitation of this study is that it is descriptive in nature. Assessing the true function of GATA6 in the interstitial stromal cell population would require chromatin studies and/or conditional GATA6 knockout mice in this cell population, a genetic manipulation that may be confounded given GATA6’s role in urinary tract developmental³⁴. Furthermore, while there is limited evidence of a major contribution for GATA6 + large cavity macrophages in acute or chronic kidney disease, this does not completely rule out a role for these cells especially since they have been shown to lose their defining markers once they infiltrate into tissues¹⁹. Meticulous early time course experiments and lineage tracing studies to track GATA6 + macrophages, as reported by other groups^35,36 would be needed to determine if these potentially reparative macrophages are attracted to the kidney during injury and disease. Finally, the kidney biopsy cohort was small, limiting the generalizability of these findings across the spectrum of human CKD. In addition, the results were not validated using human scRNA-seq or snRNA-seq datasets and therefore do not definitively establish cell type-specific GATA6 expression in the human kidney. Consequently, GATA6 expression in other renal cell populations cannot be excluded.

Conclusions

Data from this study show that there is limited evidence of a major contribution for GATA6 + large cavity macrophages during experimental kidney injury and human IgA nephropathy and hypertensive/vascular nephrosclerosis. GATA6 is primarily a marker of non-immune interstitial stromal cells in these experimental and human kidney diseases. While GATA6 + interstitial stromal cells are associated with chronic kidney disease and fibrosis, the function of GATA6 in this renal cell population remains to be fully determined.

Methods

Antibodies

All antibodies used in this study are listed in Table 1.

Table 1.

Antibodies and reagents.

Target	Source	Cat/Ref# (clone)	Dilution (technique)
GATA6	CST	5851S (D61E4)	1:300 (IF and WB)
GAPDH	CST	2118S (14C10)	1:1000 (WB)
GATA6	R&D system	AF1700	1:20 (IF)
CD45	BD Pharmingen	550,539 (30-F11)	1:50 (IF)
F4/80	Abcam	Ab6640 (A3-1)	1:100 (IF)
MHC-II	Invitrogen	14–5321-82 (M5/114.15.2)	1:200 (IF)
CD206	R&D systems	AF2535-SP	1:100 (IF)
αSMA	Abcam	ab7817 (1A4)	1:200 (IF)
PDGFRβ	CST	4564S (C82A3))	1:100 (IF)
RGS4	Proteintech	14,530–1-AP	1:100 (IF)
NG2	R&D systems	MAB6689 (546,930)	1:20 (IF)
CD45 PE	Bio Legend	103,105 (30-F11)	1:100 (Flow)
CD45 PE/Cy7	Bio Legend	103,113, Cl: 30-F11	1:100 (Flow)
F4/80 PE	BD Pharmingen	565,410 (T45-234)	1:100 (Flow)
CD11b BV605	Bio Legend	101,257 (M1/70)	1:100 (Flow)
Ly6G Percep/Cy5.5	Bio Legend	127,616 (1A8)	1:100 (Flow)
CD16/32	Bio Legend	101,330 (93)	1:100 (Flow)
Zombie Aqua	Bio Legend	423,101	1:500 (Flow
CD45	Bio Legend	304,002 (HI30)	1:50 (IF)
MHC-II	Abcam	ab7856 (CR3/43)	1:200 (IF)

Mice and experimental disease models

C57Bl/6 wild type mice and Gata6^H2B-Venus reporter mice (transgenic mice that express yellow fluorescent Venus protein under the control of the GATA6 promote¹¹, kindly provided by Dr. P. Kubes, University of Calgary) on a C57Bl/6 background were generated from an in-house breeding program. All mice were housed and fed under standard conditions at the Heath Science Animal Resource Centre (HSARC), University of Calgary. Mouse studies were approved by the University of Calgary Animal Care Committee in accordance with Canadian Council on Animal Care (Protocol# AC22-0201). Research was carried out following the guidelines set forth by the Animal Care Committee at the University of Calgary.

Normal or control kidney samples were from mice with no treatment or sham manipulation. Diseased kidney samples were collected from mice undergoing experimental unilateral ischemia reperfusion injury (IRI), non-reversal unilateral urethral obstruction (UUO), and nephrotoxic serum (NTS) glomerulonephritis. Unilateral kidney IRI was performed as previously described³⁷. Mice were anesthetized with an intraperitoneal injection of 100 mg/kg ketamine and 10 mg/kg xylazine and supplemented with inhaled isoflurane. Mice were then placed on a warming plastic pad and a small left lateral abdominal skin incision was made. The left kidney was carefully isolated and a clamp applied to the renal artery and vein for 30 min. At the end of ischemia, the clamp was removed, the kidney inspected to confirm reperfusion, and the surgical site closed. Mice were then allowed to recover on a warming pad. UUO was performed as previously described³⁸. Similar to the IRI model, mice were anesthetized and the left kidney and ureter exposed. The ureter was ligated with a suture and the surgical site closed. For the NTS model, female mice were intravenously injected with heat inactivated sheep anti-glomerular basement membrane (GBM) nephrotoxic serum as previously described²⁸.

For the AKI-to-CKD transition model, kidney samples were collected at 24 h, 3 days, 7 days and 14 days after IRI. For the UUO model, kidney samples were collected at 7 days post obstruction. For the NTS model, mouse kidney samples were collected after 10 days of NTS-treatment. Both male and female mice were used equally at 8–12 weeks of age in all models except for the NTS model where only female mice were used due to the lack of phenotype in NTS-treated males²⁸.

Human kidney samples

Formalin fixed paraffin-embedded (FFPE) normal human kidney samples were collected from the non-diseased cortical margins of nephrectomy samples slated for disposal from patients undergoing medically indicated kidney surgery. The study protocol (REB15-1158) was approved by the Conjoint Health Research Ethics Board at the University of Calgary in accordance with the Tri-Council Policy Statement: Ethical Conduct for Research Involving Human Participants, International Conference on Harmonization E6, Good Clinical Practice, Alberta’s Health Information Act and relevant sections of the Canada Food and Drug Act and its Regulations. The research was carried out following the guidelines set forth by the Conjoint Health Research Ethics Board at the University of Calgary. Since all human samples were de-identified by the pathology staff, a waiver of consent was granted by the Conjoint Health Research Ethics Board at the University of Calgary. Research staff had no access to any identifying or medical information from the patients from which the nephrectomy samples originated.

FFPE sections of human IgA nephropathy or hypertensive/vascular nephrosclerosis were randomly chosen from kidney biopsies stored in the Biobank for the Molecular Classification of Kidney Disease (BMCKD)³², University of Calgary. Biopsies with diabetes, interstitial nephritis or any other secondary diagnosis were excluded. The BMCKD operates under a protocol (REB15-1026) approved by the Conjoint Health Research Ethics Board at the University of Calgary in accordance with the Tri-Council Policy Statement: Ethical Conduct for Research Involving Human Participants, International Conference on Harmonization E6, Good Clinical Practice, Alberta’s Health Information Act and relevant sections of the Canada Food and Drug Act and its Regulations. The research was carried out following the guidelines set forth by the Conjoint Health Research Ethics Board at the University of Calgary. All human samples in the BMCKD are de-identified with a waiver of consent granted by the Conjoint Health Research Ethics Board at the University of Calgary. Research staff had no access to any identifying or medical information from the human samples obtained from the BMCKD.

Histopathology

Five µm sections from mouse FFPE kidney tissue were stained using hematoxylin and eosin (H&E), periodic acid Schiff (PAS), and Masson trichrome staining using conventional protocols. Stained sections were imaged using an Aperio Scanscope ATII device (Leica). H&E, PAS and Masson trichrome histopathology data of human kidney biopsies were retrieved from the BMCKD Aperio image database.

Immunofluorescence microscopy

Five µm sections from FFPE kidney tissue were deparaffinized followed by antigen retrieval with citrate buffer (pH 6.0) or Tris/ethylenediaminetetraacetic acid (Tris/EDTA, 100 mM Tris 10 mM EDTA, pH 9.0) buffer in a steamer for 30 min at 95° C. After cooling, sections were washed in 1 × phosphate buffered saline (PBS) and then blocked with: 5% goat or donkey serum plus 0.3% Triton-X 100 for GATA6 staining; 5% donkey serum only for CD206 staining. Tissue sections were then incubated with corresponding primary antibody or isotype control overnight at 4 °C. Primary antibodies were diluted in 1% BSA with 0.3% Triton-X 100 (for GATA6) or 1% BSA plus 2.5% donkey serum (for CD206). Slides were then washed in 1 × PBS for a total of 15 min and incubated with goat or donkey secondary antibody for one hour at room temperature in a humidified and light protected chamber. After secondary antibody staining, tissue sections were washed in 1 × PBS and then incubated with 0.1% Sudan Black B for 20 min. Sections were mounted with Prolong Gold antifade reagent with 4′,6-diamidino-2-phenylindole (DAPI), and covered with coverslips. Tissue sections were imaged using an Olympus FV31S-SW confocal microscope. Images were analyzed and quantified using ImageJ (version 2.14.0) or QuPath (version 0.7.0). Cells were counted in at least 5 random fields of view (FOV) in each biologic replicate from images captured with 20X objective lens and 30 × zoom. Data points represent mean number or percentage of cells per FOV in each biologic replicate. For the mouse IF data, at least 3 biologic replicates per group were included. For the human IF data, 4 normal nephrectomies, 7 IgA nephropathy, and 7 hypertensive/vascular nephrosclerosis samples were included in the analysis. Isotype control staining was included in all immunofluorescence microscopy experiments to confirm antibody specificity (not shown).

Immunoblotting

Snap frozen tissue kidney samples from mice were used for immunoblotting. Tissue was lysed in 150ul RIPA buffer (50 mM Tris, 150 mM NaCl, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate (SDS), 1% Nonidet P-40, pH 7.4) containing proteinase and phosphatase inhibitors. Tissues were mechanically homogenized and centrifuged at 13,000 rpm at 4 °C. The supernatant was collected and total protein concentration determined using the Pierce Bicinchoninic acid (BCA) protein assay (Thermo) or Bradford protein assay (Biorad). Protein lysates were mixed with 3 × sample buffer (188 mM Tris/HCl pH 6.8, 3% SDS, 30% glycerol, 0.01% bromophenol blue, 15% β-mercaptoethanol). Lysates were heated for 5 min at 95 °C, vortexed and briefly centrifuged prior to loading on 8–10% SDS–polyacrylamide gels. Lysate loading volume was adjusted relative to the total protein concentration and/or the intensity of the loading control. Proteins were separated by electrophoresis and transferred to nitrocellulose membranes. Membranes were incubated with a blocking buffer consisting of 5% milk proteins in 1 × tris-buffered saline with 0.1% Tween-20 (TBST) for one hour at room temperature. Membranes were then incubated with primary antibody diluted in the same blocking buffer: anti-mouse GATA6 (Cell Signaling Technology), and anti-mouse GAPDH (Cell Signaling Technology). Membranes were incubated with primary antibodies overnight at 4 °C. Following washes in TBST, membranes were incubated with appropriate horseradish peroxidase (HRP)-conjugated secondary antibodies for one hour at room temperature, washed and proteins imaged using enhanced chemiluminescence (ECL) substrate and a Bio-Rad ChemiDoc device.

Flow cytometry

Male and female Gata6^H2B-Venus mice were used to isolate CD45 + and GATA6 + cells from normal, and IRI kidneys at 24 h, 3 days, 7 days, and 14 days. Whole kidneys dissected into smaller pieces were placed in 1 ml of RPMI media containing MACS dissociation enzymes A, R&D (Miltenyi) and digested at 37 °C while shaking at 650 rpm. Cell suspensions were then passed through a pre-wetted 100 μm strainer and then centrifuged at 750xg for 5 min at 5 °C. The cell pellet was resuspended in 1 ml red blood cell lysis solution (ammonium chloride) for 3 min and then the cells were washed in 1 × PBS and passed through a 40um strainer. Cells were centrifuged at 750xg for 5 min at 5 °C and the pellet was resuspended with 1 ml of fluorescent activated cell sorter (FACS) buffer (1% BSA, 2 mM EDTA in PBS). Samples for flow cytometry were aliquoted into prelabelled microcentrifuge tubes, centrifuged at 5000 rpm for 2 min and cells resuspended and incubated for 10 min at room temperature in FcγR (CD16/32) prepared in FACS buffer. Samples were then washed with 1 × PBS and stained with Zombie Aqua for 10 min at room temperature. After another round of washing and centrifugation, samples were incubated for 20 min at 4 °C with a cocktail of conjugated antibodies each diluted 1:100 with FACS buffer. Antibodies included PE/Cy7-conjugated anti-CD45, PE-conjugated anti-F4/80, BV605-conjugated anti-CD11b, Percep/Cy5.5-conjugated anti-Ly6G. Samples were washed with FACS buffer, centrifuged, fixed with 4% paraformaldehyde, washed and resuspended with FACS buffer and stored at 4 °C overnight. Samples were analyzed using Attune NxT Acoustic Focusing Cytometer, and cells were analyzed after Zombie Aqua exclusion using FlowJo software.

nCounter transcriptomic analysis

CD45 + and GATA6 + (Venus +) cells were isolated from the kidney and sorted by flow cytometry. Samples from normal (n = 3) and 14 days post-IRI (n = 3), were aliquoted and pooled, centrifuged as indicated above and resuspended with FcγR (CD16/32) for 5 min at room temperature. After washing and centrifugation, samples were incubated for 10 min at room temperature with PE-conjugated CD45 at 1:100 dilution in FACS buffer. After one round of washing, cells were resuspended with 1 ml of FACS buffer and stained with Sytox blue for 5 min just before sorting. Cells were sorted according to CD45 and Venus labelling using BD FACS Aria III Cytometer. After sorting, total cell number was determined using a hemocytometer. Using the total cell number, cell concentration was adjusted to a value of 15,000 cells/ul or 5,000 cells/ul for cell number greater than 150,000 cells or less than 150,000 cells, respectively. Cells were centrifuged at 6000 rpm for 15 min at 4 °C and the pellet was lysed with RLT buffer (Qiagen, #79,216) as per the manufacturer’s protocol. Lysed samples were vortexed for 1 min, centrifuged briefly and stored at -80 °C. RNA containing lysates were used as a direct input for the nCounter analysis.

For each sample, mRNA expression was assayed using the Immunology (NS_Immunology_Mm_C2269) and Fibrosis (NS_Mm_Fibrosis_V2.0) panels, from NanoString Technologies. nCounter preparation station and nCounter digital analyzer for target gene hybridization was carried out according to NanoString Technologies protocol. Raw data from each sample was imported into and analyzed using nSolver 4.0 software (NanoString Technologies). Quality control parameters were assessed using default background threshold and normalization code sets. Background threshold, negative control count selection and geometric mean were set based on the negative probes. Normalization was carried out using 6 positive probe sets with geometric mean and a normalization code set content with standard selection and housekeeping genes with an average count greater than 100 and a 25 difference between % coefficient variation of the housekeeping genes. Differential expressed gene analysis was executed using the normalized data set. Simple analysis such as heatmaps was carried out using histogram filters with 20 as a lower bound for gene count.

Single cell RNA sequencing analysis

GATA6 mRNA expression in the kidney was determined from an established leukocyte-enriched single cell RNA-seq (scRNA-seq) dataset from our group (www.ncbi.nlm.nih.gov/geo, accession GSE299409)²⁵. The annotated dataset was imported into Scanpy (v1.9.6), and the original cell type annotations provided by the authors were retained. Data were normalized to 10,000 counts per cell, log-transformed, and 3,000 highly variable genes were selected using replicate information as a batch variable. Principal component analysis was performed. A neighborhood graph and UMAP embedding were generated using the Harmony-corrected principal components. Gata6 expression was subsequently visualized across annotated cell populations using UMAP and dot plot analyses.

To validate Gata6 expression in mouse kidney at 14 days post IRI, the publicly available single-nucleus RNA-seq dataset from Kirita et al. was re-analyzed (www.ncbi.nlm.nih.gov/geo, accession GSE139107)²⁶. The annotated dataset analyzed as above except Harmony integration was used to correct for inter-replicate variation.

Statistical analysis

Statistical analysis was performed using GraphPad Prism 9. Data is shown as mean ± SD and compared using unpaired t-test, one-way ANOVA and Spearman’s rank correlation as appropriate.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

S.G.B conducted the majority of experiments, analyzed data and wrote the manuscript. E.E.E., A.L., G.A. and K.C. conducted experiments, acquired data and reviewed the manuscript. J.C. co-supervised S.G.B. and E.E.E., designed experiments and reviewed the manuscript. D.A.M oversaw the design and conduct of the entire project, wrote the manuscript.

Funding

This work was supported by the Canadian Institutes for Health Research and the Biobank for the Molecular Classification of Kidney Disease at the University of Calgary. E.E.E. was funded by a Snyder Institute for Chronic Diseases Beverly-Phillips Fellowship.

Data availability

The data used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Human ethics

The protocols (REB15-1158, REB15-1026) were approved by the Conjoint Health Research Ethics Board at the University of Calgary in accordance with the Tri-Council Policy Statement: Ethical Conduct for Research Involving Human Participants, International Conference on Harmonization E6, Good Clinical Practice, Alberta’s Health Information Act and relevant sections of the Canada Food and Drug Act a0nd its Regulations. The research was carried out following the guidelines set forth by the Conjoint Health Research Ethics Board at the University of Calgary. Since all human samples were de-identified, waiver of consent was granted by the Conjoint Health Research Ethics Board at the University of Calgary.

Animal ethics

Mouse studies were approved by the University of Calgary Animal Care Committee in accordance with Canadian Council on Animal Care. Research was carried out following the guidelines set forth by the Animal Care Committee at the University of Calgary (Protocol# AC22-0201).