Immune cell composition in normal human kidneys

An understanding of immunological mechanisms in kidney diseases has advanced using mouse kidneys. However, the profiling of immune cell subsets in human kidneys remains undetermined, particularly compared with mouse kidneys. Normal human kidneys were obtained from radically nephrectomised patients with urogenital malignancy (n = 15). Subsequently, human kidney immune cell subsets were analysed using multicolor flow cytometry and compared with subsets from C57BL/6 or BALB/c mice under specific pathogen-free conditions. Twenty kidney sections from healthy kidney donors or subjects without specific renal lesions were additionally analysed by immunohistochemistry. In human kidneys, 47% ± 12% (maximum 63%) of immune cells were CD3+ T cells. Kidney CD4+ and CD8+ T cells comprised 44% and 56% of total T cells. Of these, 47% ± 15% of T cells displayed an effector memory phenotype (CCR7− CD45RA− CD69−), and 48% ± 19% were kidney-resident cells (CCR7− CD45RA− CD69+). However, the proportions of human CD14+ and CD16+ myeloid cells were approximately 10% of total immune cells. A predominance of CD3+ T cells and a low proportion of CD14+ or CD68+ myeloid cells were also identified in healthy human kidney sections. In mouse kidneys, kidney-resident macrophages (CD11blow F4/80high) were the most predominant subset (up to 50%) but the proportion of CD3+ T cells was less than 20%. These results will be of use in studies in which mouse results are translated into human cases under homeostatic conditions or with disease.


Scientific RepoRtS
| (2020) 10:15678 | https://doi.org/10.1038/s41598-020-72821-x www.nature.com/scientificreports/ to myeloid cells is similar between human and mouse kidneys, the former of which were obtained from patients who underwent radical nephrectomy. The results showed that human kidneys predominantly harboured T cells, especially effector memory and resident memory cell subsets. These results will be of use in studies in which mouse results are translated into human cases under homeostatic conditions or with disease.

Results
Immune cell subsets in human kidneys. Fifteen kidney tissues from radical nephrectomy cases were analysed by flow cytometry. The mean age of patients was 69 ± 10 years. Ten patients (66.7%) were male. The mean values of serum creatinine and the estimated glomerular filtration rate calculated by the Chronic Kidney Disease Epidemiology Collaboration equation 16 were 1.0 ± 0.3 mg/dl and 73.6 ± 15.2 ml/min/1.73 m 2 , respectively. Thirteen patients (86.7%) had estimated glomerular filtration rates greater than 60 ml/min/1.73 m 2 . Three patients (20%) had proteinuria by dipstick test, and all the levels were 1+. Other baseline characteristics are shown in Supplementary Table S1. Histologic findings are shown in Supplementary Table S2. Flow cytometric analysis of immune cell subsets in kidney tissues from patients who underwent radical nephrectomy was performed. Figure 1a shows the representative gating strategy for T cell, B cell, and natural killer cell subsets. The proportion of CD3 + T cells was 47.4% ± 11.6%, of which approximately 44% and 56% were CD4 + and CD8 + T cells, respectively. The mean ratio of CD4 + per CD8 + T cells was 0.9 ± 0.4. This low CD4/CD8 ratio was the reverse of values observed in human peripheral blood 17 .
The gating strategy for myeloid cells including monocytes/macrophages, classical dendritic cells (cDCs), and neutrophils is shown in Fig. 2a. The proportion of the CD14 + monocyte/macrophage subset was 10.2% ± 4.7%.   . 2b). There were almost no CD64 − HLA-DR + cells among CD14 + cells. The proportions of cDCs and neutrophils were 1.1% ± 0.6% and 11.5% ± 5.8%, respectively. Collectively, the most abundant immune cell subset in human kidneys was CD3 + T cells. This trend did not differ between male and female subjects or was not dependent on kidney dysfunction (see Supplementary Fig. S1).
immunostaining analysis of human kidney sections. Pre-analytic procedures such as digestion might affect the above flow cytometric results. For sensitivity analysis, kidney sections from healthy donors (i.e., zero-time biopsy) and subjects without specific renal lesions (each n = 10) were evaluated. CD3 + , CD68 + , and CD14 + cells in the interstitial area were counted after excluding cells within vessels, tubules, and glomeruli. Figure 3a is a representative image of sections from healthy donors. Compared with frequently observed CD3 + cells, CD68 + or CD14 + cells were rarely seen. When stained cells were counted, the number of CD3 + cells was higher than those of CD68 + and CD14 + cells (Fig. 3c). This trend remained consistent in subjects without specific renal lesions (Fig. 3b,d). These results supported the flow cytometric results where CD3 + T cells were dominant in human kidneys compared with CD14 + monocytes and macrophages. When these interstitial immune cells were stained in normal kidney tissues from nephrectomised patients, the residency of CD3 + T cells was higher than that of CD68 + or CD14 + cells (see Supplementary Fig. S2).
Immune cell subsets in mouse kidneys. Because human samples were obtained from older patients, 1-year-old mice were analysed in addition to 8-week-old mice. Harvested kidneys from 8-week-and 1-year-old C57BL/6 mice were evaluated after perfusion with 10 ml of cold PBS via the left chamber of the heart. Figure 4a,b show the representative gating strategy for T cells and their proportions. Total CD3 + T cells accounted for 11.4% ± 2.1% and 14.2% ± 1.3% in 8-week-and 1-year-old mice, respectively. Among the CD3 + T cells in 8-week-old mice, CD4 + and CD8 + T cells accounted for 60% and 40%, respectively. For 1-year-old mice, CD4 + and CD8 + T cells accounted for 55% and 45%, respectively. The most abundant T cell subset was CD62L − CD44 − cells in 8-week-old mice, whereas CD62L − CD44 + cells were most abundant in 1-year-old mice (Fig. 4b). The T RM subset was less than 0.5% of CD45 + immune cells. Regarding other T cell subsets, Treg, γδT, and natural killer T (NKT) cells were less than 1% of CD45 + immune cells, except in 1-year-old mice where 5% of cells were NKT cells (Fig. 4c). The proportions of NK and B cells were 7.8% ± 1.1% and 21.4% ± 0.7%, respectively, in 8-week-old mice, and 6.5% ± 0.7% and 17.9% ± 5.2% in 1-year-old mice (Fig. 4d).
The gating strategy and proportions of macrophages, monocytes, DCs, and neutrophils are shown in Fig. 5a,b. The proportions of macrophages were 45.4% ± 4.0% and 39.2% ± 4.0% in 8-week-and 1-year-old mice, respectively. When macrophages were categorized into kidney-resident (rMac: CD11b low F4/80 high ) and kidney-infiltrating (iMac: CD11b high F4/80 low ) subsets 20,21 , more than 90% were rMac (Fig. 5b). Other myeloid subsets such as neutrophils and dendritic cells were less than 1% of CD45 + cells. When 8-week-old BALB/c mice were examined, macrophages were the most abundant subset at 52%, of which more than 90% were rMac (see Supplementary  Fig. S3). Collectively, macrophages were the most abundant immune cell subset in the mouse kidney.
When immune subsets were analysed in mice that had not been perfused with PBS, the proportions of macrophages and monocytes were 30.9% ± 4.5% and 12.6% ± 1.1%, respectively in 8-week-old mice, and 26.9% ± 3.1% and 11.8% ± 1.2%, respectively in 1-year-old mice. The proportions of CD3 + T cells were 16% and 20% in 8-weekand 1-year-old mice, respectively. www.nature.com/scientificreports/ In summary, the proportion of T cells was higher than that of monocytes/macrophages in human kidneys, in contrast to mouse kidneys (Fig. 6a). Similarly, the ratio of CD3 + T cells to monocytes/macrophages was higher in human kidneys than in mouse kidneys (Fig. 6b). Representative immunofluorescence images of human and mouse kidneys show that human kidneys harboured more CD3 + T cells than CD14 + monocytes/macrophages in contrast to mouse kidneys, which harboured more GFP + monocytes/macrophages (Fig. 6c,d, and Supplementary Videos S1 and S2).

Discussion
Mouse studies are invaluable for investigating immune-mediated homeostasis and disease, but their translation to human conditions requires further validation because mice, particularly those bred under specific conditions (e.g., specific pathogen-free conditions), might have a different immune and inflammatory milieu compared with humans. The present study investigated whether human kidneys harboured different immune cell subsets from mouse kidneys. Of note, the ratio of CD3 + T cells to monocytes/macrophages was higher in human kidneys than in mouse kidneys.
Kidneys have a distinct immune system, and thus, subsets of resident immune cells are different from those in other organs 22,23 . Kidneys contain specialized rMac that monitor and scavenge the endothelial transport of immune complexes, and subsequently trigger type III hypersensitivity responses 20 . These phenomena may be crucial in diseases attributable to the insufficient clearance of immune complexes 24 . The proangiogenic activity of kidney-resident macrophages during renal artery stenosis was also reported 25 . This frontline and potential homeostatic activity of rMac may be related to their high proportion in mouse kidneys. However, the residency and proportion of rMac have not been fully evaluated in human kidneys. Recent single-cell RNA sequencing www.nature.com/scientificreports/ data suggested that markers such as CD74 and CD81 might characterize human kidney-rMac 26 . However, these markers are also expressed on other myeloid subsets of that dataset or on non-myeloid cells 27 . Accordingly, the present study used conventional myeloid markers such as CD14, CD64, HLA-DR, and CD68 to identify macrophages and monocytes, not rMac alone. Nevertheless, fewer numbers of macrophages and/or monocytes were present in human kidneys compared with mouse kidneys, and this might result in a high frequency of other immune subsets such as T cells in human kidneys. The present exploratory study did not determine the mechanisms related to the high abundance of T cells within human kidneys. Decreased numbers of other immune cell subsets might lead to an increase in T cells. In addition to this simple hypothesis, we focused on the relatively high proportion of T RM . Non-lymphoid tissues such as kidneys harbour certain populations of memory T cells that are not present in the circulation (referred to as T RM ), which are different from circulating T cells 28 . The cytokine milieu after inflammatory insults, such as infections or ischemic damage, drives the development of T RM , and thus, mouse kidneys in the present study had few T RM in the absence of kidney insults. In contrast to mouse conditions, our study subjects might have experienced subclinical injury although most of their kidney functions were within the normal range. Furthermore, continuous contact with uremic toxins, spontaneously-dead parenchymal cells or debris from tubules and glomeruli over the years might affect the development of the T RM subset 29 . The different splenic immune cell profiles between wild and laboratory mice supports the hypothesis that the environmental milieu alters the renal immune cell composition 30 . Intriguingly, CD49a + cells were the predominant CD4 + and CD8 + T RM subsets, but CD103 was not the primary marker of kidney T RM . This feature is different from the phenotype and distribution  18 . The characteristics of renal epithelial cells and their adhesion molecules may determine the markers of T RM , which will be addressed in another project.

The high proportion of T cells in human kidneys may have implications in clinical situations. The roles of recipient-originated T cells have been much studied, but it is unknown whether effector T cells or T RM cells from kidney donors have a role in the rejection process after kidney transplantation.
Although their specificity against recipient antigens may be undetermined or irrelevant, they have the potential to be activated by bystander effects 31 . Long duration of cold ischemia, surgical stress, and other systemic inflammatory cascades after donor death are sufficient to activate donor-resident T cells, which may induce delayed graft function 32 . A previous study reported that donor T cells were associated with the transplant outcome of lungs 33 , and therefore the presence or abundance of donor T cells may be related to kidney transplant outcome. A previous study identified the presence of donor-derived CD8 + T RM cells was associated with graft failure in the first month after transplantation 34 . Additionally, an abundance of T cells may affect the surveillance of kidney cancers because this cancer type is typically immunogenic and T cell-dependent 35 . Possibly because of these characteristics, immunotherapy has emerged as a first-line agent for patients with renal cell carcinoma 36 .
The proportion of renal B cells was much lower than other immune cell subsets such as T cells. These results did not exclude the clinical importance of renal B cells, because many B cells can infiltrate into tissues during disease or inflammation 37 . The intrarenal B-cell infiltrates are associated with outcomes of lupus nephritis, anti-neutrophil cytoplasmic autoantibody vasculitis, and kidney transplants [38][39][40] . Furthermore, most immune reactions consisting of B cells are evoked within germinal centers in lymphoid organs 41 ; however, the present study did not explore the renal lymph nodes. Future studies on inflammation in human kidneys or human renal lymph nodes will clarify the role of tissue-localized B cells.
The popularity of mice has increased in research to understand immune reactions in homeostatic and inflammatory kidneys. Despite the worth of mouse studies, results from these investigations should be taken with caution before translating results to human conditions in which T cells are more abundant. Future studies will determine the reasons for the high abundance of renal T cells and their clinical implications under homeostatic and disease conditions.

Methods
Animals. C57BL/6 and BALB/c wild type male mice, and LysM Cre and ROSA mTmG male mice were purchased from the Jackson Laboratory (ME, USA). The ROSA mTmG mouse is a reporter mouse expressing membrane tdTomato in all cells except those induced to express Cre recombinase, which changes reporter expression to membrane GFP. Accordingly, LysM Cre -ROSA mTmG mice display GFP fluorescence in lysozyme + cells, such as macrophages, and tdTomato fluorescence in other cells. Mice were housed under specific pathogen-free conditions at the Seoul National University College of Medicine. All experiments were approved by the Seoul National University Institutional Animal Care and Use Committee (no. SNU-150611- [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19] and in accordance with the guidelines. Human kidneys. Human kidney tissues were obtained from patients who underwent radical nephrectomy because of urogenital malignancy (renal cell carcinoma [n = 9], urothelial carcinoma [n = 5], and metastasis of signet ring cell carcinoma [n = 1]), when there was no evidence of hydronephrosis or infectious disease. Normal tissues, including the cortex and medulla, were located at the opposite pole of the tumours, and non-intrusion www.nature.com/scientificreports/ for 10 min or Fc receptor-binding inhibitor antibody for 20 min (Thermo Fisher Scientific, CA, USA), and then incubated with primary antibodies. Alternatively, following surface staining, cells were incubated with fixationpermeabilization buffer, washed with permeabilization buffer (Fixation/Permeabilization Solution Kit; BD Biosciences, CA, USA), and then incubated with antibodies against intracellular antigens. In cell population with low numbers (e.g., plasma cells), isotype controls were used for gating (see Supplementary Fig. S4). Samples were processed by a BD LSRFortessa (BD Biosciences) and analysed with FlowJo software (FlowJo, LLC, OR, USA). The antibodies used for flow cytometry are listed in Supplementary Table S3. immunohistochemistry. Biopsied kidney tissues (4-μm thick sections) from healthy kidney donors and subjects without specific renal lesions were stained with anti-CD3, anti-CD68, and anti-CD14 antibodies, and counterstained with haematoxylin. Immunohistochemistry was conducted using a Ventana Benchmark XT automated staining system (Ventana Medical Systems, Tucson, AZ, USA). Positivity was quantified using pathology slide-viewing software (Aperio ImageScope; Leica Biosystems, Wetzlar, Germany). Twenty random fields were evaluated at 20 × magnification.
Immunofluorescence staining. Mouse  The staining of 3D images was performed by an electrotransport staining method using a C-stain device (Crayon Technologies). As a reflective index matching procedure, tissues were incubated in refractive index matching solution (88% [w/v] iohexol, 0.1% Tween-20 in PBS) for 0.5-2 h. Images were acquired using a Nikon C2si microscope (Nikon, Japan) with a Plan-Apochromat 10 × lens (numerical aperture = 0.5, working distance = 5.5 mm) with a 3 × zoom. Tissue was immersed in the refractive index matching solution during image acquisition. Three-dimensional rendering was performed using Imaris software (version 6.0; Bitplane AG, Switzerland).

Statistical analysis.
All analyses and calculations were performed using GraphPad Prism (version 7.0; GraphPad Software, Inc., CA, USA). The results are expressed as the mean ± standard deviation (or standard error of the mean in the Figures). Differences between groups were evaluated using the Student's t-test. P < 0.05 was considered statistically significant.

Data availability
The datasets used or analysed during the current study can be obtained on reasonable request with the permission from the corresponding author. www.nature.com/scientificreports/ Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.