Introduction

Acute and chronic forms of kidney injury constitute a major health concern as they are associated with an increased mortality rate1,2,3. Cells of the mononuclear phagocyte (MNP) system, including monocytes, dendritic cells (DC) and macrophages, are on one side major contributors in disease progression after kidney injury by driving inflammation but simultaneously mount tissue repair and resolution of inflammation4,5,6,7. This functional diversity is reflected in a wide range of phenotypical characteristics and has made the identification of functional subunits in this complex MNP network challenging8,9,10.

In order to characterize distinct MNP subsets in the kidney, flow cytometric approaches have utilized surface markers CD11b, F4/80, Ly6C, and/or CD11c for distinction of at least three11,12,13,14,15,16 or even up to five unique subsets17,18. Drawing conclusions about the mechanistic relevance of these MNP subsets in disease models warrants careful consideration as one cannot necessarily imply their functional uniformity. Indeed, single-cell RNA sequencing revealed multimodal expression of pro- and anti-inflammatory genes among individual MNP subsets13,19, indicating additional layers of complexity in these subsets. This level of heterogeneity extends also into mechanistical studies, as the same F4/80high MNP subset has been implicated both in progression from acute to chronic kidney injury20 but also in recovery from acute kidney injury15. Results from depletion experiments affecting whole MNP subsets via clodronate liposomes or promotor-specific diphteria toxin receptor have also been rather inconclusive so far21,22,23,24, raising the need for a more granular analysis.

Kidney MNPs are often categorized into a pro-inflammatory or wound-healing cluster in order to characterize their role after kidney injury. This functional dichotomy is for example reflected in the M1/M2 paradigm, which comprises a M1 component with pro-inflammatory cytokine and chemokine secretion and a M2 component with immune-regulatory, wound healing and fibrotic properties25,26,27. In this context, CD86 and MHCII expressing cells have been associated with histological and functional injury, while CD206 expressing cells are associated with fibrotic and reparative processes28,29,30. Such binary distinctions have been used frequently to determine the overall inflammatory state of renal MNPs but often on preselected subsets or without consideration of different MNP subsets.

In order to surmount the limitations in granularity of the above ascribed methods, we aimed to establish an easily accessible flow cytometric method that combines MNP subset distinction and surface marker-based functional distinction in order to comprehensively understand renal MNP complexity Furthermore, we aimed to employ this newly established method to characterize MNP subsets in several preclinical kidney injury models which are heavily associated with MNP infiltrates.

Results

Five renal MNP subsets are defined by distinct surface marker expression and accumulate after kidney injury

Dissection of the multi-facetted nature of MNPs has been challenging and often been restricted to either phenotypical or functional distinction via flow cytometry, which we sought to combine. For the first part of our flow cytometric method we adopted a phenotypical characterization method for renal MNPs from Kawakami et al.17 because it successfully segregates five MNP subsets with the use of only few surface markers. Following this strategy, we segregated five unique MNP subsets in murine kidneys with the surface markers F4/80, CD11b and CD11c (Fig. 1A). These markers are commonly used among others like Ly6C or CX3CR1 to differentiate MNP subsets in the kidney. In line with Kawakami et al., in naïve kidneys, kidney resident F4/80high macrophages (MNP subset 3) and CD11bhigh MNPs (subsets 1 and 2) were more abundant than DC-like CD11bmediumCD11chigh (subset 4) and CD11blowCD11cmedium (subset 5) cells (Fig. 1A,B). To our knowledge the method by Kawakami et al. has not been used in physiological models of kidney injury so far. To test how MNP subset dynamics may be influenced by kidney injury we therefore analyzed kidney MNPs isolated from Col4a3−/− mice with Alport syndrome (Fig. 1A,B). While MNP subsets 1, 2 and 3 were already detected in naïve murine kidneys in relatively large numbers, subsets 4 and 5 became clearly apparent in kidneys from Col4a3−/− mice (Fig. 1B). We confirmed the uniqueness of these five subsets by assessing the expression of other distinct surface markers on these cells (Fig. 1C,D): By nature of our gating strategy, subset 3 had the highest expression of the classical macrophage marker F4/80, which also displayed intermediate expression on parts of subset 2. Expression of the inflammatory monocyte marker Ly6C was restricted to subset 2. Subsets 1, 2 and 3 had also notable expression of the chemokine receptor CX3CR1. CD103 is an integrin that can be found on conventional type 1 DCs (cDC1) and was restricted to subset 4. Fluorescence minus one (FMO) controls are available in Supplementary Figure S1. These data demonstrate that in line with the strategy by Kawakami et al. we were able to distinguish five distinct renal MNP subsets with unique surface marker expression. Moreover, all five MNP subsets were dynamically increased in diseased kidneys from Col4a3−/− mice.

Figure 1
figure 1

Markers F4/80, CD11b, and CD11c distinguish five distinct MNP subsets in the murine kidney. (A) A gating strategy for five renal MNP subsets was adopted from Kawakami et al. and representative FACS plots are shown of a naïve mouse and 7 weeks old Col4a3−/− mice with Alport syndrome as an example for diseased state. (B) Quantification of cell numbers in the five MNP subsets for naïve (n = 8) and Col4a3−/− mice (n = 11). Shown are pooled data from two independent experiments. y = ln(y) transformed cell numbers were compared by unpaired t-test *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. (C) Representative histograms for expression of surface markers F4/80, Ly6C, CX3CR1 and C103 in all five MNP subsets. (D) Quantification of geometric mean fluorescence intensity (MFI) of surface markers. MFI were compared by Kruskal–Wallis with post-hoc Dunn’s multiple comparisons test *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Full size image

CD206, CD86 and MHCII expression differentiate functionally distinct subsets

To further separate the five renal MNP subsets into functional subunits we sought to utilize surface markers that have been associated with functional distinctiveness before. In order to identify suitable markers for this, we referred to markers previously used to characterize in vitro polarized M1- and M2-like cells, as they represent two functionally distinct entitiesFlow cytometry

For flow cytometry, single-cell suspensions of the kidney were made by digesting kidney tissue with the Multi Tissue Dissciation Kit 1 (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the user manual. Following filtration through a 70 µM mesh, erythrocytes were lysed using BD Pharm Lyse (BD Biosciences) and washed with autoMACS Running Buffer (Miltenyi Biotec). Fc receptors were blocked with CD16/CD32 (BD Biosciences) for 10 min and extracellular surface markers were stained with an antibody cocktail containing FITC anti-mouse CD86 (clone GL-1, BioLegend, San Diego, CA), BV51 anti-mouse CX3CR1 (clone SA011F11, BioLegend), PE anti-mouse F4/80 (clone T45-2342, BD Biosciences, FranklinLakes, NJ), PerCP-Cy5.5 anti-mouse CD11b (clone M1/70, BD), PE-Cy7 anti-mouse CD11c (clone HL3, BD), APC anti-mouse Ly-6G (clone 1A8, BD), APC-Cy7 anti-mouse CD45 (clone 30-F11, BD), V500 anti-mouse I-A/I-E (clone M5/114.15.2, BD), FITC anti-mouse Ly-6C (clone AL-21, BD), FITC anti-mouse CD103 (clone M290, BD), FITC Rat IgG2a, κ isotype control (clone R35-95, BD), PE-Cy7 Hamster IgG1, λ1 isotype control (clone G235-2356, BD), BV421 Rat IgG2a, κ isotype control (clone R35-95, BD) each at a 1:100 dilution. For intracellular CD206 staining, cells were fixated for 10 min with Leucoperm reagent A and after a washing step permeabilized with Leucoperm reagent B (Bio-Rad, Hercules, CA) containing 1:100 BV421 anti-mouse CD206 (clone C068C2, BioLegend). After washing, cells were analyzed on a BD FACSVerse Flow Cytometer (BD Biosciences). Data analysis and graph generation was performed using FlowJo 7.6 software (TreeStar, Ashland, OR).

Macrophage isolation and polarization

Bone marrow-derived macrophages (BMDM) were acquired by flushing out bone marrow by gravitational force in a microcentrifuge from femurs and tibia. Following erythrocyte lysis using BD Pharm Lyse (BD Biosciences), cells were filtered through a 70 µM mash and washed with Dulbecco’s modified Eagle’s medium (DMEM) HAM’s F12 (Thermo Fisher Scientific) supplemented with 10%FCS and 1% penicillin–streptomycin (both Thermo Fisher Scientific). 3 × 10e5 cells were seeded into Corning CellBIND 12-well plates (Corning, New York, NY) in full DMEM supplemented with 20 ng/ml macrophage colony-stimulating factor (M-CSF; Peprotech, London, UK) and cultured at 5% CO2 at 37 °C. After 3 days half of the medium was refreshed and on day 5 medium was replaced with full DMEM for 48 h to obtain M0, stimulated with lipopolysaccharide (LPS; 2,5 µg/ml, Sigma-Aldrich, St. Louis, MO) for 2 h to obtain M1, or stimulated with IL-4 and IL-13 (both 10 ng/ml, Thermo Fisher Scientific) for 48 h to obtain M2. For qPCR analysis cell were washed in PBS and dissolved in RLT buffer (Qiagen, Hilden, Germany). Cells for flow cytometry were scraped, washed in PBS and submitted to FACS staining as described above.

Ex vivo phagocytosis assay

Kidneys from naïve C57Bl/6 mice, after 24 h IRI and 3 days UUO were digested as described above and renal MNPs were sorted using CD11b- and CD11c-MicroBeads on LD columns (both Miltenyi Biotec) according to the instruction manual. Phagocytic capacity was assessed using the Phagocytosis Assay Kit (IgG PE) (Cayman Chemical, Ann Arbor, MI) according to the instruction manual. Briefly, sorted cells were incubated in RPMI containing PE-labeled latex beads for 2 h at 37 °C and washed in assay buffer. In the FACS panel for the phagocytosis assay APC-Ly-6G and BV421-CD206 were replaced by BV421 Rat Anti-Mouse F4/80 (clone T45-2342, BD) and APC anti-mouse CD206 (clone C068C2, BioLegend). Staining, including intracellular staining of CD206, were performed as described in the flow cytometry methods. The percentage of PE + phagocytic cells was determined for each MNP subset and the relative change between subsets was calculated for each replicate.

Real-time PCR analysis

RNA was extracted with a RNeasy Mini kit (Qiagen) and submitted to reverse transcription. Gene expression analysis was determined by quantitative real-time Taqman PCR using an ABI 7900HT Fast Real Time PCR System (Thermo Fisher Scientific) and normalized to RPL32 as housekeeper, which shows in our experience least regulation in various in vitro and in vivo models of the cardiovascular indication. The following primers and FAM/TAMRA-labelled probes were used: RPL32 Fw: ACCGAAAAGCCATTGTAGAAAGA; Rev: CCTGGCGTTGGGATTGG; probe: CAGCACAGCTGGCCATC; CD86 Fw: GAGTTTCCATCTGCTCAAACG; Rev: ACTTAGAGGCTGTGTTGCTG; probe: CCTGCTAGGCTGATTCGGCTTCT; SOCS3 Fw: GAAGATTCCGCTGGTACTGAG; Rev: GCTGGGTCACTTTCTCATAGG; probe: CCGACAAAGATGCTGGAGGGTGG; TNFα Fw: CTTCTGTCTACTGAACTTCGGG; Rev: CAGGCTTGTCACTCGAATTTTG; probe: ATCTGAGTGTGAGGGTCTGGGC; Mrc1 Fw: ATGGATGTTGATGGCTACTGG; Rev: TTCTGACTCTGGACACTTGC; probe: ACGAAATCCCTGCTACTGAACCTCC; CD200r1 Fw: GAGAAAAGGTACCGAGTGAGC; Rev: ATCAGTACAACTTGACCCAGC; probe: TGTTTTGCTTTTGGAGAACTTCTGCCC. Gene expression was normalized to RPL32 as housekee** gene. Data are expressed using the 2-ΔΔCTmethod, and mRNA expression is are calculated as fold over basal.

Statistics

All statistics and respective graphs were generated using GraphPad Prism 8.0.2 for Windows (GraphPad Software, San Diego, CA) and data presented as Mean ± S.E.M. In order to achieve normal distribution of MNP cell numbers, data were y = ln(y) transformed before analysis. Multiple groups were tested for significance using one-way ANOVA (normal data) or Kruskal–Wallis (non-normal data) followed by Tukey’s, Dunnett’s or Dunn’s multiple comparisons test as indicated in figure legends. MNP subsets at different time points of the IRI and UUO model were compared using two-way ANOVA with Tukey’s multiple comparisons test. For paired tests in the ex vivo phagocytosis assay we used Friedman with post-hoc Dunn’s multiple comparisons for multiple data sets.