Loss of CD11b Exacerbates Murine Complement-Mediated Tubulointerstitial Nephritis

Acute complement activation occurs in the tubulointerstitium (TI) of kidneys transplanted from Crry−/−C3−/− mice into complement-sufficient wildtype mice, followed by marked inflammatory cell infiltration, tubular damage and interstitial fibrosis. We postulated iC3b-CD11b interactions were critical in this TI nephritis model. We transplanted Crry−/−C3−/− mouse kidneys into CD11b−/− and wildtype C57BL/6 mice. Surprisingly, there was greater inflammation in Crry−/−C3−/− kidneys in CD11b−/− recipients compared to those in wildtype hosts. Kidneys in CD11b−/− recipients had large numbers of CD11b−Ly6ChiCCR2hiF4/80+ cells consistent with inflammatory (M1) macrophages recruited from circulating monocytes of the host CD11b−/− animal. There was also an expanded population of CD11b+CD11c+Ly6C−F4/80hi cells. Since these cells were CD11b+, they must have originated from the transplanted kidney; their surface protein expression and appearance within the kidney were consistent with the intrinsic renal mononuclear cellular population. These cells were markedly expanded relative to all relevant controls, including the contralateral donor kidney and Crry−/−C3−/− mouse kidneys in CD11b+/+ wildtype recipients. Direct evidence for their in situ proliferation was the presence of nuclear Ki67 and PCNA in CD11b+F4/80+ cells. Thus, in this experimental model in which there is unrestricted C3 activation, CD11b+ monocytes limit their own infiltration into the kidney and prevent proliferation of endogenous mononuclear cells. This suggests a role for outside-in iC3b-CD11b signals in limiting intrinsic organ inflammation.


Introduction
Activation of complement through its three pathways leads to generation of C3 and C5 products. These act on a limited set of cellular receptors. C3a and C5a receptors are rhodopsin-like Class A GTP-binding protein-coupled receptors while those that bind C3b and derivatives are termed complement receptors. The latter include the heterodimeric b 2 integrins, Itgam (CR3, a M b 2 , CD11b/CD18) and Itgax (CR4, a X b 2 , CD11c/CD18).
The unique rodent complement regulator, CR1-related gene y (Crry) is a structural and functional homologue to human CR1 [1]. Crry is present in endothelial and epithelial cells of the renal tubulointerstitium (TI) in a distribution comparable to membrane cofactor protein in human beings [2,3]. The relevance of Crry in the TI was first shown by Nomura, Matsuo et al. in rats using neutralizing antibodies [4]. A series of studies from Thurman et al. have shown that the normal polarization of Crry to the basolateral aspect of mouse tubules is lost in ischemia, which leads to unrestricted alternative pathway activation and acute kidney injury upon reperfusion [5,6,7]. This appears to be relevant to acute kidney injury (tubular necrosis) in human beings [8].

Ethics Statement
All animal experimental procedures were carried out in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committees of the Universities at Buffalo and Chicago.
The model of TI nephritis relies upon transplantation of kidneys from Crry 2/2 C3 2/2 mice into C3-sufficient animals [9,10]. Kidneys from Crry 2/2 C3 2/2 mice were transplanted into wildtype or CD11b 2/2 mice (n = 5 each). Both donor and recipient mice were backcrossed onto normal C57BL/6 mice for at least 10 generations. Mice between 8 to 10 weeks old were used as kidney donors or transplant recipients. To limit any potential variation within the model, the entire cohort of mice for the study was raised contemporaneously. Transplant recipients alternated between wildtype and CD11b 2/2 mice.
Kidney donors were anesthetized and the donor left kidney was removed with artery, vein and ureter attached, and preserved in cold saline on ice. The recipient was then anesthetized and the left kidney was excised. Renal transplantation was performed with end-to-side anastomoses of the donor renal vein, artery and ureter to the recipient inferior vena cava, aorta and bladder, respectively [14]. Total cold ischemic time ranged between 45 and 60 minutes. The left native kidney was removed at the time of renal transplantation. Blood was taken at the time of transplant and then twice over 7 days. Animals were sacrificed 7 days posttransplantation.
In our past control experiments establishing the general transplantation model, wildtype kidneys were transplanted into wildtype recipients (n = 34), along with removal of the second native kidney 7 days after transplantation. In these animals, blood urea nitrogen concentrations were 26.560.8 mg/dl 21 days post-transplantation.

Measurements from Tissue
Four-mm sections of 4% paraformaldehyde-fixed paraffinembedded kidney tissue were stained with periodic acid-Schiff and examined by a renal pathologist (MH) in a blinded manner. The severity of TI nephritis was graded from 0 to 4 in 0.5 increments [9].
For immunohistochemistry, zinc-fixed and paraffin-embedded kidney sections were used. Tissue sections were deparaffinized and rehydrated through xylene and serial dilutions of ethanol to deionized water. Endogenous peroxidases and biotin were blocked with 0.3% H 2 O 2 and Avidin/Biotin Blocking Kit (Vector Laboratories, Burlingame, CA, USA) followed by 10% normal mouse serum. They were incubated in antigen retrieval buffer (S1699, DAKO, Carpinteria, CA, USA) and heated in a steamer at 97uC for 20 minutes. Anti-mouse F4/80 antibody (1:200) was applied on tissue sections for 1 hour incubation at room temperature. Following washing, tissue sections were incubated with biotinylated anti-rat IgG (10 mg/ml, BA-4001, Vector Laboratories) for 30 minutes at room temperature. The antigenantibody binding was detected by Elite kit (PK-6100, Vector Laboratories) and DAB (K3468, DAKO) system. Anti-Ki67 antibody (1:200) was applied on tissue sections for overnight incubation at 4uC in a humidity chamber. Following washing, the antigen-antibody binding was detected with ImmPress AP antirabbit kit (MP-5401, Vector Laboratories) and Warp Red chromogen (WR806s, Biocare Medical, Concord, CA, USA). Tissue sections were briefly immersed in hematoxylin for counterstaining and were covered with cover glasses. By this protocol, Ki67 and F4/80 staining was indicated by red nuclear staining and brown cellular staining, respectively.
In another set of studies, slides were then incubated with rat anti-mouse CD11b, followed by goat anti-rat IgG and Streptavidin-peroxidase (Sigma/Aldrich, St. Louis, MO, USA). Specifically-bound antibodies were detected using a 3,39-diaminobenzidine (DAB)-based technique (ImmPACT DAB, Vector Laboratories). The peroxidases and biotin were again blocked as before. Slides were then incubated with anti-mouse PCNA, followed by biotinconjugated anti-rabbit IgG (Vector Laboratories) and Streptavidin-peroxidase (Sigma). Specific PCNA staining was then detected with the VIP Substrate Kit (Vector Laboratories), followed by methyl-green counter-staining (Vector Laboratories). By this protocol, PCNA and CD11b staining was indicated by purple nuclear staining and brown cellular staining, respectively.

Flow Cytometry
Peripheral blood cells were collected via orbital draw at indicated times. Red blood cells were lysed with NH 4 Cl/KHCO 3 and resuspended in FACS buffer (16 PBS, 2% calf serum, 5 mM EDTA, 0.1% sodium azide) and stained with the antibodies listed above.
Renal infiltrating cells were isolated as described previously [15]. In brief, mouse kidneys were minced and digested at 37uC for 30 min with gentle agitation with collagenase IV (2 mg/ml) in HBSS/1% (vol/vol) BSA (all from Sigma). Erythrocytes were lysed with NH 4 Cl/KHCO 3 and the cell suspension was passed through a 40 mm cell strainer (BD Biosciences). Approximately one million cells from each kidney were stained with the antibodies listed previously. Flow cytometry was performed with a FLSRII (BD Biosciences) and analyzed with FlowJo software (Tree Star, Inc., Ashland, OR, USA).

Statistical Analyses
Anderson-Darling tests were used to evaluate data normality (Minitab v. 16.2.4, State College, PA, USA). Comparisons between two groups of parametric data were made with twosample t-testing. In the text, data are presented as means 6 SEMs.

Peripheral Blood Monocytes
Circulating Gr1 + /Ly6C hi (inflammatory) monocytes [16,17] also expressed CCR2 ( Figure 3B, red circle), and both CD11b and CD11c ( Figure 3A, red ovals). In wildtype mouse recipients of Crry 2/2 C3 2/2 kidneys, there was an initial decline in these cells 1 day after transplant ( Figure 3A), while they were expanded 7 days post-transplant ( Figure 3B). Overall, CD11b 2/2 transplant recipients had qualitatively similar findings, except for the complete absence of CD11b on inflammatory monocytes ( Figure 3C, red arrows). The initial decline in inflammatory monocytes is likely attributable to their recruitment to the transplanted kidney (see below), while at later time points, they are recruited from bone marrow stores accounting for this monocytosis. Thus, local inflammation within transplanted kidneys generates a systemic inflammatory response, which does not appear to be affected to a significant extent in the absence of CD11b. The resident CD11b + cells in the kidney do not appear in the peripheral circulation.
In contrast to these infiltrating cells, those cells that did not express Ly6C/G ( Figure 4G, blue oval) were largely or exclusively mononuclear phagocyte (MPC) cells intrinsic to the kidney (discussed further below). These included cells that highly expressed F4/80 ( Figure 4H, blue ovals), along with CD11b and CD11c, and variable amounts of CCR2 ( Figure 4I). There were two Ly6C/G 2 F4/80 2 cell populations, differing by the presence or absence of CD11b ( Figure 4H, black arrows). The CD11b lo cells included a population of CD11c hi cells (data not shown).
To compare these results with our past studies showing the C3aR-dependence of the inflammation [10], we examined Crry 2/ 2 C3 2/2 mouse kidneys in C3aR 2/2 recipients for F4/80 and Ki67. As expected, there were F4/80 + cells throughout the TI as in normal kidneys but there was not significant inflammation ( Figure 5C). Interestingly, there were cells expressing Ki67, which were in tubules ( Figure 5C, black arrowheads), suggesting an ongoing repair process.

Discussion
In these studies we utilized a model of TI nephritis in which acute C3 activation in Crry 2/2 C3 2/2 kidneys generated C3aRdependent inflammation [9,10]. Here we have further characterized the TI cellular infiltrates. The Ly6C/G hi CCR2 hi F4/ 80 + CD11b + cells are consistent with inflammatory M1 macrophages. Their accumulation in the kidney can be attributed to their recruitment from circulating inflammatory Gr1 + /Ly6C hi monocytes. This cellular pool was in turn considerably expanded by ''emergency myelopoeisis'' [18]. Thus, when the recipient animal was CD11b-deficient, the Ly6C/G hi CCR2 hi F4/80 + inflammatory macrophages completely lacked CD11b.
Among the Ly6C/G 2 cells were F4/80 hi CD11b + CD11c + cells. Since these correspond well to the MPC3 population recently described by Nelson, Duffield et al [19], we will follow their nomenclature. These MPC3 cells do have several characteristics of dendritic cells, including high CX 3 CR1 and MHCII expression, phenotypic features of dendritic processes [20,21], and ability to serve as professional antigen presenting cells [22,23]. Yet, their high F4/80 expression is atypical for dendritic cells, as is also true for variable CCR2 expression, including some with high expression indistinguishable from inflammatory macrophages. On average, inflammatory macrophages had greater CD11b and lesser CD11c surface staining than MPC3 cells, but there was considerable overlap in the aggregate populations. Overall, as noted recently by a panel of experts [24], renal MPCs have considerable overlap of macrophage and dendritic cell properties, making conventional binary naming systems inadequate.
Interestingly, there were two Ly6C/G 2 F4/80 2 cell populations (i.e., depicted by the arrows in Figure 4H), which were Figure 3. Fate of Gr1 + /Ly6C hi monocytes over time post-transplantation of Crry 2/2 C3 2/2 kidneys into wildtype and CD11b 2/2 recipients. Flow cytometry was performed on peripheral blood mononuclear cells obtained from wildtype (CD11b +/+ , A, B) and CD11b 2/2 animals (C, D) prior to and 1, 6 (A, C), or 7 days (B, D) after receiving Crry 2/2 C3 2/2 kidney transplants. The CCR2 hi Ly6C/G hi CD11b + CD11c + cellular population prior to transplantation (day 0) is bounded by a red oval; its relative position is maintained in the subsequent panels for day 1 and 6 data. The positions of the analogous cellular populations in CD11b +/+ recipients are maintained in CD11b 2/2 recipients and depicted with dashed lines. The red arrows point to CCR2 hi Ly6C/G hi CD11b 2 CD11c + cells. At each time point, blood was obtained from two animals in each group; shown are representative data from one of the two. doi:10.1371/journal.pone.0092051.g003 CD11b Limits Tubulointerstitial Nephritis PLOS ONE | www.plosone.org distinguished by CD11b staining into CD11b lo and CD11b 2 cells. These are consistent with MPC4 and 5 subpopulations in the previously described study [19]. The MPC4 cells had considerable antigen-presenting potential, and were potentially CD103 + classical dendritic cells [19]. While we cannot comment on function, that these cells were expanded in Crry 2/2 C3 2/2 kidneys in CD11b 2/2 hosts (c.f., arrowhead, Figure 4K), establishes their origin from the kidney itself.
These studies are not intended to model allotransplantation, as donor and recipient mice were genetically identical (but for the targeted genes). Yet, like with allotransplantation, the lymphatics are severed in the donor kidney. There were not any CD11b + cells within the peripheral blood mononuclear cell pool in CD11b 2/2 recipients of Crry 2/2 C3 2/2 mouse kidneys. Thus, there was not emigration of any CD11b + mononuclear cell population from the CD11b-sufficient transplanted kidney.
In normal kidneys, C3 (and C5b-9) staining is evident along the basement membranes of Bowman's capsule and renal tubules [25]. This is attributable to alternative pathway activation, and is enhanced in warm ischemic-reperfusion injury [5,26]. A comparable sequence of events occurs in Crry/C3-deficient kidneys placed in a C3-sufficient environment, in which there is acute complement activation [ [9] and unpublished data]. It is likely these pathological events mirror those destined to occur in acute infections, with a rapid response by inflammatory monocytes to generate inflammatory macrophages in tissue. These cells have direct effects on infectious microorganisms, and facilitate adaptive immune responses by lymphocytes. Once the infection is cleared, there is resolution of inflammation and return of the organ to normal.
Complement activation in the renal TI from excessive activation and/or abnormal regulation appears relevant in a number of disease states. Ammonia can hydrolyze C3 to initiate the alternative pathway. Given heightened ammoniagenesis, complement activation may have a role in progression of renal diseases of diverse origins [26], a theory supported by contempo- Figure 4. Cellular characterization of tubulointerstitial infiltrates in Crry 2/2 C3 2/2 kidneys transplanted into wildtype and CD11b 2/2 recipients. Cells isolated from Crry 2/2 C3 2/2 kidneys 7 days following transplantation into wildtype (CD11b +/+ , A-C and G-I) and CD11b 2/2 animals (D-F and J-L) were analyzed by flow cytometry. In wildtype recipients, cells characteristic of M1 macrophages were CCR2 hi Ly6C/G hi (A, red oval) and F4/80 + CD11b + (B, red ovals). There were few CCR2 2 Ly6C/G lo cells (A, black oval); these were also stained for CD11b and CCR2 (C). Ly6C/G 2 cells included F4/80 hi CD11b + cells (G and H, blue ovals) which were analyzed for CCR2 and CD11c (F). There were Ly6C/G 2 F4/80 2 cells (G) that were either CD11b lo or CD11b 2 (H, black arrows). The positions of the analogous cellular populations in CD11b +/+ recipients are shown in CD11b 2/2 recipients and depicted with solid lines where their staining characteristics are comparable, and dashed lines when they are dissimilar from wildtype recipients. The red arrows point to CCR2 hi Ly6C/G hi CD11b 2 CD11c + cells (E). Data are from six separate Crry 2/2 C3 2/2 kidneys, three each in wildtype and CD11b 2/2 recipients. doi:10.1371/journal.pone.0092051.g004 CD11b Limits Tubulointerstitial Nephritis PLOS ONE | www.plosone.org rary clinical data [27,28]. Alloantibody-mediated complement activation appears etiologic in humoral renal allograft rejection [29]. Inherited and acquired abnormalities of membrane cofactor protein, a complement regulator present in the renal TI [2,30] can underlie atypical hemolytic uremic syndrome and acute kidney injury, respectively [8]. Given their similar distributions and shared functions, our findings with Crry in the kidney are relevant to these human disease states in which there is abnormal complement regulation by membrane cofactor protein [31].
While the role for b2 integrins in cellular adhesion is well established, there is growing appreciation for their involvement in cellular signaling. Anaphylatoxin and chemokine receptor activation, including C3aR, recruit G i a leading to ''inside-out'' activation of CD11b and increased avidity for ligand [32,33] such as iC3b [34]. When CD11b binds its ligand it can generate ''outside-in'' tyrosine kinase signals mediated by immunoreceptor tyrosine-based activation motif (ITAM) proteins, DAP12 and FccR; mononuclear phagocytic cells primarily rely on DAP12 [35,36]. Of considerable interest is the anti-inflammatory nature of these signals through CD11b. For example, CD11b can limit proinflammatory signals through Toll-like receptor (TLR) 4 activation by E3 ubiquitin-protein ligase CBL-B-mediated removal of activated proteins [37,38,39].
The expanded Ly6C/G 2 F4/80 hi CD11b + CD11c + MPC3 cellular population in CD11b-deficient recipients could only have come from the transplanted kidney. That this was due to intrinsic proliferation was confirmed by Ki67 and PCNA staining. These appeared to retain the original phenotypic features of MPC3 cells in the normal uninflamed kidney. Traditional views were tissue-resident macrophages had limited self-renewal properties and the expansion of macrophages in inflammation was due to recruitment from blood monocytes [40]. However, it is now clear macrophages in a variety of tissue sites have proliferative capacity [41,42,43,44,45]. Similarly to the MPC3 cells described here, these typically are F4/80 hi and are distinct from monocyte-derived macrophages.

Conclusions
In conclusion, in our study we used a model of TI nephritis, in which acute C3 activation leads to C3a generation and deposition of C3b in the TI. When CD11b is absent on monocytes, TI nephritis was worsened. This was due both to greater inflammation with Gr1 + /Ly6C hi Figure 5. Dual immunohistochemical staining showing proliferation of F4/80 + and CD11b + cells in Crry 2/2 C3 2/2 kidneys transplanted into CD11b 2/2 recipients. Shown is representative immunohistochemistry in Crry 2/2 C3 2/2 kidneys transplanted 7 days earlier into wildtype (A), CD11b 2/2 (B, C) and C3aR 2/2 recipients (D). There was greater inflammation in CD11b 2/2 recipients compared to wildtype and C3aR 2/ 2 recipients. TI F4/80 cellular staining is brown and Ki67 is red (A, B, D); CD11b cellular staining is brown and PCNA nuclear staining is purple (C). The arrows depict cells that have both staining products. Arrowheads in D depict Ki67 staining in tubular cells. Inset to C, Glomerulus from murine lupus nephritis stained for CD11b and PCNA showing distinct reaction products. In each group, separate Crry 2/2 C3 2/2 kidneys are shown magnified 2006 and 4006. doi:10.1371/journal.pone.0092051.g005 inflammatory monocyte-derived inflammatory M1 macrophages and expansion of the Ly6C/G 2 F4/80 hi intrinsic renal MPC3 cellular pool. Thus, CD11b + monocytes limit their own infiltration into the kidney and prevent proliferation of endogenous CD11b + MPC cells. This suggests a role for outside-in iC3b-CD11b signals in limiting intrinsic organ inflammation.