Monocyte Chemoattractant Protein 1–Dependent Leukocytic Infiltrates Are Responsible for Autoimmune Disease in Mrl-Fas^lpr Mice

MRL/MpJ⁺⁺ (MRL⁺⁺), MRL/MpJ-Fas^lpr/Fas^lpr (MRL-Fas^lpr), C3H/FeJ, and C57BL/6 mice were purchased from the The Jackson Laboratory. MCP-1–intact and –deficient B6/129 mice (129SV/J × C57BL/6)F₁ were constructed as described 20. Control MCP-1–intact B6/129 mice were derived from matings of mice heterozygous for the disrupted allele. All mice were maintained in a pathogen-free animal facility.

MCP-1–deficient (MCP-1^−/−) MRL-Fas^lpr mice were created by a series of genetic backcrosses using the cross-backcross-intercross scheme. MRL-Fas^lprmice were mated with MCP-1^−/− (129/Sv × C57BL/6) mice to yield heterozygous F1 offspring. We intercrossed F1 mice and screened the progeny by PCR amplification of tail genomic DNA for the Fas^lprmutation and MCP-1 using specific primers 21,24. Double homozygotes (Fas^lpr/Fas^lprMCP-1^−/−) N1F1 progeny were backcrossed to MRL-Fas^lprmice. B1 progeny, homozygous for the Fas^lprmutation and heterozygous for MCP-1 (MCP-1^+/−), were intercrossed, and mice homozygous for the disrupted MCP-1 gene were selected by PCR typing for continued backcrossing. After three generations of backcross–intercross matings, this breeding scheme generated a colony of MRL-Fas^lprmice (94% MRL-Fas^lpr background) homozygous and heterozygous for the disrupted MCP-1 gene. We analyzed the third generation, as we have previously established that there are sufficient MRL-Fas^lpr background genes to result in phenotypic changes characteristic of the wild-type MRL-Fas^lprstrain 25. In addition, we compared sex-matched littermates to minimize variability. The MCP-1^−/− MRL-Fas^lpr mice are termed MCP-1 deficient, whereas the MCP-1^+/+MCP-1^+/− MRL-Fas^lpr mice are termed MCP-1–intact MRL-Fas^lpr mice.

Urine protein levels in MCP-1–intact and –deficient MRL-Fas^lprmice were assessed semiquantitatively by dipstick analysis (Albustix; Bayer Diagnostic Division) on a monthly basis beginning at 2 mo of age. On the day of analysis, dipstick proteinuria measurements were taken from individual mice in the morning and evening. If the measurements were inconsistent, the animal was reassessed the following day.

Protruding lymph nodes (cervical, brachial, and inguinal) and skin lesions were assessed monthly beginning at 3 mo of age. Lymph node score based on palpable nodes: 0 = none; 1 = small, at one site; 2 = moderate, at two different sites; and 3 = large, at three or more different sites. Skin lesion score by gross pathology: 0 = none; 1 = small (face or ears); 2 = moderate, <2 cm (face, ears, and back); and 3 = severe, >2 cm (face, ears, and back). Splenomegaly was determined by spleen weights.

Kidneys were fixed in 10% formalin for 24 h at 4°C. Paraffin sections (4 μm) were stained with hematoxylin and periodic acid Schiffs (PAS) reagent. We evaluated glomerular pathology by assessing 50 glomerular cross sections (gcs) per kidney and scored each glomerulus on a semiquantitative scale: 0 = normal (35–40 cells/gcs); 1 = mild (glomeruli with few lesions showing slight proliferative changes, mild hypercellularity [41–50 cells/gcs], and/or minor exudation); 2 = moderate (glomeruli with moderate hypercellularity [50–60 cells/gcs], including segmental and/or diffuse proliferative changes, hyalinosis, and moderate exudate); and 3 = severe (glomeruli with segmental or global sclerosis and/or exhibiting severe hypercellularity [>60 cells/gcs], necrosis, crescent formation, and heavy exudation). Damaged tubules (percent; consisting of dilation and/or atrophy and/or necrosis) were determined in 400 randomly selected renal cortical tubules per kidney (×400). Perivascular cell accumulation was determined semiquantitatively by scoring the number of cell layers surrounding the majority of vessel walls (score: 0 = none; 1 = <5; 2 = 5–10; and 3 = >10). Scoring was evaluated using coded slides.

The lungs were inflated and fixed with 10% formalin, and paraffin sections (4 μm) were stained with hematoxylin and PAS reagent. Perivascular and peribronchiolar infiltrates were assessed semiquantitatively in >20 vessels per section and in >20 bronchioli per section (number of cell layers surrounding the majority of vessels or bronchioli: 0 = none; 1 = 1–3; 2 = 3–6; and 3 = >6).

The following primary antibodies were used for immunostaining: rat anti–mouse CD4 IgG2a clone RM4-5 (PharMingen) to detect CD4 T cells; rat anti–mouse CD8a (Ly-2) IgG2a clone 53-6.7 (PharMingen) to detect CD8 T cells; rat anti–mouse CD45R/B220 IgG2a clone RA3-6B2 (PharMingen) to detect CD4/CD8 T cells; rat anti–mouse CD21/35 IgG2b clone 7G6 (PharMingen) to detect B cells; rat anti–mouse monocyte/macrophage IgG2b clone MOMA-2 (BioSource International) and rat anti–mouse macrophage IgG2b (prepared from F4/80 hybridoma supernatant; American Type Culture Collection number HB198) to detect macrophages; rabbit anti–mouse MCP-1 IgG (Serotec Ltd.) to detect MCP-1; mouse anticytokeratin peptide 18 IgG1 clone CY-90 (Sigma Chemical Co.) to detect epithelial cells; fluorescein-conjugated mouse anti-PCNA (proliferating cell nuclear antigen) IgG1 clone 19F4 (Boehringer Mannheim) to detect proliferating cells; fluorescein-conjugated goat anti–mouse IgG (Organon Teknika) to detect mouse IgG; and fluorescein-conjugated goat anti–mouse C3 (Organon Teknika) to detect mouse complement C3. The negative isotype control antibodies for immunostaining were rat IgG2a clone R35-95, rat IgG2b clone R35-38, mouse IgG1 clone MOPC-21 (PharMingen), and normal rabbit IgG (Sigma Chemical Co.). The secondary antibodies for immunostaining were biotin-conjugated rabbit anti–rat IgG and biotin-conjugated goat anti–rabbit IgG (Vector Labs.). ELISA analysis of serum Ig (total Ig, IgM, IgG1, IgG2a, IgG2b, and IgG3) was performed using isotype-specific standards, goat anti–mouse capture antibodies, and alkaline phosphatase–conjugated goat anti–mouse detection antibodies (Southern Biotechnology Associates Inc.).

To analyze kidney- and lung-infiltrating T cells and interstitial macrophages, cryostat tissue sections (4 μm) were fixed in ethanol at 4°C for 10 min and immunostained with CD4, CD8, B220, and F4/80 antibody as previously described 26 using the avidin-biotin-peroxidase detection system (Vector Labs.). Glomerular macrophages were identified in 4-μm acetone-fixed tissue sections using MOMA-2 antibody as previously reported 16. Cells infiltrating within and around (peri) glomeruli were assessed by counting the number of labeled cells in 20 randomly selected glomeruli per section. The cells infiltrating around tubules were enumerated in 400 randomly selected tubules per section. Macrophages and T cells within glomeruli, adjacent to glomeruli, or adjacent to cortical tubules were expressed as a cell index (mean cell number/glomeruli × glomeruli/section; mean cell number/cortical tubule × cortical tubules/section). Peribronchial and perivascular macrophages and T cells (CD4, CD8, B220) in the lungs were assessed by measuring the unit area stained/unit length of bronchiole/vessel (five per animal) with a micrometer.

To detect MCP-1, formalin-fixed sections were deparaffinized and incubated with 20% normal goat serum for 30 min. Tissue sections were incubated with MCP-1 antibody (10 μg/ml) in 1% BSA overnight at 4°C. Bound primary antibody was labeled with biotin-conjugated goat anti–rabbit IgG for 1 h and subsequently detected using the avidin-biotin-peroxidase system (Vector Labs.). We confirmed MCP-1 antibody specificity using two methods: (a) MCP-1 was not detectable in tissue sections from MCP-1–deficient mice (negative control), and (b) MCP-1 was expressed in Western blots of cell lysates from LPS-stimulated bone marrow macrophages and mesangial cells in MCP-1–intact but not MCP-1–deficient mice, as previously reported 21. Glomerular MCP-1 immunostaining was assessed by the same scoring system used to detect glomerular infiltrating cells. We analyzed MCP-1 in cortical tubules in 400 tubules per cross section. We enumerated the kidney and pulmonary vessels and bronchioli that expressed MCP-1 in each section and recorded values as the percent positive.

ELISA plates were coated overnight at 4°C with 5 μg/ml goat anti–mouse Ig capture antibodies (against total Ig, IgM, IgG1, IgG2a, IgG2b, and IgG3) in 0.1 M carbonate buffer, pH 9.4. Wells were blocked for 1 h with assay diluent (2% BSA in 0.1 M borate buffer, pH 8.0). We then added Ig standards to the plates (50 μl/well), starting at 1 μg/ml and performing a series of threefold dilutions, and assessed serum samples using serial (threefold) dilutions starting at 1:100 or 1:1,000. Standards and serum samples were incubated overnight at 4°C, and bound Ig was detected with goat anti–mouse detection antibodies conjugated with alkaline phosphatase and enzymatically developed by incubating with Sigma 104 phosphatase substrate in 9.6% diethanolamine and 0.1 mM MgCl₂, pH 9.8 (Sigma Chemical Co.). Absorbance was measured at 405 nm.

To examine IgG and C3 deposits in the kidney, we incubated cryostat-sectioned tissues (4 μm) with 20% normal goat or rabbit serum (30 min), followed by fluorescein-conjugated antibodies detecting mouse IgG or mouse C3 (30 min), washing, and mounting with Vectashield (Vector Labs.). We assessed immunofluorescence staining by titrating the antibodies on serial tissue sections using twofold dilution steps (1:100–1:25,600).

Apoptotic cells were identified in tissue sections (4 μm) using the terminal deoxynucleotide transferase–mediated dUTP-biotin nick-end labeling (TUNEL) method and immunoperoxidase staining (In Situ Cell Death Detection Kit; Boehringer Mannheim). Proliferating cells were detected by PCNA immunostaining as previously described 27. Sections were counterstained with PAS reagent to determine morphology. Apoptosis in glomeruli and tubules was assessed by counting the number of TUNEL-labeled cells in 50 glomeruli per section and 500 cortical tubules per section, respectively.

Total RNA was extracted from the snap-frozen renal cortex of half a kidney using RNAzol B (Tel-Test Inc.) and reverse transcribed using oligo-dT and the Superscript II DNA preamplification kit (GIBCO BRL). The resulting reverse transcription product was the cDNA template for PCR analysis. GAPDH (glyceraldehyde-3-phosphate dehydrogenase) expression was detected as a 500-bp product resulting from PCR with specific oligonucleotide primers (antisense, 5′-CAAAGTTGTCATGGATGACC-3′; and sense, 5′-GGTGGAGGTCGGAGTCAACG-3′; reference 21). The chemokines MCP-1, MCP-3, and MCP-5 were detected as 350-, 300-, and 370-bp PCR products, respectively, using specific oligonucleotide primers (MCP-1 antisense, 5′-GCTTGAGGTGGTTGTGGAAAA-3′; MCP-1 sense, 5′-CTCACCTGCTGCTACTCATTC-3′; MCP-3 antisense, 5′-CACATTCCTACAGACAG-CTC-3′; MCP-3 sense, 5′-AGCTACAGAAGGATCACCAG-3′; MCP-5 antisense, 5′-CTCCTTATCCAGTATGGTCC-3′; and MCP-5 sense, 5′-TCTCCCTCCACCATGCAGAG-3′; references 28,29,30). The cytokines IFN-γ and CSF-1 were detected as 500- and 245-bp PCR products using specific primers (IFN-γ sense, 5′-CACGGCACAGTCATTGAAAGCC-3′; IFN-γ antisense, 5′-CTTATTGGGACAATCTCTTCCC-3′; CSF-1 sense, 5′-CACATGATTGGGAATGGACA-3′; and CSF-1 antisense, 5′-CAGCTGTTCAGGTTATTGGA-3′; references 31 and 32). The PCR primers and conditions were chosen so that the MCP and GAPDH products were amplified with equal efficiency.

Data was analyzed using the Kruskal-Wallis test for comparing survival curves, the Mann-Whitney test to compare group means, and Spearmans coefficient for correlations.

MCP-1 is upregulated in MRL-Fas^lpr kidneys, lungs, and lymph nodes during autoimmune disease. First, we detected an increase (threefold) in MCP-1 mRNA within the renal cortex of wild-type pure MRL-Fas^lpr mice as early as 2 mo of age, as compared with the C57BL/6 strain with normal kidneys. MCP-1 mRNA increased further within the renal cortex (six- to eightfold) with advancing renal injury in these MRL-Fas^lpr mice (Fig. 1). We localized MCP-1 expression to several structures within the MRL-Fas^lpr kidney and determined that the rank order of tissue expression was tubules→glomeruli→vasculature. MCP-1 in MRL-Fas^lprkidneys (5–6 mo of age) was almost exclusively localized within the TECs. More specifically, MCP-1 was within the proximal TECs (brush border histologic identification), whereas the distal TECs (histologic identification) were only weakly positive or negative. MCP-1 expression in MRL-Fas^lprglomeruli was determined to be predominantly in epithelial podocytes, based on the morphological positioning of MCP-1–stained cells in glomeruli after histologic counterstaining (Fig. 2 b and Fig. 3 a), whereas lesser amounts localized within the epithelial cells and macrophages within crescents (two to three cells per crescent). Finally, MCP-1 in the vasculature within endothelial and smooth muscle cells was minimal (Fig. 3 a, insert). It is noteworthy that few (1–2%) infiltrating cells in the renal interstitium express MCP-1. Thus, kidney MCP-1 expression is almost exclusively within parenchymal cells and in epithelial cells in particular, rather than infiltrating cells.

MCP-1 is expressed in other tissues in MRL-Fas^lpr mice during autoimmune disease. Similar to MCP-1 in the kidney, MCP-1 in the lungs is predominantly (>90%) expressed in epithelial cells (bronchioli; Fig. 3 b) and is weakly detected within vascular endothelial and interstitial cells in MRL-Fas^lpr mice (5 mo of age). Massively enlarged lymph nodes caused by an influx of T cells are characteristic of MRL-Fas^lpr autoimmune disease 23. MCP-1 is readily detected in cells surrounding lymphatic vessels within these enlarged lymph nodes (inguinal, cervical) in MRL-Fas^lpr mice (5 mo of age; Fig. 3 c). As anticipated, MCP-1 was not detected in MCP-1–deficient MRL-Fas^lprtissues (Fig. 3d–f). Taken together, these data indicate that MCP-1 is mainly expressed by parenchymal cells in multiple tissues targeted for autoimmune injury in MRL-Fas^lpr mice.

MCP-1–deficient MRL-Fas^lprmice have a prolonged life span as compared with MCP-1–intact MRL-Fas^lprstrains. The vast majority of MCP-1–deficient MRL-Fas^lprmice (75%) remained alive at 300 d, as compared with a surviving minority of MCP-1–intact (MCP-1^+/+, 17% and MCP-1^+/−, 44%) MRL-Fas^lpr mice (Fig. 4 a; P < 0.0001). Notably, the mortality (50%) of the MCP-1^+/+ MRL-Fas^lprstrain third generation was 7 mo of age, which is similar to that of the pure wild-type MRL-Fas^lprstrain (6 mo of age) and dissimilar to that of C57BL/6J and Sv/129 strains (>20 and 16 mo of age, respectively; reference 33).

It is important to note that the spot analysis for protein in fresh individual urine specimens has several limitations, including the sample size (volume) and semiquantitative measurement. However, our confidence in making inferences from this method is enhanced by the large number of mice in each group and sequential monthly analysis. With these caveats in mind, we now report that MCP-1–deficient versus MCP-1–intact MRL-Fas^lpr mice are protected from pathological proteinuria. From 2 to 8 mo of age, both the rate of increase and incidence of pathological proteinuria were diminished in MCP-1–deficient MRL-Fas^lpr mice (Fig. 4 b). For example, the vast majority (82%) of MCP-1–deficient MRL-Fas^lpr mice had normal, nonpathologic proteinuria (1+), whereas the majority (62%) of MCP-1–intact MRL-Fas^lpr mice were pathologically proteinuric (2–4+) at 8 mo of age. It should be noted that urinary protein in normal B6/129 wild-type mice is barely detectable (0–1+; data not shown). Thus, MCP-1–deficient MRL-Fas^lpr mice are protected from proteinuria.

We examined whether MCP-1 promotes autoimmune disease in the lymph nodes, spleen, and skin of MRL-Fas^lpr mice from 3 to 8 mo of age. The incidence and severity of lymphadenopathy and skin lesions was reduced but not totally eliminated in MCP-1–deficient as compared with MCP-1–intact MRL-Fas^lpr mice (Fig. 4c and Fig. d). For example, although nearly every MCP-1–intact MRL-Fas^lpr mouse had palpable lymph nodes (94%), most of the MCP-1–deficient mouse lymph nodes were not palpable at 5 mo of age (69%). Similarly, the majority (84%) of MCP-1–deficient mice were spared skin lesions, whereas most (67%) of the MCP-1–intact MRL-Fas^lpr mice had gross skin pathology at 8 mo of age. On the other hand, we did not detect a difference in splenomegaly in MCP-1–deficient (234 ± 64 mg) versus MCP-1–intact MRL-Fas^lpr mice (242 ± 143 and 289 ± 174 mg MCP-1^+/+ and MCP-1^+/−, respectively; P = 0.2; n = 6 per group) at 5 mo of age.

To determine if MCP-1 is required for kidney and lung pathology, we compared MCP-1–deficient and –intact MRL-Fas^lpr mice killed at 5 mo of age. Renal (tubular, glomerular) and pulmonary pathology was reduced in MCP-1–deficient as compared with MCP-1–intact MRL-Fas^lpr mice (Fig. 5 and Fig. 6). The reduction in tubular and glomerular pathology assessed histologically was dramatic (50–70%) in MCP-1–deficient MRL-Fas^lpr kidneys (Fig. 6, a and b; P < 0.05). In particular, MCP-1–deficient MRL-Fas^lpr mice had markedly diminished peritubular infiltrate, tubular atrophy, glomerular hypercellularity, glomerulosclerosis, and crescent formation. We further evaluated the level of tubular and glomerular damage by identifying apoptotic cells. The number of apoptotic cells was reduced in MCP-1–deficient versus MCP-1–intact MRL-Fas^lpr kidneys. The majority of apoptotic TECs diminished from 4.2 ± 0.3% per section in MCP-1–intact to 2.0 ± 0.6% per section in MCP-1–deficient MRL-Fas^lpr kidneys (P < 0.05; n = 6 per group). Similarly, the number of apoptotic glomerular cells was reduced in MCP-1–deficient versus MCP-1–intact mice (0.11 ± 0.04 cells/gcs and 0.25 ± 0.05 cells/gcs, respectively; P < 0.05; n = 6 per group). Although there was a reduction in tubular and glomerular pathology in MCP-1–deficient MRL-Fas^lpr kidneys, it was not totally prevented (Fig. 6, a and b) in comparison to age-matched kidneys from MRL⁺⁺ and C3H mice with normal kidneys (Fig. 6, a and b; P < 0.005). There was an increase in tubular and glomerular pathology in MCP-1–deficient MRL-Fas^lpr kidneys. In contrast to the protection afforded by the absence of MCP-1 in renal tubules and glomeruli of MRL-Fas^lpr mice, the perivascular infiltrates in MCP-1–deficient and –intact MRL-Fas^lpr kidneys remained similar (Fig. 6 c). Thus, MCP-1 is responsible for tubular and glomerular but not perivascular renal pathology.

Pulmonary disease was reduced in MCP-1–deficient versus MCP-1–intact MRL-Fas^lpr mice. Lung pathology in MCP-1–intact MRL-Fas^lpr mice consists of infiltrating cells predominantly surrounding bronchioles and vasculature. The numbers of cells surrounding bronchioles in MCP-1–deficient versus MCP-1–intact MRL-Fas^lpr strains was dramatically reduced (Fig. 6 d and Fig. 7d–f; P < 0.02). As in the kidney, the infiltrating cells surrounding vessels were not reduced in MCP-1–deficient versus MCP-1–intact MRL-Fas^lpr lungs (Fig. 6 e; P = 0.15).

We determined if the MCP-1–deficient MRL-Fas^lpr mice that died (25%; 300 d) developed renal injury that was as severe as that of the MCP-1–intact MRL-Fas^lpr mice that succumbed at considerably younger ages (83% mortality; 300 d; Fig. 4 a). The MCP-1–deficient MRL-Fas^lpr mice had less (40–50%) tubular and glomerular pathology compared with the MCP-1–intact MRL-Fas^lpr kidneys (P < 0.05; n = 6; data not shown). In addition, fewer (5/11) MCP-1–deficient MRL-Fas^lpr mice, as compared with MCP-1–intact MRL-Fas^lpr mice (12/16), died with severe proteinuria (3–4+). Thus, even in the MCP-1–deficient and MCP-1–intact MRL-Fas^lpr mice that eventually die, the extent of renal injury is less severe in the MCP-1–deficient MRL-Fas^lpr strain.

There was a reduction in the number of macrophages and T cells (CD4 and CD8 but not B220) in MCP-1–deficient versus MCP-1–intact MRL-Fas^lpr kidneys. The majority of macrophages and T cells localized in the interstitium adjacent to tubules and glomeruli, whereas fewer were identified within glomeruli. We noted the largest (50–72%) decline in macrophages and T cells in the MCP-1–deficient MRL-Fas^lpr strain surrounding tubules (P < 0.01; Fig. 7, a–c), a reduction that correlated with the extent of tubular injury (r ≥ 0.80; P < 0.005). Periglomerular macrophages and CD4 but not CD8 T cells were also diminished (45–53%, P < 0.05) in MCP-1–deficient MRL-Fas^lpr mice (Fig. 7 a). Although intraglomerular macrophages were reduced (34%) in MCP-1–deficient MRL-Fas^lpr kidneys (P < 0.005; Fig. 7 a), the numbers of intraglomerular T cells (CD4, CD8, B220) were not diminished (P = 0.2). The reduction in glomerular macrophages correlated with diminished glomerular morphologic damage (intraglomerular, r = 0.78; periglomerular, r = 0.77; P < 0.005). The absence of MCP-1 in MRL-Fas^lpr lungs reduced the number of infiltrating macrophages surrounding bronchioli but not surrounding vessels (Fig. 7d–f). In contrast to the reduction in the number of kidney-infiltrating T cells in MCP-1–deficient versus –intact MRL-Fas^lpr mice, the number of T cells (CD4, CD8, B220) in the lungs was not reduced (P = 0.2; Fig. 7 d).

We determined if the decreased accumulation of macrophages and T cells in MCP-1–deficient MRL-Fas^lpr kidneys and lungs at 5 mo of age was a result of a decline in local proliferation using in situ detection of PCNA. First, few of the cells surrounding tubules (<1%), glomeruli (<1%), and bronchioli (<10%) were proliferating in the MCP-1–intact MRL-Fas^lpr strain. The number of PCNA⁺ cells was not reduced in the MCP-1–deficient MRL-Fas^lpr strain (P = 0.2; n = 6). Second, there were substantially more proliferating cells surrounding vessels (>20%) compared with proliferating cells surrounding glomeruli, tubules, and bronchioles in MCP-1–intact MRL-Fas^lpr kidneys and lungs (P < 0.002), and this number did not decrease in the MCP-1–deficient MRL-Fas^lpr strain (P = 0.1; n = 6 per group). Of course, it must be appreciated that the proliferation measurements in tissue sections are a reflection of the level of cell division at the time these tissues were removed. Thus, the reduction in kidney- and lung-infiltrating cells in MCP-1–deficient MRL-Fas^lpr mice does not appear to result from decreased local proliferation.

MCPs 1–5 are ligands for the MCP-1 receptor (CCR2; reference 22). To examine whether CCR2 ligands other than MCP-1 are upregulated after renal injury, we examined MCP-3 and MCP-5 in MRL-Fas^lpr kidneys. Renal cortical transcripts of MCP-3 and MCP-5 in MRL-Fas^lpras compared with C57BL/6 mice were increased (twofold) in advance of overt renal pathology (2 mo of age) and rose further (fourfold) as renal pathology advanced (6 mo of age; Fig. 1). We probed for MCP expression in MCP-1–deficient MRL-Fas^lpr kidneys. As anticipated, MCP-1 transcripts were not detected in the MCP-1–deficient MRL-Fas^lpr renal cortex. On the other hand, MCP-3 and MCP-5 transcripts were detected in MCP-1–deficient MRL-Fas^lpr mice (Fig. 8). However, amounts of MCP-3 and MCP-5 transcripts were 50% lower in MCP-1–deficient versus MCP-1–intact MRL-Fas^lprkidneys (P < 0.005; Fig. 8). Thus, MRL-Fas^lprmice lacking MCP-1 have reduced intrarenal productions of ligands (MCP-3, MCP-5) that bind to CCR2.

We previously established that CSF-1 and IFN-γ transcripts that are upregulated with advancing renal injury in MRL-Fas^lpr mice are required for autoimmune kidney disease 25,34,35. In MCP-1–deficient MRL-Fas^lpr kidneys, CSF-1 and IFN-γ transcripts were reduced as compared with the MCP-1 intact MRL-Fas^lpr strain (Fig. 8). In contrast, CSF-1 and IFN-γ transcripts in MCP-1–deficient and –intact B6/129 mice with normal kidneys were barely detectable (Fig. 8).

To determine if MCP-1 alters the antibody isotype profile in MRL-Fas^lpr autoimmune disease, we evaluated serum levels of IgGs (total Ig, IgM, IgG1, IgG2a, IgG2b, IgG3) in MCP-1–deficient and –intact MRL-Fas^lpr mice. We did not detect differences in serum IgGs (amount and isotype) in MCP-1–deficient and –intact MRL-Fas^lpr strains (Fig. 9 a). In addition, we did not detect an alteration in the amount and distribution of Igs within MCP-1–deficient and –intact MRL-Fas^lpr kidneys (Fig. 9 b). Similarly, complement (C3) deposition was not reduced in MCP-1–deficient as compared with MCP-1–intact MRL-Fas^lpr glomeruli (data not shown). Thus, MCP-1 does not regulate circulating or kidney-depositing IgGs in MRL-Fas^lpr mice.

In this report, we tested the hypothesis that a specific chemokine, MCP-1, is required for autoimmune tissue injury in MRL-Fas^lpr mice. We now report that MRL-Fas^lpr mice genetically deficient in MCP-1 are partially protected from autoimmune disease, including injury to the kidney, lung, and skin and lymphadenopathy, resulting in a prolonged life span. Protection against tissue injury in MCP-1–deficient MRL-Fas^lpr mice results from a reduced infiltration of leukocytes (macrophages and T cells) toward parenchymal cells that no longer express MCP-1 and a diminution in cytokines known to promote tissue injury. Furthermore, we determined that the accumulation of these cells is a result of recruitment, but not proliferation, toward the parenchymal cells in proportion to MCP-1 expression. Thus, MCP-1 is responsible for recruiting macrophages and T cells into multiple tissues, including the kidney, lung, and lymph nodes, in MRL-Fas^lpr mice, which in turn results in autoimmune tissue destruction.

During inflammation, multiple parenchymal and leukocyte cell types express MCP-1 36. Our challenge was to identify the cell types and locations expressing MCP-1 and determine the impact of cells producing MCP-1 on disease progression in MRL-Fas^lpr mice. We now report that MCP-1 is primarily expressed by parenchymal cells, and not infiltrating cells, in MRL-Fas^lpr mice, beginning before and then increasing during autoimmune disease. We determined that parenchymal epithelial cells express far more MCP-1 than other parenchymal cell types (endothelial cells, smooth muscle cells, and mesangial cells). Furthermore, we established that the accumulation of macrophages and T cells was directly proportional to MCP-1 expression by these parenchymal cells. This is consistent with our previous findings in a rapid, induced form of kidney damage, NSN, consisting of glomerular and tubular pathology 21. In NSN, we noted that MCP-1, abundantly expressed in TECs and barely detected in glomeruli, was responsible for tubular but not glomerular injury. We established that macrophages were recruited toward tubules expressing MCP-1 and after activation released molecules that induced TEC apoptosis 21. By comparison, the importance of MCP-1 in progressive autoimmune disease in MRL-Fas^lpr mice is broader than in NSN. First, MCP-1 is responsible for the accumulation of not only macrophages but also T cells within the kidney (tubules and glomeruli) and other tissues (lungs and lymph nodes) during autoimmune disease in MRL-Fas^lpr mice. For example, massively enlarged lymph nodes, composed of T cells, are characteristic of MCP-1–intact MRL-Fas^lpr mice 37. The most convincing evidence for MCP-1 fostering T cell accumulation is highlighted by the substantial decrease in the incidence of lymphadenopathy and lymph node size (>50% reduced at 8 mo of age) in MCP-1–deficient MRL-Fas^lpr mice. However, the numbers of T cells in the MCP-1–deficient MRL-Fas^lpr lymph nodes are not reduced to normal levels. In addition, MCP-1 is not required for splenomegaly in MRL-Fas^lprspleens, as MCP-1–intact and MCP-1–deficient MRL-Fas^lpr spleens were similar. Thus, MCP-1 is responsible in part for lymphadenopathy, but other molecules, perhaps chemokines, are required for splenomegaly in the MRL-Fas^lpr strain. Other candidates that are likely to be involved in splenomegaly include the β-chemokines macrophage inflammatory protein (MIP)-1α and MIP-1β, reported to foster T cell trafficking (CD4, CD8) into lymph nodes after induced hypersensitivity 38. In addition, we suggest that the broader impact of MCP-1 on kidney injury in the MRL-Fas^lpr mouse versus NSN is related to the progressive accumulation of MCP-1–dependent leukocytes over a longer period of disease manifestation in the MRL-Fas^lpr kidney and other tissues. Furthermore, comparison of MCP-1 immunostaining and mRNA levels in the two models indicates that tubules and glomeruli in MCP-1–intact MRL-Fas^lpr mice (5–6 mo) express higher amounts of MCP-1 than in NSN (7 d). Disease in NSN is mostly limited to the kidney, whereas MRL-Fas^lpr mice have multiple tissues targeted for destruction, each expressing MCP-1. Thus, by gene target–deleting MCP-1 in MRL-Fas^lpr mice, we determined that MCP-1 is responsible for recruiting macrophages and T cells to numerous tissues, each undergoing autoimmune disease.

The accumulation of macrophages and T cells surrounding parenchymal tissue in MRL-Fas^lpr mice results from recruitment and/or proliferation of these leukocytes 26,27,34,35,39. As MCP-1 induces IL-2 production by T cells 14, it is possible that MCP-1 promotes T cell proliferation in the kidney, lung, and lymph nodes. However, as few infiltrating cells are proliferating (PCNA⁺) at sites of MCP-1 expression within the kidney, lung, and enlarged lymph nodes in MRL-Fas^lpr mice, we conclude that the primary action of MCP-1 is to recruit macrophages and T cells into tissues and thereby promote autoimmune disease.

To further support the concept that MCP-1 initiates kidney injury via recruitment and not other immune events, we investigated whether MCP-1 caused an Ig isotype switch and alteration in Ig deposition in the kidney. Elevated Ig levels in MRL-Fas^lpr mice are dependent on T cell 40 and B cell 41,42 events that are independent of MCP-1 in MRL-Fas^lpr mice. In addition, isotypes such as IgG3 compromise glomerular function 43. We did not detect any difference in the serum Ig isotypes nor in the amount or location of Igs in the kidney. It is interesting to note that despite the similarly high levels of serum and glomerular Igs, the glomeruli are better preserved functionally (loss of protein) and structurally in MCP-1–deficient versus MCP-1–intact MRL-Fas^lpr mice. This suggests that, even when the Igs are deposited within the kidney, MCP-1 is required to attract leukocytes into the kidney to initiate glomerular and tubular/interstitial renal disease in MRL-Fas^lpr mice.

Protection from injury in MCP-1–deficient MRL-Fas^lpr kidneys is associated with reduced number of transcripts of the nephritogenic cytokines (IFN-γ and CSF-1) and other chemokines in addition to MCP-1 that bind to the CCR2 receptor (MCP-3 and MCP-5). This could be either directly or indirectly related to the MCP-1 deletion in MRL-Fas^lpr mice. For example, MCP-1 immunomodulation of CD4 cells is related to the release of IFN-γ 14. We suggest that after MCP-1–dependent recruitment of leukocytes into the kidney, a cascade of events triggers the production of IFN-γ, CSF-1, MCP-3, and MCP-5, which each contribute to autoimmune kidney damage. For example, we have previously established that CSF-1, a macrophage growth factor, is responsible for promoting macrophage- and T cell–initiated kidney injury 35, whereas IFN-γ, with more complex actions, either thwarts 27 or fosters 25,44 injury depending on when it is expressed during disease. In addition, MCP-3 and MCP-5 may be responsible for the migration of leukocytes into the kidney 45. We therefore speculate that the increase in MCP-1 is a proximal stimulus responsible for recruiting macrophages and T cells, which in turn are responsible for triggering the production of cytokines and other chemokines that lead to tissue injury.

Although MCP-1 depletion dampens injury in multiple tissues during autoimmune attack, these tissues are not totally protected. The kidneys, lungs, spleen, lymph nodes, and skin do not remain normal. The renal tubule/interstitium, glomeruli, lungs, and lymph nodes in MCP-1–deficient MRL-Fas^lpr mice are still invaded by leukocytes, although in far lower numbers than in MCP-1–intact MRL-Fas^lpr mice. Additionally, the limited number of MCP-1–deficient MRL-Fas^lpr mice that do not survive kidney glomerular and tubular disease and proteinuria is far lower than younger MCP-1–intact MRL-Fas^lpr mice that succumb. In contrast, vascular disease is not prevented in MCP-1–deficient kidneys and lungs. The accumulation of leukocytes around vessels, almost exclusively T cells, is just as abundant in MCP-1–deficient versus –intact MRL-Fas^lpr kidneys and lungs. There are several possible explanations for this finding. Other chemokines may be responsible for recruiting these infiltrating leukocytes. For example, MCP-1 is weakly expressed in the vascular wall and perivascular leukocytes, whereas RANTES (regulated upon activation, normal T cell expressed and secreted) is abundant in both areas (data not shown). In addition, many perivascular leukocytes in the kidney and lung are proliferating (PCNA⁺) and, therefore, perivascular leukocytes may accumulate because of local proliferation. It is worth noting that perivasculitis in the kidney contributes to mortality in wild-type MRL-Fas^lpr mice 45,46 and, therefore, may be at least partially compromising in several tissues in MCP-1–deficient MRL-Fas^lpr mice. Finally, as the accumulation of leukocytes in the renal interstitium is focal in both MCP-1–intact and –deficient MRL-Fas^lpr mice, we suggest that molecules responsible for adhesion, cell activation, proliferation, and death, including integrins, selectins, and cytokines, may contribute to this process. Furthermore, we cannot rule out the possibility that other chemokines, in addition to MCP-1, enhance leukocytic recruitment. Thus, to achieve more complete protection from autoimmune disease in the MRL-Fas^lpr strain, we will have to identify additional therapeutic targets.

In conclusion, we suggest that MCP-1 is a therapeutic target to combat autoimmune/inflammatory diseases triggered by tissue leukocytic invasion. Our data further suggest that eliminating MCP-1 expression does not confer total protection. It is thus critical to identify the other therapeutic targets responsible for leukocyte invasion and expansion to confer total protection.

We wish to thank Dr. Hyung-Jin Yoon for assisting with the construction of the MCP-1–deficient MRL-Fas^lprstrain.

This work was supported in part by National Institutes of Health grants DK-36149 and DK 52369 (to V.R. Kelley) and a research fellowship from the National Kidney Foundation (to G.H. Tesch).