Bacterial CpG-DNA accelerates Alport glomerulosclerosis by inducing an M1 macrophage phenotype and tumor necrosis factor-α-mediated podocyte loss

http://www.kidney-international.org

original article

& 2011 International Society of Nephrology

Bacterial CpG-DNA accelerates Alport glomerulosclerosis by inducing an M1 macrophage phenotype and tumor necrosis factor-a-mediated podocyte loss Mi Ryu1, Onkar P. Kulkarni1, Ewa Radomska1, Nicolai Miosge2, Oliver Gross3 and Hans-Joachim Anders1 1

Nephrologisches Zentrum, Medizinische Poliklinik der LMU, University of Mu¨nchen, Mu¨nchen, Germany; 2Department of Prosthodontics, Georg August University, Go¨ttingen, Germany and 3Department of Nephrology and Rheumatology, Georg August University, Go¨ttingen, Germany

Loss of function mutations in the a3 or a4 chain of type IV collagen cause Alport nephropathy, characterized by progressive glomerulosclerosis. While studying the mechanisms that determine disease progression, we found that the evolution of kidney disease in Col4a3-deficient mice was associated with an influx of immune cell subsets including nonactivated macrophages. This suggested that intrarenal inflammation might accelerate Alport nephropathy. A possible mechanism might be the wellknown enhancement of immune recognition by bacterial products. We found that exposure to bacterial endotoxin from 4 to 6 weeks of age did not affect disease progression, whereas an equipotent dose of cytosine–guanine (CpG)-DNA, a synthetic mimic of bacterial DNA, accelerated all aspects of Alport nephropathy and reduced the overall lifespan of Col4a3-deficient mice. This effect was associated with a significant increase of renal CD11b þ /Ly6Chi macrophages, intrarenal production of inducible nitric oxide synthase, tumor necrosis factor (TNF)-a, interleukin-12, and CXCL10, and loss of podocytes. TNF-a was essential for acceleration of Alport nephropathy, as etanercept (a soluble TNF-a receptor) entirely abrogated the CpG-DNA effect. Thus, systemic exposure to CpG-DNA induces classically activated (M1) macrophages that enhance intrarenal inflammation and disease progression. Hence, factors that modulate the phenotype of renal macrophages can affect the progression of Alport nephropathy and, potentially, other types of chronic kidney diseases. Kidney International (2011) 79, 189–198; doi:10.1038/ki.2010.373; published online 20 October 2010 KEYWORDS: Alport; chronic kidney disease; fibrosis; glomerulosclerosis; macrophages; Toll-like receptors

Correspondence: Hans-Joachim Anders, Nephrologisches Zentrum, Medizinische Poliklinik der LMU, University of Mu¨nchen, Pettenkoferstraße 8a, Mu¨nchen 80336, Germany. E-mail: [email protected] Received 1 February 2010; revised 13 July 2010; accepted 27 July 2010; published online 20 October 2010 Kidney International (2011) 79, 189–198

Alport disease results from genetic mutations in genes that encode for the a3, a4, or a5 chains of type IV collagen (Col4), which affects the proper assembly of glomerular basement membranes.1 The abnormal membrane structure activates adjacent glomerular cells; hence, Alport nephropathy is characterized by an increase in glomerular matrix, podocyte detachment, and subsequent glomerular scarring.2,3 Similar to most other types of chronic kidney disease, later stages of Alport nephropathy are associated with major tubulointerstitial pathology, including mixed interstitial leukocyte infiltrates and interstitial fibrosis.2,4 In Col4a3deficient mice, a genetic model of autosomal recessive Alport syndrome,5 disease progression is also associated with leukocytic cell infiltrates in the glomerular and the interstitial renal compartment.6 In most types of acute and chronic kidney diseases, leukocyte infiltrates have been shown to contribute to disease progression by secreting proinflammatory and profibrotic mediators.7–9 This is evidenced by experimental reduction of renal leukocytes, for example, by irradiation, targeted deletion, specific toxins, or by chemokine antagonism, which was shown to protect from renal pathology and to improve renal function.10–13 These data should also apply to Alport nephropathy, because irradiation and blockade of chemokine receptor CCR1-mediated renal macrophage recruitment both proved to reduce renal fibrosis and prolonged lifespan in Col4a3-deficient mice.14,15 However, blockade of the proinflammatory CC-chemokine, MCP1/CCL2, had no effect on kidney disease and survival of Col4a3-deficient mice,16 although MCP-1/CCL2 blockade or MCP-1/CCL2 deficiency prevents macrophage-mediated renal injury in multiple models of chronic kidney disease.17 We speculated that MCP-1/CCL2 blockade failed to delay Alport nephropathy because chemokine receptor CCR2 þ macrophages are almost absent in kidneys of Col4a3-deficient mice.16 In fact, CCR2 is expressed only by classically activated (M1) macrophages that are readily recruited from the circulating monocytes into sites of intense inflammation.18,19 However, the extent of tissue inflammation seems to be 189

original article

RESULTS Renal leukocyte subsets in Col4a3-deficient mice

CD45 þ leukocytes accumulate in glomeruli and in the tubulointerstitial compartment during the progression of Alport nephropathy in Col4a3-deficient mice (Figure 1a). Hence, we first characterized the intrarenal leukocyte subsets by flow cytometry. Kidneys of Col4a3-deficient mice contained increased numbers of CD45 þ leukocytes from 6 weeks of age (Figure 1b). Gating CD45 þ cells with additional surface markers allowed us to quantify leukocyte subsets in renal cell suspensions. F4/80 þ /CD11c þ and F4/80/CD11c þ dendritic cells, as well as F4/80 þ /CD11c macrophages were the largest leukocyte populations in kidneys of Col4a3-deficient mice (Figure 1b). Almost all the renal macrophages were negative for Ly6C, a marker of proinflammatory macrophages (data not shown). CD3 þ /CD4 þ and CD4 þ /CD8 þ T cells and Ly6Gpositive neutrophils were also present in kidneys of Col4a3deficient mice at 6 weeks (Figure 1b). Thus, the progression of Alport nephropathy is associated with infiltrates of 190

Col4a3 +/+

Col4a3 –/–

15

4 Weeks 6 Weeks WT

*** 10

***

4 2

***

*

*

**

ph ils F4 /8 0+ D C s F4 /8 0– D C M s ac ro ph ag es

tro

ce lls N

eu

T C D 8+

C D 4+

T

ce lls

0 C D 45 +

Total kidney cells (%)

rather low in Alport nephropathy; hence, we speculated that macrophages in Alport kidneys might display rather noninflammatory phenotypes. Some tumors or, in the late phase of wound healing, tissue macrophage phenotypes have been reported to lack CCR2 expression and to display different functional properties, referred to as the alternatively activated (M2) macrophages and immunoregulatory macrophages.20–23 In fact, the phenotype or activation state of renal macrophages is thought to significantly affect the extent of renal inflammation and tissue damage.24 For example, ex vivo primed M1 or M2 macrophages maintained their phenotype after adoptive transfer into mice with adriamycin-induced glomerular sclerosis.25 M1 macrophage infusion aggravated renal damage and dysfunction, whereas M2 macrophage infusion reduced renal pathology and improved renal function. These studies were consistent with our previous finding that macrophage activation toward the M1 phenotype, for example, with Met-RANTES or with bacterial DNA, aggravates immune complex glomerulonephritis.26,27 Therefore, we hypothesized that the activation state of renal macrophages would affect disease progression of Alport nephropathy of Col4a3-deficient mice. We speculated that classically activated macrophages would increase renal inflammation and accelerate the progression of Alport nephropathy. To address this topic, we exposed Col4a3deficient mice to a series of injections with lipopolysaccharide (LPS) or to unmethylated cytosine–guanine (CpG) oligodeoxynucleotides. CpG-DNA was selected because it mimicks the potential of bacterial DNA to selectively activate macrophages and dendritic cells through Toll-like receptor (TLR)9.28 LPS mimicks infection with Gram-negative bacteria and activates macrophages through TLR4. By using this study design, we also addressed the question whether infections affect the progression of Alport nephropathy, for example, by modulating intrarenal inflammation.

M Ryu et al.: CpG-DNA in Alport glomerulosclerosis

Figure 1 | Infiltrating leukocytes in murine Alport nephropathy. (a) Renal sections were obtained from 6-week-old wild-type (WT) and Col4a3-deficient mice and stained for CD45 to identify infiltrating leukocytes. Representative images are shown and glomeruli are indicated by asterisks and CD45 þ cells are indicated by arrows. Note that CD45 þ cells infiltrate glomeruli and the interstitial compartment of Col4a3-deficient mice. Original magnification  200. (b) Renal cell suspensions were prepared for leukocyte flow cytometry from 4- and 6-week-old wild-type and Col4a3-deficient mice, as described in Materials and Methods. Data represent the percentage of all kidney cells±s.e.m. from 6 to 7 mice in each group. *Po0.05, **Po0.01, ***Po0.001 wild-type versus 6-week-old Col4a3-deficient mice. DC, dendritic cells.

leukocytes, especially of macrophages and dendritic cells, which do not display markers of classically activated phenotypes. Systemic exposure to CpG-DNA shortens the lifespan of Col4a3-deficient mice

CpG-DNA and LPS are known triggers of leukocyte activation. CpG-DNA activates murine B cells, macrophages, and dendritic cells through the endosomal TLR9/MyD88 signaling pathway.29 In contrast, LPS activates leukocytes as well as nonimmune cells through TLR4 on the cell surface, which can activate MyD88 as well as TRIF signaling.30 Col4a3-deficient and wild-type mice received seven injections of CpG-DNA or LPS on alternate days starting from 4 weeks of age at a dose of 40 or 10 mg, respectively. These doses were selected because they induced similar plasma levels of interleukin-6 (IL-6) and MCP-1/CCL2 at 6 h after single injection into 4-week-old Col4a3-deficient mice (data not shown). Following up all mice until death revealed that CpG-DNA significantly reduced the lifespan of Col4a3deficient mice, as documented by a mean survival of 46 days versus 76 days in saline-treated Col4a3-deficient mice (Figure 2). In contrast, LPS had no effect on the survival of Kidney International (2011) 79, 189–198

original article

M Ryu et al.: CpG-DNA in Alport glomerulosclerosis

Does the effect of CpG-DNA on the lifespan of Col4a3deficient mice relate to an acceleration of Alport nephropathy? To answer this question, we treated a second cohort of mice and performed a cross sectional analysis of all groups at 6 weeks of age. Renal sections from all groups of mice were stained with Periodic acid-Schiff and the extent of glomerulosclerosis was quantified by a semiquantitative sclerosis score ranging from 0 to 4, as described in the Materials and Methods. In Col4a3-deficient mice, CpG-DNA significantly increased the percentages of glomeruli with a sclerosis score of 3 or 4 and reduced those with a score of 0 and 1, indicating a shift toward more severe glomerular lesions (Figure 3a and c). This was associated with a significant reduction in the number of podocytes per glomerulus (Figure 3b and d). In contrast, LPS did not affect the aforementioned glomerular abnormalities in Col4a3-deficient mice (Figure 3). Neither

0 1 2 3 4

100 80 60

†‡ †‡

40 20

†‡ †‡

0 Saline

LPS Wild type

CpG

Saline

LPS Col4a3 –/–

CpG

25 WT-1-positive cells

CpG-DNA, but not LPS, aggravates Alport nephropathy in Col4a3-deficient mice

CpG-DNA nor LPS injections affected the normal glomerular structure in wild-type mice (Figure 3a). At the ultrastructural level, the characteristic abnormalities of Alport nephropathy, that is, glomerular basement membrane irregularities and podocyte foot process effacement, became evident in Col4a3deficient mice (Figure 4a). Additional extensive accumulation of extracellular matrix and podocyte damage was seen in

Glomerulosclerosis (%)

Col4a3-deficient mice. In addition, CpG-DNA and LPS injections both did not affect the 100% survival in wildtype mice until 150 days of age. Thus, systemic exposure to CpG-DNA, but not to an equipotent dose of LPS, shortens the lifespan of Col4a3-deficient mice.

20

*

15

*

10

*†

5

100 0 WT

Survival (%)

75

Saline LPS Col4a3 –/–

CpG

P = 0.006 50 Col4a3 –/– + saline Col4a3 –/– + LPS Col4a3 –/– + CpG WT + saline WT + LPS WT + CpG

25

0 0

20

WT saline

40

60

80

Col saline

100

Days

Figure 2 | Life span of Col4a3-deficient mice. Wild-type (WT) and Col4a3-deficient mice were treated with saline, lipopolysaccharide (LPS), or cytosine–guanine (CpG)-DNA from 4 to 6 weeks of age, as indicated. Survival time is illustrated as Kaplan–Meier curve. Col LPS

Figure 3 | Renal histopathology in 6-week-old wild-type (WT) and Col4a3-deficient mice. (a) The extent of glomerulosclerosis was evaluated in mice of all groups by using a semiquantitative sclerosis score ranging from 0 (no sclerosis) to 4 (global sclerosis), as described in Materials and Methods. Data represent means±s.e.m. wPo0.05/zPo0.01 versus saline-treated mice of the same genotype. (b) The number of glomerular podocytes was quantified by counting WT-1 þ cells in 15 glomeruli per section. Data represent means±s.e.m., *Po0.05 versus wild type, wPo0.05 versus saline-treated mice of the same genotype. Renal sections of wild-type and Col4a3-deficient mice were stained with periodic acid-Schiff solution (c) or anti-WT-1 (d). The figure includes representative images of single glomeruli from mice of all groups as indicated; original magnification  400. CpG, cytosine–guanine; LPS, lipopolysaccharide. Kidney International (2011) 79, 189–198

Col CpG

WT saline

Col saline

Col LPS

Col CpG

191

original article

M Ryu et al.: CpG-DNA in Alport glomerulosclerosis

Urinary albumin/creatinine

15

WT saline

*

10

*

**

Saline

LPS

5

0

Col saline

WT

WT saline

Col saline

Blood urea nitrogen (mg/dl)

100

CpG

*

80

**

60 40 20 0 WT

Col LPS

Col CpG

Figure 4 | Glomerular ultrastructure in 6-week-old Col4a3deficient mice. Electron microscopy was applied to assess the glomerular ultrastructure in 6-weeks-old wild-type (WT) and Col4a3-deficient mice. (a) Note that lack of Col4a3 is associated with irregular GBM thickness (c) and podocyte foot process effacement ( ), as compared with age-matched wild-type mice (original magnification  25 000). (b) At lower magnification (  3150), changes in glomerular matrix and capillary structure between wild-type and Col4a3-deficient mice become visible. Note that exposure of Col4a3-deficient mice to CpG-DNA caused extensive accumulation of extracellular matrix (*) with major capillary and podocyte abnormalities, that is, glomerulosclerosis. By contrast, LPS exposure rather caused activation of glomerular endothelial cells, evident by tubular extensions into the capillary lumen (c). CpG, cytosine–guanine; GBM, glomerular basement membrane; LPS, lipopolysaccharide.

CpG-DNA-injected Col4a3-deficient mice only (Figure 4b). In contrast, LPS rather caused activation of glomerular endothelial cells, as illustrated by luminal tubular extensions of endothelial cells (Figure 4b). These morphological changes were associated with an increase of albuminuria and blood urea nitrogen levels in CpG-DNA-treated Col4a3-deficient mice (Figure 5). Together, the effect of CpG-DNA on lifespan is associated with an acceleration of Alport nephropathy. CpG-DNA increases the numbers of activated renal macrophages in Col4a3-deficient mice

Among those cells that infiltrate the kidney in Alport nephropathy, only macrophages and dendritic cells express 192

Saline

LPS

CpG

Figure 5 | Proteinuria and blood urea nitrogen in wild-type (WT) and Col4a3-deficient mice. Proteinuria (a) and blood urea nitrogen (b) were determined from 6-week-old wild-type and Col4a3-deficient mice of all groups and calculated as described in Materials and Methods. Data represent means±s.e.m. *Po0.05, **Po0.01 versus wild-type mice. CpG, cytosine–guanine; LPS, lipopolysaccharide.

TLR9.29 Hence, we speculated that CpG-DNA injections may aggravate Alport nephropathy by altering numbers of these cells in kidneys of Col4a3-deficient mice. To address this issue, we performed quantitative flow cytometry of renal cell suspensions from mice of all groups at 6 weeks of age. CpGDNA significantly increased the amount of renal F4/80 þ / CD11c macrophages (and CD8 þ T cells), but not those of renal CD11c þ dendritic cells in Col4a3-deficient mice (Figure 6). This was consistent with increased numbers of interstitial F4/80 þ cells as well as Mac2 þ glomerular macrophages in renal sections of Col4a3-deficient mice, as determined by immunostaining (Figure 7). In contrast, LPS injections did not affect renal leukocyte numbers in Col4a3deficient mice (Figures 6 and 7), and neither CpG-DNA nor LPS injection affected the numbers of renal leukocytes in wild-type mice (data not shown). Did CpG-DNA injections also affect the functional phenotype of intrarenal macrophages? We addressed this question in various ways. We used, for example, flow cytometry to quantify the numbers of Ly6Chi intrarenal macrophages, which represent a classically activated proinflammatory (M1) phenotype.19,23 CpG-DNA injections significantly increased the numbers of intrarenal Ly6Chi-positive macrophages, as compared with LPS- or saline-treated Col4a3-deficient mice (Figure 6b). In summary, CpG-DNA accelerates kidney disease in Kidney International (2011) 79, 189–198

original article

25 20 15 10

profiling and immunostaining support the fact that CpG-DNA accelerates kidney disease in Col4a3-deficient mice by modulating the phenotype of renal macrophages.

Saline LPS CpG WT

**

*

6 4 2 0

es ag ph

ac ro M

F4

/8 0

–D

+D

10 CD11b+Ly6Chi (%)

C s

s C

ils /8 0 F4

N

eu

T

tro

ph

ce l

ls 8+ D C

C

D

4+

C

T

D

ce l

45

ls

** +

Total kidney cells (%)

M Ryu et al.: CpG-DNA in Alport glomerulosclerosis

*

8 6 4 2 0 WT

Saline

LPS

CpG

Figure 6 | Cytosine–guanine (CpG)-DNA alters renal leukocyte infiltrate in 6-week-old Col4a3-deficient mice. (a) Renal cell suspensions were prepared for leukocyte flow cytometry from 6-week-old wild-type (WT) and Col4a3-deficient mice of all groups. The marker CD11c identified dendritic cells (DCs), macrophages were F4/80 þ /CD11c negative. Data represent the percentage of all kidney cells±s.e.m. from 5 to 7 mice in each group. *Po0.05, **Po0.01 CpG-DNA versus saline-treated Col4a3-deficient mice. (b) The percentage of Ly6Chi cells gated on CD11b þ population in kidney. *Po0.05 CpG-DNA versus all groups.

The effect of CpG-DNA on Alport nephropathy is mediated by TNF-a

Because the effects of CpG-DNA on kidney disease in Col4a3deficient mice were associated with a marked induction of TNF-a, we questioned whether this accounts for the aggravation of Alport nephropathy. Thus, we treated additional Col4a3-deficient mice with CpG-DNA together with or without the TNF-a antagonist, etanercept. Etanercept completely abrogated the effect of CpG-DNA on the acceleration of glomerulosclerosis, as glomerular sclerosis scores were similar or even lower as compared with salinetreated Col4a3-deficient mice (Figure 9a and b). Similarly, etanercept significantly increased the numbers of WT-1 þ podocytes in glomerular tufts (Figure 9c and d) and significantly reduced the albumin–creatinine ratio in CpGDNA-treated Col4a3-deficient mice (Figure 10a). The premature mortality of CpG-DNA-treated Col4a3-deficient mice was also improved from a mean lethality at 46 days to 77 days by etanercept treatment, which was identical to saline-treated Col4a3-deficient mice (Figure 10b). Thus, CpG-DNA accelerates Alport nephropathy in Col4a3-deficient mice by activating renal macrophages to secrete TNF-a. DISCUSSION

Col4a3-deficient mice in association with increased numbers of classically activated renal macrophages. CpG-DNA and LPS differently modulate the renal mRNA expression of macrophage markers

As another way to characterize macrophage polarization, we quantified a number of macrophage activation markers by real-time PCR using renal RNA extracts from mice of all treatment groups. In Col4a3-deficient mice, CpG-DNA injection increased renal mRNA levels of the following M1 (classically activated) macrophage markers: tumor necrosis factor (TNF-a), iNOS, IL-12, and CXCL10 (Figure 8a). All of these markers were rather downregulated in LPS-treated Col4a3-deficient mice. CpG-DNA injections also induced some of the markers that refer to M2 (alternatively activated) macrophages, such as arginine-1, TGF-b, and Fizz-1 (Figure 8b). All aforementioned results were confirmed by real-time PCR carried out on mRNA isolates from CD11b þ cells that we isolated using magnetic beads from mice of all groups (see Supplementary Figure S1 online). The predominant intrarenal induction of, for example, TNF-a and iNOS mRNA by CpG-DNA was consistent with an increased glomerular (and interstitial) staining intensity by anti-iNOS IgG and anti-TNF IgG on renal sections of CpG-DNA-treated, but not of saline- or LPS-treated Col4a3-deficient mice (Figure 7). Interestingly, TNF-a immunostaining was also positive on glomerular podocytes (Figure 7). Taken together, mRNA Kidney International (2011) 79, 189–198

Col4a3-deficient mice represent a suitable model to study the role of environmental factors in Alport syndrome, because they replicate the genetic and phenotypical characteristics of the human Alport syndrome.5 Here, we show that exposure to CpG-DNA, a synthetic mimic of bacterial DNA ligating TLR9, increased numbers and activation state of intrarenal macrophages, a phenomenon that was associated with a robust acceleration of disease progression. These data suggest first, that the natural course of Alport nephropathy involves a low degree of intrarenal inflammation associated with rather alternatively-activated macrophage infiltrates. Second, environmental stimuli such as microbial DNA can drive disease progression by enhancing inflammation by modulating intrarenal macrophage numbers as well as their phenotype. Third, TNF-a induction is a major element of CpG-DNAinduced acceleration of Alport nephropathy. The phenotypical diversity of tissue macrophages has evolved as a new research field, but classifying macrophage phenotypes remains a matter of debate.19,20,22 Classically activated (M1) macrophages were defined as an IL-12, IL-23, and major histocompatibility complex II-expressing phenotype on interferon-g and TNF-a stimulation.18 In contrast, IL-4, IL-13, or IL-10 stimulation induces alternatively activated (M2) macrophages that do not produce the aforementioned proinflammatory mediators but rather express IL-10, arginase-1, Fizz-1, and the mannose receptor-1.18 Alternatively activated macrophages seem to predominate the kidneys of Col4a3-deficient mice and may 193

original article

F4/80

iNOS

TNF

Col4a3 CpG

Col4a3 LPS

Col4a3 saline

WT saline

MAC 2

M Ryu et al.: CpG-DNA in Alport glomerulosclerosis

Figure 7 | Immunostaining of renal sections. Renal sections of wild-type (WT) and Col4a3-deficient mice were stained with appropriate antibodies as indicated. The figure includes representative images of mice from all groups, original magnification  200 or  400.

194

Kidney International (2011) 79, 189–198

original article

50

CpG Glomerulosclerosis (%)

3.5 3.0 2.5 2.0 1.5 1.0 0.5

40

20

0 1 2 3 4

† †‡

30

‡

20

‡

†‡

WT-1- positive cells

LPS

4.0

0 TNF-α

iNos

IL-12

50 45 40 35 30 25 4

CXCL10

LPS CpG

5

Saline

Col4a3 CpG

IL-6

CpG +etanercept

Col4a3 CpG+etanercept

CCL5

CpG

Col4a3 CpG

CCL2

10

0 Saline

0.0

15

‡

10

Col4a3 CpG+etanercept

Kidney mRNA ratio relative to saline in Col4a3-deficient mice (6 weeks)

Kidney mRNA ratio relative to saline in Col4a3-deficient mice (6 weeks)

M Ryu et al.: CpG-DNA in Alport glomerulosclerosis

CpG

CpG +etanercept

3 2 1 0 Mrc-1

Arg-1

TGF-β1

CCL22

Fizz-1

Ym1

IL-10

Figure 8 | Renal mRNA expression of macrophage phenotype markers in Col4a3-deficient mice. Total kidney mRNA was prepared from mice of all groups for real-time reverse transcriptase PCR. Markers of classically activated (M1) (a) or alternatively activated (M2) (b) macrophages were quantified and were calculated as ratio per respective 18s rRNA expression. The graph shows the mRNA levels of lipopolysaccharide (LPS)- (white bars) or cytosine–guanine (CpG)-DNA-injected mice (black bars) relative to the levels of saline-treated mice to indicate induction or suppression of each marker.

contribute to remodeling of the extracellular matrix. For example, macrophage metalloelastase was shown to contribute to the digestive damage of the glomerular basement membrane in Col4a3-deficient mice.31 In contrast, we found only very few CD11b þ /Ly6Chi macrophages in kidneys of Col4a3-deficient mice at 6 weeks of age. This may explain why blocking the CC-chemokine, MCP-1/CCL2, does not modulate murine Alport nephropathy,16 although MCP-1/ CCL2 blockade reduces the progression of most types of experimental kidney diseases by preventing the recruitment and activation of CCR2 þ M1 macrophages.18,24 Our data show that systemic exposure to CpG-DNA is more potent than LPS to modulate intrarenal macrophages toward a proinflammatory phenotype, as evidenced by increased numbers of CD11b þ /Ly6Chi macrophages and renal mRNA expression of M1 macrophage markers in CpG-DNA-treated Col4a3-deficient mice. These results are consistent with the effects of CpG-DNA on cultured macrophages23 and the effects of CpG-DNA injections in murine models of immune complex glomerulonephritis.27,32 Although such models involve effects of CpG-DNA on humoral immunity that compromise conclusions on the role of glomerular macrophage phenotypes, a recent study has documented the role of activated macrophages in focal segmental glomerulosclerosis. Wang et al.25 observed that injection of ex vivo CpG-DNAprimed macrophages accelerated adriamycin nephropathy, Kidney International (2011) 79, 189–198

Figure 9 | Effect of tumor necrosis factor-a (TNF-a) blockade with etanercept on cytosine–guanine (CpG)-DNA-induced aggravation of Alport nephropathy. (a) The extent of glomerulosclerosis was evaluated in mice of all groups by using a semiquantitative sclerosis score ranging from 0 (no sclerosis) to 4 (global sclerosis), as described in Materials and Methods. Data represent means±s.e.m. wPo0.05 versus saline-treated/zPo0.05 versus CpG-DNA-treated Col4a3-deficient mice. (b) Renal sections of Col4a3-deficient mice were stained with periodic acid-Schiff solution. The figure illustrates representative images of single glomeruli from mice of all groups as indicated; original magnification  400. (c) The number of glomerular podocytes was quantified by counting WT-1 þ cells in 15 glomeruli per section. Data represent means±s.e.m., wPo0.05 versus CpG-DNAtreated Col4a3-deficient mice. (d) Renal sections of Col4a3deficient mice were stained with anti-WT-1. The figure illustrates representative images of single glomeruli from mice of all groups as indicated; original magnification  400.

while injection of resting macrophages had no effect. Our data are consistent with these results, although we had injected CpG-DNA directly into the mice, thereby mimicking an environmental and systemic trigger of macrophage activation. Interestingly, CpG-DNA also induced M2 macrophage markers such as arginase-1, TGF-b, mannose receptor-1, Fizz-1, and IL-10. Macrophage diversity is obviously more complex in tissue environments and is likely to extend beyond the M1/M2 paradigm.19 Irrespective of such classifying categories, our data show that TNF-a induction is functionally crucial during CpG-DNA-accelerated Alport nephropathy. TNF-a release by classically activated macrophages accounts for inflammation and tissue damage in chronic rheumatic diseases, psoriasis, and Crohn’s disease. Our finding that CpG-DNA increased renal TNF-a 195

Urinary albumin/creatinine

original article

M Ryu et al.: CpG-DNA in Alport glomerulosclerosis

20 P <0.01

P <0.05

15 10 5 0 WT

CpG CpG+etanercept

100

Survival (%)

75 P = 0.002

50

Col4a3 –/– + saline

25

CpG-DNA CpG-DNA+etanercept

0 0

20

40

60

80

100

Days

Figure 10 | Effect of tumor necrosis factor-a (TNF-a) blockade on proteinuria and lifespan of cytosine–guanine (CpG)-DNAaccelerated Alport nephropathy. (a) Proteinuria was determined in 6-week-old wild-type (WT) and Col4a3-deficient mice, the latter being treated either with CpG-DNA only or CpGDNA and etanercept. Data represent means of urinary albumin/ creatinine ratios±s.e.m. (b) CpG-DNA-treated Col4a3-deficient mice were injected with etanercept from 4 to 6 weeks of age as indicated. Survival time is illustrated as Kaplan–Meier curve.

expression and that TNF-a blockade prevents the acceleration of Alport nephropathy is consistent with the protective effect of TNF-a inhibitors in renal disease models that are associated with the presence of classically activated macrophages.33–36 In addition, macrophage-derived TNF mediates podocyte activation and dedifferentiation in vitro.37 This should explain our finding of TNF-a expression in glomeruli and podocyte loss after CpG-DNA injection and the protection from CpG-DNA-induced acceleration of glomerulosclerosis after TNF-a blockade. In contrast, LPS can also trigger macrophage TNF-a release, for example, in the initial phase of Gram-negative sepsis, but the induction of TLR4 signaling inhibitors account for LPS tolerance after repetitive LPS exposure.38 In summary, the progression of murine Alport nephropathy is associated with macrophage infiltrates of noninflammatory phenotypes. Exposure to bacterial CpG-DNA accelerates podocyte loss and glomerulosclerosis in Col4a3deficient mice by increasing the numbers of classically activated macrophages that release proinflammatory mediators such as TNF-a. We conclude that environmental stimuli that modulate renal macrophages toward a proinflammatory phenotype accelerate the progression of glomerulosclerosis such as that in Alport nephropathy. 196

MATERIALS AND METHODS Animal studies Col4a3-deficient and wild-type littermate mice in identical SvJ129 genetic backgrounds were selected from the available Alport disease models because of their consistent phenotype leading to uremic death at B10 weeks of age.5 All mice were bred under specificpathogen-free housing conditions and genotyped as described.39 At the age of 4 weeks, groups of Col4a3-deficient and wild-type mice (n ¼ 12) started to receive seven intraperitoneal injections on alternate days of 100 ml normal saline (vehicle), 10 mg of ultrapure LPS (Invivogen, San Diego, CA) in vehicle, or 40 mg endotoxin-free CpG-phosphothioate oligonucleotide 1668 (50 -TCGATGACGTTC CTGATGCT-30 ; TIB-Molbiol, Berlin, Germany) in vehicle. The doses of both TLR agonists were selected because they induced identical serum levels of IL-6 and MCP-1/CCL2, at 6 h after intraperitoneal injection in Col4a3-deficient and wild-type mice. One cohort of mice was bled and killed at 6 weeks of age. Another cohort was followed until death in order to determine lifespan. An additional cohort of CpG-DNA-treated Col4a3-deficient mice (n ¼ 6–9) was intraperitoneally injected with either saline or 100 mg etanercept (Wyeth, Taplow, Maidenhead, UK) every other day from week 4 to 6 to block TNF-a.33,36 All experimental procedures were performed according to the German animal care and ethics legislation and were approved by the local government authorities. Histopathological evaluation Parts of the kidneys were fixed in 10% formalin in PBS and embedded in paraffin. Sections (2 mm) were stained with periodic acid-Schiff reagent. Glomerular sclerotic lesions were assessed using a semiquantitative score by a blinded observer as follows: 0 ¼ no lesion, 1 ¼ o25% sclerotic, 2 ¼ 25–49% sclerotic, 3 ¼ 50–74% sclerotic, 4 ¼ 75–100% sclerotic, respectively. A total of 15 glomeruli were analyzed per section.39 For ultrastructural analysis, kidney specimens of 6-week-old treated and control mice were dissected and fixed in 0.1 mol/l PBS buffer (pH 7.4) supplemented with 0.4% paraformaldehyde and 2% glutaraldehyde. Polymerization (using fresh Epon resin) was carried out at 60 1C for 24 h as described.40 Immunohistochemistry All immunohistological studies were performed on paraffinembedded sections as described.39 The following antibodies were used as primary antibodies: rat anti-Mac2 (glomerular macrophages; Cederlane, Ontario, Canada; 1:50), rat anti-F4/80 (interstitial macrophages; Serotec, Oxford, UK; 1:50), rat anti-CD45 (leukocytes; Dianova, Hamburg, Germany; 1:25), rat anti-Wilms tumor (WT)-1 (podocytes; Santa Cruz, Santa Cruz, CA; 1:200), goat anti-mTNF-a (R&D Systems, Minneapolis, MN; 1:800). Glomerular cell counts evaluated only cells of the tuft from 15 glomeruli, and interstitial cells were counted in 15 high-power fields (400  ). Renal function parameters Urine albumin/creatinine ratios and blood urea nitrogen levels were determined as previously described.26,41 Blood and urine samples were collected from each animal at the end of the study. Urine albumin concentration was measured using a commercial mouse albumin ELISA quantitation kit (Bethyl Laboratories, Montgomery, TX) and urinary creatinine concentrations were determined according to the Jaffe´ method (DiaSys Diagnostic Systems, Holzheim, Germany). The values of urinary albumin and creatinine Kidney International (2011) 79, 189–198

original article

M Ryu et al.: CpG-DNA in Alport glomerulosclerosis

were converted to the same unit and albumin value (mg/dl) was divided by creatinine value (mg/dl). Blood urea nitrogen in plasma was measured using a commercially available kit (DiaSys Diagnostic Systems). Isolation and flow cytometry of renal leukocytes Renal cell suspensions were prepared and stained for flow cytometry using a FACScalibur machine (BD, Heidelberg, Germany) as previously described.42 In brief, kidneys were mechanically disrupted and incubated in 1  Hanks’ balanced salt solution containing 1 mg/ml collagenase type I and 0.1 mg/ml deoxyribonuclease type I (Sigma-Aldrich, Steinheim, Germany) for 20 min at 37 1C. After washes, tissues were incubated in 5 ml of 2 mmol/l EDTA in 1  Hanks’ balanced salt solution (without calcium and magnesium) for 20 min at 37 1C. The supernatant containing isolated cells was kept on ice, and for a second enzyme step, the remaining pellet was incubated in 5 ml of 1 mg/ml collagenase I in 1  Hanks’ balanced salt solution for 20 min at 37 1C. The suspension was subsequently passed through a 19- and 27-gauge needle, and pooled with the first supernatant from the EDTA incubation. Cells were filtered through a 70-mm cell strainer (BD) and washed twice in PBS. All washing steps were performed in FACS buffer (PBS containing 0.2% of BSA and 0.1% of NaAzide). Renal leukocytes were characterized by using the following antibodies: phycoerythrin-conjugated anti-CD45 (clone 30-F11, BD), phycoerythrin- or allophyocyanin-conjugated anti-CD11b (clone M1/70, BD), FITC-conjugated anti-CD3e (clone 145-2C11, BD), antiCD11c (clone N418, Serotec), anti-Ly6G (clone 1A8), anti-Ly6C

(clone AL-21, BD), allophyocyanin-conjugated anti-CD4 (clone RM4-5, BD), F4/80 (clone CL:A3-1, Serotec), anti-Ly6G and C (clone RB6-8C5, BD), and PerCp-conjugated anti-CD8a (clone 536.7, BD). For some studies renal CD11b þ macrophages were isolated from renal cell suspensions by using microbead-conjugated antibodies (Miltenyi Biotech, Bergisch-Gladbach, Germany). Real-time quantitative reverse transcriptase PCR Total RNA was isolated from kidney or cells using an RNA extraction kit (QIAGEN, Hilden, Germany) according to the manufacturer’s instructions, and RNA quality was assessed using agarose gels. After isolation of RNA, cDNA was generated using reverse transcriptase (Superscript II; Invitrogen, Carlsbad, CA). SYBR Green dye detection system was used for quantitative real-time PCR on Light Cycler 480 (Roche, Mannheim, Germany) using 18s rRNA as a housekeeper gene. Gene-specific primers (300 nM, Metabion, Martinsried, Germany) were used as listed in Table 1. Controls consisting of ddH2O were negative for target and housekeeper genes. To reduce the risk of false-positive Cp, the high confidence algorithm was used. The melting curves profiles were analyzed for every sample to detect eventual unspecific products or primer dimers. Statistical analysis Paired Student’s t-test was used for the comparison of single groups. A value of Po0.05 was considered to indicate statistical significance. Survival curves were compared by Kaplan–Meier analysis by logrank two-tailed testing. DISCLOSURE

All the authors declared no competing interests.

Table 1 | Oligonucleotide primers used for SYBR-Green RT-PCR Gene

Accession

Sequence (50 -30 )

18s

NR_003278

Arg1

NM_007482

Ccl2

NM_011333

Ccl5

NM_013653

Ccl22

NM_009137

Cxcl10

NM_021274

Forward: GCAATTATTCCCCATGAACG Reverse: AGGGCCTCACTAAACCATCC Forward: AGAGATTATCGGAGCGCCTT Reverse: TTTTTCCAGCAGACCAGCTT Forward: CCTGCTGTTCACAGTTGCC Reverse: ATTGGGATCATCTTGCTGGT Forward: CCACTTCTTCTCTGGGTTGG Reverse: GTGCCCACGTCAAGGAGTAT Forward: TCTGGACCTCAAAATCCTGC Reverse: TGGAGTAGCTTCTTCACCCA Forward: GGCTGGTCACCTTTCAGAAG Reverse: ATGGATGGACAGCAGAGAGC Forward: CCCTTCTCATCTGCATCTCC Reverse: CTGGATTGGCAAGAAGTTCC Forward: TGATGCACTTGCAGAAAACA Reverse: ACCAGAGGAAATTTTCAATAGGC Forward: ATCGATTTCTCCCCTGTGAA Reverse: TGTCAAATTCATTCATGGCCT Forward: AGCAGTAGCAGTTCCCCTGA Reverse: AGTCCCTTTGGTCCAGTGTG Forward: ATATATAAACAAGAATGGTGGGCAGT Reverse: TCCATCCAAATGAATTTCTTATCC Forward: TTCTGTGCTGTCCCAGTGAG Reverse: TGAAGAAAACCCCTTGTGCT Forward: GGAGAGCCCTGGATACCAAC Reverse: CAACCCAGGTCCTTCCTAAA Forward: CCACCACGCTCTTCTGTCTAC Reverse: AGGGTCTGGGCCATAGAACT Forward: TCTGGGTACAAGATCCCTGAA Reverse: TTTCTCCAGTGTAGCCATCCTT

Fizz-1 (Retnla) NM_020509 IL6

NM_031168

IL10

NM_010548

IL12 (Il12p40) NM_008352 Mrc1

NM_008625

Nos2 (iNos)

NM_010927

Tgf-b1

NM_011577

Tnf-a

NM_013693

Ym1 (Chi3l3)

NM_009892

Abbreviation: RT-PCR, reverse transcriptase PCR. Kidney International (2011) 79, 189–198

ACKNOWLEDGMENTS

The work was supported by the Association pour l‘Information et la Recherche sur les maladies re´nale Ge´ne´tiques, France. Parts of this project were prepared as doctoral thesis at the Medical Faculty, University of Munich, Germany, by MR. The authors thank Dan Draganovic for his expert technical assistance. SUPPLEMENTARY MATERIAL Figure S1. mRNA profile of renal CD11b þ macrophage in Col4a3deficient mice. Supplementary material is linked to the online version of the paper at http://www.nature.com/ki REFERENCES 1.

2. 3.

4. 5.

6.

7.

Hudson BG, Reeders ST, Tryggvason K. Type IV collagen: structure, gene organization, and role in human diseases. Molecular basis of Goodpasture and Alport syndromes and diffuse leiomyomatosis. J Biol Chem 1993; 268: 26033–26036. Heidet L, Gubler MC. The renal lesions of Alport syndrome. J Am Soc Nephrol 2009; 20: 1210–1215. Hudson BG, Tryggvason K, Sundaramoorthy M et al. Alport’s syndrome, Goodpasture’s syndrome, and type IV collagen. N Engl J Med 2003; 348: 2543–2556. Jedlicka J, Soleiman A, Draganovici D et al. Interstitial inflammation in Alport syndrome. Hum Pathol 2009; 41: 582–593. Cosgrove D, Kalluri R, Miner JH et al. Choosing a mouse model to study the molecular pathobiology of Alport glomerulonephritis. Kidney Int 2007; 71: 615–618. Rodgers KD, Rao V, Meehan DT et al. Monocytes may promote myofibroblast accumulation and apoptosis in Alport renal fibrosis. Kidney Int 2003; 63: 1338–1355. Eardley KS, Kubal C, Zehnder D et al. The role of capillary density, macrophage infiltration and interstitial scarring in the pathogenesis of human chronic kidney disease. Kidney Int 2008; 74: 495–504.

197

original article

8. Wilson HM, Walbaum D, Rees AJ. Macrophages and the kidney. Curr Opin Nephrol Hypertens 2004; 13: 285–290. 9. Boor P, Sebekova K, Ostendorf T et al. Treatment targets in renal fibrosis. Nephrol Dial Transplant 2007; 22: 3391–3407. 10. Kitching AR, Holdsworth SR, Hickey MJ. Targeting leukocytes in immune glomerular diseases. Curr Med Chem 2008; 15: 448–458. 11. Ferenbach D, Kluth DC, Hughes J. Inflammatory cells in renal injury and repair. Semin Nephrol 2007; 27: 250–259. 12. Sean Eardley K, Cockwell P. Macrophages and progressive tubulointerstitial disease. Kidney Int 2005; 68: 437–455. 13. Anders HJ, Ninichuk V, Schlondorff D. Progression of kidney disease: blocking leukocyte recruitment with chemokine receptor CCR1 antagonists. Kidney Int 2006; 69: 29–32. 14. Katayama K, Kawano M, Naito I et al. Irradiation prolongs survival of Alport mice. J Am Soc Nephrol 2008; 19: 1692–1700. 15. Ninichuk V, Gross O, Reichel C et al. Delayed chemokine receptor 1 blockade prolongs survival in collagen 4A3-deficient mice with Alport disease. J Am Soc Nephrol 2005; 16: 977–985. 16. Clauss S, Gross O, Kulkarni O et al. Ccl2/Mcp-1 blockade reduces glomerular and interstitial macrophages but does not ameliorate renal pathology in collagen4A3-deficient mice with autosomal recessive Alport nephropathy. J Pathol 2009; 218: 40–47. 17. Vielhauer V, Anders HJ. Chemokines and chemokine receptors as therapeutic targets in chronic kidney disease. Front Biosci (Schol Ed) 2009; 1: 1–12. 18. Mantovani A, Sica A, Sozzani S et al. The chemokine system in diverse forms of macrophage activation and polarization. Trends Immunol 2004; 25: 677–686. 19. Mosser DM, Edwards JP. Exploring the full spectrum of macrophage activation. Nat Rev Immunol 2008; 8: 958–969. 20. Varin A, Gordon S. Alternative activation of macrophages: Immune function and cellular biology. Immunobiology 2009; 214: 630–641. 21. Gabrilovich DI, Nagaraj S. Myeloid-derived suppressor cells as regulators of the immune system. Nat Rev Immunol 2009; 9: 162–174. 22. Pollard JW. Trophic macrophages in development and disease. Nat Rev Immunol 2009; 9: 259–270. 23. Mantovani A, Sica A, Locati M. Macrophage polarization comes of age. Immunity 2005; 23: 344–346. 24. Ricardo SD, van Goor H, Eddy AA. Macrophage diversity in renal injury and repair. J Clin Invest 2008; 118: 3522–3530. 25. Wang Y, Wang YP, Zheng G et al. Ex vivo programmed macrophages ameliorate experimental chronic inflammatory renal disease. Kidney Int 2007; 72: 290–299. 26. Anders HJ, Frink M, Linde Y et al. CC chemokine ligand 5/RANTES chemokine antagonists aggravate glomerulonephritis despite reduction of glomerular leukocyte infiltration. J Immunol 2003; 170: 5658–5666. 27. Anders HJ, Banas B, Linde Y et al. Bacterial CpG-DNA aggravates immune complex glomerulonephritis: role of TLR9-mediated expression of

198

M Ryu et al.: CpG-DNA in Alport glomerulosclerosis

28. 29. 30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

chemokines and chemokine receptors. J Am Soc Nephrol 2003; 14: 317–326. Anders HJ. Innate pathogen recognition in the kidney: toll-like receptors, NOD-like receptors, and RIG-like helicases. Kidney Int 2007; 72: 1051–1056. Hemmi H, Takeuchi O, Kawai T et al. A Toll-like receptor recognizes bacterial DNA. Nature 2000; 408: 740–745. Poltorak A, He X, Smirnova I et al. Defective LPS signaling in C3H/HeJ and C57BL/10ScCr mice: mutations in Tlr4 gene. Science 1998; 282: 2085–2088. Rao VH, Meehan DT, Delimont D et al. Role for macrophage metalloelastase in glomerular basement membrane damage associated with alport syndrome. Am J Pathol 2006; 169: 32–46. Anders HJ, Vielhauer V, Eis V et al. Activation of toll-like receptor-9 induces progression of renal disease in MRL-Fas(lpr) mice. FASEB J 2004; 18: 534–536. Karkar AM, Smith J, Pusey CD. Prevention and treatment of experimental crescentic glomerulonephritis by blocking tumour necrosis factor-alpha. Nephrol Dial Transplant 2001; 16: 518–524. Choi DE, Jeong JY, Lim BJ et al. Pretreatment with the tumor nerosis factor-alpha blocker etanercept attenuated ischemia-reperfusion renal injury. Transplant Proc 2009; 41: 3590–3596. Elmarakby AA, Quigley JE, Pollock DM et al. Tumor necrosis factor alpha blockade increases renal Cyp2c23 expression and slows the progression of renal damage in salt-sensitive hypertension. Hypertension 2006; 47: 557–562. Lech M, Avila-Ferrufino A, Allam R et al. Resident dendritic cells prevent postischemic acute renal failure by help of single Ig IL-1 receptor-related protein. J Immunol 2009; 183: 4109–4118. Ikezumi Y, Suzuki T, Karasawa T et al. Activated macrophages downregulate podocyte nephrin and podocin expression via stress-activated protein kinases. Biochem Biophys Res Commun 2008; 376: 706–711. Biswas SK, Lopez-Collazo E. Endotoxin tolerance: new mechanisms, molecules and clinical significance. Trends Immunol 2009; 30: 475–487. Ninichuk V, Gross O, Segerer S et al. Multipotent mesenchymal stem cells reduce interstitial fibrosis but do not delay progression of chronic kidney disease in collagen4A3-deficient mice. Kidney Int 2006; 70: 121–129. Miosge N, Klenczar C, Herken R et al. Organization of the myotendinous junction is dependent on the presence of alpha7beta1 integrin. Lab Invest 1999; 79: 1591–1599. Sayyed SG, Hagele H, Kulkarni OP et al. Podocytes produce homeostatic chemokine stromal cell-derived factor-1/CXCL12, which contributes to glomerulosclerosis, podocyte loss and albuminuria in a mouse model of type 2 diabetes. Diabetologia 2009; 52: 2445–2454. Vielhauer V, Anders HJ, Perez de Lema G et al. Phenotyping renal leukocyte subsets by four-color flow cytometry: characterization of chemokine receptor expression. Nephron Exp Nephrol 2003; 93: e63.

Kidney International (2011) 79, 189–198