No difference in renal injury and fibrosis between wild-type and NOD1/NOD2 double knockout mice with chronic kidney disease induced by ureteral obstruction

NOD1 and NOD2 are members of the cytoplasmic PRR family of NLRs. PRRs are important in mediating a rapid response to pathogens via recognition of several highly conserved pathogen- associated molecular patterns (PAMP). In addition to PAMPs various (endogenous) damage-associated molecular patterns (DAMP) or stress signals have been identified that can initiate sterile inflammation [1]. Upon renal injury DAMPs are released such as biglycan, high-mobility group box 1 (HMGB1), and hyaluronic acid that can signal via TLRs and NLRs [2‐5]. NOD1 and NOD2 detect specific substructures from bacterial peptidoglycan (PGN). NOD1 senses Gram⁻-derived PGN containing Tri-DAP [6, 7], while NOD2 senses Gram⁻- and Gram⁺-derived PGN containing MDP [8, 9]. In line, we found that NOD1/2 are involved in the development of acute renal disease during septic shock induced by bacterial components [10]. In addition to bacterial structures, Sabbah et al. reported the activation of NOD2 by single-stranded RNA viruses [11]. Recently, activation of NOD1 and NOD2 by the non-pathogenic derived cell permeable small molecule DMXAA was reported [12]. As far as we know, no endogenous DAMPs for NOD1 and NOD2 are described. However, based on their structural and functional similarities with other NLR family members and TLRs it could be speculated that NOD1 and NOD2 are also activated by currently unknown endogenous ligands. In line with this reasoning, Shigeoka et al. showed that NOD1 and NOD2 participate in acute renal ischemia reperfusion injury (IRI), suggesting that these receptors are able to respond to endogenous ligands released upon IRI [13]. NOD1 is widely expressed in many cell types and organs including the tubular epithelial cells (TEC) in human and mouse kidney [13‐16]. NOD2 is expressed on murine TECs, mesangial cells, podocytes and on human TECs and glomerular endothelial cells [13, 17, 18]. Given the expression of NOD1 and NOD2 in the kidney and more specific in TEC and the fact that NOD1/2 play a deleterious role in acute kidney disease could suggest that these PRRs contribute to the pathogenesis of chronic renal damage as well. PRRs like NLRP3 and TLR4 have already been shown to play a role in obstructive nephropathy [19‐21]. However, nothing is known about the role of NOD1 and NOD2 in inflammation and fibrosis during obstructive nephropathy. In the present study we therefore investigated the role of NOD1 and NOD2 in a model of obstructive nephropathy induced by ureteral obstruction.

The role of NOD1 and NOD2 in CKD was investigated using the mouse model UUO. First we analysed the expression of NOD1 and NOD2 mRNA in WT kidney at several time points after UUO. We found expression of both transcripts in the kidney (Fig. 1), which were not altered during the development of obstructive nephropathy. Tubular injury, as assessed by scoring PAS-D-stained kidney sections, increased markedly after UUO with a similar degree of damage in WT and NOD1/2 DKO at all time points examined (Fig. 2a, c). In line with the injury score, there were no differences in KIM-1 expression between the WT and the KO mice at all time points (Fig. 2b). KIM-1 was declined at day 14 of UUO possibly as an adaptation to prevent or slow down KIM-1 mediated chronic inflammation and renal fibrosis, as described previously [23]. Tubulointerstitial injury in obstructed kidneys can result in an imbalance between TEC apoptosis and proliferation. Apoptosis and proliferation of TEC was increased at all investigated time points after obstruction (Fig. 3). However, no difference between WT and NOD1/2 DKO mice was observed. Fibrosis was determined by collagen type I (Fig. 4a) and total collagen deposition (Additional file 2: Figure S2). In both WT and NOD1/2 DKO obstructed kidneys fibrosis increased progressively, however no difference between the WT and NOD1/2 DKO mice was observed. Next we analysed the amount of myofibroblasts by αSMA immunohistochemistry (Fig. 4b). In line with tubular injury and fibrosis, the amount of myofibroblasts increased after UUO. Although myofibroblast accumulation was lower in NOD1/2 DKO mice 3 days following ureteral obstruction compared to WT mice, no differences were found after 7 and 14 days (Fig. 4b). One of the early events in progressive renal injury is the induction of chemokines and the subsequent recruitment of inflammatory cells. The granulocyte chemoattractant KC (Fig. 5a) and the monocyte chemoattractant MCP-1 (Fig. 5b) increased significantly following obstruction in both WT and NOD1/2 DKO kidneys. Except for a slight but significant higher MCP-1 level in kidneys from NOD1/2 DKO mice compared with WT 7 days following obstruction, no difference in KC and MCP-1 levels were observed between the WT and NOD1/2 DKO mice. The influx of granulocytes (Fig. 5c, e) and accumulation of macrophages (Fig. 5d, f) increased in the obstructed WT and NOD1/2 DKO kidneys, yet there were no differences in these parameters between the WT and NOD1/2 DKO mice, 3 and 14 days post UUO. Seven days following obstruction there was a slight but significant decrease in granulocyte influx and a significant increase in macrophage accumulation in NOD1/2 DKO kidneys compared with WT kidneys. To get more insight in the activation of the NF-kB signaling pathway, we measured the downstream cytokines IL-1b and TNF-α in total kidney homogenates. This revealed no significant differences between the WT and NOD1/2 DKO mice, expect for IL-1b day 3 (Additional file 3: Figure S3). Together these results reveal that there are no or minor changes in the inflammatory response, renal damage and fibrosis following CKD induced by ureteral obstruction in NOD1/2 DKO mice compared with WT mice.

To investigate the role of NOD1 and NOD2 in chronic renal inflammation, injury and fibrosis we subjected NOD1/2 DKO and WT mice at various time points to UUO. UUO initiates a sequence of events in the obstructed kidney, including interstitial inflammation and TEC death, ultimately leading to renal fibrosis which is the final common pathway for numerous forms of progressive renal disease. Recently, the role of the PRRs TLR2, TLR9, TLR4 and NLRP3 in progressive renal injury was investigated. Although TLR2 initiates the inflammatory response during obstructive nephropathy, it does not play a significant role in the development of renal progressive injury and fibrosis [3, 24]. Similarly, TLR9 was not involved in the pathogenesis of UUO [24]. On the other hand, TLR4 attenuates tubular damage and does contribute to renal fibrosis during obstructive nephropathy as demonstrated by an increased injury score and decreased collagen deposition in TLR4-deficient mice [20, 25]. Other work implied a central role for NLRP3 in renal inflammation, fibrosis and tubular damage at different phases of UUO [19, 21]. Apparently, different PRR members have unique response during obstructive nephropathy that lead to a profoundly different outcome of local injury and tubulointerstitial inflammation and fibrosis.

From our study we conclude that NOD1 and NOD2 do not play a significant role in the development of tubulointerstitial fibrosis and inflammation nor in the progression of renal damage after UUO-induced injury. No differences were observed between WT and NOD1/2 DKO obstructed kidneys regarding tubular injury score, apoptosis, proliferation and myofibroblast accumulation. A marginal effect of NOD1/2 deficiency could be detected in the inflammatory response during obstructive nephropathy. Slightly more MCP-1 and concomitant increased macrophage accumulation was observed in the NOD1/2 DKO kidney 7 days following obstruction, while granulocyte influx was lower at this time point. The majority of infiltrating leukocytes into the UUO-damaged kidney are macrophages, which produce cytokines responsible for tubular apoptosis and fibroblast proliferation and activation. However, enhanced macrophage accumulation did not affect the progression of renal fibrosis in NOD1/2 DKO mice. In another study on progressive kidney disease, namely diabetes, it was shown that NOD2 is upregulated and promoted the transcription of extracellular matrix genes and renal injury by inducing inflammation and podocyte insulin resistance [18]. Moreover, in sepsis- and ischemia-induced acute kidney disease models, NOD1/2 DKO mice were demonstrated to be protected against renal disease [10, 13]. Apparently, as it was the case for TLR2 [3, 26], NOD1 and NOD2 are involved in the initiation of inflammation but do not necessarily contribute to further renal damage and fibrosis. Considering that we found no difference between WT and NOD1/2 DKO mice in renal pathology, one could speculate that despite their structural and functional similarities with other PRRs, NOD1 and NOD2 are not activated by DAMPs that are released after UUO. This would be in line with current literature in which solely bacterial ligands and not DAMPs are described to activate NOD1 and NOD2. We anticipate that the difference in renal pathology between WT and NOD1/2DKO mice in the renal ischemia reperfusion injury model might be due to translocation of bacterial products across the leaky intestinal barrier that activate NOD1/2 resulting in inflammation-associated nephropathy. The phenomenon of intestinal barrier disruption is known to occur after renal ischemia reperfusion injury [27] but is not described for UUO.

This study however, has some limitations which have to be pointed out. By using double knockout mice we cannot rule out the possibility that either NOD1 or NOD2 play a different role in UUO and compensate each other. Alternatively, the function of NOD1 or NOD2 might be masked in our knockout model by redundancy or compensatory mechanism. Our experiments tested moreover only unilateral ureteral obstruction with its own advantages and limitations [28] and no other animal models of chronic kidney disease and fibrosis. Other possible roles of NOD1/NOD2 in these disorders that may be activated under different circumstances remain therefore to be tested.

Taken together, our data do not show a functional role for NOD1/2 in kidney injury and fibrosis following chronic kidney disease induced by ureteral obstruction and suggest that similar to infection, different forms of sterile kidney disease will be sensed by different PRRs triggering different signalling pathways which culminate in different kidney disease outcomes.

Together these results reveal that there are no or minor changes in the inflammatory response, renal damage and fibrosis following obstructive nephropathy induced by ureteral obstruction in NOD1/2 DKO mice compared with WT mice.