Leptospiral outer membrane protein LipL32 induces inflammation and kidney injury in zebrafish larvae

Leptospirosis is an often overlooked cause of acute kidney injury that can lead to multiple organ failure and even death. The principle protein that conserved in many pathogenic leptospires is the outer membrane protein LipL32. However, the role of LipL32 in the pathogenesis of renal injury in leptospirosis is not entirely clear. Here we studied the effects of LipL32 on the developing kidney in zebrafish larvae. Incubation of zebrafish larvae with Leptospira santarosai serovar Shermani induced acute tubular injury predominantly in the proximal pronephric ducts. Furthermore, microinjection of lipl32 mRNA or recombinant LipL32 protein into zebrafish larvae increased macrophage accumulation and disrupted the basolateral location of NA-K-ATPase in pronephric ducts. These changes led to substantial impairment of the pronephric kidney structure. We further demonstrated that morpholino knockdown of tlr2, but not tlr4, reduced the LipL32-induced leukocyte infiltration and kidney injury. These data demonstrate that LipL32 contributes to the renal pathology in leptospirosis and gives some clues to the potential virulence of LipL32. Our results support the use of zebrafish as a model organism for studying the disease mechanism of leptospiral infection. This model might permit the future exploration of the virulence and molecular pathways of different leptospiral outer membrane proteins.

Structurally, LipL32 contains the calcium-bound dock which is crucial for its interaction to host cells 21,22 . LipL32 recognizes and adheres to the individual components of extracellular matrix (ECM) including laminin, collagen I, and collagen V of the host cell 23,24 . However, the in vivo role of LipL32 in the pathogenesis of renal injury remains to be determined.
Zebrafish (Danio rerio) has been a vertebrate model for studying infectious diseases such as Streptococcus, Salmonella, and Candida albican infection [25][26][27] . These studies have demonstrated novel disease mechanisms that are not likely seen in other animal models. The advantages of zebrafish models include the ease of genetic manipulation, the ability to keep in germ-free conditions, and the resemblance of physiology to that of mammals 28,29 . There are only sparse data available on the use of zebrafish larvae for leptospirosis studies. An earlier study showed that microinjection of L. interrogans serovar Copenhageni stably infect zebrafish embryos 30 . The accumulation of macrophages surrounding the pronephric region suggests a role of leptospiral infection in causing kidney damage. However, the effects of leptospiral infection on the pronephric kidneys of zebrafish remain uninvestigated.
In this study, we have utilized a zebrafish model to study the effects of leptospiral infection and its major outer membrane protein, LipL32, on the developing kidneys. We demonstrated that LipL32 induced inflammation, reversed the polarity of NA-K-ATPase, and resulted in kidney injury. In addition, we found that LipL32 induced inflammation through TLR2 but not TLR4 by specifically knocking down these receptors in zebrafish larvae. These results may improve our understanding of the cellular and molecular mechanisms of the pathogenesis of leptospirosis.

Leptospira infection caused tubular injuries in zebrafish larvae. To examine the virulence of
Leptospira infection in zebrafish larvae, we first immersed zebrafish larvae to L. Shermani and L. Patoc from 24 hpf for 5 days (Fig. 1A). L. Shermani is the most frequently isolated pathogenic serovar in Taiwan and L. Patoc is a non-pathogenic leptospira 2 . The survival of zebrafish larvae treated with L. Shermani was 78.0% (85/109), whereas the survival of those treated with L. Patoc and E3 buffer control was 81.7% (89/109) and 85.3% (93/110), respectively. There was no significant difference in survival between groups (log-rank test, P = 0.4248). These results suggest that incubating zebrafish larvae with L. Shermani at a concentration of up to 1 × 10 6 CFU/ml had no significant effects on larva viability.
To confirm the actual infection of zebrafish larvae with L. Shermani, we then investigated the expression of LipL32 in these larvae using Reverse Transcriptase PCR (RT-PCR). As shown in Fig. 1B, lipl32 mRNA was detected in the zebrafish larvae that were incubated with L. Shermani but not in those treated with L. Patoc. The expression of lipl32 mRNA appeared as early as two hours after incubation (26 hpf) and persisted until 48 hpf. We also examined the flagellin gene flaB mRNA, which showed a similar expression pattern but persisted until 72 hpf ( Fig. 1B and Supplementary Figure S1) 31 . These results indicate a successful infection of L. Shermani in the zebrafish larvae using the immersion method.
We then examined the pathological changes in pronephric kidneys after L. Shermani and L. Patoc infection. Compared with control larvae ( Fig. 2A,D,G) and L. Patoc-infected larvae (Fig. 2B,E,H), the L. Shermani-infected larvae (Fig. 2C,F,I) showed marked swelling of the pronephric epithelial cells, predominantly in the proximal parts of pronephric ducts on histological examination. These results were confirmed by quantifying the area of pronephric ducts by image analysis on histology sections (Fig. 2J-L). Thus, our results demonstrated that infection with the pathogenic L. Shermani, which expresses the outer membrane protein LipL32, could cause acute tubular injury in zebrafish larvae. It seems unlikely that these changes are due to non-specific cytotoxicity because the larval survival was not jeopardized by current experimental conditions. Ectopic expression of lipl32 mRNA in zebrafish larvae. We then sought to examine the pathological role of LipL32 in the developing kidneys by ectopic expression of the protein in zebrafish larvae. To this end, zebrafish embryos were firstly microinjected with a capped RNA of lipl32 (myc-tagged) in the one-cell stage. As shown in Fig. 3A, in situ hybridization of lipl32 mRNA at 24 hpf revealed a widespread expression of LipL32 including the regions of pronephros and the posterior blood island, indicating a successful injection of lipl32 mRNA. We further confirmed the colocalization of the lipl32 mRNA expression with wt1b:GFP fluorescence in pronephric tubules and ducts, suggesting that the expression of LipL32 may occur in the pronephric kidneys. Similarly, immunostaining for the Myc-tagged epitope in larvae at 48 hpf demonstrated that LipL32 expression can be detected in the pronephros (Fig. 3B). We also found colocalization of Myc and wt1b:GFP signals in the pronephric tubules and ducts (Fig. 3B). These results enabled us to examine the specific effects of LipL32 on the pronephros in our model. The ectopic expression of LipL32 was further confirmed by Western blotting with an  Figure S1). We next examined whether the ectopic expression of LipL32 induces pronephric duct damages. As shown in Fig. 3D, edematous degeneration and swelling of pronephric epithelium were observed in the lipl32 mRNA-injected larvae, suggesting that LipL32 may directly contribute to acute tubular injury in leptospiral infection.
Expression of lipl32 mRNA impaired kidney development. We next tested whether ectopic expression of lipl32 mRNA in zebrafish embryo perturbs the development of pronephros. As shown in Fig. 4A, the percentage of pronephric kidney abnormalities (mild and severe) was significantly increased in lipl32 mRNA-injected larvae compared with control (30/35 vs 3/30, P < 0.0001). To investigate the underlying mechanism of kidney injury, we performed in situ hybridization for a macrophage marker, l-plastin. The expression of l-plastin was markedly increased in lipl32 mRNA-injected larvae and the staining was merged to the wt1b:GFP signal (Fig. 4B), suggesting that these inflammatory cells infiltrated the pronephric kidneys. Meanwhile, we tested if LipL32 could affect NA-K-ATPase or cilia in renal epithelial cells. As illustrated in Fig. 4C, the expression of LipL32 disoriented the basolateral location of NA-K-ATPase to the apical membrane of the renal epithelial cells. In contrast, immunostaining for the primary cilia, which is responsible for the mechanosensation of renal tubular epithelium, did not show any structural abnormality in lipl32 mRNA-injected larvae (Fig. 4D). These results demonstrate that LipL32 may cause pronephric kidney injury by inducing inflammation and mislocalization of NA-K-ATPase. We also found that the glomerular filtration rates, as estimated by the excretion of injected 10-kDa rhodamine dextran, were significantly reduced in the lipl32 mRNA-injected larvae compared to controls (Fig. 4E), indicating a functional impairment of the affected pronephric kidneys.
LipL32 triggered pronephric inflammation through the TL2R pathway. We then investigated whether LipL32 triggers inflammation through the TLR2 pathway according to our previous in vitro studies 19,20,32 . The retention rates of 10-kDa rhodamine dextran as measured from the boxed region of the posterior pronephric ducts 6 hours after pericardial injection were significantly reduced in the lipl32 mRNA-injected larvae compared to myc-mRNA controls (* * P < 0.01, n = 10 to 11 in each group, scale bar, 100 μ m). Transverse sections demonstrate retention of rhodamine fluorescence in the lumen of pronephric ducts (circle) in lipl32 mRNA-injected larvae (scale bar, 5 μ m).
As shown in Fig. 5A,B, the ectopic expression of lipl32 mRNA triggers the accumulation of l-plastin positive cells in the posterior blood island and the region surrounding pronephric ducts. The inflammatory response was significantly abolished (P < 0.0001) in the tlr-2 knockdown larvae but remained unchanged with concomitant tlr4a and tlr4b knockdown. Knockdown of the tlr2 downstream transducers myeloid differentiation factor 88 (myd88) also blocked (P < 0.0001) the lipl32-mediated l-plastin expression. Quantitative real-time RT-PCR confirmed that the expression of l-plastin mRNA was reduced significantly after knocking down tlr2 (P < 0.05) rather than tlr4a and tlr4b (Fig. 5C).
Injection of LipL32 protein through the tail vein. We finally investigated the virulence of LipL32 protein by direct injection of a recombinant LipL32 protein into the tail vein of zebrafish larvae at 24 hpf. Successful injection of the recombinant protein to the circulation was confirmed by the presence of co-injected rhodamine fluorescence. As shown in Fig. 6A, the frequency of pronephric kidney deformity was significantly increased compared to control larvae (16/21 vs 7/26, P < 0.01). Injection of LipL32 protein also disrupted the expression of NA-K-ATPase, which was partially translocated from the basolateral membrane to the apical cell membrane (Fig. 6B). Furthermore, the expression of l-plastin mRNA in the pronephric area as shown by in situ hybridization was markedly increased after injection of LipL32 and was significantly suppressed (P < 0.0001) with morpholino knockdown of tlr2 but not of tlr4a and tlr4b (Fig. 6C). These data confirmed that circulating LipL32 can induce inflammation and kidney malformation in zebrafish larvae.
TLR2 blockade ameliorated the kidney injury induced by LipL32. We further determined the potential rescue effects of TLR blockade on the LipL32-induced injury to the pronephros. As shown in Fig. 7, the frequencies of kidney deformities caused by the ectopic expression of lipl32 mRNA in wt1b:GFP larvae were significantly reduced by microinjection of tlr2 MOs (P < 0.05) but not tlr4a and tlr4b MOs (P = 0.2275). These results suggest that LipL32 might induce kidney injury predominantly through a TLR2-dependent pathway in zebrafish larvae.

Discussion
Here we used a zebrafish model for investigating the potential virulence mechanisms of LipL32 in leptospirosis. We found that L. Shermani infection induced pronephric kidney injury in zebrafish larvae. Meanwhile, ectopic expression of LipL32 in zebrafish larvae prompted macrophage infiltration to the pronephric kidneys through the innate immunity TLR2 pathway, disrupted the basolateral NA-K-ATPase location, and impaired the normal kidney structure. These results support the use of zebrafish as a model organism for leptospiral infection 30 .
Zebrafish has been used as a model organism for various diseases and may provide a unique opportunity to observe early pathological changes. Zebrafish has a fully functional innate immune system by 32 hpf with the development of circulating systems and phagocytes 30 . The pattern-recognition receptor system in zebrafish is highly similar to that in other vertebrates, which makes it a useful tool for examining host-microbe interactions 33,34 . An earlier study that injected L. interrogans into the caudal vein of zebrafish larvae has shown that leptospires are readily ingested by macrophages, which are later found specifically accumulated in the ventral wall of the dorsal aorta, equivalent to the aorta-gonad-mesonephros (AGM) in mammals 30 . In this current study, we further demonstrated that overexpression of LipL32 induced accumulation of macrophages in the posterior blood island and the pronephros, suggesting that LipL32 may participate in guiding macrophages to the target organs. Thus, LipL32 might be a potential leptospiral adhesin that is important in the specific interactions to host substrates 35 .
LipL32 is a major immunogen during leptospiral infection in humans and mammals 9 . In hamster model of leptospirosis, immunohistochemical analyses of L. kirschneri-infected kidneys have demonstrated the intense LipL32 expression in proximal tubule cells and the interstitium 16 . Vaccination with LipL32 reduces the severity of kidney inflammation in the hamsters infected with L. interrogans serovar Canicola in hamsters 36 . In this study, we demonstrated that the ectopic expression of LipL32 directly induced an inflammatory reaction and caused kidney injury, suggesting that LipL32 is a virulence factor and plays an important role in mediating the pathogenesis of renal injury in leptospirosis 37 .
Disturbances of renal electrolyte handling may occur in acute leptospirosis. This could lead to decreased renal water absorption and renal salt losing commonly seen as non-oliguric renal failure in leptospirosis 10 . Previous studies have demonstrated that the expression and function of NA-K-ATPase are inhibited by L. interrogans glycoproteins [38][39][40] . The addition of GLP extracted from L. interrogans specifically inhibits the Na-K-ATPase in rabbit kidney epithelial cells 39 . Downregulation of NA-K-ATPase and other ion transporters have been found in L. interrogans-infected mice 41 . Basolaterally located Na-K-ATPase is essential for transfer of salt and water across epithelium to interstitium and its dislocation has been related to alveolar edema in leptospirosis 11 . Our data are consistent with these findings and provide in vivo evidence that LipL32 also perturbed the expression of NA-K-ATPase in renal epithelial cells. Whether LipL32 might interact with NA-K-ATPase directly warrants future studies. The TLR pathways have been shown to play a pivotal role in the pathogenesis of leptospirosis 42 . It has been demonstrated that TLR2-deficient mice showed a reduced cytokine response to leptospiral LPS challenge, and LPS activates macrophages through TLR2 in cultured monocytes 9 . Our previous studies have shown that LipL32 activates TLR2 and downstream genes in cultured mouse medullary ascending limb cells and proximal tubule cells 19,43 . Experimental evidence from an atomic force microscope study suggests that LipL32 binds directly to TLR2, but not TLR4, on the renal cell surface 32 . In the current study, we further demonstrated that LipL32 induced leukocyte infiltration through the TLR2 pathway in vivo. These finding strongly support that LipL32 binds directly to TLR2 to mediate its virulence functions. Future research could also explore the potential LipL32 ligands other than immune receptors 44,45 .
Our results are different to a previous study which showed that a LipL32 mutant constructed in L. interrogans was still able to infect hamsters and rats 46 . This discrepancy may be explained partly by a high degree of redundancy in leptospiral proteins involved in adhesion, survival, and renal colonization 47 . However, by direct expression of LipL32 in the zebrafish model, our data demonstrated clearly the virulence of LipL32 in causing host inflammation and target organ damage. It is also worth noting that other outer membrane proteins such as Loa22, OmpL1, p31/LipL45, and LenA have been regarded as potential leptospiral adhesins and might warrant further investigation 35,48 .
In conclusion, we have shown that LipL32 promotes inflammation and induces kidney injury in zebrafish larvae. This investigation supports the hypothesis that LipL32 contributes to the pathogenesis of renal injury in leptospirosis. Our results support the use of zebrafish as a model organism for studying the disease mechanism of Microinjection of lipl32 mRNA. An myc epitope-tagged pCS2-MT-LipL32 plasmid was generated by cloning lipl32 cDNA (798 bp, amino acids 1-266) from pathogenic L. santarosai serovar Shermani genomic DNA, into the Cla1/EcoR1 sites of the pCS2-MT plasmid (Promega). The lipl32 sequence was confirmed by DNA sequencing. In vitro transcribed capped RNA of lipl32 (myc-tagged, 600 ng/μ l, 4 nl) was microinjected in the one-cell stage of zebrafish embryos according to standard protocols 52 .
Microinjection of LipL32 protein. Detoxified LipL32 protein (amino acids 21-272) was prepared as described previously 20,32 . Larvae were anesthetized using 0.02% tricaine. LipL32 recombinant protein (0.07mg/ml, 4 nl) was injected into the caudal vein at 24 hpf together with a 10-kDa rhodamine-dextran (Invitrogen) to confirm the injection into blood circulation. Control larvae were injected with phosphate-buffered saline (PBS) or red fluorescent protein (RFP). Morphological abnormalities in pronephros. Pronephros was observed in living transgenic wt1b:GFP zebrafish larvae under fluorescence microscopy at 48 hpf 55 . Developmental abnormalities of the pronephros including glomerular malformation and tubular dilatation were categorized as normal, mild (glomerular malformation or a loss of tubular angle) or severe deformity (glomerular malformation and a loss of tubular angle) 50,56,57 . Histology analysis. Zebrafish larvae were fixed in 4% paraformaldehyde in PBS overnight at 4 °C, embedded in glycol methacrylate (JB-4; Polyscience), and cut at 5 μ m. Slides were stained with Hematoxylin and Eosin (H&E). Quantitative image analysis for the area of pronephric ducts was performed using ImageJ software.

Immunohistochemistry.
Immunostaining was performed in whole-mount larvae as described previously 58 . A mouse anti-NA-K-ATPase alpha-subunit (a5) antibody (1:200, Developmental Studies Hybridoma Bank) was used to label the location of NA-K-ATPase in the pronephric epithelial cells. A mouse anti-acetylated alpha-tubulin antibody (1:200, Sigma) was used to label the primary cilia of pronephric ducts. A mouse anti-c-myc (9E10) antibody (1:500, Covance) was used to label the Myc-tag of LipL32 protein expression. A rabbit anti-GFP antibody (1:500, Invitrogen) was used to label the pronephric GFP expression in the wt1b:GFP larvae. The secondary antibodies used were Alexa Fluor 488-or 594-conjugated goat anti-mouse IgG and goat anti-rabbit IgG (1:500, Invitrogen). The immunostained larvae were then embedded in OTC medium and cryosectioned. Sections were mounted in Vectashield (Vector Laboratories) with DAPI and observed under fluorescence microscopy.
Rhodamine-dextran injection. For renal function estimation, a 10-kDa rhodamine-dextran (Invitrogen) was injected into the pericardium of zebrafish larvae at 48 hpf after anesthesia with tricaine 59 . Images of larvae were collected under fluorescence microscopy 6 hours after injection. A region of 200 × 100 pixels containing the pronephric duct and cloaca was selected and the fluorescence intensity was measured using ImageJ software. Whole-mount in situ hybridization. Larvae were fixed overnight in 4% paraformaldehyde at 4 °C. Whole-mount in situ hybridization was performed according to published protocols 49 . Antisense digoxigenin-labeled probes were made for lipl32 and l-plastin from corresponding cDNA constructs.
RT-PCR. The expression of lipl32 and flaB mRNAs was examined by RT-PCR to confirm the infection of pathogenic Leptospira species according to the previously described methods 31,60 . Larvae were mixed and ground in liquid nitrogen for total RNA extraction using TRIzol Reagent (Life Technologies). We obtained cDNA from the total RNA using the First Strand cDNA synthesis kit for RT-PCR (AMV) (Roche Diagnostics) following the manufacturer's instructions. The primers used were: lipl32, forward 5′-CGCTGAAATGGGAGTTCGTATGATT-3′ and reverse 5′ -CCAACAGATGCAACGAAAGATCCTTT-3′ 60 ; flaB, forward 5′ -CTCACCGTTCTCT AAAGTTCAAC-3′ and reverse 5′ -TGAATTCGGTTTCATATTTGCC-3′ 31 . Zebrafish ß-actin was used as internal control.
Real-time RT-PCR. Real-time quantitative RT-PCR was performed with the ABI PRISM 7700 sequence detection system (Applied Biosystems) in the presence of SYBR Green. The primers used to amplify zebrafish l-plastin were: forward 5′ -CCGCTACGACCTGCTGAAG-3′ , reverse: 5′ -GCGCCATCGAGATTGCAT-3′ . The zebrafish ribosomal protein S18 (rps18) was used as internal control. The reaction was performed in duplicate for each sample.
Western blot analysis. Proteins extracted from whole larvae were analyzed by Western blotting following the standard method 49 . Antibodies used include anti-tubulin (1:8000, clone DM1a, Sigma) and a purified custom-made rabbit polyclonal antibody (1:10000) raised against the LipL32 protein (amino acids 21-272).

Statistical analysis.
Continuous variables are expressed as mean ± SEM and compared by using Student's t-test or one-way ANOVA followed by Newman-Keuls Test. Categorical variables were analyzed using the Fisher's exact or Chi-square test. Cumulative survival curves were generated by the Kaplan-Meier method and compared with the log-rank test. The standard error of the survival curves was calculated using the Greenwood method. P-values < 0.05 are considered statistically significant. All analyses were performed using the Graphpad Prism 5.1 (Graphpad, La Jolla, CA).