Drug screening in zebrafish larvae reveals inflammation-related modulators of secondary damage after spinal cord injury in mice

Abstract

Prolonged inflammation after spinal cord injury is detrimental to recovery. To find pharmacological modulators of the inflammation response, we designed a rapid drug screening paradigm in larval zebrafish followed by testing of hit compounds in a mouse spinal cord injury model. Methods: We used reduced il-1β linked green fluorescent protein (GFP) reporter gene expression as a read-out for reduced inflammation in a screen of 1081 compounds in larval zebrafish. Hit drugs were tested in a moderate contusion model in mice for cytokine regulation, and improved tissue preservation and locomotor recovery. Results: Three compounds robustly reduced il-1β expression in zebrafish. Cimetidine, an over-the-counter H2 receptor antagonist, also reduced the number of pro-inflammatory neutrophils and rescued recovery after injury in a zebrafish mutant with prolonged inflammation. Cimetidine action on il-1β expression levels was abolished by somatic mutation of H2 receptor hrh2b, suggesting specific action. In mice, systemic treatment with Cimetidine led to significantly improved recovery of locomotor behavior as compared to controls, accompanied by decreased neuronal tissue loss and a shift towards a pro-regenerative profile of cytokine gene expression. Conclusion: Our screen revealed H2 receptor signaling as a promising target for future therapeutic interventions in spinal cord injury. This work highlights the usefulness of the zebrafish model for rapid screening of drug libraries to identify therapeutics to treat mammalian spinal cord injury.

Keywords: Betazole, Bortezomib, Sildenafil, Cimetidine; chronic inflammation; histamine receptors; irf8; zebrafish.

Figure 1

Screening of an Il-1β reporter fish reveals drugs that reduce endogenous Il-1β mRNA expression. A: Schematic representation of the drug screening workflow. Briefly, 3 dpf larvae were pre-treated with various compounds for 4 hours, followed by manual injury. Next day, larvae that showed an overall decrease in GFP signal (pre-screening) were fixed, fluorescence was quantified (screening step) and positive compounds were re-tested (validation) and mechanistically tested by qRT-PCR for il-1β mRNA. The numbers of tested compounds after each screening step are depicted in red. B: Representative images of 16-18 hpl drug-treated larvae are shown, including negative (DMSO) and positive control (dexamethasone). C: Quantification of the overall GFP fluorescence signal relative to DMSO-treated control is shown for the validation step. All five compounds elicit a ≥50% drop in the overall GFP signal. Dexamethasone treatment (positive control) induced an approx. 50% decrease in overall GFP signal versus control. D: Three out of five compounds lead to a reduction in the relative il-1β mRNA expression in wild-type larvae (50 treated or control larvae were pooled per measurement). Cimetidine (10 µM) and sildenafil citrate lead to a 40% reduction in il-1β mRNA expression, while bortezomib leads to a 70% reduction. Scale bars: 100 µm

Figure 2

Cimetidine reduces the number of neutrophils in the injury site. A: Representative images of the lateral view of the injury site in the il-1β:GFP line treated with cimetidine and DMSO-control are shown. Broken blue lines outline spinal cord (SC) and lesion site (LS). B, C: At 2 hpl, the total number of neutrophils (B, mpx+, t-test, p* = 0.0143), and the number of double positive neutrophils (C, GFP⁺/Mpx⁺ t-test, p* = 0.0279) are reduced by cimetidine treatment. D, E: At 24 hpl, the number of double-positive cells is not altered by cimetidine treatment. Broken yellow lines outline the area of quantification. Scale bar: 100 µm.

Figure 3

Cimetidine partially rescues impaired axonal regeneration in irf8 mutant zebrafish. A: Representative images of the lateral view of the injury site in cimetidine- and DMSO-treated larvae are shown. An axonal connection across the injury is visualized with DsRed, expressed under a neuronal promoter. Presence of an axonal bridge is indicated in the images. B: Quantification by scoring of animals for the presence of an axonal bridge indicates that Cimetidine increases the proportion of irf8 mutant larvae with an axonal bridge (two-way ANOVA, interaction effect F(1, 16) = 10.86, p* = 0.0046; Tukey's multiple comparison test, irf8 mutant, DMSO vs irf8 mutant, cimetidine p* = 0.0215, irf8, DMSO vs wt, DMSO p** = 0.0019, irf8, DMSO vs wt, cimetidine p* = 0.0295), but does not affect wild-type regeneration. C: Measuring the thickness of the axonal bridge yielded similar results for the experimental groups (two-way ANOVA, interaction effect F(1, 125) = 11.60, p*** = 0.0009; Tukey's multiple comparison test, irf8 mutant, DMSO vs irf8 mutant, cimetidine p* = 0.0112, irf8 mutant, DMSO vs wt, DMSO p** = 0.0074). Scale bar: 100 µm

Figure 4

Pharmacological manipulations HH2R receptor activity indicates positive regulation of il-1β expression by the receptors. A: Representative images of the injury site in il-1β:GFP line treated with 10µM cimetidine, or DMSO are shown. B: Quantification of the overall GFP fluorescence relative to DMSO-treated larvae indicates a dose-dependent reduction of GFP fluorescence by Cimetidine treatment (Kruskal-Wallis test with Dunn's multiple comparison test, p*** = 0.0003; DMSO vs 10 µM cimetidine, p* = 0.0250; DMSO vs 50 µM cimetidine, p*** = 0.0008). C: Representative images of the injury site in il-1β:GFP line treated with 50 µM Betazole, or DMSO are shown. D: Quantification of the overall GFP fluorescence relative to DMSO-treated larvae reveals a dose-dependent increase of GFP fluorescence by Betazole treatment (Kruskal-Wallis test with Dunn's multiple comparison test, p* = 0.0145, DMSO vs 50 µM Betazole, p* = 0.0422; DMSO vs 100 µM Betazole, p* = 0.0473). Scale bars: 100 µm.

Figure 5

Somatic receptor mutation indicates that Cimetidine acts through hrh2b in zebrafish. A: Examples of gels used to assess the efficiency of injected CrRNA by RFLP are shown. Each lane represents one embryo. In uninjected controls, a complete band shift is observed, indicating activity of the indicated restriction enzyme, whereas in 8 haCR-injected animals almost no digest occurs, indicating successful mutation of the recognition site. B: A schematic timeline of the experimental design, combining somatic mutation with drug treatment is depicted. C: Lateral view of the spinal cord injury sites in somatic mutant larvae with and without Cimetidine treatment are shown. D: Quantification of the overall GFP fluorescence of the cimetidine-treated somatic receptor mutants relative to DMSO-treated control indicate that in condition in which hrh2b haCRs were injected the effect of Cimetidine is abolished. (uninjected condition, Mann-Whitney tests, p* = 0.0373; hrh2a CrRNA condition, t-test, p*** = 0.0005). E: Cimetidine treatment leads to an increase in the relative hrh2b mRNA expression in lesioned wild-type larvae (50 treated or control larvae were pooled per measurement; Kruskal-Wallis test, p* = 0.0242 with Dunn's multiple comparisons test: injured, DMSO vs injured, cimetidine p* = 0.0324). hrh2a mRNA expression is unaffected by cimetidine treatment. Scale bars: 100µm.

Figure 6

Cimetidine improves locomotor recovery after spinal injury in mice. A: Locomotor recovery was assessed using the 9-point Basso Mouse Scale (BMS) analysis. Mice treated with Cimetidine at 15mg/kg under the same conditions showed statistically significant improvement in locomotor control starting at day 21 post-injury (Two-way RM-ANOVA; interaction effect F_(8,160) = 4.610, p < 0.0001; with post-hoc Bonferroni multiple comparison test; p _(21d) = 0.0058, p _(28d) = 0.0225, n = 11 mice per group). B: DigiGait analysis shows significant improvement in propulsion duration (i) and decrease in step angle (ii) with Cimetidine treatment both indicators of improved gait control. (Vehicle vs. Cimetidine: two-tailed Mann Whitney U-test, propulsion duration: p** = 0.0016; step angle: p* = 0.0279; n = 9-10 mice per group).

Figure 7

Treatment with Cimetidine reduces secondary damage induced by spinal cord contusion injury in mice. A: Confocal images of the spinal cord at 28-day survival after injury stained with anti-GFAP are shown. Note that the size and extent of the lesion is smaller in the Cimetidine treated group compared to the vehicle-treated control (quantified in C). B: Luxol fast blue (LFB) stained sections of the spinal cord 28 days after SCI. Note greater sparing of myelin LFB staining in the Cimetidine-treated group (quantified in D). C: Quantification of tissue loss is shown. Note although the mean values in the Cimetidine treated group are smaller than in the vehicle-treated group, the value reaches statistical significance only at - 500 µm rostral to the lesion due to the variability between animals (Two-way RM-ANOVA; group effect F_(1,8) = 3.188, p = 0.1120; with Sidak's multiple comparison test; p* _(-500) = 0.0372, n = 5 mice per group. Note also that the longitudinal extent of the lesion is smaller in the Cimetidine-treated group. D: Mice treated with Cimetidine show a significant reduction of myelin loss (LFB staining) from the center of the lesion to 600 μm rostrally (Two-way RM-ANOVA; group effect F_(1,7) = 12.35, p = 0.0098; with post-hoc Bonferroni multiple comparison test; p* _(-600µm) = 0.0438, p* _(-400µm) = 0.0204, p** _(-200µm) = 0.0084, p* _(center) = 0.0103, n = 4-5 mice per group). Scale bars = 200 µm.

Figure 8

Cimetidine treatment leads to higher density of neurons and greater serotonergic innervation after spinal cord injury in mice. A: Single plane confocal images of spinal cord cross-sections 28 days after spinal cord contusion injury stained with anti-NeuN antibody to label neurons. Note the greater sparing of neurons in the Cimetidine-treated animal compared to the vehicle group. B: Quantification of NeuN labeling in the ventral horn is shown. Mice treated with Cimetidine showed increased survival of NeuN+ neurons in the ventral horn of the spinal cord as compared to controls on either side of the lesion center (Two-way RM-ANOVA; group effect F_(1,7) = 20.65, p = 0.0027; with post-hoc Bonferroni multiple comparison test; p* _(-1000µm) = 0.0448, p* _(-500µm) = 0.0496, p* _(+500µm) = 0.0172, p** _(+1000µm) = 0.0024, n = 4-5 mice per group.) C: Increased serotonergic innervation of the ventral horn treated with Cimetidine is observed at 1mm caudal to the lesion center by 5-HT immunohistochemistry. D: Quantification by optical density measurements shows that there is significantly greater 5-HT innervation in Cimetidine treated mice than in vehicle-treated controls. Mann-Whitney unpaired t test, p ≤ 0.0001, n = 9 per group; Scale bar = 200 µm and 100 µm for A and C, respectively.

Figure 9

Cimetidine changes the cytokine profile of the injury site in mice. A: qRT-PCR shows significant decrease in il-1β mRNA expression after 1 day of Cimetidine treatment (one-way ANOVA with Tukey's multiple comparison test, p**** < 0.0001, vehicle vs. cimetidine p* = 0.0219). B: Results of the RT² profiler cytokine array (PAMM-150Z) after 10 days of Cimetidine treatment depicted as a heat map indicating genes with decreased (green) and increased (red) transcript levels (quantification shown in C and D) in individual mice in the three indicated groups is shown. C: Quantification of genes with decreased transcript levels is shown for B. D: Quantification of genes with increased transcript levels is shown for B. Statistics for C,D: vehicle vs. cimetidine: two-tailed Mann Whitney U-test for all comparisons; Bmp2: p** = 0.0012; Bmp4: p** = 0.009; Ccl3: p*** = 0.0002; Ccl4: p**** ≤ 0.0001; Cd40lg: p* = 0.0238; Cntf: p** = 0.0013; Csf1: p** = 0.0023; Il11: p*** = 0.0007; Il12α: p* = 0.0219; Il1α: p** = 0.0069; Il21: p** = 0.0073; Il4: p* = 0.0257; Tnfrsf11b: p* = 0.0104; Cxcl3: p* = 0.0469; Il16: p** = 0.0065; Lif: p* = 0.0148; Lta: p* = 0.0264; Pf4: p* = 0.0418; Thpo: p* = 0.0425; Tnfsf11: p*** = 0.0005; n = 4-5 mice per group.