The corticosteroid compounds prednisolone and vamorolone do not alter the nociception phenotype and exacerbate liver injury in sickle cell mice

Abstract

Clinicians often hesitate prescribing corticosteroids to treat corticosteroid-responsive conditions in sickle cell disease (SCD) patients because their use can be associated with complications (increased hospital readmission, rebound pain, strokes, avascular necrosis, acute chest syndrome). Consequently, SCD patients may receive suboptimal treatment for corticosteroid-responsive conditions. We conducted a preclinical trial of dissociative (vamorolone) and conventional (prednisolone) corticosteroid compounds to evaluate their effects on nociception phenotype, inflammation, and organ dysfunction in SCD mice. Prednisolone and vamorolone had no significant effects on nociception phenotype or anemia in homozygous mice. Conversely, prednisolone and vamorolone significantly decreased white blood cell counts and hepatic inflammation. Interestingly, the effects of vamorolone were milder than those of prednisolone, as vamorolone yielded less attenuation of hepatic inflammation compared to prednisolone. Compared to controls and heterozygotes, homozygotes had significant liver necrosis, which was significantly exacerbated by prednisolone and vamorolone despite decreased hepatic inflammation. These hepatic histopathologic changes were associated with increases in transaminases and alkaline phosphatase. Together, these results suggest that, even in the setting of decreasing hepatic inflammation, prednisolone and vamorolone were associated with significant hepatic toxicity in SCD mice. These findings raise the possibility that hepatic function deterioration could occur with the use of corticosteroids (conventional and dissociative) in SCD.

Conflict of interest statement

Jesse Damsker is an employee of ReveraGen BioPharma Inc. All other authors report no conflict of interest.

Figures

Figure 1

Effect of prednisolone and vamorolone on nociception. Data are presented as means and standard errors of percentage changes from baseline by genotype and treatment for hot plate (A) and tail flick latencies (B), current thresholds in response to 2000 (C), 250 (D) or 5 Hz (E) stimulations. Hetero indicates heterozygotes and homo homozygotes (N = 13–18 per each of 9 treatment groups for all outcome measurements). Regardless of genotype and sex, six-week treatment with vehicle, prednisolone, or vamorolone yielded similar percent changes from baseline on hot plate and tail flick latencies and on current thresholds in response to 2000 or 5 Hz stimulations (all p > 0.05, A,B,C and E). Regarding 250 Hz current thresholds, the effects of prednisolone and vamorolone, compared with vehicle, varied according to genotype as there was a significant treatment*genotype interaction (p = 0.014) although neither the main effect of treatment or genotype reached statistical significance (p = 0.07 and p = 0.18, respectively, (D). As shown in (D), the pattern of percent change from baseline over treatment groups is markedly different for each genotype.

Figure 2

Effect of prednisolone and vamorolone on hematologic parameters and spleen size. Data are presented as means and standard errors. Hetero indicates heterozygotes and homo homozygotes (N = 13–17 per each of 9 treatment groups for all outcome measurements). (A) Homozygous mice, independent of treatment, had increased white blood cell counts p < 0.001 compared to control animals. Regardless of genotype, prednisolone- and vamorolone-treated mice had significantly lower white blood cell counts compared with vehicle-treated animals (p < 0.001). The effects of treatment on the number of neutrophils varied according to genotype (p = 0.045 for the treatment*genotype interaction term). (B) The number of neutrophils was significantly higher in homozygous mice as compared to controls (p = 0.006) and heterozygous mice (p = 0.006) for all treatment groups, however no main effect of treatment was observed (p = 0.21). (C) Homozygous mice had significantly greater lymphocytes than either control (p < 0.001) or heterozygous mice (p < 0.001) and both, prednisolone (p = 0.027) and vamorolone-treatment (p < 0.001) were associated with decreased lymphocyte counts compared with vehicle. (D) Homozygotes, independent of treatment, had lower red cell counts, compared to controls and heterozygotes (p < 0.001). Treatment with prednisolone and vamorolone had no effect on red cell counts, regardless of genotype (p = 0.69). (E) Homozygous mice had significantly increased spleen-to-body weight ratio compared with heterozygotes and control animals (all p < 0.001) independent of treatment although the effects of treatment varied according to genotype (p = 0.010 for treatment*genotype interaction). Mice of all genotypes treated with prednisolone had significantly lower spleen to body weight ratios than vehicle treated mice (p = 0.006), however no significant difference was observed between vamorolone and vehicle treated mice independent of genotype.

Figure 3

Effect of prednisolone and vamorolone on liver function tests and malondialdehyde levels. Data are presented as means and standard errors of alanine aminotransferase (ALT, A), aspartate aminotransferase (AST, B), alkaline phosphatase (ALK, C), and malondialdehyde (D). Hetero indicates heterozygotes and homo homozygotes (N = 5–7 per each of 9 treatment groups for all outcome measures). Among vehicle-treated mice, compared with controls, homozygotes had significantly higher plasma levels of ALT (p < 0.001, A), AST (p = 0.001, B), and ALK (p < 0.001, C). Overall, the effects of vamorolone and prednisolone on ALK, ALT, and AST were similar (p = 0.071, p = 0.27, p = 0.40 respectively). Additionally, independent of genotype, compared with vehicle, vamorolone and prednisolone treatment was associated with increased ALK and AST (p = 0.018 and p = 0.042 respectively for main treatment effect). (D) Homozygotes had higher levels of liver malondialdehyde formation compared with control and heterozygous mice (p < 0.001). Additionally, independent of genotype, there was no effect of treatment on malondialdehyde formation in liver homogenates (p = 0.66).

Figure 4

Effect of prednisolone and vamorolone on liver histopathology. Representative hematoxylin and eosin-stained sections from each genotype and respective treatment group (N = 14–20 per each of 9 treatment groups). In contrast to controls and heterozygotes, homozygous mice had significant hepatic necrosis (top row: Vehicle/Prednisolone/Vamorolone: 5x; middle row: Vehicle/Prednisolone/Vamorolone: 5x; bottom row; Vehicle: 10x, Prednisolone: 4x, Vamorolone: 3x). In homozygotes, hepatic necrosis was characterized by areas of patchy, often confluent, well-demarcated zone III (centrilobular) coagulative necrosis with admixed hemosiderin deposition and inflammation. Pools of sickled red blood cells were readily identified within the hepatic vasculature, leading to congestion and occasional occlusion within areas of necrosis.

Figure 5

Effect of prednisolone and vamorolone on quantitative hepatic necrosis. Box plots show median, interquartile range, whiskers show 10th and 90th percentiles and points reflect 5th and 95th percentiles of quantitative liver necrosis in all experimental groups according to genotype and treatment. Ve represents vehicle, P prednisolone, and Va vamorolone. (N = 14–20 per each of 9 treatment groups). When comparing the percentage of hepatic necrosis between treatments in homozygous mice, a significant increase in hepatic necrosis was observed in vamorolone treated homozygotes compared to both vehicle- (p = 0.005) and prednisolone-treated (p = 0.008) animals. There was also a trend towards increased liver necrosis in prednisolone-treated homozygotes compared to vehicle-treated mice even though this increase in hepatic necrosis did not reach statistical significance (p = 0.15).

Figure 6

Effect of prednisolone and vamorolone on hepatic inflammation in homozygous sickle cell mice. Representative hematoxylin and eosin (H&E) slides from homozygotes shown in original magnification 10x (top row) and 20x (bottom row). N = 14–20 per each of 9 treatment groups. There was evidence of significant hepatic inflammation in homozygotes and the degree of inflammation varied between treatment groups. The infiltrates were composed predominantly of lymphocytes and macrophages, with only rare plasma cells and eosinophils. Neutrophils were found associated mainly with areas of necrosis. The degree of inflammation was worst in vehicle-treated homozygotes, less intense in vamorolone-treated and with the mildest degree seen in prednisolone-treated homozygotes.