MicroRNA profiling identifies a novel compound with antidepressant properties

Abstract

Patients with traumatic brain injury (TBI) are frequently diagnosed with depression. Together, these two leading causes of death and disability significantly contribute to the global burden of healthcare costs. However, there are no drug treatments for TBI and antidepressants are considered off-label for depression in patients with TBI. In molecular profiling studies of rat hippocampus after experimental TBI, we found that TBI altered the expression of a subset of small, non-coding, microRNAs (miRNAs). One known neuroprotective compound (17β-estradiol, E2), and two experimental neuroprotective compounds (JM6 and PMI-006), reversed the effects of TBI on miRNAs. Subsequent in silico analyses revealed that the injury-altered miRNAs were predicted to regulate genes involved in depression. Thus, we hypothesized that drug-induced miRNA profiles can be used to identify compounds with antidepressant properties. To confirm this hypothesis, we examined miRNA expression in hippocampi of injured rats treated with one of three known antidepressants (imipramine, fluoxetine and sertraline). Bioinformatic analyses revealed that TBI, potentially via its effects on multiple regulatory miRNAs, dysregulated transcriptional networks involved in neuroplasticity, neurogenesis, and circadian rhythms- networks known to adversely affect mood, cognition and memory. As did E2, JM6, and PMI-006, all three antidepressants reversed the effects of TBI on multiple injury-altered miRNAs. Furthermore, JM6 reduced TBI-induced inflammation in the hippocampus and depression-like behavior in the forced swim test; these are both properties of classic antidepressant drugs. Our results support the hypothesis that miRNA expression signatures can identify neuroprotective and antidepressant properties of novel compounds and that there is substantial overlap between neuroprotection and antidepressant properties.

Fig 1. Schematic of experimental workflow and pilot studies.

(A) The original hypothesis that diverse drugs with neuroprotective properties share common molecular mechanisms was validated by analysis of genome-wide microRNA expression profiles induced by three different drugs in rat brains after TBI. Unexpectedly, we found these microRNA profiles to be similar to those induced by antidepressant drugs. (B) Stereological assessment of neuronal injury after JM6 treatment reveals neuroprotective effect of JM6. JM6 treated rats had reduced numbers of FJC+ neurons in the CA1 (p = 0.19) and CA3 (p = 0.07) sub-regions of the hippocampus. The overall average (CA1 and CA3 taken together) was significantly different *p = .038 vs untreated. (C) Immunohistochemistry using CD11b (OX-42) shows JM6 reduced microglial activation after TBI. Increased microglial activation after TBI (left) is ameliorated by treatment with JM6 (right) in the rat hippocampus (upper) and cortex (lower). Note, in sham-injured brains (not shown), staining was undetectable.

Fig 2. Agilent microRNA microarray analysis.

(A) Heatmaps of genome-wide miRNA profiles of control and TBI rats with and without treatment with either E2 (upper), JM6 (middle) or PMI-006 (bottom), demonstrated that after TBI, each drug restored miRNA expression profiles to patterns that are similar to sham-injured (no TBI) rats. (B) Dynamic principal component analysis (PCA; left) and hierarchical clustering (right) in Qlucore Omics Explorer showing that injured rats treated with JM6 exhibit miRNA expression patterns that resemble those of sham-injured rats. PCA shows that Naïve, Sham, TBI and TBI+JM6 are all distinguishable one from another.

Fig 3. Ingenuity pathway analysis (IPA) of JM6 effects on Huntington’s disease (HD) pathway.

(A) JM6 treatment downregulates the expression of HD genes associated with cell death, calpain and caspase (green) and upregulates a gene associated with cell survival, Akt (red). (B) In silico analysis of drug-affected miRNAs after JM6 treatment, miRNAs affected by TBI and returned to sham patterns by JM6 treatment (shown enlarged) are associated with depression-linked genes. (C) List of genes associated with depression are regulated by broadly conserved miRNAs affected by TBI and/or TBI+ JM6.

Fig 4. Ingenuity pathway analysis after JM6 treatment.

(A) JM6 treatment significantly affects genes in pathways linked to depression, including axonal guidance, CREB and Huntington’s disease signaling. (B) The top immune response pathways affected by JM6 treatment are also linked to depression, including CXCR, IL-4 and NF-κB activation by virus signaling. Threshold indicates minimum significance level [–log (p-value) from Fisher’s exact test]. Ratio refers to the number of molecules from the dataset that map to the pathway listed, divided by the total number of molecules that define the canonical pathway from within the IPA knowledgebase. (C) JM6 influences the synaptic long-term depression pathway by upregulating AMPA receptor signaling.

Fig 5. Heatmap showing hippocampal miRNA expression profiles of antidepressant-treated injured rats.

Prozac (fluoxetine), imipramine and Zoloft (sertraline) restored TBI-induced miRNA expression profiles to those of sham-injured control rats. Probes differentially expressed in at least one of seven comparisons are displayed as normalized to the median expression across all 14 samples (72 probes, >1.2Fold, p<0.05).

Fig 6. Heatmaps of drug-induced gene expression (GE), performed two years apart.

(A) Hippocampal GE profiles from injured rats treated with fluoxetine, imipramine or sertraline, [data from 2014 (right), is from the same rats as in Fig 5] are similar to hippocampal GE profiles of rats treated with JM6, E2 or PMI, data from 2012 (left). Genes > 1.5 fold in the same direction in at least five drug treatments vs TBI and p-value < .05 in at least one drug treatment vs. TBI are displayed as normalized to the mean expression of the study specific TBI samples (69 probes). (B) The FST was used to test the antidepressant efficacy of JM6. First imipramine was tested to validate the depression assay. Imipramine (15 mg/kg) decreased immobility and increased climbing, in the forced swim test (FST). (C) While there were no significant effects on duration of climbing, swimming or immobility, JM6 had dose-dependent antidepressant effects on latency to immobility, The FST validates antidepressant efficacy of JM6 *p < .05.