Male C57BL/6 mice at 8–10 weeks of age were obtained from The Jackson Laboratory (Bar Harbor, ME). Mice were subjected to TBI at 3–4 months of age, treated for 2 weeks, and later underwent behavioral assessments (Fig. 1A). Mice tested for anxiety corresponded to some of the animals for which we previously reported visual deficits.²³ Other cohorts of mice were used to examine depression and fear. The numbers of mice per group for each experiment are provided in the relevant figure legends. Animal studies were performed in accordance with University of Tennessee Health Science Center Institutional Animal Care and Use Committee–approved and Department of Defense–approved protocols and complied with National Institutes of Health and Society for Neuroscience guidelines.

Our focal cranial blast model of mild TBI has been previously described in detail.¹⁹ In brief, mice were anesthetized with avertin (400 mg/kg, injected intraperitoneally [i.p.]) and inserted into protective tubing that shielded the animal, except for a 7.5-mm diameter area on the left side of the head (Fig. 1B). A single blast of 50 pounds per square inch (psi) above atmospheric pressure, which produces widespread axonal injury but no overt brain contusion,^19,26 was delivered to the mouse head. Sham blast (0-psi) mice were handled in the same way, but with a metal plate blocking the blast wave from reaching the animal. After blast, mice were kept warm and recovered from anesthesia in 15–30 min. Acetaminophen was provided in drinking water at 1.6 mg/mL for 24 h before and after blast.

Raloxifene (Sigma-Aldrich, St. Louis, MO), diluted in vehicle containing ethanol/Cremophor/0.9% saline (5:5:90), was administered at a dose of 5 or 10 mg/kg of body weight, as in our past studies.^23–25 Mice were injected i.p. with raloxifene or vehicle, beginning 2 h after blast, and again at approximately the same time every day (±1 h) for the next 14 days (15 doses in total). For brevity, mice that received blast and vehicle will be referred to as blast-vehicle, sham mice that received vehicle as sham-vehicle or simply sham, and mice that received blast and drug as blast-ral5 or blast-ral10, depending on the dose.

The tail suspension test was used to assess the characteristic of depression termed behavioral despair, as described in our past studies.^16,17 Mice suspended by the tail eventually stop attempting to escape and become immobile, with a depressive-like state indicated by a longer duration of immobility compared to control animals. Each mouse was suspended by its tail and video recorded. Immobility was analyzed over a 5-min period using automated software (FreezeFrame; Coulbourn, Whitehall, PA).

Fear responses were examined, as previously described,^16,17,26 using a fear-conditioning chamber (Model ENV-008; MED Associates Inc., Fairfax, VT) and automated software (FreezeFrame; Coulbourn). Mice were acclimated to the chamber for 4 min and then received five training trials, each consisting of a 12-kH tone conditioned stimulus (CS) for 30 sec, coterminating with a 0.250-sec, 0.4-mA foot shock (unconditioned stimulus; US), with 3.5 min between trials. Mice were returned to the chamber the following day, and contextual fear was assessed for 3 min. This was followed by fifteen 20-sec presentations of the CS alone, with 2 min between presentation onsets, to evaluate retention of the conditioned fear. Contextual and conditioned fear responses were measured in 20-sec time blocks.

Rodent behavior in a light-dark box was first used to test the effectiveness of anxiolytics such as benzodiazepines.^27,28 The basis of light-dark box testing is that rodents prefer dark, enclosed areas where they can hide, with more time spent in the dark compartment of the testing arena indicative of greater anxiety. As described previously,²³ the light-dark box consisted of two equally sized compartments—an open chamber with clear walls and an enclosed chamber with black walls—connected by an opening. The test arena was covered by a black drape during the testing, to minimize outside distractions. The enclosed chamber contained a light bulb that could provide illumination at 500 or 1000 lux, and the illumination of the open chamber ranged from slightly >0 lux to a maximum of 8 lux when the enclosed chamber was at 1000 lux. Each test began with 5 min of no light in the enclosed chamber (0 lux), followed by 5 min of 500 lux and another 5 min of 1000 lux. Infrared laser beams detected the mouse location, and software measured time spent in each compartment. Given that mice are averse to bright light, they spend less time in the enclosed chamber as its illumination increases. More time spent in the enclosed arena compared to sham mice, particularly as the illumination increases, is indicative of heightened anxiety.

Data were analyzed with one-way analysis of variance (ANOVA), followed by post hoc comparisons, using SPSS software (SPSS, Inc., Chicago, IL). The specific corrections for multiple comparisons are provided in the figure legends.

As in our past studies,^16,17 mice subjected to focal cranial blast TBI showed a longer duration of immobility, indicative of increased depression, over the 5-min testing period compared to sham-vehicle mice (p = 0.0002; Fig. 2). The amount of immobility for blast-ral5 and blast-ral10 mice was significantly less than for blast-vehicle mice (p = 0.023 and p = 0.031, respectively) and nearly the same as in sham-vehicle mice. Thus, both doses of raloxifene rescued the depression-like state that is otherwise produced by focal cranial blast.

Sham-vehicle mice spent roughly equal amounts of time in the two chambers at the start of the testing session. They decreased their occupancy of the enclosed chamber as its illumination was increased (Fig. 4A), because of the aversiveness of the increasingly brighter lights. Blast-vehicle mice consistently spent a greater proportion of their time than sham-vehicle mice in the enclosed chamber at each of the three illumination levels, although the differences were not statistically significant. However, combining the data across illumination levels (Fig. 4B) revealed that blast-vehicle mice occupancy of the enclosed chamber was significantly greater than for sham-vehicle mice (p = 0.014), suggesting that blast-vehicle mice exhibit heightened anxiety. Occupancy of the enclosed chamber by blast-ral5 mice was significantly less than for blast-vehicle mice (p = 0.025) and similar to sham-vehicle mice, showing that 5 mg/kg of raloxifene reversed the increase in anxiety produced by focal cranial blast. Blast-ral10 mice did not differ significantly from sham-vehicle mice or blast-vehicle mice, suggesting that the higher raloxifene dose partially rescued the increase in anxiety.

Numerous brain regions are involved in regulating mood, including pre-frontal, insular and anterior cingulate cortex, hippocampus, several amygdala subnuclei, nucleus accumbens, and the bed nucleus of the stria terminalis.^11,32–35 The precise brain regions that cause depression, fear, and anxiety after mild TBI, and the nature of the alterations in them, are not well understood. Thus, determining how raloxifene acts on specific regions and cells within the brain to yield functional rescue, though desirable, is beyond the scope of the current study.

Mild TBI is characterized by widespread axonal injury.^5,36,37 For our experimental model, focal cranial blast, damaged axons are abundant in white matter tracts after a few days^19,20,23 and degenerating axons are visualizable after a few weeks.¹⁹ Such axonal injury may disturb communication between (and within) brain regions that regulate emotional state and thereby lead to a disordered mood.^14,15 Neuronal loss¹⁸ and altered synaptic signaling ensuing from mild TBI may similarly disrupt communication between brain regions. The reduction in axon and neuron loss with raloxifene treatment,²³ as well as the possible restoration of normal synaptic signaling, would then lessen the amount of disruption and thereby mitigate emotional deficits.

The axonal injury occurring after mild TBI sets off a series of secondary processes, one of the more prominent being microglial activation. The initial microglial responses are typically proinflammatory and, as such, cause further damage.^38–41 We have, in fact, found activated microglia alongside damaged axons in white matter tracts^19,20,23 and in some brain regions^18,26 a few days after focal cranial blast, and increased expression of proinflammatory M1 markers.^18,20,23 Consistent with a role of activated microglia, brain levels of proinflammatory cytokines are elevated in rodent models of depression, fear, and anxiety and with depression in humans.^31,42–48 Thus, microglial activation in mood-regulating brain regions may contribute to progression of the injury after mild TBI and exacerbate the outcome. As an aside, it should be noted that our experimental approach does not cause contusive injury to the brain, and our previous morphological studies have not revealed any obvious macrophage invasion of the brain parenchyma.

Microglia rapidly upregulate CB2 expression when neighboring cells are damaged, allowing drugs that bind CB2 receptors to specifically target microglia.^40,49–52 CB2 inverse agonists stabilize CB2 receptors, which are otherwise constitutively active, in an inactive state, thereby biasing microglia away from the proinflammatory M1 state, toward the protective M2 state. CB2 inverse agonists thus derive their benefit from both types of actions.^21,53,54 Our previous studies have demonstrated that the loss of optic nerve axons that otherwise ensues after focal cranial blast TBI is significantly reduced by 2 weeks of treatment with either raloxifene²³ or the CB2 inverse agonist, SMM-189,²⁰ suggesting that microglia can drive axonal degeneration.

In addition, SMM-189 mitigates abnormalities in oscillatory neuronal activity recorded from the hippocampus and pre-frontal cortex⁵⁵ and rescues around half the neuron loss in cortex and striatum,¹⁸ supporting a role for microglia in these processes. More relevant to the work reported here, we have also shown that SMM-189 normalizes depression and contextual fear and diminishes conditioned fear after focal cranial blast.¹⁷ Further, SMM-189, the CB2 inverse agonist SR144528, and raloxifene yield a similar rescue of the contrast-sensitivity deficit produced by focal cranial blast, consistent with the idea that raloxifene rescue of visual deficits is attributable to effects on CB2 receptors.^17,20,,23

Although we have previously shown that raloxifene benefit for visual contrast sensitivity does not depend on its estrogenic effects²³ and that raloxifene modulates microglia after brain and ocular injury,^23–25 we cannot rule out the possibility that the rescue of emotional disorders reported here is partly, or entirely, attributable to raloxifene action at estrogen receptors. Estrogen is known to exert neuroprotective effects through multiple mechanisms, including inhibiting apoptosis and decreasing neuroinflammation.⁵⁶ Estrogen receptors are expressed by microglia, neurons, and astrocytes in diverse brain regions, including the amygdala, pre-frontal cortex, and hippocampus.⁵⁷ Raloxifene, like other SERMS, can elicit either agonist or antagonist effects at estrogen receptors, depending on the tissue, cell type, and context.^58,59

An unresolved issue from the present study concerns differences in results for the two raloxifene doses. Raloxifene at 5 and 10 mg/kg provided similar benefit for depression and contextual fear, but the lower dose was better for anxiety, whereas the higher dose was better for conditioned fear. We previously observed dose differences in other functional and structural assessments and on the expression of M1- and M2-state microglial markers,^23–25 suggesting that these may stem from variability in the extent and timing of microglial CB2 upregulation in different brain regions because of variation in severity of the injury in those regions. Regardless of the results in animal studies, optimizing the dosage, dosing frequency, and treatment window would be important components of any initial clinical testing. Although we typically administer raloxifene starting 2 h after injury, delaying treatment up until day 3 after focal cranial blast TBI is still effective,²³ adding to the feasibility of raloxifene therapy. That a treatment period of only 2 weeks provides benefit for many of the end-points in the current and our previous studies^23–25 suggests that a lengthy period of treatment would not be necessary.

An important drawback in beginning to assess the efficacy of SMM-189 and SR144528 for mild TBI in patients is that neither has undergone phase 1 safety trials. By contrast, given that raloxifene is already FDA approved, phase 2 clinical trials could start at any time. Raloxifene has been used since the late 1990s for treating post-menopausal osteoporosis. It can be taken orally, is also safe and effective in men, and has no evident adverse hormonal side effects.^60–63 Interestingly, daily treatment with 120 mg of raloxifene (twice the dose used for osteoporosis) improves attention and memory in schizophrenic patients,⁶³ indicating sufficient brain penetration for achieving benefit. Although we have administered raloxifene at yet higher doses, those doses are known to be safe in humans.⁵⁸

We have previously demonstrated that raloxifene improves visual outcomes in mice after mild TBI produced by focal cranial blast,²³ mild TBI produced by an impact to the head,²⁴ and after ocular blast injury.²⁵ Here, we expand upon our evidence for raloxifene benefit, showing that raloxifene reduces the depression, heightened fearfulness, and anxiety ensuing from focal cranial blast TBI. Given that mild TBI in humans often leads to a similar set of persistent, debilitating symptoms, a regimen of raloxifene delivered in the early aftermath of the injury may help reduce adverse emotional outcomes, as well as visual impairments. Taken together, our findings strongly support the consideration of raloxifene for treating mild TBI in humans.