The present study demonstrates that minocycline exerted direct protection on neurons, in the absence of astrocyte participation, against ischemic stroke. An equally important finding is that minocycline not only promoted dose-dependent neuroprotective effects, but also induced toxicity at a high dose for both neurons and astrocytes. Both sets of in vitro and in vivo studies corroborated such neuroprotection and toxicity profile of minocycline. In addition, in vitro mechanistic studies revealed that a major therapeutic pathway, by which minocycline prevented the ischemic cell death, is via an anti-apoptotic mechanism. Parallel in vivo data showed that low dose, but not high dose, minocycline attenuated stroke-induced behavioral deficits, decreased apoptotic cell death and reduced cerebral infarction. The intravenous route and the post-stroke delivery further advance the utility of minocycline in the clinic.
To date, the primary CNS mechanism implicated in minocycline neuroprotection is the drug's highly potent inhibitory effect on microglial activation, which is achieved by blocking the phosphorylation of p38 and the translocation of 5-Lipoxygenase into the nucleus, thereby preventing the release of cytokines and the induction of inflammation [15, 40, 43–45]. On the other hand, recent evidence has shown that minocycline in the periphery affords protective effects on kidney cells against ischemia via the apoptotic Bcl-2/cytochrome c pathway . We report here that minocycline also promoted protection against ischemia in the CNS by arresting apoptotic Bcl-2/cytochrome c pathway. In our in vitro OGD condition, cultured neurons and astrocytes underwent apoptosis-like cell death as revealed by induction of caspase 3/7 activity and DNA fragmentation (TUNEL positive cells). Treatment with low, but not high dose minocycline abrogated apoptosis characterized by reduced caspase 3/7 activity and decreased number of TUNEL positive cells. Of note, such blockade of OGD-induced apoptosis by low dose minocycline only occurred in cultured neurons and was not evident in cultured astrocytes.
In stroke brains, increased chemokine mRNA expression displays a biphasic profile, being found initially in neurons, then subsequently in astrocytes . Of interest, high levels of chemokines were found in areas of gliosis surrounding recent infarcts [47–49] and correlated with the accumulation of macrophage/microglia in the ischemic lesion, suggesting chemokine's role in the recruitment of inflammatory cells into the brain in response to stroke [49–52]. Based on the above observations, suppressing chemokine elevation during its initial onset in neurons, before astrocytes become involved in this inflammation-ischemia-triggered secondary cell death, may provide better therapeutic outcome than treatment regimen targeting astrocytes. Although in recent years enhancing astrocyte survival has been suggested as an alternative protective approach against ischemic damage [53, 54], therapeutic strategies that confer direct neuronal protection are likely to improve clinical prognosis. The present results indicate that minocycline, in addition to its established anti-microglial activity, could directly protect neurons via an anti-apoptotic mechanism.
To further clarify the anti-apoptotic features of minocycline, we examined the alterations in expression of apoptosis-related components, specifically the cell survival-enhancing Bcl-2/cytochrome c pathway. Our results revealed that low dose minocycline protected again neurons, but not astrocytes against OGD by elevating Bcl-2 expression and consequently strengthening the anchor of cytochrome c to the mitochondria. We extend here the participation of Bcl-2/cytochrome c pathway in minocycline's direct protection of OGD-exposed neurons, previously shown in ischemic kidney cells .
To reveal the possible toxic side effects of minocycline, we similarly examined cell survival and apoptosis in OGD-exposed cultured neurons and astrocytes treated with high dose (100 μM) minocycline. Minocycline at a high dose was toxic as revealed by markedly reduced cell survival of both OGD-exposed neurons and astrocytes compared to vehicle treated OGD-exposed cells. Moreover, relative to vehicle treated OGD exposed cells, high dose minocycline did not elevate Bcl-2 expression, but increased caspase 3/7 activity, as well as the number of TUNEL positive cells in the ischemic striatum.
In parallel to the toxicity profile of minocycline observed in the in vitro OGD condition, high dose minocycline exacerbated both behavioral and histological deficits in stroke animals. In contrast, low dose minocycline increased Bcl-2, but decreased TUNEL positive cells in the ischemic peri-infarct area. Moreover, low dose minocycline-treated animals exhibited a pattern of Bcl-2 expression that was only found in neurons, but not in astrocytes, further supporting the neuroprotective mechanism whereby minocycline exerted anti-apoptotic effects directly on neurons.
In previous reports, therapeutic efficacy in different animal models of neurological disorders was consistently observed when minocycline was administered 3 mg/kg-45 mg/kg either intravenously or intraperitoneally [37–39]. Recent studies have suggested that depending on the animal species (i.e., mice), minocycline may confer neurotoxicity in experimental ischemia [19, 38, 42] and Parkinson's disease . These reports and the present data, taken together, further clarified the toxicity profile of minocycline highlighting critical factors including type of cell line (for in vitro studies), experimental injury paradigm, and dosage, as well as delivery route of minocycline [19, 37–39].
The observed dose-dependent protection of neurons over astrocytes by low dose minocycline, and the neurotoxic effects of high dose minocycline provide guidance in designing the clinical protocol for stroke patients. Because astrocytes play a crucial role in blood brain barrier maintenance, a perturbed astrocyte viability, as seen with high dose minocycline, may compromise the barrier that could allow inflammatory cells to penetrate the CNS and exacerbate the stroke deficits. Indeed, most of the animals that received high dose minocycline exhibited severe edema. The establishment of an effective dose range that confers protection on neurons, while not disrupting astrocytes, would perhaps lead to improved therapeutic outcome of minocycline.
Minocycline's inability to protect astrocytes or to increase Bcl-2 expression in these cells in vitro seems to be the most original finding of this study. Our approach to use low doses and high doses to show minocycline's protection versus toxicity in the same in vitro and in vivo stroke models is clinically relevant since the drug is already in clinical trials. At first glance, the choice for the present high doses of minocycline (100 μM or 100 mg/kg) would seem extremely high, considering that in a clinical trial  multiple sclerosis patients who received orally 200 mg minocycline daily dose (equals 2.85 mg/kg for 70 kg person, serum levels reach maximum of 4 mg/l = 8 uM) during a 6-month period exhibited no observable significant side effects. However, our recent study clearly demonstrates that a 3 mg/kg intravenous dose of minocycline is required to obtain serum levels in rats similar to that achieved in humans after a standard 200 mg dose , suggesting differences in the drug metabolism between rats and humans. Accordingly, the rationale for selecting the present doses of minocycline is based on our studies [37, 39] and those of others [13, 28, 34, 55, 56] indicating that these doses correspond to the clinically relevant doses of minocycline in stroke rodent models. In addition, we extended the high dose range to reveal the toxicity profile of minocycline. Indeed, a multiple high dose minocycline injection regimen, involving subcutaneous 135 mg/kg over 2 days followed by 68 mg/kg over the succeeding two days, was recently shown to exacerbate the striatal damage produced by hypoxic-ischemic injury in rats . Depending on the dose and route of delivery, discordant results and conclusions accompany the actions of minocycline in various stroke and neurodegeneration models. The present data underscore that the minocycline dose is critical as it might attenuate or worsen the stroke outcome. While many studies have pursued intraperitoneal or subcutaneous injections of high dose minocycline in order to promote neuroprotection, we show here that robust neuroprotective effects in acute stroke can be achieved with intravenous low dose minocycline, thereby circumventing the toxicity now increasingly being recognized with high dose minocycline. This neuroprotective action of low dose minocycline at a clinically suitable dosing regimen advances the entry of this drug for phase I human stroke trials. In view of a recent clinical trial showing that the relatively high dose of 400 mg/day for 9 months minocycline led to an accelerated deterioration in the amyotrophic lateral sclerosis functional rating scale, accompanied by gastrointestinal and neurological adverse events , a more careful consideration of minocycline dose is indicated as similarly shown in the present study.