Treating Traumatic Brain Injury with Minocycline

Abstract

Traumatic brain injury (TBI) results in both rapid and delayed brain damage. The speed, complexity, and persistence of TBI present large obstacles to drug development. Preclinical studies from multiple laboratories have tested the FDA-approved anti-microbial drug minocycline (MINO) to treat traumatic brain injury. At concentrations greater than needed for anti-microbial action, MINO readily inhibits microglial activation. MINO has additional pleotropic effects including anti-inflammatory, anti-oxidant, and anti-apoptotic activities. MINO inhibits multiple proteins that promote brain injury including metalloproteases, caspases, calpain, and polyADP-ribose-polymerase-1. At these elevated doses, MINO is well tolerated and enters the brain even when the blood–brain barrier is intact. Most preclinical studies with a first dose of MINO at less than 1 h after injury have shown improved multiple outcomes after TBI. Fewer studies with more delayed dosing have yielded similar results. A small number of clinical trials for TBI have established the safety of MINO and suggested some drug efficacy. Studies are also ongoing that either improve MINO pharmacology or combine MINO with other drugs to increase its therapeutic efficacy against TBI. This review builds upon a previous, recent review by some of the authors (Lawless and Bergold, Neural Regen Res 17:2589–92, 2022). The present review includes the additional preclinical studies examining the efficacy of minocycline in preclinical TBI models. This review also includes recommendations for a clinical trial to test MINO to treat TBI.

Introduction

Approximately 2.5 million cases of traumatic brain injury (TBI) per year in the US result in over 50,000 fatalities [2]. In spite of its importance for public health, no drugs have been approved to treat TBI. Multiple challenges have hindered drug development. Head trauma produces a heterogenous injury that depends upon its severity and how the trauma was produced. Regardless of this injury heterogeneity, TBI produces both rapid and delayed brain damage [3, 4]. Primary injury includes mechanical deformation of the brain that produces focal lesions, edema and hemorrhage, axonal shearing, and demyelination [3, 4]. Cytogenic and vasogenic edema leads to increased intracerebral pressure and further necrotic and apoptotic cell loss, as well as loss of axons, myelin and oligodendrocytes. Thus, both gray and white matter injury occurs rapidly and evolves for weeks post-injury [5].

Primary injury induces a subsequent gray and white matter damage that lasts for weeks to months [3–5]. Multiple mechanisms underlie this secondary injury. Impaired ATP production from damaged mitochondria produces neuronal depolarization and increased levels of intracellular calcium. Mitochondrial damage also results in excessive production of reactive oxygen species. Excitotoxic neuronal loss results from increased synaptic release of glutamate. Damage to cerebral circulation leads to hypoxia, hypoglycemia, and breakdown of the blood-brain barrier. Axons and oligodendrocytes are also damaged leading to decreased axonal transport and loss of myelin [5, 6]. Damage associated molecular patterns (DAMPS) are released from damaged and lysed cells that promote chemokine and cytokine release. Astrocytes and microglia are rapidly activated by chemokines and cytokines that promotes further blood brain barrier breakdown and infiltration of inflammatory cells from the peripheral circulation [7].

Early injury also leads to multiple forms of regulated necrosis [8]. Necroptosis is triggered by TNF family receptors that activate receptor interacting protein kinases (RIPK) 1 and 3. Inflammasome activation can produce pyroptosis that releases proinflammatory cytokines and DAMPS through newly formed membrane pores. Pore formation may also result in cell lysis. Excess lipid peroxidation can deplete glutathione and increase free intracellular iron levels leading to ferroptosis. Parthanatos arises from NAD⁺ depletion and energy failure arising from DNA damage activating poly (ADP-ribose) polymerase-1 (PARP-1). TBI also induces cytophillin D to open the mitochondrial permeability transition pore leading to mitochondrial dysfunction, oxidative stress, energy failure and finally, necrosis.

Primary injury is rapidly followed by a secondary injury that includes glutamate excitotoxicity, impaired intracellular calcium homeostasis, mitochondrial dysfunction, energy failure, free radical formation, apoptosis, blood–brain barrier breakdown, and chronic neuroinflammation [3, 7]. Apoptosis can occur for days to months after injury [6]. All these processes may result in a long-lasting progressive atrophy of gray and white matter [6]. TBI severity also ranges from mild to severe. Mild TBI is characterized by the absence of radiological abnormalities. Moderate to severe TBI is characterized by loss of consciousness, hematomas, hemorrhage, contusions, and radiological abnormalities [3]. Regardless of the severity of the injury, TBI rapidly triggers multiple pathophysiological mechanisms that are subsequently altered and amplified over time.

The pathophysiology of TBI in the first few hours to days is a major determinant of time to first dose since drug targets may rapidly arise, dissipate, or change [9]. In contrast, pathophysiology weeks to months after TBI may be more stable or not change at all. The unavoidable hours that elapse between injury and treatment indicate that only delayed damage can be targeted for drugs to treat TBI [9]. The rapid and complex progression of TBI presents large obstacles to drug development [9].

The FDA-approved drug minocycline (MINO) has undergone extensive preclinical testing to examine its efficacy to treat TBI (Tables 1 and 2). Many, but not all, preclinical studies show that MINO improves outcomes after experimental TBI. In contrast, limited clinical trials for MINO to treat TBI have been inconclusive (Table 3). MINO also failed multiple clinical trials for a variety of neurological and psychological diseases [10].

Table 1.

Preclinical TBI studies using minocycline

Study	Rigor	Species, sex, age	TBI model, controls	Dose, time to first dose	Outcomes
Sanchez-Meija, 2001 [18]	Randomization, yes; blinded, NS	Mouse, age NS, sex NS	Moderate CCI, sham-CCI	45 mg/kg, 12 h pre-injury 90 mg/kg 30 min 12 h and 24 h post-injury 90 mg/kg, 30 min 12 h and 24 h post-injury	1 D rotarod $\uparrow$ 4 D rotarod $\uparrow$ 4 h IL-1β $\uparrow$ 4 D LV $\uparrow$ 1 D caspase-1,3 $\uparrow$
Sheng, 2006 [40]	Randomization, yes; blinded, yes	Rat, male, 200–250 g	Moderate CCI, sham-CHI	Pretreatment 45/mg/kg twice daily 2 D and 1 D, once 30 min; posttreatment 1 h, twice daily 1 D, 2 D	1 D LV $\uparrow$ 4 D LV $\to$ 1–14 D NSS $\to$ 4–5 D inclined plane $\uparrow$ 7 D LV $\uparrow$ 7 D CA1 neurons $\uparrow$
Bye, 2007 [19]	Randomization, NS; blinded, NS	Mouse, male, 12–14 weeks	Moderate CHI, sham-CHI;	45 mg/kg 30 min PI, 45 mg/kg every 12 h until sacrifice	1–4 D NSS $\to$ 1 D beam test $\uparrow$ 1 D LV $\uparrow$ 4 D LV $\to$ 4 D apoptosis $\to$ 4 D MP/MG $\uparrow$ 4 D neutrophils $\to$ 4 h cytokines (IL-1 β, IL-6, G-CSF, MCP-1, CXCL4) $\uparrow$
Homsi, 2009 [20]	Randomization, NS; blinded, NS	Mouse, male, 28–30 g	Moderate WD, sham-WD-saline, WD-saline	45 mg/kg 5 min	6 h IL-1β$\to$ 6 h MMP-9 $\to$
		Mouse, male, 28–30 g		90 mg/kg 5 min	6 h IL-1β$\uparrow$ MMP-9 $\uparrow$
		Mouse, male, 28–30 g		45 or 90 mg/kg 5 min, 45 mg/kg 3 h	6 h GSH $\to$
		Mouse, male, 28–30 g		90 mg/kg 5 min, 45 mg/kg 3, 9 h	1 D edema $\to$ 1 D MMP-9 $\uparrow$ 1 D string test $\uparrow$
Abdel-Baki, 2009 [62]	Randomization, NS; blinded, NS	Rat, male 250–300 g	Moderate CCI, sham-CCI	45 mg/kg 1 h, 1 D, 2 D	7 D APA acquisition $\uparrow$ 12 D APA retention $\to$ 14 D myelin $\uparrow$
Homsi, 2010 [46]	Randomization, yes; blinded, yes	Mouse, male, 28–30 g	Moderate WD, WD-saline	90 mg/kg 5 min, 45 mg/kg 3 and 9 h	1 D MG $\uparrow$ 1 D LV $\uparrow$ 2–84 D open field $\uparrow$ 2–84 D body weight $\uparrow$
Siopi, 2011 [21]	Randomization, yes; blinding, NS	Mouse, male, 28–30 g	Moderate WD, naïve, WD-vehicle	90 mg/kg, 5 min, 45 mg/kg 3 h, 9 h	1 D sAAP$\alpha \uparrow$ 84 D corpus callosum volume $\uparrow$ 84 D thalamus volume $\uparrow$ 84 D lateral ventricle volume $\uparrow$ 84 D GFAP $\uparrow$ 84 D MP/MG $\uparrow$
Siopi, 2012 [45]	Randomization, NS; blinding, yes	Mouse, male, 28–30 g	Moderate WD, naïve, WD-vehicle	90 mg/kg, 5 min, 45 mg/kg 3 h, 9 h	7–84 D odorant avoidance $\uparrow$ 84 D olfactory bulb surface area $\uparrow$ 21–91 D NORT $\uparrow$ 21–91 D elevated plus maze $\to$ 21–91 D elevated zero maze $\to$
Ng, 2012 [41]	Randomization, yes blinding, yes	Mouse, sex, NS, 28–34 g	Moderate WD, sham WD	45 mg/kg 30 min, 45 mg/kg every 12 h for 7 D	3–42 D NSS $\uparrow$ 7 D MP/MG $\uparrow$ 7 D neurogenesis $\to$
Kovesdi, 2012 [31]	Randomization, NS; blinding, NS	Rat, male 245–265 g	Mild blast, sham-blast	50 mg/kg 1 h, 1–4 D	9 D elevated plus maze $\to$ 46 D elevated plus maze $\to$ 10 D BM $\to$ 47 D BM $\to$ 51 D c-reactive peptide $\uparrow$ 51 D CCL2 $\uparrow$ 51 D claudin 5 $\uparrow$ 51 D neuronal-specific enolase $\uparrow$ 51 D neurofilament heavy chain $\uparrow$ 51 D tau $\uparrow$ 51 D S100B $\uparrow$ 51 D GFAP $\uparrow$ 51 D corticosteroid $\uparrow$ 51 D vascular endothelial growth factor receptor 2 $\uparrow$
Haber, 2013 [22]	Randomization, NS; blinding, NS	Rat 250,300	Moderate CCI, sham-CHI	45 mg/kg 1 h, 1 D, 2 D	7 D conflict active place avoidance $\uparrow$ 7 D spaced active place avoidance $\to$ 2 D MG activation $\uparrow$
Lam, 2013 [23]	Randomization, NS; blinding, yes	Rat	Moderate CCI, CCI- vehicle	25 mg/kg 1–7 D	7 D–168 D MP/MG $\uparrow$ 7 D GFAP $\uparrow$ 12 D GFAP $\to$ 168 D GFAP $\uparrow$ 168 D LV $\to$ 56 D vermicelli handling $\to$ 56 D thigmotaxis $\uparrow$ 56 D MWM $\uparrow$
Vonder Haar, 2014 [39]	Randomization, yes; blinding, NS	Rat, male 350 g	Moderate CCI, sham-CCI	50 mg/kg 1 h, every 12 h for 3 D	7–16 D grid walk $\uparrow$ 7 D rotarod $\to$ 15 D MWM $\to$ 25 D LV $\uparrow$
Lopez-Rodriguez, 2015 [24]	Randomization, yes; blinding, yes	Mouse, male, 28–30 g	Moderate WD, naive	90 mg/kg, 5 min, 45 mg/kg 3 h, 9 h	1 D edema $\uparrow$ 1 D NSS $\to$ 1 D MP/MG $\uparrow$ 1 D β-APP $\uparrow$
Shochat, 2015 [63]	Randomization, NS; blinding, NS	Mouse, male, 40 g	Mild WD	45 mg/kg 20 min	1 h oxyhemoglobin $\uparrow$ 1 h total hemoglobin $\uparrow$ 1 h arterial oxygen saturation $\uparrow$ 1 h edema $\uparrow$
Hanlon, 2017 [64]	Randomization, yes; blinding, yes	Rat, male and female, 11 D	Moderate CCI, sham-CCI	45 mg/kg 5 min, every 12 h for 3 D	3 D MP/MG $\uparrow$ 7 D MP/MG $\to$ 3 D cortical Fluoro-Jade positive cells $\uparrow$ 7 D cortical Fluoro-Jade positive cells $\to$ 13 D MWM $\to$
Chhor, 2017 [65]	Randomization, yes; blinding, yes	Mouse, male female, 4–5 g	Mild WD, sham-WD	45 mg/kg 5 min, 1 D, 2 D	1 D inflammatory cytokines $\uparrow$ 1 D ventricular volume $\uparrow$ 5 D ventricular volume $\to$ 1 D apoptosis cortex, hippocampus, striatum $\uparrow$ 1 D MP/MG $\uparrow$ 5 D LV $\to$ 5 D myelin $\to$
Haber, 2018 [66]	Randomization, yes; blinding, no	Rat, male 250–300 g	Moderate CCI, sham-CCI	45 mg/kg 1 h, 1 D, 2 D	14 D myelin 14 D $\uparrow$ 14 D oligodendrocytes 14–4 D $\uparrow$ 14 D oligodendrocyte apoptosis $\uparrow$ 4 D MP/MG activation $\uparrow$ 7 D MP/MG activation $\uparrow$,
Sangobowale, 2018a [43]	Randomization, yes; Blinded, yes	Rat, male 250–300 g	Moderate CCI, sham-CCI	22.5 mg/kg 6 h, 1 D and 2 D	7 D BM $\uparrow$, 7 D APA, entrances $\uparrow$ 7 D APA time to first entrance $\to$ 14 D hippocampal MAP2 $\to$ 14 D myelin $\uparrow$
				22.5 mg/kg 12, 1 D and 2 D	7 D BM $\uparrow$, 7 D APA entrances $\uparrow$ 7 D APA time to first entrance $\to$
				22.5 mg/kg 1 D, 2 D and 3 D	7 D BM $\uparrow$ 7 D APA, entrances $\to$ 7 D APA time to first entrance $\to$ 14 D hippocampal MAP2 $\to$ 14 D myelin $\to$
		Mouse, male 28–30 g	Moderate CHI, sham-CHI	22.5 mg/kg 12 h, 1 D, 2 D	7 D BM $\uparrow$ 7 D APA entrances $\uparrow$ 7 D APA time to first entrance $\to$ 14 D hippocampal MAP2 $\to$ 14 D myelin $\uparrow$
				22.5 mg/kg 1 D, 2 D, 3 D	7 D 12–24 h, BM $\uparrow$, 7 D 24 h APA, entrances $\to$ 7 D 24 h APA, time to first entrance $\to$ 14 D hippocampal MAP2 14 D $\to$ 14 D myelin $\to$
Sangobowale, 2018b [49]	Randomization, yes; blinded, yes	Mouse, male 28–30 g	Moderate CHI, sham-CHI	22.5 mg/kg 12 h, 1 D, 2 D	14 D oligodendrocytes $\uparrow$
Simon, 2018 [67]	Randomization, yes; blinded, NS	Rat, 35–40 g	Moderate CCI, sham-CCI	90 mg/kg 10 min and 20 h	1 D high mobility group B1 $\uparrow$ 7 D MP/MG $\uparrow$ 7 D Fluoro-Jade positive cells, $\to$ 14 D thalamic neurons $\uparrow$ 14 D LV $\to$ 5 D balance beam, inclined plane $\uparrow$ 14 D MWM $\uparrow$
Taylor, 2018 [51]	Randomization, yes; blinded, NS	Rat, male female, 60–70 D	Moderate CCI, sham-CCI	50 mg/kg 1 h 1 D, 2 D, 3 D	Il-1 β 35 D male $\to$, female $\to$ Il-6 35 D male $\to$, female $\to$ TNF$\alpha$ 35 D male $\to$, female $\to$
Zhang, 2020 [14]	Randomization, NS; blinded, NS	Rat, male, 150–180 g	Moderate CCI, sham-CCI	10 mg/kg 12 h and daily 2–7 D	5–14 D body weight $\to$ 7–14 D foot fault $\to$ 7–14 D cylinder test $\to$ 7–14 D wire hang $\to$ 21 D MWM $\to$
				20 mg/kg 12 h and daily 2–7 D	5–14 D body weight 20 mg/kg $\to$ 7–14 D foot fault $\uparrow$ 7–14 D cylinder test $\uparrow$ 7–14 D wire hang $\to$ 21 D MWM $\uparrow$ 7 D hippocampal neurons $\uparrow$ 7 D cortical nissl postive cells $\uparrow$ 7 D serum iron $\uparrow$ 7 D CSF iron $\uparrow$ 7 D brain tissue iron $\uparrow$ 7 D ferritin $\uparrow$ 7 D transferrin receptor 1 $\uparrow$ 7 D divalent metal transporter 1 $\uparrow$ 7 D ferroportin 1 $\uparrow$ 7 D hepcidin $\uparrow$
				40 mg/kg 12 h and daily 2–7 D	5–14 D body weight $\uparrow$ 7–14 D foot fault $\uparrow$ 7–14 D cylinder test $\uparrow$ 7–14 D wire hang $\to$ 21 D MWM $\uparrow$
Pernici, 2020 [38]	Randomization, yes; blinded, yes randomization, NS; blinded, NS	Mouse, male female, 20–30 g	Mild FPI. sham-FPI	45 mg/kg 45 min 2 D, 3 D	2–7 D NSS $\uparrow$ 2–7 D rotarod $\to$ 60 D NORT $\to$ 60 D open field $\to$ 60 D tail suspension $\to$ 30 D axon loss $\uparrow$
				45 mg/kg 3 D, 4 D or 5 D	2–7 D NSS $\to$ 2–7 D rotarod $\to$ 60 D NORT $\to$ 60 D open field $\to$ 60 D tail suspension $\to$ 30 D axon loss $\uparrow$
He, 2021 [32]	Randomization, yes; blinded, NS	Rat, male 250–300 g	Moderate CCI, sham-CCI	40 mg/kg 5 min prior to injury	3 D cortical apoptosis $\uparrow$ 3 D bax $\uparrow$ 3 D cleaved caspase 3 $\uparrow$ 3 D Bcl-2 $\uparrow$
Wang, 2021 [68]	Randomization, yes; blinded, NS	Mouse, male, 20–35 g	Moderate CCI, sham-CCI	20 mg/kg 30 min, 1 D, 2 D, 3 D	3 D Garcia neurobehavioral score $\uparrow$ 1 D edema $\uparrow$ 1 D occuldin 1 $\uparrow$ 1 D C/EBP homologous protein $\uparrow$ 1 D growth related protein73 $\uparrow$ 1 D β -catenin $\uparrow$ 3 D IL-6 $\uparrow$ 3 D TNF$\alpha \uparrow$ 3 D apoptosis $\uparrow$
Hiskens, 2021 [69]	Randomization, yes; blinded, yes	Mouse, male	Repetitive WD, sham-repetitive WD	50 mg/kg	1 D, 4 D righting reflex $\uparrow$ 4 D MWM latency $\to$ 90 D MWM latency $\uparrow$ 90 D MWM probe trial $\to$ 4 D cortical, hippocampal TNF$\alpha$ mRNA $\uparrow$ 4 D cortical, hippocampal tau mRNA $\uparrow$ 4 D cortical, hippocampal AMPA subunit receptor 1 mRNA $\uparrow$ 4 D hippocampal neurofilament light chain mRNA $\uparrow$ 90 D cortical, hippocampal tar binding protein mRNA $\uparrow$ 90 D cortical, hippocampal TNFa mRNA $\uparrow$ 90 D hippocampal Iba-1 mRNA $\uparrow$
Lu, 2022 [70]	Randomization, NS; blinded, NS	Mouse, male, 20–25 g	Moderate CCI, sham-CCI	45 mg/kg 30 min	3 D cortical Nissl $\uparrow$ 3 D cortical Fluoro-Jade $\uparrow$ 3 D cortical TUNEL $\uparrow$ 3 D cortical cleaved caspase-3 $\uparrow$ 3 D cortical Bcl-2/Bax ratio $\uparrow$ 3 D cortical MMP-9 $\uparrow$ 3 D BBB integrity (sodium fluorescein $\uparrow$) (albumin $\uparrow$) (ZO-1 $\uparrow$) (occludin $\uparrow$) (claudin-5 $\uparrow$) (aquaporin-4 $\uparrow$) 3 D edema $\uparrow$ 3 D cortical GFAP $\uparrow$ 3 D apoptotic astrocytes $\uparrow$
Pechacek, 2022 [48]	Randomization, yes; blinded, yes	Rat, 4–6 months	Bilateral frontal controlled cortical impact TBI	45 mg/kg, 1 h, then every 12 h for 5 D	prelimbic cortex, orbitofrontal cortex, hippocampus Iba-1 $\to$ cortex, hippocampus, TNF$\alpha$ $\to$ lesion volume $\to$ 5-choice serial reaction time task $\to$
				45 mg/kg, 9 weeks, then every 12 h for 5 D	lesion volume $\to$ 5-choice serial reaction time task $\to$
Celorrio, 2022 [71]	Randomization, yes; blinded, yes	Mice, male, 8 weeks old	Moderate CCI, sham-CCI	45 mg/kg, 22–25 h, 2 D, 3 D	CA3 neuronal loss, 7 D $\to$ lesion volume, 7 D $\to$
				90 mg/kg, 22–25 h, 2 D, 3 D	CA3 neuronal loss, 7 D $\uparrow$ lesion volume, 7 D $\to$
				90 mg/kg, 22–25 h, 45 mg/kg 22–25 followed by 5 additional doses every 12 h	CA3 neuronal loss, 7 D $\uparrow$ lesion volume, 7 D $\to$ microglia 7 D, total cells $\uparrow$ MHCII+ cells $\uparrow$ microglial morphology $\uparrow$ peripheral lymphocytes $\uparrow$ peripheral monocytes $\uparrow$ myelin $\uparrow$ oligodendrocyte precursor cell number $\to$, morphology $\uparrow$
Perumal, 2023a [72]	Randomization, no; blinded, no	Rat, male 250 g	Moderate blast, sham-blast	I.V. 3 mg/kg minocycline, 4 h daily 2–4 D	30 D auditory brainstem response, 7, $\uparrow$ auditory cortex, Iba-1 $\uparrow$ GFAP $\uparrow$, NMDAR1 $\uparrow$, GABAA $\to$, NMDAR1 $\uparrow$ inferior colliculus Iba-1 $\to$, GFAP $\to$, GABAA $\to$, NMDAR1 $\to$
				I.V. 3 mg/kg transferrin tagged MINO, 4 h, daily 2–4 D	30 D auditory brainstem response, 7, $\uparrow$ auditory cortex, Iba-1 $\uparrow$ GFAP $\uparrow$, GABAA $\to$, NMDAR1 $\uparrow$, Inferior colliculus Iba-1 $\to$, GFAP $\uparrow$, GABAA $\to$, NMDAR1 $\uparrow$
				I.V. 3 mg/kg, PEGylated MINO, 4 h, daily 2–4 D	30 D auditory brainstem response, 7, $\uparrow$ auditory cortex, Iba-1 $\to$ GFAP $\to$, GABAA $\to$, NMDAR1 $\uparrow$ inferior colliculus Iba-1 $\to$, GFAP $\to$, GABAA $\to$, NMDAR1 $\uparrow$
Perumal, 2023b [73]	Randomization, no; blinded, no	Rat, male 250 g	Moderate blast, sham-blast
Noriega-Navarro, 2023 [74]	Randomization, no; blinded, no	Rat, male, 250–300 g	Moderate-CCI, sham-CCI	Intracortical into motor cortex, 160 µg/4 µl 15 min post-injury	Square beam, $\to$ round beam, pan (rat grimace scale) $\to$, CA1 neuronal loss $\uparrow$, CA2/3 neuronal loss $\to$, dentate gyrus neuronal loss $\uparrow$, striatal neuronal loss $\uparrow$
Bai, 2023 [75]	Randomization, no; blinded, no	Mouse, male and female 8–14 weeks	Penetrating TBI	I.P., 5 mg/ml daily for 3 D	TNF$\alpha$ $\uparrow$, IL-6 $\uparrow$, microglia $\uparrow$

Rigor for each study is provided followed by species, injury model plus severity and control groups, and details of dosing of MINO. Time after injury is followed by whether MINO improved ($\uparrow$) or had no difference ($\to$) in the experimental outcome

APA active place avoidance, BM Barnes maze, CCI Controlled cortical injury, CHI Closed head injury, FPI Fluid percussion injury, GSH Glutathione, GFAP Glial fibrillary acid protein, Iba-1 Ionized calcium-binding adaptor molecule 1, MWM Morris water maze, MMP-9 metalloprotease 9, LV lesion volume, MP/MG macrophage/microglia, NMDAR1 N-methyl-d-aspartate receptor subunit 1, NORT Novel object recognition, NS not specified, NSS neurological severity score, TNF a Tumor Necrosis factor a, WD weight drop

Eight additional preclinical studies published between 2021 and 2023 have been added that were not present in Lawless et al. 2022 [1]

Table 2.

Preclinical studies of drug combinations with MINO

Study	Second drug	Randomization	Blinded	Species, sex, age	Model	Time to first dose, dose	Outcomes
Kelso, 2011 [60]	Melatonin	NS	NS	Rat, male, 225–275 g	Moderate CCI, No sham-CCI control	5 min 60, or 90 min MINO 40 mg/kg, melatonin 5 mg/kg	7 D LV $\to$ 5 D MWM $\to$ 7 D MP/MG $\to$
Vonder Haar, 2014 [39]	Simvastatin	NS	NS	Rat, male, 350 g	Moderate CCI, sham-CCI control	2.5, 12, 24, 36, 48, 60, 72 h MINO 60 mg/kg simvastatin 10 mg/kg	7–16 D locomotor placing test $\to$ 7–12 D rotarod $\to$ 7–16 D MWM $\to$ 7–16 D MWM working memory $\to$
Lam, 2013 [23]	Botulinum toxin, and physical exercise	NS	NS	Rat male 250–300 g	Moderate CCI, sham-CCI	25 mg/kg MINO every 12 h for 7, 12	7 D, 12, 16 D MP/MG $\uparrow$ 7 D, 12, 16 D GFAP $\uparrow$ 56 D MWM $\uparrow$ 56 D paw use cylinder task $\to$ 56 D vermicelli handling test $\to$ 56 D LV $\to$
						25 mg/kg MINO every 12 h for 7, 12, and 16 D 2 doses of 1.25 U botox into 4 left forelimb muscles	56 D MWM $\uparrow$
						2 doses of 1.25 U botox into 4 left forelimb muscles	56 D MWM $\to$
						25 mg/kg MINO every 12 h for 7, 12 physical exercise	56 D paw use cylinder task $\uparrow$ 56 D vermicelli handling test $\uparrow$ 56 D LV $\to$
Abdel Baki, 2009 [62]	N-acetylcysteine	NS	NS	Rat, male 250–300 g	Moderate CCI, sham-CCI	45 mg/kg MINO, 150 mg/kg NAC 1 h, 1 D, 2 D	8–23 D spaced APA, entrances $\uparrow$ 8–23 D spaced APA, time to first entrance $\uparrow$ 14 D myelin $\uparrow$
			NS	Rat, male 250–300 g	Moderate CCI, sham-CCI	45 mg/kg MINO, 150 mg/kg NAC 3 h preinjury	1 h IL-1β $\uparrow$
Haber, 2013 [22]	N-acetylcysteine	NS	NS	Rat 250–300 g	Moderate CCI, sham-CHI-saline, CCI-saline,	45 mg/kg MINO, 150 mg/kg 1 h, 1 D, 2 D	7 D conflict APA, entrances $\uparrow$ 8–23 D spaced APA, time to first entrance $\uparrow$ 2 D MP/MG $\uparrow$ 2 D glial fibrillary acid protein $\uparrow$
Haber, 2018 [66]	N-acetylcysteine	Yes	No	Rat, male 250–300 g	Moderate CCI, sham-CCI, CCI-saline	45 mg/k45 mg/kg MINO, 150 mg/kg 1 h, 1 D, 2 D	14 D myelin $\uparrow$ 1–4 D oligodendrocytes $\uparrow$ 14 D oligodendrocyte apoptosis $\uparrow$ 4 D MP/MG activation $\uparrow$ 7 D MP/MG activation $\uparrow$ CD86 expression $\uparrow$
Sangobowale, 2018a [43]	N-acetylcysteine	Yes,	Yes	Rat, male 250–300 g	Moderate CCI, sham-CCI, CCI-saline	22.5 mg/kg MINO, 75 mg/kg NAC 6 h followed by 1 D and 2 D	7 D BM $\uparrow$ 7 D APA, entrances 6 h dosing $\uparrow$ 7 D APA, time to first entrance $\uparrow$ 14 D hippocampal MAP2 $\uparrow$
						22.5 mg/kg MINO, 75 mg/kg NAC 12 followed by 2 D and 3 D	7 D BM $\uparrow$ 7 D APA, entrances $\uparrow$ 7 D APA, time to first entrance $\uparrow$ 14 D hippocampal MAP2 $\uparrow$ 14 D myelin $\uparrow$
						22.5 mg/kg MINO, 75 mg/kg NAC 1 D, 2 D and 3 D	7 D BM $\uparrow$ 7 D APA, entrances $\to$ 7 D APA, time to first entrance $\to$ 14 D hippocampal MAP2 $\uparrow$ 14 D myelin content $\to$
				Mouse, male 28–30 g	Moderate CHI, sham-CHI, CHI-saline	22.5 mg/kg MINO, 75 mg/kg NAC 12 h, by 2 D and 3 D	7 D BM $\uparrow$ 7 D APA, entrances $\uparrow$ 7 D APA, time to first entrance $\to$ 14 D hippocampal MAP2 $\uparrow$ 14 D myelin $\uparrow$
						22.5 mg/kg MINO, 75 mg/kg NAC 1 D, 2 D and 3 D	7 D $\uparrow$ 7 D APA, entrances $\to$ 7 D APA, time to first entrance $\to$ 14 D hippocampal MAP2 $\to$ 14 D myelin $\to$
Sangobowale, 2018b [49]	N-acetylcysteine	Yes	Yes	Mouse, male 28–30 g	Moderate CHI, sham-CHI, CHI-saline	22.5 mg/kg MINO, 75 mg/kg NAC 12 h, by 2 D and 3 D	2–14 D oligodendrocytes $\uparrow$ 14 D oligodendrocyte apoptosis $\uparrow$
Whitney et al. 2021 [61]	N-acetylcysteine	Yes	Yes	Mouse, male 28–30 g	Moderate CHI, sham-CHI, CHI-saline	22.5 mg/kg, 75 mg/kg 3 D, 4 D, 5 D	7 D BM $\uparrow$ 7 D BM probe trial $\uparrow$, 14 D contralesional CA3 size, number, complexity $\uparrow$ 14 D ipsilateral CA1 size, number, complexity $\uparrow$ 14 D CA1 synapse density $\uparrow$

Study rigor is followed by species, model plus injury severity and control groups, and details of drug dosing. Outcomes are described by time of assessment, outcome type, and whether MINO improved ($\uparrow$) or was indifferent ($\to$) to the experimental outcome APA Active place avoidance, BM Barnes maze, CCI Controlled cortical injury, CHI Closed head injury, MWM Morris water maze, MG macrophage/microglia, NS not specified

Table 3.

Clinical trials of MINO to treat TBI

Author	TBI severity	Design	Enrolled	Dose, duration	Time of first dose	Findings
Meythalar, 2019 [76]	Moderate-severe TBI (Glasgow Coma Score 3–9)	Dose escalation	15	800 mg IV loading followed by 200 mg or 400 mg twice daily for 7 days	Within 6 h post-injury	Safe in TBI population; higher doses trended toward improved disability rating score. No statistical significance
Scott, 2018 [54]	Moderate-severe TBI (Mayo classification)	Randomized clinical trial	15	100 mg PO twice daily for 12 weeks	At least 6 months post-injury	Lowered chronic microglial activation Increased serum neurofilament light protein
Koulaeinejad, 2019 [52]	Moderate-severe TBI (Glasgow Coma Score < 12)	Randomized, double blind, placebo. Includes both	34	100 mg twice daily for 7 D	Within 24 h post-injury	Lowered serum S100B, trending (p < 0.1), lowered serum neuronal specific enolase (p = 0.01), no effect, Glasgow Outcome Scale-Extended

Clinical trial design is provided followed by the number of subjects enrolled, dose, and duration, and the most salient findings

This review also evaluates combinations of MINO with other drugs. Combinations can act in an additive or synergistic manner to increase drug efficacy through and reduce potential unwanted effects [11]. The efficacy of drug combinations is determined by comparison to the efficacy of the individual drugs. With these caveats in mind, this review first evaluates the preclinical studies of MINO and the few clinical trials using MINO to treat TBI will also be summarized.

Search Methodology

The search string “Minocycline traumatic brain injury” received 93 articles from PubMed on 21 June 2023. Forty-five articles were excluded: 12 were earlier reviews, 32 studies whose focus were not TBI or did not dose MINO as a single drug, and one study was retracted. The remaining 48 articles are summarized in Tables 1, 2, and 3.

Minocycline

Minocycline was first introduced in 1967 as a second-generation tetracycline derivative [12]. MINO is a lipophilic semisynthetic tetracycline derivative developed to enter the brain to treat bacterial meningitis [10]. At concentrations higher than needed for anti-microbial action, MINO is an anti-inflammatory and anti-apoptotic that also chelates iron [13, 14].

In humans, MINO has a half-life of 13 h and penetrates into most tissues with CSF levels approximately 20% of serum levels [15]. MINO has close to 100% bioavailability, and absorption is unaffected by food. MINO is largely eliminated in the feces with some renal excretion. Serum concentrations are unaffected by renal impairment [15]. MINO is metabolized to 9-hydroxyminocycline, mono-N-demethylated minocycline, or 4-epiminocycline. It is unknown if these metabolites retain the non-antimicrobial activities of MINO [15].

MINO has improved outcomes in experimental animal models of ischemia, Parkinson’s disease, Huntington’s disease, amyotrophic lateral sclerosis, Alzheimer’s disease, multiple sclerosis, and spinal cord injury [10, 16]. There is broad agreement about the anti-inflammatory action of MINO. MINO inhibits multiple proinflammatory cytokines (IL-1β, IL-6, granulocyte-colony stimulating factor, CCL8, CXCL4) [17].

MINO has multiple modes of action that likely contribute to its therapeutic efficacy in animal models of TBI (Fig. 1). MINO potently inhibits or modulates microglial polarization [18–27]. Modulation of cytokine expression likely underlies MINO’s action on microglia. The endocannabinoid system may play a role since CB1 or CB2 cannabinoid receptor inhibitors alter microglial activation and decrease the therapeutic efficacy of MINO after experimental TBI [24]. MINO represses the M1 microglia through inhibition of the MAPK pathway and NF-κB as well as inducing M2 microglia via upregulation of the TrkB/BDNF [28]. A caveat of these studies is the use of the oversimplified division of activated microglia into M1 (pro-inflammatory) and M2 (anti-inflammatory) types. While this M1/M2 scheme has provided some insight, newer studies employing single-cell transcriptomics will provide a more detailed understanding of MINO action on microglia [29, 30]. MINO also prevents astrocyte activation [21, 23, 31].

Fig. 1.

Pleotropic drug actions of minocycline. Minocycline has potent anti-inflammatory action by inhibiting proinflammatory cytokines. This results in modulating microglial polarization and astrocyte activation. Minocycline prevents apoptosis by regulating BCL-2 family members and inhibiting caspase. Inhibition of PARP-1 and iron chelation potentially prevents programmed necrotic cell death. Minocycline also prevents lipid peroxidation and metalloprotease-9 activation

MINO may be a potential inhibitor of ferroptosis for it chelates Fe⁺² and Fe⁺³ in vitro and lowers iron levels in CSF 7–9 days after experimental TBI [14, 20, 32]. MINO also lowers iron levels in stroke models [33]. MINO inhibits apoptosis by inhibiting caspase activation as well as increasing the expression ratio of the anti-apoptotic Bcl-2 to the pro-apoptotic Bax [18, 26]. MINO also inhibits metalloprotease 9 [34]. MINO inhibits poly(ADP-ribose) polymerase-1 in vitro [35]. PARP-1 activation depletes NAD⁺, resulting in energy failure and parthanosis [8]. MINO also prevents lipid peroxidation in models of TBI and spinal cord injury suggesting that it may prevent ferroptosis [8, 34, 36, 37]. MINO also prevents calpain activation in models of spinal cord injury [38].

Preclinical Testing of Minocycline in TBI Animal Models

Despite this broad consensus about its multiple modes of neuroprotection, early dosing of MINO produced highly disparate results on behavioral testing in multiple TBI models (Table 1). While early MINO dosing improved motor outcomes in multiple studies [18, 39], other studies report no effect [23, 40, 41]. Early MINO dosing did not have a uniform effect on neurological severity score [19, 24, 39, 41, 42] or cognitive tasks such as Morris water maze or Barnes Maze [14, 23, 25, 40, 43, 44]. MINO enabled acquisition and retention of active place avoidance, a cognitive task with higher cognitive demand than either Morris water maze or Barnes Maze [45]. MINO, however, had no effect when a demand on long-term memory was increased by expanding the intertrial interval from 1 h to 1 day [22]. This suggests that MINO did not improve long-term memory impairments. MINO did not consistently improve recognition of novel objects [39, 46]. MINO improved odor recognition and inhibited olfactory bulb injury [46]. MINO was unable to alter anxiety deficits on tests of affect, yet did reduce anxiety in a mild blast TBI model [20, 31, 46].

MINO did not consistently prevent gray matter injury proximal to a site of traumatic injury [18, 19, 25, 32, 40, 47]. In contrast, multiple studies report neuroprotection in brain regions distal to the injury site [14, 18, 25, 43]. MINO effectively prevents white matter injury [22, 45]. MINO induced oligodendrocyte precursor cells to repopulate and remyelinate injured corpus callosum [26]. Depending upon the study, prevention of gray and white matter injury by MINO either did [14, 18, 20, 22, 23, 43–48] or did not [24, 25, 31, 39, 40, 49] improve brain function.

Drugs to treat TBI need a favorable therapeutic time window [9]. The extent of a clinically useful time window is not well-defined; phase 3 clinical trials for TBI (PROTECT III) and (SYNAPSE) had an average lag between injury and treatment of 3.6 to 7 h [9]. Most preclinical studies included in this review initiate MINO dosing at early times after treatment. These studies provide information about potential drug mechanisms targeting early events after TBI and provide an initial proof-of-principal that these drugs may be effective. Early dosing of a large number of drugs are effective against experimental TBI, yet far fewer have shown efficacy with clinically relevant delayed dosing [9].

Most of the studies examining therapeutic effects, initiate MINO dosing at less than 1-h post-injury. Such early dosing likely has little clinical relevance. Fewer studies examined the efficacy of MINO dosed at later time points. Kovesdi et al. reported in a rat blast model that a first MINO dose at 4-h post-injury did not restore performance on acquisition in Barnes maze testing [31]. A first dose of MINO at 6-, 12-, or 24-h post-injury restored Barnes maze in two rodent TBI models [44]. Beginning MINO at 12-h post-injury also improved acquisition of Morris water maze and active place avoidance in a rat CCI model [14, 44]. These data strongly suggest that MINO improved acquisition of hippocampal-dependent tasks with a clinically relevant time window. A first dose of MINO at 12-h post-injury improved beam walk and cylinder tasks yet had no effect on grip strength in wire hang [14]. MINO dosing beginning at 6 h or more prevented loss of hippocampal dendrites while protecting oligodendrocytes and increasing myelin content [44, 50]. These data show that MINO retains efficacy when first dosed at 4 h or more after injury. Dosing MINO beginning at 24-h post-injury had no effect on myelin loss or deficits in active place avoidance [44]. Nonetheless, MINO may retain substantial efficacy when dosed many hours after injury.

A caveat of most animal studies is that MINO is administered via intraperitoneal injection. Due to its high hydrophobicity, intraperitoneally injected MINO yields variable plasma levels [51]. More uniform dosing is achieved with intravenous injection. Intravenous injection of transferrin-conjugated MINO yielded better brain penetration as well as improved outcomes than unconjugated MINO [48]. MINO has good oral bioavailability, yet no preclinical study has evaluated the efficacy of orally administered MINO [10]. The efficacy of MINO in females is largely unknown [52]. A second caveat of preclinical studies is that they are all performed on rodents; the efficacy of MINO may differ in large animals with gyrencephalic brains.

Clinical Tests of Minocycline for TBI

MINO has shown mild efficacy in early clinical trials for TBI (Table 2). Beginning MINO treatment within 1-day post-TBI significantly lowered the serum neuronal specific enolase and trended to decreasing serum S100B [53]. Glasgow Outcome Score-Extended showed no improvement. A separate, dose escalation study demonstrated that MINO could be safely dosed. Higher doses of MINO could be safely administered that produced a trend toward improved disability rating scores [54]. Administration of MINO 6 months after moderate to severe TBI reduced binding of the PET ligand, ¹¹C-PBR28, that assesses microglial activation [55]. The reduction of ¹¹C-PBR28 binding was particularly marked in white matter regions with DT-MRI abnormalities. These data suggest MINO suppressed chronic microglial activation in areas of ongoing white matter injury. Despite this reduction of microglial activation, MINO increased plasma levels of neurofilament light chain, a marker of axonal injury [56]. A similar paradoxical effect was observed in a trial of MINO in the condition “clinically isolated syndrome,” which is a first clinical episode that leads to multiple sclerosis in the majority of cases [57]. Despite increasing plasma neurofilament light chain MINO had had a therapeutic effect on both MRI outcomes in clinically isolated syndrome and early multiple sclerosis. This suggests that plasma neurofilament light chain levels do not capture the therapeutic efficacy of MINO. These limited studies provide evidence for further clinical studies to study MINO.

Daily dosing of MINO between 200 and 400 mg yielded plasma levels ranging from 3.5 to 7.0 mg/L [58, 59]; a single 90 mg/kg intraperitoneal dose of MINO resulted in a plasma level of 20 mg/L in rats which is higher than the drug levels in the clinic [60]. In contrast to most preclinical studies, only Lam et al. and Sangobowale et al. have used MINO doses that likely produce drug levels comparable to those used in clinical trials [44, 50].

A recent phase 2 trial (multi-arm multi-stage adaptive platform trial (APT) for the acute treatment of traumatic brain injury (APT-TBI-01, ClinicalTrials.gov ID NCT05826912) has been designed to test the efficacy of MINO, atorvastatin and candesartan to treat mild and moderate TBI (post-resuscitation Glasgow Coma Scale between 9 and 15). The MINO treatment group receives two loading 200 mg doses of oral MINO on day 1 followed by 6 days of 100 mg twice daily. A second cohort receives 7 days of placebo. APT-TBI-01 does not specify a cutoff based upon time after injury enrollment, which is a potential confounder since preclinical studies suggests a loss of efficacy when a first MINO dose is later than 12-h post-injury. The primary outcome of APT-TBI-01 is the Glasgow Outcome Scale-Extended (GOSE-TBI) that is assessed between 2 weeks and 3 months post-injury. Plasma levels of glial fibrillary acidic protein (GFAP) and neurofilament light chain in plasma are measured 2 weeks post-injury. Assessment of neurofilament light chain levels is potentially problematic since two earlier studies report that MINO treatment either produces no change or increased neurofilament light chain [56, 57]. The positron emission tomography radiotracer ¹¹C-PBR28 is a potential alternative measure to test efficacy. [¹¹C]-PBR28 binds to the 18 kD translocator protein (TSPO) that increases expression in activated microglia, reactive astrocytes, and vascular endothelium. Multiple preclinical studies show that MINO suppresses microglial activation (Table 1). MINO lowered [¹¹C]-PBR28 binding in patients with amyotrophic sclerosis [61].

Combinations of Drugs with MINO

Multiple studies have examined drug combinations containing MINO (Table 3). MINO was combined with the anti-oxidant melatonin in a rat CCI model of TBI, yet all outcome measures showed no improvement with MINO, melatonin alone nor the combination [62]. MINO was also combined with simvastatin, which has anti-inflammatory activity along with its ability to lower cholesterol levels [40]. Neither the individual drugs nor the combination improved motor or cognitive function. When dosed individually, both drugs altered multiple brain transcripts, but an effect of the combination on transcription was not tested [40]. Lam et al. examined the efficacy of combining MINO with botulinum toxin–induced constraint physical therapy [23]. MINO alone improved performance on Morris water maze yet provided no improvement on two motor tasks or lesion volume. Combining MINO with botulinum toxin yielded similar improvements in Morris water maze and motor function, but lesion volume was unaffected.

MINO has also been combined with the anti-oxidant N-acetyl cysteine (NAC) [45]. In a rat CCI model, MINO restored acquisition, but not retention of an active place avoidance task [45]. NAC alone did not improve task acquisition. Combining both drugs restored both acquisition and retention of active place avoidance suggesting synergistic improvement of memory [45]. Increasing the intertrial interval from 10 min to 1 day increases the demand for long-term memory to acquire active place avoidance. Treatment with MINO plus NAC improved the 1-day interval version of active place avoidance while the individual drugs had no effect [22]. Lowering the dose of both MINO (22.5 mg/kg) and NAC (75 mg/kg) resulted in no loss of efficacy [44]. The combination of MINO (22.5 mg/kg) plus NAC (75 mg/kg) had a therapeutic time window of at least 12 h [44]. CHI lowers hippocampal levels of MAP2. MAP2 levels increase in injured mice when the combination was dosed beginning at six or 12 hours after injury. This suggests synergy since the individual drugs had no effect on MAP2 levels [44]. Myelin levels in corpus callosum were restored when MINO or MINO plus NAC was initiated 12 hours post-injury; at 24 hours post-injury, only the combination was effective. These data suggest drug indifference when first dosed at 12 hours, but drug synergy when the time to first dose was increased to 24 hours. Loss of oligodendrocytes in both the rat CCI and mouse CHI models was more effectively prevented by MINO plus NAC than the individual drugs [26, 50]. Analysis of microglial activation also revealed synergistic action. In the rat CCI model, dosing of MINO one hour post injury lowered microglial expression of Iba-1, Arg^HI, and Fzz-1 at 2 days post-injury, suggesting suppression of microglial activation. One hour NAC dosing had no effect. Dosing of MINO plus NAC one hour post-injury increased microglial expression of Iba-1, Arg^HI, and Fzz-1 at 2 days post-injury that lowered significantly at 4 days. The differential effect of the combination to the individual drugs suggests a synergistic modulation of microglial activation [26]. MINO plus NAC, but not the individual drugs, induced transcription of CD40 and CD86 that promote myelin repair [26]. Thus, the synergistic action of MINO plus NAC has been demonstrated in two TBI models with a variety of therapeutic outcomes.

The MINO plus NAC combination retained therapeutic efficacy when first dosed 3 days post-injury. In the CHI model, injured mice could acquire and retain Barnes maze when MINO plus NAC was first dosed at 3 days post injury. In hippocampal CA1, a first dose of MINO plus NAC 3 days post-injury blocked bilateral neuronal and synaptic loss and maintained neuronal morphology. A similar treatment, however, was only effective on contralesional CA3 neurons [63].

Conclusions

At present, no drugs are available to treat TBI; it remains unknown how much preclinical efficacy a drug needs to be translated to the clinic. MINO has multiple modes of action which have resulted in multiple demonstrations of therapeutic efficacy in preclinical testing (Table 1). MINO has shown promise including some evidence of efficacy in small clinical trials for TBI (Table 2). MINO may become more potent with improved drug delivery or combining it with additional drugs (Table 3). More studies are needed on dosing MINO with a clinically useful therapeutic time window and in gyrencephalic TBI animal models to better design clinical trials for TBI.

Funding

This work was funded by an award (NS108190) to P.J.B.

Declarations

Conflict of Interest

The authors declare no competing interests.