Molecular changes evoked by the beta-lactam antibiotic ceftriaxone across rodent models of substance use disorder and neurological disease

Abstract

Ceftriaxone is a beta-lactam antibiotic that increases the expression of the major glutamate transporter, GLT-1. As such, ceftriaxone ameliorates symptoms across multiple rodent models of neurological diseases and substance use disorders. However, the mechanism behind GLT-1 upregulation is unknown. The present review synthesizes this literature in order to identify commonalities in molecular changes. We find that ceftriaxone (200 mg/kg for at least two days) consistently restores GLT-1 expression in multiple rodent models of neurological disease, especially when GLT-1 is decreased in the disease model. The same dose given to healthy/drug-naive rodents does not reliably upregulate GLT-1 in any brain region except the hippocampus. Increased GLT-1 expression does not consistently arise from transcriptional regulation, and is likely to be due to trafficking changes. In addition to altered transporter expression, ceftriaxone ameliorates neuropathologies (e.g. tau, amyloid beta, cell death) and aberrant protein expression associated with a number of neurological disease models. Taken together, these results indicate that ceftriaxone remains a strong candidate for treatment of multiple disorders in the clinic.

1. Introduction

Ceftriaxone is a third-generation cephalosporin antibiotic that has also been found to modulate the glutamate neurotransmitter system in the central nervous system. Cephalosporin antibiotics are derived from the fungus Acremonium and are members of the family of beta-lactam antibiotics, which share the common feature of a beta-lactam ring. Beta-lactam antibiotics are used to treat a number of gram-positive and gram-negative bacteria (Reiner et al., 1980). Ceftriaxone was developed by Hoffmann-La Roche and the first publication on this new antibiotic in 1980 reports that it possesses greater in vivo activity and a longer half-life than other cephalosporins (Reiner et al., 1980). The plasma half-life of ceftriaxone in humans is approximately 6–8 h, which is 2–10 fold higher than that of other cephalosporins and allows for once/daily dosing (Patel et al., 1981; Seddon et al., 1980). In the rat, the plasma half-life is ~35 min (Reiner et al., 1980). Ceftriaxone can pass through the blood-brain-barrier, and as such is prescribed to treat meningitis (Beskid et al., 1981).

In 2005, 1040 Food and Drug Administration (FDA)-approved medications were screened for their ability to increase central nervous system (CNS) expression of the major glial glutamate transporter, GLT-1 (EAAT2 in humans), finding that 15 members of the beta-lactam family of antibiotics upregulate GLT-1 in vitro (Rothstein et al., 2005). Ceftriaxone was identified as the most potent GLT-1 up-regulator and also ameliorated GLT-1 deficits in vivo, in the superoxide dismutase 1 (SOD1) mouse model of Amyotrophic Lateral Sclerosis (ALS). This finding was followed by a number of studies (over 100 to date) examining the ability of ceftriaxone to alter behavior and GLT-1 expression throughout the brain in rodent models of neurological diseases and disorders of motivation. While the ability of ceftriaxone to restore GLT-1 expression is consistently detected across multiple rodent models of neurological disease, inconsistent findings have been reported in healthy or intact rodents and cell culture.

The goal of the present review is to bring together the large body of literature that investigates the ability of ceftriaxone to ameliorate symptoms and neuropathologies associated with a number of neurological diseases and substance use disorder. Because the mechanism of action behind the ability of ceftriaxone to alter neurobiology has not been reliably demonstrated across models, we hypothesized that a review of the literature would yield commonalities in the dose and brain region dependent effects of ceftriaxone on molecular changes across disorders. Thus, we consider dose and length of ceftriaxone treatment, species, brain region, and neurological disease model to uncover the molecular mechanism of action of ceftriaxone. We restrict our discussion to papers in which molecular consequences of ceftriaxone treatment are investigated; many more papers with solely behavioral endpoints are not discussed here.

1.1. Glutamate transporters

The excitatory amino acid transporters (EAATs) are a family of proteins that remove glutamate from the synapse via coupling the transport of Na⁺ and K⁺ down their concentration gradients, which drives glutamate into the cell (Anderson and Swanson, 2000). There are five EAATs in the mammalian nervous system, named EAAT1–5. EAAT1 (rodent GLAST) and EAAT2 (rodent GLT-1) are located mainly on astrocytes (Rothstein et al., 1994). EAAT3 (EAAC1) shows only moderate brain expression in comparison with GLT-1 and GLAST (Haugeto et al., 1996). EAAT4 and EAAT5 have little expression outside of the cerebellum and retina, respectively (Danbolt, 2001). Quantitative measurements of transporter densities (Lehre and Danbolt, 1998) and glutamate uptake (Rothstein et al., 1996) indicate that GLAST and GLT-1 are the most abundant glutamate transporters in brain tissue. Furthermore, GLT-1 is responsible for 90% of the total CNS glutamate uptake (Haugeto et al., 1996; Tanaka et al., 1997). GLT-1 expression is highest on portions of the glial membrane which face the neuropil (Cholet et al., 2002), allowing GLT-1 to prevent almost all glutamate spillover from the synapse into the extrasynaptic compartment (Danbolt, 2001; Diamond and Jahr, 2000).

A second type of sodium-independent glutamate transport system exchanges extracellular cystine for intracellular glutamate. The system responsible is referred to as the cystine-glutamate exchanger or system x_C-. System x_C- is composed of a 4FH2 heavy chain that is characteristic of many amino acid transporters and the specific catalytic subunit xCT (Sato et al., 1999). System x_C- is an important contributor to the maintenance of basal extracellular glutamate levels in several brain regions and inhibitors of this system decrease these levels by 60% in the nucleus accumbens (NA) (Baker et al., 2002) and by approximately 25% in the hippocampus (De Bundel et al., 2011), with no effect in the prefrontal cortex (PFC) (Melendez et al., 2005). Genetic knockdown of xCT by ~50% in the NA results in a corresponding decrease in basal glutamate levels in that brain region (LaCrosse et al., 2017).

Co-regulation of sodium-dependent glutamate transport and glutamate export via system xc- has been observed (Bannai, 1986), and thus many studies have investigated the ability of ceftriaxone to affect both sodium-dependent and -independent transport and corresponding GLT-1 and xCT expression.

1.2. Ceftriaxone pharmacokinetics in humans and rodents

When ceftriaxone is used as an anti-microbial agent in humans the recommended daily dose is 2 g/day for up to 4–6 weeks (Baddour et al., 2005). Ceftriaxone must be administered intravenously (IV) or intraperitoneally (IP) as there is no orally bioavailable formulation of ceftriaxone. The US Food and Drug Administration endorses the use of Body Surface Area (m²) to convert doses from different species to the human equivalent dose, with a standard formula (see Reagan-Shaw et al., 2008). Using this formula, the daily dose of ceftriaxone administered to the rat would be 200 mg/kg, which is the most common dose used in the publications reviewed here. The rat equivalent dose is 50% of that of the mouse. Thus, 200 mg/kg ceftriaxone administered to rat is equivalent to 400 mg/kg administered to the mouse. Long-term (weeks to months) administration of ceftriaxone at these doses has been found to be safe and well-tolerated in humans and rodents (Baddour et al., 2005; Ratti et al., 2015). However, it should be noted that long-term use of antibiotics contributes to the development of antibiotic resistant bacteria and alters the gut microbiome (Holota et al., 2019; Pletz et al., 2004). Understanding the molecular features that make ceftriaxone treatment successful at ameliorating symptoms of neurological diseases can help the design of ceftriaxone-like drugs lacking antimicrobial properties to avoid such side effects.

2. Ceftriaxone’s effects on GLT-1 expression and function in cell culture

Rothstein et al., (2005) was the first to identify ceftriaxone as having the ability to upregulate GLT-1. Upon screening 1040 FDA-approved drugs and nutriceuticals in an in vitro model, it was determined that 15 members of the beta-lactam family increase the expression of GLT-1. Organotypic spinal cord slice cultures were prepared from postnatal day 9 rats and after 5–7 days of each drug treatment (100 μM, added biweekly), tissue was harvested and western blots for GLT-1 were conducted. For ceftriaxone, 3.5 μM was determined to be the EC₅₀, and upregulation of GLT-1 was observed after only 48 hours of treatment in spinal cord cultures as well as in stable cell lines of primary human fetal astrocytes (PHFA). The human EAAT2 promoter fragment was activated by ceftriaxone and amoxicillin, but not by the antibiotic vancomycin or by glutamate itself. These effects were dose dependent, seen as early as 48 hours after drug administration, and persisted for at least 7 days in vitro (Rothstein et al., 2005). Supporting these results, in PHFA treated with 10 μM ceftriaxone for 2 days, ceftriaxone increases EAAT2 (but not EAAT1) mRNA and protein levels, leading to increased glutamate uptake (Lee et al., 2008). Ceftriaxone was found to increase EAAT2 transcription through the nuclear factor-kappaB (NF-κB) signaling pathway (Lee et al., 2008).

HT22 cells and astrocytes co-cultured with neurons were used to examine the ability of ceftriaxone to upregulate both EAATs and system xc- (Lewerenz et al., 2009). After 7 days of 300 μM ceftriaxone, a 37% increase in sodium-dependent glutamate uptake (EAAT activity) and 40% increase in sodium-independent glutamate uptake (system xc- activity) was observed in HT22 cells. Interestingly, mRNA expression of EAAT1, 2 and 3 were increased by ceftriaxone. In astrocytes co-cultured with neurons, 7 days of treatment with both 30 and 300 μM ceftriaxone increase EAAT activity by 20%. While EAAT1 mRNA and protein expression are not increased in PHFA treated with 10 μM ceftriaxone for 2 days (Lee et al., 2008), it may be that longer treatment with >10 μM ceftriaxone is needed to see this effect or that HT22 cells are uniquely sensitive to the ability of ceftriaxone to increase expression of multiple glutamate transporter subtypes.

In cortical astrocyte cultures from C57Bl mice, 24 hours of ceftriaxone (10 μM) increases GLT-1 and GLAST mRNA and protein expression (Benkler et al., 2013). However, at least one study failed to find an effect of ceftriaxone on glutamate uptake in vitro. Ceftriaxone (100 μM) was applied for 48 h to primary cortical astrocytes obtained from Sprague-Dawley rat pups. While total protein expression of GLT-1 was increased by ceftriaxone, membrane bound GLT-1 was unchanged, likely explaining the finding that glutamate uptake was also not altered by ceftriaxone (Zhang et al., 2015). Interestingly, GLT-1 mRNA was also not increased by ceftriaxone, and thus the mechanism behind increased total protein GLT-1 expression is unclear. When the cytotoxin 1-methyl-4-phenylpyridinium (MPP⁺) was added to cultures, this resulted in decreased membrane-bound GLT-1 and glutamate transport. Ceftriaxone (100 but not 1 or 10 μM) increased glutamate transport in MPP⁺ treated cultures and expression of membrane-bound GLT-1 but not GLT-1 mRNA or total protein expression. This effect of ceftriaxone was found to be mediated by its ability to suppress NF-κB/JNK/c-Jun signaling following MPP⁺ treatment (Zhang et al., 2015). These results in vitro are consistent with other findings reviewed below regarding the inability of ceftriaxone to alter GLT-1 expression/mRNA in ex vivo tissue from “healthy/control” rodents treated with ceftriaxone only once/day (e.g. Knackstedt et al., 2010a).

3. Ceftriaxone’s effects on GLT-1 expression and function in healthy rodents

Upon finding an effect of ceftriaxone on GLT-1 expression in a cell culture model. Rothstein et al. (2005) confirmed upregulation in vivo. In normal rats (strain not specified) injected with ceftriaxone (200 mg/kg for 5–7 days or 3 months), protein expression of GLT-1, but not EAAT1, 3 or 4, increased in the hippocampus and spinal cord (Rothstein et al., 2005). In mice (strain not specified), glutamate uptake was found to be increased in ex vivo cortical tissue after 7 days of ceftriaxone treatment (Rothstein et al., 2005).

In male Balb-c mice, ceftriaxone (200 mg/kg/day for 9 days) increases GLT-1 immunoreactivity in the CA1, CA3, and the dentate gyrus (DG) of the hippocampus (Karaman et al., 2013). In the DG, increased GLT-1 expression is also found after 50 mg/kg of ceftriaxone. GLT-1 is not increased after a single ceftriaxone (200 mg/kg) injection, nor does any dose of ceftriaxone affect learning in the Morris water maze (Karaman et al., 2013). In male Wistar rats, 200 mg/kg ceftriaxone (8 or 14 days) increases GLT-1 immunoreactivity in both hippocampal total protein and specifically in glial fibrillary acid protein (GFAP+) cells (Hsu et al., 2015; Omrani et al., 2009). This upregulation of GLT-1 is accompanied by impaired mGluR-dependent long-term depression (LTD) in mossy fibre-CA3 synapses. Glutamate uptake is upregulated in ex vivo CA1 hippocampal tissue from male Wistar rats treated with ceftriaxone (200 mg/kg for 5 days) (Hu et al., 2015). Only one paper failed to find an effect of ceftriaxone (200 mg/kg for 5 days) on GLT-1 expression in the hippocampus; in this study rats were decapitated 4 days after the last ceftriaxone injection and not within 24 hours as in the other studies discussed above (Krzyzanowska et al., 2016).

In the dorsal striatum of male Sprague-Dawley rats, ceftriaxone (200 mg/kg for 8 days) increases GLT-1 expression in the membrane fraction (Chotibut et al., 2014), and increases glutamate uptake as assessed with no-net-flux microdialysis (Miller et al., 2008). In male Wistar rats administered ceftriaxone (200 mg/kg for 5 days), GLT-1 mRNA expression was found to be increased in the dorsal striatum and frontal cortex; however this did not correspond to changes in GLT-1 protein expression in the same animals (Krzyzanowska et al., 2016). The same paper reported that ceftriaxone upregulates xCT mRNA and protein expression in the dorsal striatum but not in the frontal cortex or hippocampus in healthy rats. In the ventral striatum (NA core), 200 mg/kg does not increase GLT-1 (or xCT) in a membrane fraction (Garcia et al., 2019; Knackstedt et al., 2010a), nor does it increase glutamate uptake or system xc- activity (Trantham-Davidson et al., 2012). At least one study has also failed to find an effect of ceftriaxone (200 mg/kg) on GLT-1 (and xCT) expression in the PFC of male Sprague-Dawley rats (Garcia et al., 2019). Taken together, while ceftriaxone consistently increases GLT-1 expression in the hippocampus of healthy rodents, the same is not observed in other brain regions. At this time, it is not clear why this may be, although the hippocampus has greater GLT-1 expression density compared to the majority of other brain regions (Danbolt, 2001), and this may make it more sensitive to the upregulation by ceftriaxone. Differential concentrations of ceftriaxone may also be attained across the brain; however, this has not been thoroughly assessed.

4. The effects of ceftriaxone on molecular and behavioral markers in neurological disease

4. 1. Amyotrophic lateral sclerosis (ALS)

After establishing that ceftriaxone upregulates GLT-1 function and expression in vitro and in healthy rodents, Rothstein et al. (2005) tested ceftriaxone in the G93A SOD1 mouse model of ALS. Mice treated with ceftriaxone (200 mg/kg) or vehicle starting at 12 weeks of age, when symptoms such as reduced grip strength emerge, resulted in preservation of muscle strength and body weight by ceftriaxone, but only for 4–6 weeks. Overall survival was shown to be increased by 10 days and GLT-1 expression was increased in the spinal cord (Rothstein et al., 2005). However, when ceftriaxone treatment (100 mg/kg) was initiated at 13 weeks of age, no increase in survival time was found (Kong et al., 2012). This could be due to both the lower dose of ceftriaxone used and the timing. GLT-1 expression was not assessed after the lower dose of ceftriaxone, and so it is not possible to draw conclusions regarding the necessity of such upregulation in survival in this model. Interestingly, in cortical cultures from newborn hemizygous SOD1G93A transgenic mice, ceftriaxone (10 μM) does not upregulate GLT-1 but increases GFAP and the growth factors brain derived neurotrophic factor (BDNF) and glial cell-derived neurotrophic factor (GDNF) (Benkler et al., 2013). Ceftriaxone may be ineffective in this model because it was only applied for 24 h. The slow neuronal death observed following axotomy in adult mice is considered a model for progressive neurodegenerative disorders. Male C67BL/6 mice received right hypoglossal nerve axotomy and ceftriaxone (200 mg/kg) treatment began the day of surgery and continued for 28 days, at which point the mice were sacrificed. In this paradigm, ceftriaxone increases cell survival in the hypoglossal nucleus and expression of GLT-1 in the astrocytes of this region (Yamada and Jinno, 2011).

A randomized, double-blind, placebo-controlled clinical trial for ceftriaxone as a treatment for ALS found no benefits of long-term ceftriaxone on patient survival (Cudkowicz et al., 2014). Patients were randomized to receive ceftriaxone (2 or 4 g/day) or placebo beginning on average 1.5 years after symptoms began and a half year after diagnosis. While symptoms progressed more slowly in the 4 g/day group during the first 20 weeks (Stage 2) of treatment, at the end of one year (Stage 3) there were no group differences in a cohort that included patients from Stage 2 as well as newly recruited patients. Ceftriaxone did not increase survival and had a modest effect on symptoms in patients. However, ALS is a heterogenous disease and less than 10% of all ALS patients have the familial SOD1 mutation that the rodent model is based upon. As the patients in this trial were not recruited based on possession of this mutation, it is possible that ceftriaxone may have had an effect in Stage 2 because more patients had a form of ALS that was responsive to ceftriaxone treatment.

4.2. Huntington’s Disease (HD)

Huntington’s disease (HD) is an autosomal dominant disorder that can be modeled with transgenic mice such as the R6/2 strain. These mice (Miller et al., 2008), as well as HD patients themselves (Hassel et al., 2008) display reduced glutamate uptake in striatum and PFC. In 8-week old R6/2 mice exhibiting behavioral deficits (e.g. twitching, reduced motor flexibility), 5 days of 200 mg/kg ceftriaxone attenuates these deficits and increases GLT-1 protein expression in the dorsal striatum (Miller et al., 2008). The same benefits were seen when ceftriaxone treatment was initiated at 12 weeks of age, potentially indicating that ceftriaxone therapy would be beneficial in older HD patients (Sari et al., 2010).

4.3. Parkinson’s Disease (PD)

The MPTP (1-methyl-4phenyl-1,2,3,6 tetrahydropyridine) rat model of PD causes symptoms by destroying dopaminergic neurons and inducing neurodegeneration. MPTP is typically administered intracranially, into the substantia nigra pars compacta (SNc). This induces impaired motor behavior and cognition as well as a number of pathological brain responses (Bisht et al., 2014; Ho et al., 2014; Huang et al., 2015). In this model, ceftriaxone (100 and 200 mg/kg for 14 days) improves cognitive deficits as assessed via T-maze (working memory) and novel object recognition. This protective effect is observed when ceftriaxone treatment is initiated both prior to and 3 days after MPTP treatment, with one exception: the 100 mg/kg dose does not improve recognition memory when initiated 3 days after MPTP (Ho et al., 2014; Hsu et al., 2015). Both doses of ceftriaxone also ameliorate motor deficits and decrease oxidative damage in the striatum and cortex (Bisht et al., 2014; Ho et al., 2014; Hsu et al., 2015). Both doses of ceftriaxone, whether initiated 5 days prior to or 3 days after MPTP, increase GLT-1 expression in GFAP+ cells (likely astrocytes) in both MPTP-lesioned groups in the dorsal striatum (Hsu et al., 2015). In the hippocampus, only 200 mg/kg ceftriaxone increases GLT-1 expression, but it does so in both MPTP-lesioned and sham-operated rats. When ceftriaxone administration is started one day after MPTP, ceftriaxone completely prevents the MPTP-induced decreased density of dopaminergic neurons in the SNc and ameliorates the MPTP-induced decrease in tyrosine hydroxylase (TH) immunoreactivity in the striatum (Ho et al., 2014). Ceftriaxone also decreases pro-inflammatory cytokines tumor necrosis factor-α (TNF-α) and interleukin- β (IL-β) (Bisht et al., 2014).

Because higher doses of ceftriaxone were shown to be effective at ameliorating brain and behavior deficits in the MPTP model, Huang et al. (2015) sought to determine the effects of a sub-threshold dose of ceftriaxone (5 mg/kg/day) in the MPTP model, finding that it was able to increase performance in the T-maze and novel object recognition test. MPTP lesioning induced degeneration in the nigrostriatal dopaminergic system, as shown through immunohistochemical staining for TH. Fourteen days of ceftriaxone treatment resulted in a higher density of dopaminergic neurons in the SNc and dorsal striatum compared to the vehicle-treated MPTP group (Huang et al., 2015). However, this study did not measure levels of GLT-1 expression and thus it is not clear if these changes are found in the presence of GLT-1 upregulation.

Another rodent model of PD utilizes intracerebral infusion of the neurotoxin 6-hydroxydopamine (6-OHDA) which destroys nigrostriatal dopaminergic neurons and leads to behavioral deficits, motor dysfunction, and neurodegeneration similar to that observed in PD. Both male and female Sprague-Dawley rats treated with striatal 6-OHDA lesion show decreased GLT-1 expression, striatal glutamate uptake and a significant decrease in striatal TH protein (Chotibut et al., 2014; Leung et al., 2012). Ceftriaxone (200 mg/kg) administered daily for 8 days beginning on the day of lesioning restores GLT-1 expression and increases striatal glutamate uptake by 30% compared to vehicle. When administered prior to lesioning in female rats, the same dose of ceftriaxone is also able to rescue GLT-1 expression in the dorsal striatum. Ceftriaxone attenuates 6-OHDA-induced striatal TH protein loss and reduces amphetamine-induced rotation behavior, an index of nigrostriatal neuron loss, both when treatment begins 7 days prior to and on the day of lesioning (Chotibut et al., 2014; Leung et al., 2012). Furthermore, ceftriaxone reduces TH phosphorylation, a calcium-dependent target specific for nigrostriatal neurons, compared to vehicle groups (Chotibut et al., 2014). Ceftriaxone has no effects on striatal GLAST expression (Chotibut et al., 2014).

4.4. Alzheimer’s Disease (AD)

The APPPS1 transgenic mouse model of AD exhibits decreased GLT-1 expression and chronically elevated glutamate levels in the area surrounding amyloid deposits. Treating APPPS1 mice with 200 mg/kg ceftriaxone for 5 consecutive days restores GLT-1 expression levels nearly to control levels and normalizes glutamate levels in the area surrounding amyloid deposits and amyloid plaque. This effect was observed in the hindlimb region of the somatosensory cortex and primary visual cortex (Hefendehl et al, 2016).

3xTg-AD mice are triple transgenic mice models of Alzheimer’s Disease which exhibit increased production of amyloid beta (Aβ), increased ratio of Aβ42/Aβ40, and tau pathology. Glutamate transport has been observed to be decreased in the brains of deceased AD patients (Masliah et al., 1996) and 3xTG-AD mice also exhibit decreased hippocampal GLT-1 expression in the detergent-soluble fraction that worsens with age; this is due to the toxic effect of Aβ species on astrocytes (Zumkehr et al., 2015). 3xTg-AD mice treated daily with 200 mg/kg ceftriaxone for 2 months show increased hippocampal GLT-1 expression and decreased tau pathology that is accompanied by improvements in spatial memory retention and recognition memory in the novel object recognition task (Zumkehr et al., 2015). Ceftriaxone also reduces insoluble tau proteins in the cortex. However, ceftriaxone does not affect the expression of amyloid precursor protein (APP) or proteins that regulate its processing (C99 and C83).

OXYS rats are a transgenic rat model of sporadic AD characterized by early neurodegeneration, cognitive decline, Aβ deposits, and increased tau phosphorylation. In these rats, ceftriaxone affects the expression of genes involved in the pathogenesis of AD. In OXYS rats, 100 mg/kg/day of ceftriaxone for 36 days reverses cognitive and neuronal deficits (Tikhonova et al., 2018). Ceftriaxone also decreases Bace1 (involved in Aβ production) and Ace2 (enzyme involved in Aβ degradation) mRNA levels in the hypothalamus of OXYS rats. It also decreases levels of Aktb (β-actin gene) mRNA levels in the frontal cortex. Moreover, ceftriaxone increases mRNA expression of several enzymes involved in Aβ degradation: Ece1 in the striatum and Ide and Mme in the amygdala. Aktb expression was also increased by ceftriaxone in the striatum of OXYS rats. Additionally, ceftriaxone treatment increased Epo (gene for a hormone involved in clearance of Aβ in the blood brain barrier) mRNA levels in the amygdala (Tikhonova et al, 2018). However, this study did not measure levels of GLT-1 expression.

Taken together, the few rodent studies conducted on the topic indicate that ceftriaxone ameliorates a number of brain pathologies and cognitive deficits observed in clinical Alzheimer’s cases. Ceftriaxone increases GLT-1 expression and function in several brain regions, particularly in regions surrounding amyloid plaques. The direct relationship between GLT-1 and amyloid has yet to be determined.

4.5. Ischemia

Ceftriaxone has been demonstrated to be protective in different rodent models of brain ischemia. In a rat model of short-term global cerebral ischemia and reperfusion, ceftriaxone (100 mg/kg) administered once 2 h before ischemia improves the morphological structure of microvessels and neurons (Altaş et al., 2013). The authors propose that this improvement may be due to the ceftriaxone antioxidant activity in the brain and its influence on the oxidative stress parameters [a reduction of malondialdehyde (MDA) levels and an increase of superoxide dismutase (SOD) and glutathione peroxidase (GPx) activity] (Altaş et al., 2013). Anti-ischemic properties of ceftriaxone are supported by another study in a rat model of global brain ischemia showing that ceftriaxone pre-treatment (50, 100 and 200 mg/kg) alleviates delayed neuronal death of hippocampal pyramidal neurons in the CA1 region in a dose-dependent manner (Hu et al., 2015). This is accompanied by increased GLT-1 expression, while administration of dihydrokainate (a selective inhibitor of GLT-1) diminishes this effect (Hu et al., 2015). Additionally, antisense mediated-knockdown of GLT-1 also prevents ceftriaxone from delayed neuronal death in this model (Hu et al., 2015). These data support the participation of upregulation of GLT-1 in the neuroprotective effects of ceftriaxone.

Ceftriaxone pre-treatment (200 mg/kg/day for 5 days) also reduces delayed CA1 neuronal death and increases GLT-1 expression in the hippocampus in a rat model of forebrain ischemia (oxygen glucose deprivation, OGD) after 7 days of reperfusion (Ouyang et al., 2007). In hippocampal astrocyte cultures subjected to OGD followed by 24 h recovery, ceftriaxone (100 μM for 5 days) reduced CA1 neuronal loss and up-regulated GLT-1 expression, but had no effect on control cultures not exposed to OGD (Ouyang et al., 2007). However, when acute hippocampal slices were obtained from rats treated with ceftriaxone for 5 days (200 mg/kg), then subjected to OGD, ceftriaxone increased the activity, but not expression, of glutamate transporters in the CA1 hippocampal region (Lipski et al., 2007). This difference may be due to the different methods (in vivo vs. in vitro application of ceftriaxone and western blotting vs. immunohistochemistry). Higher levels of GLT-1 mRNA and protein, but not GLAST, were found after ceftriaxone preconditioning (200 mg/kg/day; 5 days) in 14 postnatal day rats subjected to hypoxic ischemia (Mimura et al., 2011). Additionally, ceftriaxone pre-treatment reduced infarct volume and ameliorated the markers of apoptotic processes at this time, suggesting a potential role in the prevention of neonatal encephalopathy (Mimura et al., 2011). These data parallel observations in the hippocampus of rats at 14 postnatal days after perinatal hypoxic-ischemic encephalopathy, when apoptotic cells were decreased after administration of ceftriaxone (200 mg/kg) for 48 h prior to experimental ischemia (Lai et al., 2011). Ceftriaxone pre-treatment diminished the brain injury scores and restored myelination in the external capsule of rats at 14 postnatal day. At 7 postnatal days, the ceftriaxone treated group showed an increase in GLT-1 protein expression in the cerebral cortex without changes in the expression of this protein in the corpus callosum, hippocampus and striatum and at 23–24 days, attenuated learning and memory deficits (Lai et al., 2011). Neonatal rat pups pretreated with ceftriaxone (200 mg/kg/day; 3 days) 1 h before dexamethasone application on postnatal days 1–3 showed attenuated hypoxic-ischemic brain injury and increased GLT-1 protein expression in the frontal cortex, hippocampus, and striatum, without changes in brain GLAST levels (Chang et al., 2013).

In a rat model of focal cerebral ischemia and reperfusion induced by middle cerebral artery occlusion (MCAO), increased GLT-1 expression was observed in the cortex (Krzyzanowska et al., 2016; Verma et al., 2010), dorsal striatum and hippocampus (Krzyzanowska et al., 2016) in rats that were pre-treated with ceftriaxone (200 mg/kg/day) for 5 days prior to MCAO. In the same rats, ceftriaxone treatment increased both GLT-1 and xCT mRNA levels in the frontal cortex (Krzyzanowska et al., 2016), and glutamine synthetase activity in the cortex (Verma et al., 2010), while a reduction in xCT mRNA levels in the hippocampus and dorsal striatum was shown in ischemic animals treated with ceftriaxone (Krzyzanowska et al., 2016). In this model, ceftriaxone pre-treatment improves the neurological deficit scores and decreases infarct volumes 24 h after reperfusion (Chu et al., 2007; Krzyzanowska et al., 2016; Thöne-Reineke et al., 2008; Verma et al., 2010), decreases early mortality and neurological deficits (Krzyzanowska et al., 2016; Thöne-Reineke et al., 2008; Verma et al., 2010). Ceftriaxone (200 mg/kg/day) administered for 5 consecutive days with the last dose administered 3 days prior to the MCAO reduced glutamate levels in the frontal cortex and hippocampus and normalized the reduced levels of GLT-1 expression in astrocytes in both the frontal cortex and hippocampus. In the same brain regions, xCT expression on astrocytes and neurons was reduced (Krzyżanowska et al., 2017).

These data support the idea that upregulation of GLT-1 is a therapeutic target for neuroprotection following cerebral ischemic injury. However, it is unlikely that in the clinic, patients would be pre-treated with ceftriaxone prior to ischemia. In a more translational rodent model of ischemia, when a single injection of ceftriaxone (200 mg/kg) was administered 90 min after MCAO, increased neuronal survival and GLT-1 activity was observed, but GLT-1 expression was not changed (Thöne-Reineke et al., 2008). However, other studies have found that post-injury treatment with ceftriaxone did not reduce the infarct volume (200 mg/kg/day, 30 min after the ischemic insult for 3 days; Chu et al., 2007), did not ameliorate the sensory motor dysfunction induced by ischemia (100 mg/kg once 2 h after reperfusion; Verma et al., 2010), and even exacerbated initial deficits including skill learning rate in reaching performance (200 mg/kg/day, starting 3 days after lesion for 5 days; Kim and Jones, 2013). One clinical case in which prophylactic pre-treatment with ceftriaxone would potentially prevent ischemia is in the case of cerebral vein occlusion during a brain surgery. Inui et al. (2013) pretreated rats with 200 mg/kg ceftriaxone for 5 days prior to cortical vein occlusion, finding that ceftriaxone reduces infarct volume and blocking GLT-1 activity with DHK prevents this benefit.

On the whole, the evidence from preclinical models indicates that ceftriaxone may provide a neuroprotective effect against brain ischemia and during the acute phase of ischemia. In cases when there is a higher risk of occlusion, such as during a brain surgery, a prophylactic regimen of ceftriaxone may be warranted. However, it seems that a potential clinical application of this drug in the post-acute period is limited.

4.6. Epilepsy & seizure

Several studies find that GLT-1 expression and function is decreased in tissue from patients with temporal lobe epilepsy (Proper et al., 2002). The same is observed in some preclinical models of seizure disorder (Wong et al., 2003), while GLT-1 knockout mice show lethal spontaneous seizures and mortality (Tanaka et al., 1997). One week of ceftriaxone (200 mg/kg/day) after traumatic brain injury (TBI) in rats restores reduced GLT-1 expression, decreases the level of regional GFAP expression in the lesioned cortex and reduces post-traumatic seizure frequency and duration (Goodrich et al., 2013). A newer study showed that in a mouse model of viral-induced epilepsy using the Theiler’s Murine Encephalomyelitis Virus (TMEV), acute increases in cortical GLT-1 and xCT expression are observed during seizures, while cortical GLT-1 levels are reduced after seizures (Loewen et al., 2019). Ceftriaxone (400 mg/kg) had no effect on total number of seizures or cumulative seizure burden, and did not reduce neuronal injury or gliosis in the hippocampus and cortex in TMEV-infected animals despite its ability to increase hippocampal GLT-1 protein expression level (Loewen et al., 2019). A possible explanation for these results may be the lack of adequate time of ceftriaxone administration, as ceftriaxone was administered for the first time immediately after the first seizures were induced. Supporting this idea, in a mouse model of Tuberous Sclerosis Complex (Tsc1^GFAPCKO mice), early treatment with ceftriaxone (200 mg/kg/day, 7 days) prior to the onset of epilepsy increases GLT-1 protein expression, thereby reducing glutamate levels, neuronal death and seizure frequency (Zeng et al., 2010). Survival was also improved in this condition. However, late application of ceftriaxone (after onset of epilepsy) increases GLT-1 expression levels in the hippocampus, but fails to decrease seizures (Zeng et al., 2010). Ceftriaxone (200 mg/kg/day, 7 days) pretreatment prevents the increase in vulnerability to kainic acid-induced seizures and increase in extracellular glutamate levels in the hippocampus after 3,4-methylenedioxy-methamphetamine (MDMA) administration (Huff et al., 2016). Interestingly, ceftriaxone abolishes the MDMA-induced loss of glutamic acid decarboxylase (GAD) 67-IR neurons in regioins of the hippocampus (Huff et al., 2016). These data support a neuroprotective effect of ceftriaxone on glutamate-dependent excitotoxicity and a reduction in seizure susceptibility when ceftriaxone is administered prior to the insult.

Oxidative stress is another possible mechanism involved in the progression of epileptic-seizures and several studies showed the beneficial effects of ceftriaxone on oxidative stress homeostasis. In a mouse model of epilepsy, ceftriaxone (200 mg/kg/day, 6 days) provides significant protective actions against pentylenetetrazole (PTZ)-induced generalized clonic and clonic-tonic convulsions, as well as convulsion-induced mortality within 30 min of PTZ administration (Jelenkovic et al., 2008). The same was observed in rats treated with 200 mg/kg or 400 mg/kg only once, 60 min prior to the PTZ administration (Uyanikgil et al., 2016). The protective effects of ceftriaxone for oxidative stress damage due to PTZ administration have also been shown in the hippocampus, where reduced lipid peroxidation and increased catalase activity and glutathione levels are found (Hussein et al., 2016). Additionally, ceftriaxone might exert neuroprotective actions through another mechanism such as inactivation of connexins 43 hemichannels as down regulation of connexins 43 expression in hippocampal regions was seen in ceftriaxone treated animals (Hussein et al., 2016). Treatment with ceftriaxone (100 or 200 mg/kg) diminishes PTZ-induced mean kindling score and cognitive impairment (Soni et al., 2015). Additionally, ceftriaxone pretreatment attenuates PTZ-induced increase of MDA and nitrite levels and decrease of glutathione levels in the cortex and subcortical brain regions in rats (Soni et al., 2015).

In conclusion, ceftriaxone reduces seizures and normalizes glutamate and oxidative stress homeostasis in preclinical studies, while not being effective in the TMEV model. The timing of ceftriaxone treatment initiation may play a role in its effectiveness, with reduced ability to influence GLT-1 and seizure frequency when treatment is started after seizure activity in some, but not all models.

5. The effects of ceftriaxone on molecular and behavioral markers in rodent models of substance use disorders

5. 1. Cocaine

Due to the well-established role for glutamate in regulating cocaine-seeking, ceftriaxone has been tested in several rodent models of cocaine use disorder. The first experiments published utilized the extinction-reinstatement model of cocaine relapse, in which rats are trained to self-administer intravenous cocaine in an operant chamber. Cocaine infusions are typically accompanied by presentation of drug-paired cues such as a stimulus light and tone. Self-administration is followed by extinction training, in which the instrumental response made to obtain cocaine is no longer accompanied by drug or drug-associated cues resulting decreased responding. In male Sprague Dawley rats, subchronic ceftriaxone (200 mg/kg once daily for 7 days) administered after each daily extinction session attenuates the reinstatement of the cocaine-seeking response that is primed by both drug-associated cues and cocaine itself (Knackstedt et al., 2010a). This paper was the first to report that cocaine decreases the expression of both GLT-1 in the NA core and that ceftriaxone restores this expression (Knackstedt, et al. 2010a). Importantly, in cocaine-naïve rats, ceftriaxone does not increase NA core GLT-1 or xCT. The effect of ceftriaxone on reinstatement and GLT-1 expression following cocaine is dose-dependent, as 100 and 200 mg/kg upregulate GLT-1 and attenuate cue-primed reinstatement, while 5 days of 50 mg/kg administration does not (Sari et al., 2009). Glutamate uptake assays further confirm that ceftriaxone (200 mg/kg for 5 days) restores both GLT-1 and system xc- function in the NA core after cocaine self-administration, resulting in increased basal glutamate levels and attenuated glutamate efflux during a cocaine-primed reinstatement test (Trantham-Davidson et al., 2012). Antisense-mediated knockdown of GLT-1 or xCT in the NA core reduces GLT-1 levels and prevents ceftriaxone from attenuating reinstatement (LaCrosse, et al. 2017). However, xCT knockdown in the NA core also results in decreased GLT-1 expression in rats treated with ceftriaxone and thus a necessary role for xCT upregulation alone has not yet been established. On the other hand, NA core GLT-1 knockdown does not alter xCT expression. Thus, ceftriaxone administered during extinction training for 5–7 days consistently attenuates the reinstatement of cocaine seeking and this effect is dependent on GLT-1 upregulation in the NA core.

Ceftriaxone’s ability to attenuate reinstatement and restore GLT-1 and xCT expression persists at least two weeks beyond acute treatment (Sondheimer and Knackstedt, 2011). When administered for 5 days prior to the initiation of cocaine self-administration, ceftriaxone (200 mg/kg) does not alter the acquisition or maintenance of cocaine self-administration (Sondheimer and Knackstedt, 2011). Ceftriaxone treatment was discontinued during the two weeks of extinction training, but still resulted in both increased NA core GLT-1 expression and attenuated cue- and cocaine-primed reinstatement, indicating a long-term protection from relapse (Sondheimer and Knackstedt, 2011). While Sondheimer and Knackstedt (2011) found no effects of ceftriaxone on the acquisition of self-administration in rats, the same dose of ceftriaxone administered to mice prevents both the acquisition of cocaine self-administration and the motivation to seek cocaine on a progressive-ratio schedule (Ward et al., 2011). Thus, the effects of ceftriaxone on cocaine intake may be species-specific.

When extinction training is not employed during cocaine abstinence, ceftriaxone continues to attenuate cocaine-seeking after rats are placed back into the operant chamber for the first time in 21–45 days (Fischer et al., 2013; LaCrosse et al., 2016). Ceftriaxone (200 mg/kg for 5 days) attenuates cue-primed cocaine seeking and increases both NA core and shell GLT-1 expression after 45 days of abstinence without extinction. However, when pharmacological inhibition of GLT-1 is employed, only NA core and not shell GLT-1 is necessary for ceftriaxone to attenuate cue-primed relapse (Fischer et al., 2013). After a shorter period of abstinence (21 days), ceftriaxone (100 and 200 mg/kg for 6 days) reduces context-primed relapse to cocaine seeking but only 200 mg/kg increases GLT-1 expression in the NA core (LaCrosse et al., 2016). Thus, ceftriaxone has protective effects against relapse and GLT-1 expression in the absence of extinction training. NA core GLT-1 upregulation is necessary for the attenuation of relapse in this model.

Comorbidities between cocaine use and alcohol use/stress can alter ceftriaxone’s ability to attenuate cocaine seeking. Ceftriaxone (200 mg/kg for 5–7 days) is not effective in preventing cue+cocaine-induced reinstatement in Sprague Dawley rats with a history of both cocaine self-administration and alcohol consumption. This may be due to the fact that rats with a combined cocaine and alcohol use history exhibit greater NA core GLT-1 surface expression than rats that were only exposed to cocaine, and thus ceftriaxone treatment would not be indicated (Stennett et al., 2019). Ceftriaxone restores NA core GLT-1 expression and function and reduces stress-induced locomotion in Sprague Dawley rats exposed to acute immobilization stress, although it does not reverse the α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)/ N-methyl-d-aspartate (NMDA) ratio increase initiated by acute stress (Garcia-Keller et al., 2016). Additionally, in a rat model of comorbid PTSD+cocaine use disorder, ceftriaxone restores NA core GLT-1 and xCT in Sprague Dawley rats following cocaine self-administration regardless of stress-susceptibility. However, ceftriaxone (200 mg/kg for 5 days) has less of an effect on cue-primed reinstatement in stress-susceptible rats relative to stress-resilient and unstressed controls, indicating that restoring GLT-1 and xCT expression is not sufficient to attenuate reinstatement following stress (Schwendt et al., 2018).

In addition to altering behavior and protein expression after cocaine self-administration, ceftriaxone pretreatment (200 mg/kg) for 5 or 10 days decreases locomotion and stereotypy following non-contingent cocaine administration in Sprague Dawley rats (Barr et al., 2015; Sondheimer and Knackstedt, 2011). Ceftriaxone (200 mg/kg for 10 days) also attenuates the increase in NA dopamine release produced by acute cocaine injection, an effect not prevented by pharmacological blockade of GLT-1 (Barr et al., 2015). Further linking the effects of ceftriaxone on cocaine-induced locomotion to dopamine transmission, increased Akt/GSK3 signaling and decreased expression of the dopamine transporter, TH and phosphorylated α-synuclein is observed in the NA after 10 days of ceftriaxone (Barr et al., 2015). In a conditioned place preference (CPP) model, ceftriaxone administered during a drug-free period (200 mg/kg for 7 days) reduces time spent by Wistar rats in the cocaine-paired chamber (Niedzielska-Andres et al., 2019). In this model, vehicle-treated rats display decreased GLT-1 protein expression in a NA core membrane fraction; ceftriaxone prevents this decrease. While there was a trend for similar effects for xCT expression this did not attain significance. Interestingly, in the NA, neither cocaine- nor ceftriaxone-induced changes in GLT-1 are not accompanied by changes in expression of NF-κB which regulates its transcription (Niedzielska-Andres et al., 2019). This agrees with the lack of effect of either cocaine or ceftriaxone 200 mg/kg for 6 days on GLT-1 mRNA in the NA core after cocaine self-administration and extinction (LaCrosse et al., 2017).

Ceftriaxone alters expression of other glutamatergic proteins and GLT-1 in the NA core and brain regions outside the NA after cocaine. While ceftriaxone (100 and 200 mg/kg) increases GLT-1 in the PFC following 5 days of extinction training with ceftriaxone administered immediately following the extinction session (Sari et al., 2009), following 21 days of abstinence, ceftriaxone (100 and 200 mg/kg) does not increase GLT-1 in the PFC (LaCrosse et al., 2016). This is consistent with the idea that implementation of extinction training alters glutamate neuroplasticity in several brain regions (Knackstedt et al., 2010b; Peters et al., 2008). In Wistar rats, ten days of ceftriaxone treatment (200 mg/kg) has no significant effect on GLT-1 or GLAST expression in the dorsal striatum following non-contingent cocaine injections (Parikh et al., 2014). Ceftriaxone reverses GFAP (glial fibrillary acidic protein) reduction in the NA core following cocaine self-administration but has no effect on volume or surface area of astrocytes, which are major sites of extracellular GLT-1 expression (Scofield et al., 2016). Ceftriaxone reduces total protein mGlu5 expression in the NA core of female rats, regardless of estrous phase, an effect not found in male rats (Bechard et al., 2018). Cocaine has been found to increase expression of the GluA1 subunit of the AMPA receptor in the NA core, which results in increased calcium influx in response to glutamate binding and potentiated cocaine seeking (Conrad et al., 2008). Both 100 and 200 mg/kg ceftriaxone reduce GluA1 subunit expression and have no effect on GluA2 subunit expression in the NA core after abstinence without extinction (LaCrosse, et al. 2016). Seven days of ceftriaxone following non-contingent cocaine in a CPP model increases hippocampal GLT-1 and NF-κB, increases dorsal striatum xCT and NF-κB, and decreases nuclear factor erythroid 2-related factor 2 (Nrf2) in the prefrontal cortex, relative to vehicle-treated rats (Niedzielska-Andres et al., 2019).

To summarize, ceftriaxone consistently increases GLT-1 expression in the NA core while also attenuating reinstatement to cocaine seeking. Ceftriaxone’s capacity to restore GLT-1 dysfunction persists across stress phenotypes and despite acute stress. Many, but not all, studies find that cocaine decreases and ceftriaxone increases xCT expression in the NA core. Furthermore, ceftriaxone acts on several other proteins that mediate cocaine reinstatement, such as GluA1 in males and females and mGlu5 in females.

5.2. Other Psychostimulants: Nicotine and Amphetamine

The effects of ceftriaxone on GLT-1/xCT expression after administration of other psychostimulants has only been assessed in four studies. In male Sprague-Dawley rats that self-administered IV amphetamine, 6 days of ceftriaxone (200 mg/kg) prevents cue-primed reinstatement of amphetamine-seeking only in rats housed in an enriched environment, and not in those housed in standard or impoverished conditions (Garcia et al., 2019). Ceftriaxone also attenuates amphetamine-primed reinstatement in rats housed in enriched conditions. When these rats were euthanized 5 days after the reinstatement test, GLT-1 and xCT expression were unchanged in the membrane fraction of PFC or NA core tissue (Garcia et al., 2019). There are several potential reasons for such effect, including the analysis after reinstatement testing and the fact that amphetamine itself does not reduce GLT-1 expression (Garcia et al., 2019). In male Wistar rats receiving 4 non-contingent injections of methamphetamine (METH) and euthanized 48 h later, GLT-1 is decreased by METH in the NA, medial PFC (mPFC), dorsal striatum and hippocampus; this expression is restored by treatment with ceftriaxone during the withdrawal period (Alshehri et al., 2017; Althobaiti et al., 2016). In these rats, xCT expression is not affected by METH but is increased by ceftriaxone in both NA and mPFC. No changes in xCT are observed following METH or ceftriaxone in either the dorsal striatum or hippocampus and no changes in GLAST are observed in any of the four brain regions. METH decreases tissue content of DA in the NA and glutamate in the PFC that is restored by ceftriaxone (Althobaiti et al., 2016). In a CPP model, ceftriaxone (200 mg/kg) attenuates the reinstatement of a place preference for METH in male Sprague-Dawley rats (Abulseoud et al., 2012). These rats were euthanized 72 h after the reinstatement test, revealing that ceftriaxone increases GLT-1 mRNA expression in the mPFC but not in the NA. The same pattern was observed when rats were euthanized 4 h after the last METH injection during conditioning (Alshehri et al., 2017). Male P rats consuming a sweetened nicotine solution show increased NA and PFC GLT-1 expression after 7 days of 100 mg/kg ceftriaxone. While this dose of ceftriaxone reduces consumption of the sweetened nicotine solution, it also reduces consumption of sucrose alone (Sari et al., 2016). Thus, ceftriaxone shows efficacy at reducing the motivation to seek other psychostimulants in addition to cocaine.

5.3. Alcohol

In alcohol-preferring “P” rats bred to consume higher than average amounts of unsweetened alcohol, five or more days of 100 and 200 mg/kg/day ceftriaxone decreases continuous alcohol consumption in the home cage while increasing total protein expression of GLT-1a, GLT-1b and xCT in both the PFC and NA (Alhaddad et al., 2014; Das et al., 2015; Rao et al., 2015; Sari et al., 2011). Lower doses of ceftriaxone (25 and 50 mg/kg) decrease alcohol intake in P rats in the absence of changes in NA or PFC GLT-1 expression (Qrunfleh et al., 2013; Sari et al., 2011). Similarly, while only 2 days of ceftriaxone administration (100 mg/kg) decreases drinking and increases GLT-1/xCT protein expression in the PFC, it does not affect GLT-1/xCT expression in the NA (Rao et al., 2015). Ceftriaxone also attenuates relapse-like resumption of alcohol consumption or operant seeking after a period of abstinence in both continuously drinking P rats and outbred Sprague-Dawley rats given intermittent access (Qrunfleh et al., 2013; Rao and Sari, 2014) (Weiland et al., 2015). In these experiments, rats were treated with 50, 100 or 200 mg/kg ceftriaxone for at least 5 days prior to relapse testing; all three doses were found to be effective.

In contrast to findings in continuously drinking P rats, accumbal GLT-1 and xCT expression are not decreased after 7 weeks of intermittent access to alcohol in outbred male Sprague-Dawley rats (Pati et al., 2016). Despite this, ceftriaxone (200 mg/kg) decreases alcohol consumption in this model after just two days of ceftriaxone administration and lasting for at least three days after cessation of ceftriaxone administration (Stennett et al., 2017). The amount of alcohol (g/kg/day) consumed by P rats and Sprague Dawley rats (prior to ceftriaxone administration) in the studies reviewed here is similar, and thus dose of alcohol is likely not the source of differences in GLT-1/xCT adaptations in these models. The intermittent nature of alcohol consumption with periods of withdrawal may underlie these differences, as could genetic differences between the two strains of rat. Despite differences in alcohol-induced GLT-1/xCT expression between the two models, ceftriaxone is capable of reducing alcohol consumption and increasing xCT expression in both models. This reduction in consumption is evident after only two days of administration when GLT-1 and xCT levels in the NA are not yet restored, indicating a dissociation between the two phenomena. However, this remains to be tested explicitly.

In addition to xCT and GLT-1, other proteins have been assessed after ceftriaxone administration in alcohol models. Ceftriaxone (100 mg/kg) administered for 2 or 5 days to alcohol-consuming P rats increases nuclear translocation of NFκB in the NA and PFC (Alhaddad et al., 2014). This is observed in the absence of increased GLT-1 in the NA after only 2 days of ceftriaxone, possibly dissociating NF-κB -from ceftriaxone’s mechanism of action. No change in NA or PFC expression of another glutamate transporter, GLAST, is observed after ceftriaxone (100 mg/kg) following alcohol consumption (Alhaddad et al., 2014).

In conclusion, ceftriaxone reduces alcohol consumption and seeking across different rodent models and strains. Such decreases are not always accompanied by increased expression of GLT-1 or xCT, indicating that these adaptations may not be necessary for decreased alcohol consumption. Furthermore, in both P rats and Sprague-Dawley rats, ceftriaxone increases GLT-1 and xCT in the PFC and NA in the absence of decreased expression of these proteins after alcohol consumption (Alhaddad et al., 2014; Stennett et al., 2017), akin to observations after METH. This is interesting in light of the finding that ceftriaxone does not increase GLT-1 and xCT in the NA of drug-naïve rats (Knackstedt et al., 2010a), indicating that a history of alcohol consumption may alter the NA and PFC in ways other than GLT-1/xCT expression that make it possible for ceftriaxone to alter GLT-1/xCT expression.

5.4. Opioids

Chronic morphine administration in mice reduces GLT-1 expression in the spinal cord of mice, an effect reversed by 7 days of ceftriaxone (200 mg/kg; Chen et al., 2012). Ceftriaxone (200 mg/kg) prevents the reinstatement of hydrocodone CPP in P-rats while increasing xCT expression in the NA and hippocampus but not PFC or amygdala (Alshehri et al., 2018). GLT-1 and GLAST expression is unaffected by hydrocodone or ceftriaxone in this model. Ceftriaxone (400 mg/kg) also prevents the reinstatement of morphine CPP in mice (Hearing et al., 2016). While protein expression was not assessed in this study, ceftriaxone was found to normalize the morphine-induced increased sensitivity of AMPA receptor-mediated EPSCs to Naspm, a selective antagonist of GluA2-lacking AMPA receptors, akin to the ability of ceftriaxone to reduce surface GluA1 expression after cocaine (LaCrosse et al., 2017). Following heroin self-administration and 2 weeks of extinction, decreased surface (but not total) expression of GLT-1 is decreased in the NA, leading to decreased glutamate uptake in male rats (Shen et al., 2014). Ceftriaxone (200 mg/kg for 7 days) normalizes glutamate uptake and attenuates cue-primed reinstatement of heroin-seeking. This attenuation of reinstatement is dependent on GLT-1 expression, as antisense knockdown of GLT-1 in the NA core prevents ceftriaxone from attenuating relapse. Thus, ceftriaxone attenuates the reinstatement of both heroin-seeking after operant self-administration and of opioid CPP. While GLT-1 is not decreased after only 4 non-contingent administrations of hydrocodone followed by one week of withdrawal, it is after intravenous heroin self-administration and two weeks of withdrawal. At this time, the limited data does not permit the understanding of whether these differences are caused by route of administration, response contingency of drug-delivery or dose of drug attained.

5.5. Ketamine

In adolescent male mice, chronic ketamine decreases GLT-1 and increases GFAP expression in the hippocampus, both of which are normalized by 14 days of 200 mg/kg ceftriaxone in C57BL/6Hsd mice (Dodman et al., 2015; Featherstone et al., 2012).

6. Off-target effects of ceftriaxone

Based on the ability of ceftriaxone to reduce seeking of many classes of drugs and attenuate neuropathologies, a number of studies have investigated off-target side effects of chronic ceftriaxone treatment. For example, in male Sprague-Dawley rats, subchronic ceftriaxone (200 mg/kg) does not alter spontaneous locomotion or the self-administration of regular rat chow or sucrose pellets (Knackstedt et al., 2010a; Weiland et al., 2015). In C57Bl/6 mice, 200 mg/kg ceftriaxone does not alter the consumption of sweetened condensed milk (Ward et al., 2011). However, in male P rats, ceftriaxone reduces the consumption of a sucrose solution (Sari et al., 2016). Thus, in the majority of cases, ceftriaxone has been found to be well-tolerated and without significant off-target side effects.

7. General Conclusion

Ceftriaxone increases GLT-1 expression in multiple brain regions in many rodent models of neurological and addictive diseases (see Table 1). The same is observed in vitro, where 10 μM ceftriaxone consistently increases GLT-1 expression in many cell culture models. However, 2 days of 100 μM ceftriaxone applied to cortical astrocytic cultures fails to increase glutamate uptake possibly because this dose was too high. The dose of 200 mg/kg ceftriaxone administered for at least 5 days consistently increases GLT-1 expression in vivo in rodent models of all neurological and addictive disorders reviewed here, while lower doses have been shown to increase expression only in some models and brain regions. For example, 100 mg/kg increases GLT-1 expression in the PFC and NA after continuous consumption of alcohol, but not after abstinence from cocaine (LaCrosse et al., 2016; Sari et al., 2011). The lowest dose tested, 50 mg/kg, is unable to increase GLT-1 expression in the PFC or NA after alcohol (Qrunfleh et al., 2013; Sari et al., 2011) but does so in the DG in healthy mice (Karaman et al., 2013). The hippocampal regions are more amenable to GLT-1 upregulation after ceftriaxone in the absence of disease, in terms of sensitivity to lower doses and consistent GLT-1 upregulation. In healthy rats and those consuming amphetamines or alcohol, ceftriaxone does not increase NA or PFC GLT-1 in the absence of a reduction in expression (Garcia et al., 2019; Knackstedt et al., 2010a; Stennett et al., 2017). Ceftriaxone is equally able to increase GLT-1 expression in both male and female rodents, with only one strain difference: the ability of ceftriaxone to increase GLT-1 after alcohol consumption. Another finding consistent across in vivo models is the lack of effect of ceftriaxone on GLAST expression (but see Benkler et al., 2013).

Table 1.

The effects of ceftriaxone on GLT-1 expression

Reference(s)	Model	Species/strain	Brain region (s)	GLT-1increased by Ceftriaxone?
Lee et al., 2008 Lewerenz et al., 2009 Rothstein et al., 2005	Cell culture	HT22, spinal cord, PHFA	n/a	Yes: 10–300 μM for 2–7 days
Zhang et al., 2015	Cell culture	Primary cortical astrocytes	n/a	100 μM for 48 hours Yes: Total protein No: Membrane-bound and glutamate uptake
Benkler et al., 2013	Cell culture	Primary cortical astrocytes	n/a	Yes 10 μM for 24 hours
Rothstein et al., 2005	Healthy rats	Unspecified	Hipp, spinal cord	Yes: 200 mg/kg
Hsu et al., 2015; Hu et al., 2015; Karaman et al., 2013; Omrani et al., 2009; Krzyzanowska et al., 2016	Healthy rats	Balb-c mice, Wistar rats	CA1, CA3, DG of Hippocampus	Yes: all regions after 200 mg/kg for 5–14 days; 50 mg/kg DG only; No: 200 mg/kg when killed 4 days later; No: 1 × 200 mg/kg
Chotibut et al., 2014; Krzyzanowska et al., 2016	Healthy rats	Wistar rats	Dorsal striatum	Yes: 200 mg/kg 8 days No: 200 mg/kg 5 days
Garcia et al., 2019; Knackstedt et al., 2010; Krzyzanowska et al., 2016	Healthy rats	Male Sprague Dawley and Wistar rats	Frontal cortex, NA core	No: 200 mg/kg for 5– 8 days
Rothstein et al., 2005	ALS (SOD 1 mutation)	G93A SOD1 mice	Spinal cord	Yes: 200 mg/kg for 14 days
Yamada and Jinno, 2011	ALS (axotomy)	Male C67BL/6 mice	Hypoglossal nerve	Yes: 200 mg/kg for 28 days
Miller et al., 2008	HD	Male R6/2 mice	Dorsal striatum	Yes: 200 mg/kg for 5– 8 days
Hefendehl et al, 2016	APPPS1	APPPS1 mice	Somatosensory cortex; primary visual cortex	Yes: 200 mg/kg for 5 days
Zumkehr et al, 2015	3xTg-AD	3xTg-AD mice	Hipp	Yes: 200 mg/kg/day for 2 months
Hsu et al, 2015	MPTP	Male Wistar rats	Dorsal striatum, Hipp	Yes: 100, 200 mg/kg for 9–14 days
Chotibut et al, 2014; Leung et al, 2012	6-OHDA	Male & Female Sprague Dawley rats	Dorsal striatum	Yes: 200 mg/kg for 7–8 days
Zhang et al, 2015	MPP+ in vitro model of toxicity	Primary cortical astrocytes from	n/a	Yes: 100 μM for 48 hours
Hu et al., 2015	Global brain ischemia	Male Wistar rats	Hipp	Yes: 200 mg/kg
Ouyang et al., 2007	Forebrain ischemia (OGD)	Male Sprague Dawley rats	Hipp	Yes: 200 mg/kg/day for 5 days
Lipski et al., 2007	Forebrain ischemia (OGD)	Male Wistar rats	Hipp	Yes: (only uptake), 200 mg/kg for 5 days
Mimura et al., 2011	Hypoxic ischemia	Sprague Dawley rats	Ischemic hemisphere	Yes: (mRNA and protein), 200 mg/kg for 5 days
Lai et al., 2011	Perinatal hypoxic-ischemic encephalopathy	Pregnant female Sprague Dawley rats	Cerebral cortex, Hipp, striatum	Yes: 200 mg/kg for 3 days
Chang et al., 2013	Hypoxic ischemia	Sprague Dawley rat pups	Frontal cortex, Hipp, striatum	Yes: 200 mg/kg for 3 days (postnatal days 1–3)
Verma et al., 2010	Cerebral ischemia (MCAO)	Male Sprague-Dawley rats	Cortex	Yes: 100 mg/kg for 5 days
Krzyzanowska et al., 2016	Cerebral ischemia (MCAO)	Male Wistar rats	Frontal cortex, Hipp; dorsal striatum,	Yes: 200 mg/kg for 5 days
Thone-Reinekie et al., 2008	Cerebral ischemia (MCAO)	Male Wistar rats	Hipp, amygdala, piriform cortex, striatum, hypothalamus, brain stem	No (mRNA, protein): 200 mg/kg, 90 min after MCAO
Chu et al., 2007	Cerebral ischemia (MCAO)	Male Sprague-Dawley rats	Ischemic hemisphere	Yes (mRNA, protein), 200 mg/kg for 5 days
Loewen et al., 2019	TMEV-induced epilepsy	C57BL6/J mice	Hipp (CA1)	Yes: 800 mg/kg
Zeng et al., 2010	Tuberous Sclerosis Complex	Tsc1GFAPCKO mice	Hipp, cortex,	Yes: 200 mg/kg for 1, 3, 5 weeks,
Goodrich et al., 2013	Epilepsy (LFPI)	Male Long-Evans rats	Cortex	Yes: 200 mg/kg; 1 week
Fischer et al. 2013; Knackstedt et. al., 2010, LaCrosse et al., 2016 and 2017, Sari et al., 2009; Sondheimer and Knackstedt, 2011	Cocaine relapse	Male Sprague Dawley rats	NA core and shell	Yes: 100 and 200 mg/kg but not 50 mg/kg for5–7 days after extinction; Yes: 200 mg/kg after abstinence No: 100 mg/kg after abstinence
LaCrosse et al., 2016; Sari et al., 2009	Cocaine relapse	Male Sprague-Dawley rats	PFC	Yes: 100 and 200 mg/kg for 5 days after extinction No: 100 and 200 mg/kg for 6 days after abstinence
Niedzielska-Andres et al., 2019	Cocaine CPP	Male Wistar rats	NA core	Yes: 200 mg/kg for 7 days
Garcia et al., 2019	Amphetamine relapse	Male Sprague-Dawley rats	NA core and PFC	No: 200 mg/kg for 6 days
Althobaiti et al., 2016; Alshehri et al., 2017	Non-contingent METH	Male Wistar rats	NA and mPFC	Yes: 200 mg/kg for 2 days
Alhaddad et al., 2014; Das et al., 2015; Qrunfleh et al., 2013; Rao et al., 2015; Sari et al., 2011	Alcohol consumption	Male P rats	NA core, PFC, dorsal striatum, Hipp	Yes: 100 and 200 mg/kg but not 50 mg/kg
Stennett et al., 2017	Alcohol consumption	Male Sprague Dawley rats	NA core	No: 200 mg/kg for 5 days
Chen et al., 2012	Non-contingent morphine	Male ICR mice	Spinal cord	Yes: 200 mg/kg
Alshehri et al., 2018	Hydrocodone CPP	Male P rats	NA, Hipp, PFC, amygdala	No: 200 mg/kg
Dodman et al., 2015	Ketamine CPP	Male C57BL/6Hsd mice	Hipp	Yes: 200 mg/kg

ALS- amyotrophic lateral sclerosis, CPP- conditioned place preference, HD- Huntington’s disease, Hipp- hippocampus, 6-OHDA- 6-hydroxydopamine, LFPI- lateral fluid percussion injury, MCAO- middle cerebral artery occlusion, mPFC- medial prefrontal cortex, MPP⁺- 1-methyl-4-phenylpyridinium, MPTP- 1-methyl-4phenyl-1,2,3,6 tetrahydropyridine, n/a- not applicable, NA- nucleus accumbens, OGD- oxygen glucose deprivation, PFC- prefrontal cortex, SOD- superoxide dismutase, TMEV- Theiler’s Murine Encephalomyelitis Virus.

Ceftriaxone consistently increases xCT expression in the NA and PFC after cocaine and alcohol consumption (see Table 2). Following both alcohol and METH, ceftriaxone increases xCT expression in the PFC and NA core in the absence of drug-induced deficit in expression (Alhaddad et al., 2014; Althobaiti et al., 2016; Stennett et al., 2017). However, in healthy drug-naïve rats, ceftriaxone is unable to increase xCT in the NA, PFC, or hippocampus. Thus, alcohol and METH exposure may make the NA and PFC amenable to modulating by ceftriaxone, even in the absence of xCT/GLT-1 reduction. The contribution of xCT upregulation to the ability of ceftriaxone to reduce alcohol consumption and METH reward has not yet been determined.

Table 2.

The effects of ceftriaxone on xCT expression

Reference(s)	Model	Species/strain	Brain region (s)	xCT increased by Ceftriaxone?
Lewerenz et al., 2009	Cell culture	HT22 cells	n/a	Yes: 300 μM (7 days) increases system xc-activity
Krzyzanowska et al., 2016	Healthy rats	Male Wistar rats	Dorsal striatum	Yes: 200 mg/kg for 8 days
Krzyzanowska et al., 2016	Healthy rats	Male Wistar rats	Frontal cortex, Hippocampus	No: 200 mg/kg for 8 days
Garcia et al., 2019; Knackstedt et. al., 2010	Healthy rats	Male Sprague Dawley rats	NA core	No: 200 mg/kg for 5 days
Krzyzanowska et al., 2016	Cerebral ischemia (MCAO)	Male Wistar rats	Frontal cortex	Yes (mRNA): 200 mg/kg for 5 days
Krzyzanowska et al., 2017	Cerebral ischemia (MCAO)	Male Wistar rats	Frontal cortex (astrocytes and neurons)	No: ↓ after 200 mg/kg for 5 days MCAO
Bechard et al., 2018; Knackstedt et. al., 2010; LaCrosse et al., 2017; Schwendt et al., 2018	Cocaine relapse	Male Sprague Dawley rats	NA core	Yes: 200 mg/kg for 5–7 days
Niedzielska-Andres et al., 2019	Cocaine CPP	Male Wistar rats	Dorsal striatum	Yes: 200 mg/kg for 7 days
Althobaiti et al., 2016	Non-contingent METH	Male Wistar rats	NA and mPFC	Yes: 200 mg/kg for 2 days
Alhaddad et al, 2014; Das et al, 2015; Rao et al, 2015; Sari et al, 2011	Alcohol consumption	Male P rats	NA core; PFC	Yes: 100 and 200 mg/kg but not 50 mg/kg
Stennett et al., 2017	Alcohol consumption	Male Sprague Dawley rats	NA core	Yes: 200 mg/kg
Alshehri et al, 2018	Hydrocodone CPP	Male P rats	NA, Hipp, but not PFC or amygdala	Yes: 200 mg/kg

CPP- conditioned place preference, Hipp- hippocampus, MCAO- middle cerebral artery occlusion, METH- methamphetamine, mPFC- medial prefrontal cortex, n/a- not applicable, NA- nucleus accumbens, PFC- prefrontal cortex.

While the ability of ceftriaxone to attenuate the reinstatement of cocaine-seeking has been shown to be dependent on xCT and GLT-1 upregulation in the NA, the same has not been done for alcohol-seeking. In fact, reductions in alcohol consumption are observed in the absence of measurable increases in xCT and GLT-1 expression. The necessity of GLT-1 upregulation in ceftriaxone-induced neuroprotection has only been investigated following OGD, in which antisense-mediated GLT-1 knockdown prevents the ability of ceftriaxone to reduced delayed neuronal death (Hu et al., 2015). Since ceftriaxone also ameliorates neuropathologies associated with multiple neurological disease models such as enzymes involved in amyloid beta production (see Table 3), future work should endeavor to determine the necessity of GLT-1 and xCT upregulation in these adaptations. Acute blockade of GLT-1 has been used in some studies to test the necessity of GLT-1 in mediating an effect. However, because ceftriaxone-mediated GLT-1 upregulation is chronic and stable, likely beginning 2 days after the start of administration (e.g. Althobaiti et al., 2016; Rothstein et al., 2005; Sondheimer and Knackstedt, 2011), acute pharmacological blockade is not the best method to investigate the necessity of such upregulation.

Table 3.

Other proteins affected by ceftriaxone

Reference(s)	Model	Species/strain	Brain region (s)	Effect of ceftriaxone on other proteins
Lee et al., 2008	Cell culture	PHFA	n/a	No effects of 10 μM (2 days) on GLAST
Lewerenz et al., 2009	Cell culture	HT22 cells	n/a	300 μM (7 days) increases GLAST and EAAT3 mRNA
Benkler et al., 2013	Cell culture	Primary mouse cortical astrocytes	n/a	10 μM (24 hrs) increases GLAST mRNA and protein expression
Rothstein et al., 2005	Healthy rats	Unspecified	Hippocampus, spinal cord	GLAST, EAAT3, EAAT4: no change after 200 mg/kg
Hefendehl et al, 2016	APPPS1	APPPS1 mice	Somatosensory cortex and primary visual cortex	GFAP: no change after 200 mg/kg
Zumkehr et al, 2015	3xTg-AD	3xTg-AD mice	Hippocampus	No: Amyloid precursor protein (APP), C99, C83 after 200 mg/kg ↓ Insoluble tau proteins after 200 mg/kg
Tikhonova et al, 2018	OXYS	OXYS rats	Hypothalamus, frontal cortex, striatum, amygdala	↑ Bace1 (hypothalamus), Ace2 (hypothalamus), Aktb (frontal cortex), mRNA levels after 100 mg/kg ↓ Ece1 (dorsal striatum), Ide (amygdala), Mme (amygdala), Aktb (striatum), Epo (amygdala) mRNA levels after 100 mg/kg
Bisht et al, 2014	MPTP	Male Wistar rats	Striatum, cortex	↓ TNF-α and IL-β after 100–200 mg/kg for 14 days
Ho et al, 2014; Huang et al 2015	MPTP	Male Wistar rats	Striatum, cortex	↑ TH after 5–200 mg/kg for 14–15 days
Chotibut et al, 2014	6-OHDA	Male Sprague Dawley rats	Dorsal striatum	No: GLAST ↑ TH after 200 mg/kg for 8 days
Leung et al, 2012	6-OHDA	Female Sprague Dawley rats	SNpc	↑ TH after 200 mg/kg for 8 days
Zhang et al, 2015	MPP+ in vitro model	Primary cortical astrocytes	n/a	↑ GFAP after 100 μM
Alta§ et al., 2013	Short-term global cerebral ischemia	Male Wistar rats	Brain tissues	↓ MDA levels after 100 mg/kg/once; ↑ SOD and GPx activity after 100 mg/kg for 1 day
Mimura et al., 2011	Hypoxic ischemia (postnatal day 7)	Sprague Dawleyrats	Ischemic hemisphere	GLAST: no change after 200 mg/kg for 5 days ↑ MAP-2 positive area; 200 mg/kg for 5 days ↓ TUNEL positive cells; 200 mg/kg for 5 days
Lai et al., 2011	Perinatal hypoxic-ischemic encephalopathy	Pregnant female Sprague Dawley rats	Hipp	↓ TUNEL positive cells: 100, 200 mg/kg for 3 days
Chang et al., 2013	Hypoxic ischemia	Sprague Dawley rat pups	Frontal cortex, Hipp, striatum	No: GLAST, 200 mg/kg for 3 days (postnatal days 1–3)
Verma et al., 2010	Cerebral ischemia (MCAO)	Male Sprague-Dawley rats	cortex	↑ Glutamine synthetase; 100 mg/kg for 5 days
Thone-Reineke et al., 2008	Cerebral ischemia (MCAO)	Male Wistar rats	Periinfarct zone	↑ BDNF, TrkB, NT3, but notIl-6, 200 mg/kg, 90 min after MCAO
Chu et al., 2007	Cerebral ischemia (MCAO)	Male Sprague-Dawley rats	Ischemic hemisphere	↓ TNF-α, FasL, MMP-9, activated caspase-9, 200 mg/kg for 5 days
Uyanikgil et al., 2016	PTZ-induced convulsion	Male Sprague-Dawley rats	Brain tissues	↓ MDA levels after 200 and 400 mg/kg 60 min prior to PTZ; ↑ SOD activity after 200 and 400 mg/kg 60 min prior to PTZ
Soni et al., 2015	PTZ-induced convulsion	Male Wistar rats	Cortex and subcortical region	Cef every other day for 27 days (14 injections): ↓ MDA levels 100 and 200 mg/kg ↑ nitrate levels after 100 and 200 mg/kg; ↑ glutathione levels after 100 and 200 mg/kg
Hussein et al., 2016	PTZ-induced kindling	Male Sprague-Dawley rats	Hipp	↓ MDA levels after 200 mg/kg; ↑ catalase activity after 200 mg/kg; ↑ glutathione levels after 200 mg/kg
Huff et al., 2016	Kainic acid-induced seizures	Male Sprague-Dawley rats	Hipp	↑ GAD 67-IR neurons after 200 mg/kg for 7 days
Goodrich et al., 2013	Seizure (LFPI)	Male Long-Evans rats	Cortex	↓ GFAP levels after 200 mg/kg for 7 days
Bechard et al., 2018; Lacrosse et al., 2017	Cocaine relapse	Male and female Sprague-Dawley rats	NA core	↓ GluA1 after 200 mg/kg for 5–7 days
Bechard et al., 2018	Cocaine relapse	Male and female Sprague-Dawley rats	NA core	↓ mGlu5 in female rats only
Niedzielska-Andres et al., 2019	Cocaine CPP	Male Wistar rats	Hippocampus, PFC, NA core, dorsal striatum	No change in Nrf2 after 7 days of 200 mg/kg
Niedzielska-Andres et al., 2019	Cocaine CPP	Male Wistar rats	Hippocampus, dorsal striatum, PFC, NA core	NF-κB : ↑ in hippocampus & dorsal striatum, ↓ in PFC, no change in NA after 7 days of 200 mg/kg
Althobaiti et al., 2016; Alshehri et al., 2017	Non-contingent METH	Male Wistar rats	NA and mPFC	GLAST: no change after 200 mg/kg for 2 days
Alhaddad et al., 2014;	Alcohol consumption	Male P rats	NA core; PFC	GLAST: no change after 100 mg/kg
Alshehri et al., 2018	Hydrocodone CPP	Male P rats	NA, Hipp, PFC, amygdala	GLAST: no change after 200 mg/kg
Dodman et al., 2015	Ketamine CPP	Male C57BL/6Hsd mice	Hipp	↓ GFAP after 200 mg/kg

Abbreviations: APP- amyloid precursor protein, BDNF- brain derived neurotrophic factor, CPP- conditioned place preference, GAD- glutamic acid decarboxylase, GFAP- glial fibrillary acid protein, GLAST- glutamate aspartate transporter 1, GPx- glutathione peroxidase, Hipp- hippocampus, 6-OHDA- 6- hydroxydopamine, Il-6- interleukin 6, LFPI- lateral fluid percussion injury, MAP-2- microtubule-associated protein 2, MCAO- middle cerebral artery occlusion, MDA- malondialdehyde, METH- methamphetamine, MMP-9- matrix metalloproteinase 9, mPFC- medial prefrontal cortex, MPP⁺- 1-methyl-4-phenylpyridinium, MPTP- 1-methyl-4phenyl-1,2,3,6 tetrahydropyridine, n/a- not applicable, NA- nucleus accumbens, NF-κB- nuclear factor kappaB, NT3- neurotrophin 3, Nrf2- nuclear factor erythroid 2-related factor 2, OGD- oxygen glucose deprivation, PFC- prefrontal cortex, PTZ- pentylenetetrazole, SOD- superoxide dismutase, TNF- α- tumor necrosis factor-α, TrkB- tyrosine receptor kinase B, TUNEL- transferase-mediated deoxyuridine triphosphate nick end labeling,

Another consistent finding across rodent models is the ability of ceftriaxone to alter TH (Chotibut et al., 2014; Leung et al., 2012) and dopamine signaling. Following non-contingent METH administration, NA dopamine content is reduced and is restored after ceftriaxone treatment (Althobaiti et al., 2016). Conversely, the increase in dopamine typically observed after a non-contingent cocaine injection is attenuated by ceftriaxone (Barr et al., 2015). These latter two findings are somewhat in conflict with one another, and thus more work should address the effects of ceftriaxone on dopamine transmission after chronic and acute psychostimulant administration.

While pretreatment with ceftriaxone increases GLT-1 and ameliorates neuropathologies in multiple rodent models of neurological disease, such as PD and ischemia, this does not model the clinical case when patients receive a diagnosis after exhibiting symptoms for some time. Thus, more work should be done to test the effects of ceftriaxone administered after the insult. Thus far, this has been done in the case of HD, PD and ALS: when ceftriaxone treatment is begun 3 days after a neurological insult (e.g. MPTP lesions) or 12 weeks of age in the SOD1 mutant mouse, it continues to protect against cognitive, motor and GLT-1 deficits. However, when administered 90 min after MCAO, ceftriaxone is unable to restore GLT-1 in multiple brain regions.

Several papers examined the effects of ceftriaxone on regulators of GLT-1 transcription, namely NF-κB. Ceftriaxone (100 mg/kg) administered for 2 or 5 days to alcohol-consuming P rats produced increases nuclear translocation of NFκB in the NA in the absence of increased GLT-1 (Alhaddad et al., 2014). Similarly, while GLT-1 is upregulated in the NA core after ceftriaxone in a CPP model, NFκB expression is unchanged in this region. These findings possibly dissociate NF-κB from ceftriaxone’s mechanism of action. Furthermore, some papers report the upregulation of GLT-1/xCT protein without corresponding changes in mRNA in the same animals/system (Krzyzanowska et al., 2016; LaCrosse et al., 2017; Zhang et al., 2015). Thus, while in their initial report Rothstein et al., (2005) found that ceftriaxone interacts with the GLT-1 promoter, there is no consistent evidence that ceftriaxone upregulates GLT-1 protein expression via a transcriptional mechanism.

In rodent models of ischemia and epilepsy, where basal extrasynaptic glutamate levels are elevated, ceftriaxone upregulates GLT-1 and decreases basal levels of glutamate in the frontal cortex and hippocampus (Huff et al., 2016; Krzyzanowska et al., 2017; Zeng et al., 2010). As xCT expression is reduced by Cef in the ischemia model, the effects of Cef on basal glutamate levels is likely dependent on relative changes in GLT-1 and xCT whereby decreases in xCT decrease basal glutamate levels and vice versa. For example, weeks after cessation of cocaine self-administration, nucleus accumbens basal extracellular glutamate levels are decreased, as is xCT expression (Baker et al., 2003; Knackstedt et al., 2010a). In this state, ceftriaxone increases basal glutamate and xCT expression, yet GLT-1 upregulation is observed along with increased basal glutamate. While changes in xCT expression largely map onto changes in basal extrasynaptic glutamate levels, one exception is in the case of alcohol consumption in P rats. While alcohol consumption in P rats increases basal glutamate in the NA core, ceftriaxone both increases xCT and decreases basal glutamate in this model (Das et al., 2015). Thus, at this time it is not completely understood why ceftriaxone decreases basal glutamate in animal models of some diseases/disorders and increases it in others. The simple explanation is that ceftriaxone is able to “restore glutamate homeostasis”, possibly through titration of xCT and GLT-1 levels/function in a manner that is hard to assess with current tools. Alternatively, it may be that ceftriaxone is altering a molecular substrate intermediate to GLT-1/xCT that is currently unknown.

In conclusion, synthesis of the results from a variety of in vitro and in vivo models of neurological and motivational disorders reveals key information about the mechanism of action of ceftriaxone and its potential utility in humans. Ceftriaxone consistently ameliorates deficits in GLT-1, and to an extent, xCT, expression and function. Ceftriaxone is also effective in models of neurodegeneration when administered after neurological symptoms set in, with the exception of ischemia. Thus, ceftriaxone should be explored in more clinical trials for diseases such as substance use disorders, PD, and HD. While thus far, no pharmacotherapy identified to reduce drug-seeking in the reinstatement model has had successful translation to the clinic, we propose that this may be due to the lack of testing of such compounds in currently abstinent drug users. In regards to cocaine use disorder, we propose that such clinical trials should include a supervised drug-free period during which ceftriaxone is first administered, as this is necessary for the reduction of cocaine seeking as in rodent models. Conversely, rodent data shows that ceftriaxone can reduce ongoing alcohol consumption, and this should be considered in the design of such trials for alcohol use disorder. A new clinical trial with ceftriaxone for ALS in which patients with the familial SOD1 mutation are recruited may also yield positive results.

Highlights.

• Ceftriaxone (200 mg/kg for minimum 2 days) consistently restores GLT-1 expression

• In healthy rodents ceftriaxone only reliably upregulates GLT-1 in the hippocampus

• Increased GLT-1 expression does not consistently arise from increased transcription

• Ceftriaxone alters expression of other proteins that accompany neurological disease