Abstract

1.0 Introduction

Doxorubicin (DOX), trade name Adriamycin, is a potent chemotherapeutic agent, often used synergistically with other drug regimens for the treatment of cancer. Following DOX chemotherapy, breast cancer survivors report persisting cognitive dysfunction (Aluise et al., 2010; Raffa and Tallarida, 2010; Wefel et al., 2004). Chemotherapy-induced cognitive impairment (CICI) is a term often used to indicate significant impairments in concentration, verbal learning, nonverbal learning, memory function, attention and fine motor dexterity (Ahles et al., 2002; Tangpong et al., 2006; Wefel et al., 2004).

At the present time, little is known about the direct effects of DOX treatment on cognitive circuitry in vivo. DOX and it's major metabolites do not cross the blood brain barrier, yet cognitive impairment is often reported. Therefore, a gap in knowledge exists as to the mechanisms by which DOX treatment can influence cognition. Cognitive processes are mediated by glutamatergic communication between the frontal cortex (FC) and dentate gyrus (DG) of the hippocampus. We hypothesize that a single dose of DOX can produce changes in glutamate neurotransmission in brain areas where the primary excitatory neurotransmitter is glutamate (Danbolt, 2001). DOX-induced changes in glutamate neurotransmission would confirm that a single dose of DOX is capable of altering in vivo glutamate signaling in this circuitry and identify a mechanism as to how DOX treatment could consequently contribute to CICI acutely and after a chronic dosing regimen.

Recent technological advancements have provided enzyme-based glutamate-selective microelectrode arrays (MEAs) using amperometric recordings for the sub-second measurements of glutamate neurotransmission in animal models (Burmeister et al., 2002). These techniques have been previously used to investigate phasic glutamate signaling in discrete brain area of rats and mice (Hinzman et al., 2010; Stephens et al., 2011; Thomas et al., 2009; Thomas et al., 2012). In these experiments, mice were treated with 25 mg/kg of DOX and subjected to in vivo recordings in 3 sub-regions of the mouse FC and the DG 24 h post-treatment to investigate potential acute changes in glutamate regulation (see figure 1A). In an alternate cohort of mice, the Morris Water Maze (MWM) test was used as an initial indicator of their cognitive function.

Open in a separate window

Figure 1

Twenty-four hours after DOX treatment, glutamate signaling was measured in both hemispheres of the FC and DG of the mouse hippocampus using MEAs. (A) The MEA placement is illustrated in the right hemispheres of the coronal sections (adapted from the mouse brain atlas (Paxinos and Franklin, 2001)). In the FC, recordings were first made in the cingulate (Cg) cortex prior to moving into the prelimbic (PrL), then infralimbic (Il) cortices. The lower coronal section shows the approximate placement of the MEA recording sites in the DG of the hippocampus. (B) The representative in vivo glutamate recording depicts the signaling parameters assessed in these experiments. Extracellular glutamate levels were calculated after the MEA reached a stable baseline (10-20 minutes) by subtracting recordings on the sentinel site from recordings on the glutamate-selective site (containing glutamate oxidase; GluOx). The arrow denotes the local application of exogenous glutamate where detection and subsequent clearance were recorded. Once the maximum amplitude was reached, the uptake rate constant (k_-1) and T₈₀ (the time required to clear 80% of the amplitude) were calculated from clearing signal.

2.0 Results

2.1 Glutamate Clearance Parameters in the FC and DG

Local application of exogenous glutamate was used to evaluate the effect of DOX treatment on glutamate clearance from the extracellular space of the mouse brain. Two different parameters were calculated from the resulting peak: (1) the uptake rate constant (k_-1), which is the linear fit of the first order decay of the glutamate signal (sec^-1) and (2) T₈₀ clearance time (sec), which is the time it takes for the signal to decay 80% from the maximum amplitude (see figure 1).

Recordings demonstrated a significant decline of the uptake rate constant measured at various depths in the FC (F(1,9)=7.45; p=0.02; Figure 2A), where the most profoundly impacted depth corresponded to the infralimbic cortex (-2.5 mm; p<0.01). The uptake rate constant in the averaged depths of the FC was 0.44±0.05 s^-1 for saline and 0.29±0.02 s^-1 for DOX (t₁₃=3.25; p=0.01; Figure 2B). However, no significant change in mean uptake rate constant in the DG of the hippocampus was measured, with uptake rate constants of 0.36±0.05 s^-1 for saline and 0.26 ± 0.07 s^-1 for DOX-treated mice (t₁₁=1.57; p=0.15; Figure 2C).

Open in a separate window

Figure 2

The uptake rate constant (k_-1) in the FC is significantly decreased 24 h after DOX treatment. (A) The depth profile of the k_-1 in the FC reveals depth-dependent sensitivity to DOX treatment; especially at the depth of -2.5 mm, which corresponds to the infralimbic cortex (*p<0.05). (B) When the depths were averaged, a 45% decrease of the k_-1 in DOX-treated mice is demonstrated in comparison to saline treated animals (Saline=6, DOX=9; *p<0.05). (C) In the DG, there were no significant changes in k_-1 due to DOX treatment in comparison to saline treatment (n=4/group). Statistical analyses were carried out using a RM 2-way ANOVA and a unpaired 2-tailed Student's t-test. Error bars represent the mean ± SEM.

A repeated measures two-way ANOVA on T₈₀ measures at various depths in the FC did not identify a treatment effect (F(1,9)=2.72; p=0.13; Figure 3A). Also, no differences were detected when the T_80s for each depth was averaged to a single point per animal (saline: 2.3±0.1 seconds, DOX: 3.2±0.4 seconds; t₁₃=1.63; p=0.13; n=3/group; Figure 3B). However, in the DG, the clearance time was increased by 48% in the DOX-treated group; with saline-treated mice clearing 80% of glutamate in 2.9±0.2 seconds and DOX-treated mice taking 4.3±0.4 seconds (t₁₁=2.83; p=0.02; Figure 3D). Taken together, these data indicate that DOX treatment results in decreased clearance parameters of glutamate in the FC and DG.

Open in a separate window

Figure 3

Clearance times are significantly slower in the DG after DOX treatment. (A) The depth profile of the T₈₀ shows that clearance times as a function of depth and (B) combined to represent the entire region did not differ as a function of treatment (n=6-9). (C) Representative traces (baseline subtracted out) demonstrate the changes in glutamate clearance. The black arrow is where exogenous glutamate was locally applied within the FC. (D) In the DG, a 48% increase in clearance time as a function of DOX treatment (n=7) was also measured in comparison to saline treatment (n=6; *p<0.05). Statistical analyses were carried out using a 2-way ANOVA and unpaired 2-tailed Student's t-tests. Error bars represent the mean ± SEM.

2.2 KCl-evoked Glutamate Overflow in the FC and DG

To evaluate the effect of DOX on evoked release of glutamate, local applications of an isotonic KCl solution were used to depolarize terminals in the FC and DG. Similar volumes of KCl (~50-75 nl) elicited reproducible releases of glutamate (Figure 4C), which were compared between DOX-treated and control mice. Potassium-evoked glutamate overflow in the averaged depths of the FC was not significantly increased in saline treated mice (10.85±2.5 μM) as compared to in DOX-treated mice (15.68±3.0 μM) (t₁₂=1.24; p=0.24; Figure 4A). In the DG, KCl-evoked glutamate overflow was significantly greater in DOX treated mice compared to saline treated mice (saline:8.9±1.5 μM; DOX:25.3±5.3 μM; t₁₃=2.77; p=0.02; Figure 4B). Thus, local application of KCl elicited reproducible release of glutamate in the FC and DG (Figure 4C); however it was only significantly altered in the DG of DOX treated mice.

Open in a separate window

Figure 4

Potassium-evoked glutamate overflow in increased in the DG of DOX treat mice. (A) No differences in evoked glutamate overflow were evident following DOX treatment in the prefrontal cortex (n=7/group). (B) However, in the DG (Sal=7, DOX=8), glutamate overflow was significantly greater in DOX treated mice (*p<0.05). Statistical analysis was done using an unpaired 2-tailed Student's t-test. Error bars represent the mean ± SEM. (C) Representative raw traces of potassium-evoked glutamate release in the DG. The black arrows mark where KCl was locally applied to the DG. The hash marks (//) indicate where ~100 seconds were removed to preserve space.

2.3 Cognitive function as measured in the Morris Water Maze test

The MWM test was used as an assessment of cognitive function in mice treated with either Saline or DOX. A repeated measures two-way ANOVA revealed no significant changes in distance traveled to the platform over time as a function of DOX treatment (F(1,12)=0.83; p=0.38; Figure 5A). However, both swim speed and time to platform changed as a function of DOX treatment over time. With swim speed of DOX, treated animals were slower than controls (F(1,12)=10.04; p=0.008; Figure 5B) and consequently took longer to find the platform (F(1,12)=5.15; p=0.04; Figure 5C). A post-hoc multiple comparison of swim speed over time indicated the most profound impact of DOX was over the first 2 days following treatment. These data indicated that DOX significantly decreased the speed at which the mice traveled, but the overall distance to the platform was not impacted.

Open in a separate window

Figure 5

DOX treatment impaired performance of MWM. (A) Distance traveled was similar between both groups. (B) Swim speed was significantly slower after DOX treatment. (C) The time it took to travel to the platform was significantly greater for DOX treated mice. Significance of the overall effect of DOX treatment using a repeated measures 2-way ANOVA is indicated with a black diamond (p<0.05). Significant differences indicated with the post-hoc multiple comparison are indicated with a star (*p<0.05 in comparison to control on the same day). Error bars represent the mean ± SEM.

3.0 Discussion

These experiments evaluated the effect of a single, acute DOX treatment on glutamate neurotransmission in two anatomical structures integral to normal cognition in mice (frontal cortex and hippocampus). The Morris Water Maze was employed to assess cognitive function. To the best of our knowledge, these are the first studies to directly address the in vivo properties of glutamate neurotransmission following a dose of DOX in the mouse similar to that known to cause cognitive deficits in humans. Our results showed that at 24 h post-injection, glutamate dynamics were affected in different ways depending of the brain region. Glutamate uptake was 45% slower in the FC and in the DG, glutamate took 48% longer to clear than saline-treated mice. KCl-evoked glutamate release was significantly greater in the DG of DOX treated mice, supporting that evoked release of glutamate was impacted by DOX treatment (Stephens et al., 2011). These data support the hypothesis that systemic DOX treatment changes glutamate neurotransmission within 24 h post-administration. Further, these changes may represent the beginning of alterations to glutamate neurotransmission involved in cognitive circuitry following DOX treatment that may lead to the development of CICI.

DOX-induced slower clearance of glutamate was identified by decreased k_-1 in the FC and a longer T₈₀ in the DG. Changes in glutamate clearance can result from changes in glutamate transporter trafficking, transport capacity, transporter affinity, transcription and translation, and abnormal transporter post-translational modifications (Maragakis and Rothstein, 2004). In these studies, the k_-1 best represents the velocity of transporters, whereas the T₈₀ is more representative of affinity for glutamate to the transporter. In all areas of the brain, 80-90% of glutamate transporters are located on astrocytes (Danbolt, 2001). Previous research in our lab has shown that changes in the surface expression of glutamate transporters can correlate with uptake parameters (Nickell et al., 2007). Thus, a decrease in the surface expression of GLAST and/or GLT-1, the predominant glial transporters in rodents, may explain the findings following DOX administration. The effects of DOX on glutamate clearance could also represent an adaptive change to glial elements that alter the regulation of glutamate signaling. This has been seen in our recent work involving an animal model of traumatic brain injury (Hinzman et al., 2012).

From our experience, an increase in KCl-evoked glutamate overflow could represent several alterations: an increase in presynaptic release, output of glutamate from glia, a decreased capacity for glutamate clearance and possible changes in mGluR regulation of glutamate release. Changes in glutamate overflow could also arise from impaired extracellular buffering capacity (Santhakumar et al., 2003) or vesicular loading (high quantal transmitter concentration or more released vesicles) (Watt et al., 2000). GABA-ergic modulation is also important in balancing inhibitory/excitatory signaling, and changes in GABA neurotransmission, receptor number and function could also contribute to increased glutamate overflow (Jacob et al., 2008). Slower glutamate clearance parameters measured in these experiments likely contribute to the increased glutamate overflow in the DG, however, additional pharmacological studies are needed to determine the mechanism(s).

In order to assess cognitive function after DOX treatment, mice were subjected to MWM testing for 5 consecutive days after a single DOX administration by assessing distance traveled, swim speed and time to reach the platform. DOX treatment did not significantly impact the total distance traveled to the platform, only the speed at which the mice were traveling. Slower travel likely results from the single dose DOX causing lethargy, but not necessarily impairing cognitive function. Thus, the changes we are seeing in glutamate regulation may be the beginning of the functional changes that occur following DOX treatment that later culminate in cognitive impairment. This is in keeping with slow changes that are often seen in the cognitive function of cancer survivors who have been treated with DOX (Zimmer et al., 2015).

Currently, data support that DOX does not enter the CNS directly but accumulates outside the blood-brain barrier (Tangpong et al., 2006). Despite DOX's, or its major metabolite's, inability to traverse the blood brain barrier, DOX treatment has been reported to increase tumor necrosis factor-α (TNF) in serum from 1 h to 24 h post-injection. TNF could then infiltrate the brain, causing CNS effects by TNF-initiated production of reactive oxygen and nitrogen species (Gaman et al., 2016; Gutierrez et al., 1993; Joshi et al., 2007; Joshi et al., 2010; Keeney et al., 2015; Tangpong et al., 2006; Tangpong et al., 2007). Tangpong et al. provides support for this mechanism by reporting increased levels of TNF in brain homogenates at 3 h post-injection, heavily localized to neurons in the cortex and hippocampus (Tangpong et al., 2006). Previous studies in rat organotypic hippocampal–entorhinal cortex (HEC) brain slice culture suggest that TNF inhibited glutamate clearance through a similar mechanism as the glutamate uptake inhibitors, dihydrokainate (DHK) and trans-pyrrolidine-2,4-dicarboxylic acid (t-PKC) reference (Zou and Crews, 2005). Indirect alterations in glutamate clearance can also arise from oxidative stress; where receptor mediated alterations in glutamate uptake through NF-κB have been demonstrated (Zou and Crews, 2005). Additionally, TNF has also been shown to inhibit EAAT2 (the human homolog of GLT-1) in primary normal human fetal astrocytes by interacting with transcription and EAAT2 promoter activity as well as decreasing mRNA and protein levels (Su et al., 2003). More studies are necessary to delineate mechanisms controlling glutamate transporter disruption after DOX treatment, especially additional studies that address the long-term effects and repeated administration of DOX, which would better mimic the routine clinical usage of DOX in cancer patients.

These data support that the FC and DG of mice are vulnerable to acute treatment with DOX. Clinical studies, using a variety of chemotherapy regimens, have documented long-term treatment-induced deficits in executive function and working memory in cancer survivors (Ahles and Saykin, 2007; Kesler et al., 2011; Saykin et al., 2006). Structural and functional magnetic resonance imaging (MRI) studies report volumetric reductions and reduced activation in the frontal cortex during cognitive tasks (Ahles and Saykin, 2007; McDonald, 2011; Saykin et al., 2006). Functional MRI of patients during tasks testing executive function and episodic memory at 10 years post-chemotherapy treatment indicated hypoactivation of the dorsolateral prefrontal cortex and parahippocampal gyrus, while no significant changes in the hippocampus were detected (Note: 5-fluorouracil, epirubicin, cyclophosphamide chemotherapy regimen were administered in these experiments, where epirubicin, like DOX, is an anthracycline with similar mechanisms of action) (de Ruiter et al., 2011). In a study with monozygotic twins, where one was treated with a regimen including DOX, hyperactivation of the frontal cortex and parietal cortex were apparent in the DOX-treated twin in a functional MRI testing working memory; however performance accuracy was similar (Ferguson et al., 2007). Taken together, these studies support the idea that the FC and DG are susceptible to long-term deficits as a result of chemotherapy treatment (DOX and alternative anthracyclines).

Taken together, these data support the hypothesis that systemic DOX treatment results in changes in glutamate neurotransmission in key memory areas within 24 h, without a lasting impact on spatial learning and memory. The DOX dose used in these studies was similar to the maximum single clinical therapeutic dose, supporting that these studies likely have relevance to the clinical usage of DOX. Additional studies are needed as understanding the functional effects of DOX on glutamate neurotransmission may help us understand and prevent some of the debilitating side effects of chemotherapeutic treatment in cancer survivors.

4.0 Experimental Procedures

Acknowledgments

Abbreviations

Footnotes

References