Skip to main content
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2015 Aug 22.
Published in final edited form as: Neuroscience. 2014 Jun 4;274:308–317. doi: 10.1016/j.neuroscience.2014.05.051

Astrocytes, but not microglia, are activated in oxaliplatin and bortezomib-induced peripheral neuropathy in the rat

Caleb R Robinson a, Hongmei Zhang a, Patrick M Dougherty a
PMCID: PMC4099296  NIHMSID: NIHMS607165  PMID: 24905437

Abstract

Spinal microglia are widely recognized as activated by and contributing to the generation and maintenance of inflammatory and nerve injury related chronic pain; whereas the role of spinal astrocytes have received much less attention, despite being the first glial cells identified as activated following peripheral nerve injury. Recently it was suggested that microglia do not appear to play a significant role in chemotherapy-induced peripheral neuropathy (CIPN), but in contrast astrocytes appear to have a key role. In spite of the generalizability of astrocyte recruitment across chemotherapy drugs, its correlation to the onset of the behavioral CIPN phenotype has not been determined. The astroglial and microglial markers GFAP and OX-42 were imaged here to examine glial reactivity in multiple models of CIPN over time and to contrast this response to that produced in the spinal nerve ligation model. Microglia were strongly activated following spinal nerve ligation, but not activated at any of the time points observed following chemotherapy treatments. Astrocytes were activated following both oxaliplatin and bortezomib treatment in a manner that paralleled chemotherapy-evoked behavioral changes. Both the behavioral phenotype and activation of astrocytes was prevented by co-administration of minocycline hydrochloride in both CIPN models, suggesting a common mechanism.

Keywords: astrocytes, microglia, oxaliplatin, bortezomib, minocycline

Introduction

The precise molecular pathways that govern the induction and maintenance of neuropathic pain phenotypes are not fully understood. However, several lines of evidence indicate that an interaction between sensitized spinal neurons and activated spinal glial cells mediated by the localized release of pro-inflammatory cytokines plays a critical role in this complex process (Graeber, 2010; Sivilotti and Woolf, 1994; Ji and Suter, 2007; Miller et al., 2009). Microglia in particular have been implicated as playing an important role across multiple models of chronic pain via immune or inflammatory activity (Graeber, 2010; Ji and Suter, 2007). Microglia are not distinct in this regard, however, as it has also been known that astrocytes are activated following nerve injury for over 20 years (Garrison et al., 1991). Identification of a common mechanism for glial involvement in multiple models of pain may be important for understanding how such models are developed or maintained.

Chemotherapy-induced peripheral neuropathy (CIPN) is a chronic disorder characterized by numbness, tingling, burning sensations, lack of sensation, or other dysthesias in the extremities (Dougherty et al., 2008). However, the exact symptoms and time to their development in CIPN vary from one drug to another (Cata et al., 2006; Cavaletti and Marmiroli, 2010). To better understand the pathophysiology of CIPN, and to identify new possible treatments that translate well between chemotherapeutics, it is necessary to first identify common mechanisms that contribute to or distinguish CIPN as a type of chronic pain. A recent paper highlighted one such feature in CIPN, a lack of microglial reactivity in CIPN models that is otherwise present in overt nerve injury models (Zheng, Xiao, and Bennett, 2011; Zhang et al. 2012). This finding was surprising in the context of other glial research in chronic pain, which suggested a major role for microglia in chronic pain as a whole. On the other hand, astrocytes have also been shown to be involved in and sufficient for the development and maintenance of some types of chronic pain (Gao and Ji, 2010; Hald, 2009). The involvement of astrocytes in CIPN was previously shown in our lab in the absence of microglial activation using a paclitaxel model (Zhang et al., 2012). The present study tests the generalizability of this observation to oxaliplatin and bortezomib-induced CIPN models under the hypothesis that astrocytes contribute a common activity in CIPN as a whole.

Minocycline hydrochloride has been shown to prevent the development of behavioral indicators of pain in multiple models, presumably through the inhibition of microglia (Hua et al., 2005; Ledeboer et al., 2005; Guasti et al., 2009). However, minocycline has prevented the development of CIPN symptoms in spite of the lack any sign of microglial activation (Boyette-Davis and Dougherty, 2011; Boyette-Davis et al., 2011; Cata, Weng, and Dougherty, 2008; Zheng, Xiao, and Bennett, 2011). This would suggest that it is not, as many believe, a selective inhibitor of microglia, but may act through global anti-inflammatory mechanisms. This kind of activity would certainly inhibit microglial activation, but could also prevent inflammatory mechanisms within astrocytes. Accordingly, an important follow-up to investigating astrocyte activity was to examine whether minocycline abrogated potential up-regulation of astrocytes in bortezomib- and oxaliplatin-related CIPN. Abrogation of both astrocyte up-regulation and changes to behavioral phenotype by this single agent would suggest a correlation between the two. Abrogation of mechanical sensitivity by treatment with minocycline alongside oxaliplatin has already been shown (Boyette-Davis and Dougherty, 2011), but effects of minocycline on mechanical sensitivity in bortezomib have not yet been shown. Thus, whereas the first goal of the present study was to establish a glial activation profile in bortezomib compared to oxaliplatin, the second goal was to establish whether any observed changes in mechanical sensitivity and glial activation in either model are similarly blocked by minocycline.

Experimental Procedures

Animals

All procedures were reviewed and approved by the M.D. Anderson Institutional Animal Care and Use Committee and were in accordance with the guidelines established by the NIH and the International Association for the Study of Pain. 111 Male Sprague-Dawley rats between 60-75 days of age upon beginning treatment (300-350 g) were used for all experiments. Rats were housed in a facility with a 12h light/dark cycle and were given food and water ad libitum. All efforts were taken at each stage of the experiments to limit the numbers of animals used and any discomfort to which they might be exposed.

Drugs

All drugs were administered by intraperitoneal injection in a volume of 0.5 ml. Oxaliplatin (Tocris Bioscience) was administered in dextrose vehicle at a dose of 2mg/kg on days 1, 3, 5, and 7 of experimentation for a cumulative dose of 8mg/kg as previously described (Boyette-Davis and Dougherty, 2011). Bortezomib (Millennium Pharmaceuticals) was administered in saline vehicle at a dose of 0.15 mg/kg on days 1, 3, 5, and 7 of experimentation for a cumulative dose of 0.60 mg/kg. Groups treated with minocycline hydrochloride (Sigma Aldrich) were injected daily with 25.0 mg/kg minocycline in saline vehicle beginning at day 0 and continuing daily through day 8 (one day past chemotherapy treatment) of experimentation for a cumulative dose of 225mg/kg. Control groups were injected with an equivalent volume of appropriate vehicles (saline for bortezomib or dextrose for oxaliplatin).

Surgery

As a positive control for activation of both astrocytes and microglia, 6 rats received spinal nerve ligation (SNL) surgery (Kim and Chung, 1992). The rats were anesthetized using inhaled isoflurane (3-4%) to an adequate depth, verified by loss of nociceptive and blink reflexes. The L5 spinal nerve was exposed immediately distal to the dorsal root ganglion and then ligated with 6-0 silk suture. The wound was then closed in layers using vicryl suture and the skin closed with wound clips. The animal was monitored during recovery until it resumed normal activity. Another 6 rats received sham surgery, in which the L5 nerve was exposed, but not ligated.

Behavior testing

Sensitivity to mechanical stimuli was assessed in all animals using von Frey filaments (Boyette-Davis and Dougherty, 2011; Boyette-Davis et al., 2011). Filaments calibrated to 4g, 10g, 15g, and 26g bending force were applied 6 times each to the mid-plantar surface of each hindpaw in order to determine the filament corresponding to the animal's response threshold. Application of filaments began following a half-hour habituation period with the lowest (4g) filament. This was followed by other filaments of increasing bending force until a withdrawal threshold was obtained. Rats were allowed a resting period of 5 to 10 minutes between filaments in order to minimize the possibility of responses due to anxiety during testing. Filaments were applied with steady force until bending of the filament was observed and held for approximately 1second. A response was evaluated as a rapid withdrawal of the paw. The threshold for sensitivity to mechanical stimuli was recorded as the bending force of the filament for which at least 50% of applications elicited a response. The mean of this threshold was reported for each treatment group at each time point. Error was reported as standard error of the mean (SEM), and significance was tested at critical time points (those in which the errors of the two groups did not overlap or in which there was minimal overlap of errors) using Mann-Whitney tests.

The persistence of sensation was also measured based on von Frey responses that evoked exaggerated behaviors such as prolonged lifting, shaking, or licking of the paw. Out of the responses used to determine the 50% withdrawal threshold, the number of these that evoked an exaggerated response was recorded and expressed as percent of total responses. Error was reported as standard error of the mean (SEM), and significance was tested in bortezomib and bortezomib + minocycline groups versus saline-treated controls using Mann-Whitney tests.

Tissue collection

At the conclusion of the behavioral testing, animals with confirmed CIPN or SNL mechanical hyper-responsiveness were overdosed with sodium pentobarbital (150mg), then perfused intracardially with room temperature heparinized saline, followed by cold 4% paraformaldehyde in 0.1M phosphate buffer. Spinal cords were removed and post-fixed in 4% paraformaldehyde at 4oC overnight, then moved to 15% sucrose the following day. Tissue was then moved after 24h to 30% sucrose for a minimum of 48h as a cryoprotectant. The lumbar enlargement was mounted and the L5 segment cut in a cryostat at a thickness of 30μm.

Immunohistochemistry

Spinal cord slices were washed in phosphate buffered saline (PBS) for 6 washes lasting 15 minutes each and then blocked in normal donkey serum (5% NDS and 0.2% triton X in PBS). Slices were incubated overnight at 4°C with primary antibodies against GFAP (mouse anti-rat 1:1000, Cell Signaling Technology) or OX-42 (mouse anti-rat 1:1000, AbD Serotec). Slices were washed the following day in PBS for 6 washes lasting 15 minutes each, then incubated with either FITC-conjugated donkey anti-mouse secondary antibody for the GFAP primary antibody (1:500, Jackson Immunoresearch) or Cy-3-conjugated donkey anti-mouse secondary antibody for the OX-42 primary antibody (1:500, Jackson Immunoresearch) for 2 hours at room temperature. Slices were washed for a final course of 6 washes lasting 15 minutes each, then mounted onto glass slides using Vectamount medium.

Quantification of immunohistochemistry

Slices were viewed and images captured using light and fluorescent illumination at 10X using a Nikon Eclipse E600 microscope. These images were layered with both a fluorescent and bright field image to allow toggling between filters to draw a region of interest limited to but containing the full dorsal horn. The region of interest was drawn using the light image to avoid bias from fluorescent imaging. The filter was then changed to fluorescence imaging, maintaining the location of the region of interest. A region containing only background signal was selected within the slice, and the corresponding level of fluorescence intensity was analyzed using NIS Elements software (Nikon, USA). Background was subtracted and the resulting signal was then expressed as intensity within the region of interest in pixels/μm2. For each treatment group, the mean of these values was calculated and expressed as a percent versus the control group of the same time point. Error was expressed as the combined standard error of the mean (SEM) for treatment and control groups to account for variability within controls. Significance was determined via Mann-Whitney test (α = 0.05, 0.025, 0.01).

Results

Behavior

Mechanical sensitivity as described by von Frey filament testing was taken for rats treated with bortezomib, bortezomib + minocycline, saline, oxaliplatin, oxaliplatin + minocycline, dextrose, SNL (ipsilateral and contralateral to surgery), and sham surgery (ipsilateral and contralateral to surgery). Baseline von Frey withdrawal thresholds were not significantly different between groups treated with bortezomib (21.2 ± 1.5g), saline (21.2 ±1.4g), or bortezomib + minocycline (23.3 ± 1.7g). Rats treated with bortezomib showed a steady decrease in withdrawal thresholds to von Frey stimuli starting at day 6 that became significantly different from vehicle-treated controls at day 12 (bortezomib = 10.3 ± 1.2g/ saline = 19.8 ± 2.1g) and reached peak severity at days 20 (bortezomib = 11.0 ± 2.6g/ saline = 19.4 ± 2.3g), 24 (bortezomib = 5.7 ± 1.1g/ saline = 18.6 ± 3.2g) through day 28 (bortezomib = 9.4 ± 1.3g/ saline = 18.3 ± 2.7g) (Fig. 1A). The rats showed gradual recovery in withdrawal threshold after day 41 until a complete recovery was observed at day 69 (bortezomib = 15 ± 0g/ saline = 17.8 ± 3.8g). Rats co-treated with minocycline showed no difference in withdrawal threshold versus saline-treated controls at any time point.

Figure 1.

Figure 1

Changes in mechanical sensitivity were assessed using von Frey filament withdrawal thresholds. (A) Bortezomib treated rats (n=16) showed an decrease in withdrawal thresholds compared to saline treated controls (n=15) that was prevented in rats co-treated with minocycline (n=8). (B) Rats treated with spinal nerve ligation (SNL) (n=6) showed a marked decrease in withdrawal thresholds in feet ipsilateral to ligation versus contralateral, and sham surgery rats (n=6) showed no such change. (C) Oxaliplatin treated rats (n=12) showed an decrease in withdrawal thresholds compared to dextrose treated controls (n=12) that was prevented in rats co-treated with minocycline (n=6). (D) Bortezomib treated rats (n=12) showed increases in the percent of persistent withdrawals during von Frey stimulation versus saline-treated rats (n=11) that were prevented in animals co-treated with minocycline (n=12)

Baseline von Frey filament thresholds were not significantly different between groups that were treated with oxaliplatin (20.5 ± 1.6g), dextrose (22.3 ± 1.5), or oxaliplatin + minocycline (20.5 ± 2.3g). Oxaliplatin-treated rats showed a significant decrease in threshold following treatment at day 14 (oxaliplatin = 12.0 ± 1.4g/ dextrose = 20.6 ± 2.8g) (Fig. 1B), but were not observed for recovery. As with bortezomib, co-treatment with minocycline in oxaliplatin-treated rats prevented this change (Boyette-Davis et al., 2011).

SNL rats showed a sharp decrease in response threshold in the paw ipsilateral to surgery on day 3, the first day of testing following surgery (6.0 ± 1.3g ipsilateral/ 16.0 ± 2.2g contralateral) (Fig. 1C). This increased sensitivity was maintained a week following surgery (3.5 ± 0.5g ipsilateral/ 17.0 ± 3.0g contralateral) until animals were sacrificed on day 14 (5.0 ± 1.0g ipsilateral/ 19.6 ± 2.9g contralateral). There was no change in sensitivity between ipsilateral or contralateral paws in sham surgery rats.

The percent of persistent withdrawal responses during von Frey testing was exaggerated by chemotherapy treatment, with bortezomib-treated rats showing a significantly higher percentage of exaggerated withdrawals at days 7 and 28 versus saline-treated controls (Figure 1D). This was prevented in bortezomib animals co-treated with minocycline, as this group did not differ significantly from saline-treated controls at any time point.

In summary, animals treated with bortezomib or oxaliplatin, as well as those with SNL surgeries, developed marked increases in sensitivity to von Frey filament stimulation as an assay of mechanical sensitivity. The bortezomib-treated rats were tested until the phenotype recovered to baseline, as this model has not been described in as much detail. Minocycline treatment in both bortezomib and oxaliplatin models fully prevented the onset of the mechanical sensitivity observed in those treated with chemotherapy alone.

Qualitative changes in glial morphology and distribution

Although measures used in the present study were limited to overall expression of glia, several qualitative changes in spinal glial morphology were also noted. Activated astrocytes and microglia showed increased surface marker expression, as well as an observed increase in density and number of processes and swollen cell bodies when viewed under high magnification (Figure 2). These changes were not restricted to the dorsal horn, where GFAP and OX-42 expression were quantified, but activation was distributed in an apparently homogenous manner throughout the ventral horn and other spinal cord gray matter, as well (Figures 3-5). The observed changes in morphology are notable as a trend that is not restricted to a single model, but as generalized trends in all models that showed increases in GFAP or OX-42 markers.

Figure 2.

Figure 2

Activated astrocytes and microglia are marked with greater arborization and hypertrophy versus inactive counterparts. Astrocytes shown are from day 30 saline treated (left) and bortezomib treated (right) dorsal horn. Microglia shown are from sham surgery (left) and SNL surgery (right) dorsal horn ipsilateral to surgery.

Figure 3.

Figure 3

GFAP staining intensity as represented in day 30 saline, bortezomib + minocycline, and bortezomib treated spinal cord tissue.

Figure 5.

Figure 5

OX-42 staining intensity is shown increased following spinal nerve ligation (bottom) but not 30 days following bortezomib treatment (top), or following sham surgery (middle). Inserts at show increased magnification with background subtracted to highlight microglial cell morphology in each condition.

Astrocyte expression following bortezomib treatment

Fluorescent intensity of GFAP immunohistochemistry in bortezomib-treated rats was found to be significantly higher versus saline-treated controls at multiple time points. Staining intensity for GFAP showed statistically significant increases at day 7 (111 ± 3.9% of control), day 14 (113 ± 5.1% of control), and day 30 (162 ± 11.5% of control) (Figure 3 and Figure 6A). GFAP staining intensity was not statistically significant in bortezomib at day 69 (94 ± 9.5% of control). Co-treatment of minocycline with bortezomib produced a significantly lower activation of astrocytes versus saline-treated controls at day 30 (76 ± 3.0% of control), showing a complete prevention of the astrocyte activation observed in rats treated with bortezomib alone. In short, bortezomib-treated animals showed higher GFAP staining intensity at days 7, 14, and 30, with peak intensity at day 30. This was fully prevented by the co-administration of minocycline as observed at day 30.

Figure 6.

Figure 6

Staining intensity in treated animals was expressed as a percentage of intensity versus respective controls. (A) GFAP staining in bortezomib treated rats was significantly higher at days 7 (n=6 vs. n=5), 14 (n=7 vs. n=5), and 30 (n=6 vs. n=6), but not at day 69 (n=3 vs. n=3). Co-treatment with minocycline resulted in siginificantly lower GFAP staining at day 30 (n=3). (B) GFAP staining in oxaliplatin treated rats was significantly higher at days 7 (n=9 vs. n=8) and 14 (n=9 vs. n=9), but not at day 40 (n=4 vs. n=4). Co-treatment with minocycline resulted in GFAP staining equivalent to control at day 7 (n=6) and 14 (n=6). (C) Both GFAP and OX-42 staining were higher in ipsilateral spinal cord in SNL rats (n=6) versus contralateral cord. Neither GFAP nor OX-42 were higher in ipsilateral cord in sham treated rats (n=6) versus contralateral cord. (D) OX-42 staining intensity in bortezomib treated rats was significantly lower versus control at days 7 (n=6 vs. n=5), 14 (n=7 vs. n=5), and 30 (n=8 vs. n=9).

Astrocyte expression following oxaliplatin treatment

Oxaliplatin-treated animals showed significantly higher GFAP fluorescence intensity versus controls. Staining intensity for GFAP showed statistically significant increases at day 7 (131 ± 9.4% of control) and day 14 (122 ± 4.7% of control) (Figure 4 and Figure 6B). Groups co-treated with minocycline showed an increase at day 7 (115 ± 13.5% of control) and decrease at day 14 (91 ± 20.2% of control), but neither of these findings were statistically significant. There was a decrease of GFAP intensity versus control at day 40 (87 ± 7.8% of control), but this was not found to be statistically significant. Thus, the activation of astrocytes observed in oxaliplatin-treated animals was prevented with co-treatment of minocycline to a great extent, if not fully. Because the increase in astrocyte activation from earlier time points is fully reversed by this time, and because other studies have shown that sensitivity to mechanical stimuli is abolished by this time, further time points were not examined.

Figure 4.

Figure 4

GFAP staining intensity as represented in day 7 dextrose, oxaliplatin + minocycline, and oxaliplatin treated spinal cord tissue.

Astrocyte expression following spinal nerve ligation

Astrocyte expression indicated by GFAP staining was significantly higher in SNL rats on the ipsilateral side of the spinal cord versus the contralateral side (126 ± 9.28%) (Figure 6C). Sham surgery rats showed no significant difference between halves of the spinal cord ipsilateral or contralateral to surgery (100 ± 10.3%). The overall finding was therefore in line with this model as a positive control, that there was a marked activation of astrocytes only in those animals that received the actual surgery, and this activation was significantly higher on the side ipsilateral to the ligation.

Microglial expression following spinal nerve ligation

Microglial expression of OX-42 was markedly higher in day 14 SNL rats ipsilateral to surgery versus contralateral spinal cord (130 ± 12.3%), whereas sham surgery rats showed statistically equivalent microglial expression on ipsilateral versus contralateral spinal cord (98 ± 10.4%) (Figure 5 and Figure 6C). This was therefore a trend in line with that of the SNL astrocyte data: the microglial activation occurred only in SNL rats and primarily on the side of the spinal cord ipsilateral to the ligation.

Microglial expression following bortezomib treatment

The staining intensity of OX-42 as a marker for microglial activation was slightly, though significantly, lower in bortezomib versus control at day 7 (95 ± 5.2%), 14 (96 ± 8.6%), and 30 (93 ± 4.6%) (Figure 6D). The finding that bortezomib does not activate microglia at any of the time points observed is in line with other models of chemotherapy treatment.

Discussion

The results of this study show an upregulation of GFAP staining intensity in oxaliplatin and bortezomib-treated rats that is interpreted as an increase in the activity of astrocytes. What role astrocytes may play in CIPN or other models of chronic pain is not fully understood at this time. However, one recent study reported a lack of microglial activation in a bone cancer model, concluding that microglial activation is a possible hallmark of neuropathic pain, rather than pain as a whole (Ducourneau et al. 2014). The implication of this is that whatever activity is modulated by reactive microglia may be critical to how neuropathic pain operates. The capability of astrocytes to produce proinflammatory cytokines suggests one possible explanation for astrocyte involvement (Gao and Ji, 2010; Hald, 2009). Proinflammatory cytokines in inappropriate levels are known to have a negative impact on neurons that could explain hypersensitive responses seen in CIPN (Miller et al., 2009).This possibility is further supported by the prevention of the observed astrogliosis by co-administration of minocycline, which has anti-inflammatory properties and has been shown to prevent the upregulation of proinflammatory cytokines in other models (Ledeboer et al., 2005; Mao-Ying et al., 2012). It is worth mentioning that some studies claim that minocycline is specific to microglia and has no effect on astrocyte activation (Yrjänheikki et al., 1999), whereas others show inhibition of both astrocytic and microglial activation (Ledeboer et al., 2005; Ryu et al., 2004; Sung et al., 2012; Teng et al., 2004). The findings of the current study are in support of the latter, which may be a model-dependent finding. In the present study, when minocycline was administered in addition to bortezomib, both behavioral responses to mechanical stimuli and intensity of GFAP staining were abrogated completely so that these rats were closely in line with control rats. Minocycline may act in this context as a general inhibitor of proinflammatory cytokine activity, which is characteristic of both astrocytic and microglial activation. Like other tetracyclines, minocycline is frequently evaluated as an antibiotic, but the primary clinical use of minocycline since its FDA approval has been in the treatment of inflammatory acne vulgaris (Garner et al., 2012). Cytokines of interest for future studies in CIPN may include IL-1β and TNF-α (Sung et al., 2012; Mao-Ying et al., 2012) among others, although these are not specific to astrocytes.

Explaining an astrocyte-specific activation is challenging, but is a finding that suggests a few possibilities for future study. The first of these is that upstream effectors such as damage-associated molecular pattern (DAMP) or pathogen-associated molecular pattern (PAMP) proteins bind to a receptor found on astrocytes, but not microglia. The second possibility for future investigation is that the propagation of glial activation in CIPN models is dependent on intercellular signaling via gap junctions, which would explain a lack of microglial involvement. Our lab has previously shown an upregulation of connexin 43 in support of this hypothes is in oxaliplatin (Yoon et al., 2013), but this has not yet been examined in other treatment models.

Following activation, astrocytes are capable of directly impacting synaptic transmission by a maladaptive down regulation of glutamate transporters such as GLT-1 and GLAST (Liaw et al., 2005; Cata et al., 2006). Down regulation of both GLT-1 and GLAST has been observed previously in paclitaxel-treated rats as well as in injury models (Sung, Lim, and Mao, 2003; Weng et al., 2005; Cata et al., 2006; Xin, Weng, and Dougherty, 2009; Zhang, Xin, and Dougherty, 2009) and could explain a variety of persistent or enhanced sensations reported by patients with CIPN (Cata et al., 2007; Dougherty et al., 2007; Boyette-Davis et al., 2013). In support of the possibility of glutamate transporter involvement in oxaliplatin and bortezomib, rats used in the present study have been observed to respond to von Frey stimulation with licking or shaking of the paw that suggests persistent sensation after the stimulus is removed (Fig. 1B). The data in the present study suggests at least a correlation in bortezomib between level of astrocyte activation and degree of sensitivity to mechanical stimuli, as the presence, intensity, and extinction of GFAP up regulation follows closely with the degree of mechanical hypersensitivity. Whether these increases in sensitivity to stimuli are also found in other modalities, such as thermal or cold sensation are yet to be defined, Similarly, other chemotherapeutics been have shown to reduce peripheral epidermal nerve density and to produce increased expression of numerous stress-related proteins in dorsal root ganglion cells and each of these measures were protected by co-treatment with minocycline. Finally, spinal neuron sensitization and changes in spinal expression of numerous molecules including glutamate transporters and gap junction proteins expression were altered by chemotherapy that was also prevented by co-treatment with minocycline. Thus, there numerous important follow-up studies. Yet, importantly as astrocyte activation was reduced in minocycline treated rats measures of hyper-responsiveness to mechanical stimulation also showed decreases that provides an internal validation of the behavioral measures presented here.

Oxaliplatin-treated rats showed a somewhat different involvement of astrocytes that did not correlate with behavior in the same manner as bortezomib. Oxaliplatin treatment produced an initial up regulation of GFAP that decreased over time after drug treatment ceased. While astrocytes may play a role both in the induction and maintenance of bortezomib CIPN, these data would suggest that astrocyte activity is only involved in the induction of oxaliplatin CIPN. Once again, co-treatment with minocycline showed an abrogation of astrocyte activity in all time points that chemotherapy treatment without minocycline showed an increase in GFAP intensity.

The lack of OX-42 up regulation at any time point in bortezomib was interpreted as a lack of microglial activation in line with other models of chemotherapy, as the SNL model produced a robust activation in microglia. However, some existing theories on glial interactions suggest that activation of astrocytes occurs in response to signals from already activated microglia (Zhang et al., 2010). A recent report showed activation of microglia with the Iba1 marker at day 7 in oxaliplatin (Di Cesare Mannelli et al., 2013) yet, others like the data shown here did not find microglial activation in paclitaxel-related CIPN (Zheng, Xiao, and Bennett, 2011). One possible explanation for the differences seen here and the other report where microglia were activated during oxaliplatin treatment may be that the cumulative dose of oxaliplatin used in the aforementioned study was much higher than that used in the present study. This is the second study from our lab that has demonstrated a model in which there does not appear to be microglial activation in a CIPN model (Zhang et al., 2012). It is very interesting that this finding in CIPN is similar to that observed in cancer pain models whole (Ducourneau et al. 2014) suggesting common underlying signaling mechanisms to spinal glial cells may be engaged in spite of the markedly diverse model systems.

It is also possible that other markers of microglial activity could show differing activation. Although OX-42 is a widely-accepted marker for proliferative microglia, microglial activation may not be exhaustively demonstrated by OX-42. Apart from proliferative characteristics, microglia may also show a shift in polarity from the neuroprotective M2 phenotype to the pro-inflammatory M1 phenotype (Guerrero et al., 2012; Hu et al., 2012; Boche, Perry, and Nicoll, 2013). It could be possible that examining proliferation measured by OX-42 overlooks a change in the proportion of M1 and M2 phenotypes that could contribute to inflammation-mediated damage to neurons. However, this possibility assumes that microglia would enter into an active state without proliferating, which could be problematic in the context of our current understanding of microglial response to CNS insults.

Conclusion

The lack of an observed upregulation of microglia suggests a lack of immune response in the development and maintenance of CIPN. The activation of astrocytes instead could mean many things. As previously mentioned, one possibility is that astrocytes modulate an inflammatory response through the release of proinflammatory cytokines. However, microglia are classically thought to be involved in the production of cytokines involved in such a response (Ji and Suter, 2007; Zhuang et al., 2007). This is problematic in the context of astrocyte activation unless mechanisms specific to astrocytes but not microglia are identified. Future studies should be aimed at identifying what these may be. It is also possible that astrocytes are involved in sensitizing synapses in pain pathways through their activities in glutamate transport (Zhang, Xin, and Dougherty, 2009). Studies using pharmacological blockers of astrocyte activation should also be conducted to verify a direct relationship between astrocytes and behavioral phenotypes. However, it is clear from the present data that there is a correlation between the two.

Highlights.

  • Chemotherapy causes mechanical hypersensitivity and activated astrocytes

  • Microglia were not activated by chemotherapy but were by spinal nerve ligation

  • Astrocyte activation and mechanical sensitivity were prevented with minocycline

Acknowledgments

This work was supported by NIH grant NS046606 and NCI grant CA124787.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Araque A, Navarrete A. Glial cells in neuronal network function. Philos Trans R Soc Lond B Biol Sci. 2010;365:2375–2381. doi: 10.1098/rstb.2009.0313. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Boche D, Perry VH, Nicoll JA. Review: activation patterns of microglia and their identification in the human brain. Neuropathol Appl Neurobiol. 2013;39:3–18. doi: 10.1111/nan.12011. [DOI] [PubMed] [Google Scholar]
  3. Boyette-Davis J, Dougherty PM. Protection against oxaliplatin-induced mechanical hyperalgesia and intraepidermal nerve fiber loss by minocycline. Exp Neurol. 2011;229:353–357. doi: 10.1016/j.expneurol.2011.02.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Boyette-Davis J, Xin W, Zhang H, Dougherty PM. Intraepidermal nerve fiber loss corresponds to the development of taxol-induced hyperalgesia and can be prevented by treatment with minocycline. Pain. 2011;152:308–313. doi: 10.1016/j.pain.2010.10.030. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Boyette-Davis JA, Cata JP, Driver LC, Novy DM, Bruel BM, Mooring DL, Wendelschafer-Crabb G, Kennedy WR, Dougherty PM. Persistent chemoneuropathy in patients receiving the plant alkaloids paclitaxel and vincristine. Cancer Chemother Pharmacol. 2013;71:619–626. doi: 10.1007/s00280-012-2047-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bradesi S. Role of spinal cord glia in the central processing of peripheral pain perception. Neurogastroenterol Motil. 2010;22:499–511. doi: 10.1111/j.1365-2982.2010.01491.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Cata JP, Weng HR, Lee BN, Reuben JM, Dougherty PM. Clinical and experimental findings in humans and animals with chemotherapy-induced peripheral neuropathy. Minerva Anestesiol. 2006;72:151–169. [PubMed] [Google Scholar]
  8. Cata JP, Weng HR, Chen JH, Dougherty PM. Altered discharges of spinal wide dynamic range neurons and down-regulation of glutamate transporter expression in rats with paclitaxel-induced hyperalgesia. Neuroscience. 2006;138:329–338. doi: 10.1016/j.neuroscience.2005.11.009. [DOI] [PubMed] [Google Scholar]
  9. Cata JP, Weng HR, Burton AW, Villareal H, Giralt S, Dougherty PM. Quantitative sensory findings in patients with bortezomib-induced pain. J Pain. 2007;8:296–306. doi: 10.1016/j.jpain.2006.09.014. [DOI] [PubMed] [Google Scholar]
  10. Cata JP, Weng HR, Dougherty PM. The effects of thalidomide and minocycline on taxol-induced hyperalgesia in rats. Brain Res. 2008;1229:100–110. doi: 10.1016/j.brainres.2008.07.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Cavaletti G, Marmiroli P. Chemotherapy-induced peripheral neurotoxicity. Nat Rev Neurol. 2010;6:657–666. doi: 10.1038/nrneurol.2010.160. [DOI] [PubMed] [Google Scholar]
  12. Di Cesare Mannelli L, Pacini A, Bonaccini L, Zanardelli M, Mello T, Ghelardini C. Morphologic features and glial activation in rat oxaliplatin-dependent neuropathic pain. J Pain. 2013;14:1585–1600. doi: 10.1016/j.jpain.2013.08.002. [DOI] [PubMed] [Google Scholar]
  13. Dougherty PM, Cata JP, Burton AW, Vu K, Weng HR. Dysfunction in multiple primary afferent fiber subtypes revealed by quantitative sensory testing in patients with chronic vincristine-induced pain. J Pain Symptom Manage. 2007;33:166–179. doi: 10.1016/j.jpainsymman.2006.08.006. [DOI] [PubMed] [Google Scholar]
  14. Ducourneau VR, Dolique T, Hachem-Delaunay S, Miraucourt LS, Amadio A, Blaszczyk L, Jacquot F, Ly J, Devoize L, Oliet SH, Dallel R, Mothet JP, Nagy F, Fenelon VS, Voisin DL. Cancer pain is not necessarily correlated with spinal overexpression of reactive glia markers. Pain. 2014;155:275–291. doi: 10.1016/j.pain.2013.10.008. [DOI] [PubMed] [Google Scholar]
  15. Gao YJ, Ji RR. Targeting astrocyte signaling for chronic pain. Neurotherapeutics. 2010;7:482–493. doi: 10.1016/j.nurt.2010.05.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Garner SE, Eady A, Bennett C, Newton JN, Thomas K, Popescu CM. Minocycline for acne vulgaris: efficacy and safety. Cochrane Database Syst Rev 8. 2012 doi: 10.1002/14651858.CD002086.pub2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Garrison CJ, Dougherty PM, Kajander KC, Carlton SM. Staining of glial fibrillary acidic protein (GFAP) in lumbar spinal cord increases following a sciatic nerve constriction injury. Brain Res. 1991;565:1–7. doi: 10.1016/0006-8993(91)91729-k. [DOI] [PubMed] [Google Scholar]
  18. Graeber MB. Changing face of microglia. Science. 2010;330:783–788. doi: 10.1126/science.1190929. [DOI] [PubMed] [Google Scholar]
  19. Guasti L, Richardson D, Jhaveri M, Eldeeb K, Barrett D, Elphick MR, Alexander SP, Kendall D, Michael GJ, Chapman V. Minocycline treatment inhibits microglial activation and alters spinal levels of endocannabinoids in a rat model of neuropathic pain. Mol Pain. 2009;5:35. doi: 10.1186/1744-8069-5-35. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Guerrero AR, Uchida K, Nakajima H, Watanabe S, Nakamura M, Johnson WE, Baba H. Blockade of interleukin-6 signaling inhibits the classic pathway and promotes an alternative pathway of macrophage activation after spinal cord injury in mice. J Neuroinflammation. 2012;9:40. doi: 10.1186/1742-2094-9-40. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Hald A. Spinal astrogliosis in pain models: cause and effects. Cell Mol Neurobiol. 2009;29:609–619. doi: 10.1007/s10571-009-9390-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Hu X, Li P, Guo Y, Wang H, Leak RK, Chen S, Gao Y, Chen J. Microglia/macrophage polarization dynamics reveal novel mechanism of injury expansion after focal cerebral ischemia. Stroke. 2012;43:3063–3070. doi: 10.1161/STROKEAHA.112.659656. [DOI] [PubMed] [Google Scholar]
  23. Hua XY, Svensson CI, Matsui T, Fitzsimmons B, Yaksh TL, Webb M. Intrathecal minocycline attenuates peripheral inflammation-induced hyperalgesia by inhibiting p38 MAPK in spinal microglia. Eur J Neurosci. 2005;22:2431–2440. doi: 10.1111/j.1460-9568.2005.04451.x. [DOI] [PubMed] [Google Scholar]
  24. Ji RR, Suter MR. P38 MAPK, microglial signaling, and neuropathic pain. Mol Pain. 2007;3:33. doi: 10.1186/1744-8069-3-33. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kim SH, Chung JM. An experimental model for peripheral neuropathy produced by segmental spinal nerve ligation in the rat. Pain. 1992;50:355–363. doi: 10.1016/0304-3959(92)90041-9. [DOI] [PubMed] [Google Scholar]
  26. Ledeboer A, Sloane EM, Milligan ED, Frank MG, Mahony JH, Maier SF, Watkins LR. Minocycline attenuates mechanical allodynia and proinflammatory cytokine expression in rat models of pain facilitation. Pain. 2005;115:71–83. doi: 10.1016/j.pain.2005.02.009. [DOI] [PubMed] [Google Scholar]
  27. Liaw WJ, Stephens RL, Jr, Binns BC, Chu Y, Sepkuty JP, Johns RA, Rothstein JD, Tao YX. Spinal glutamate uptake is critical for maintaining normal sensory transmission in rat spinal cord. Pain. 2005;115:60–70. doi: 10.1016/j.pain.2005.02.006. [DOI] [PubMed] [Google Scholar]
  28. Mao-Ying QL, Wang XW, Yang CJ, Li X, Mi WL, Wu GC, Wang YQ. Robust spinal neuroinflammation mediates mechanical allodynia in Walker 256 induced bone cancer rats. Mol Brain. 2012;5:16. doi: 10.1186/1756-6606-5-16. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Miller RJ, Jung H, Bhangoo SK, White FA. Cytokine and chemokine regulation of sensory neuron function. Handb Exp Pharmacol. 2009:419–449. doi: 10.1007/978-3-540-79090-7_12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Nie H, Zhang H, Weng HR. Bidrectional neuron-glia interactions triggered by deficiency of glutamate uptake at spinal sensory synapses. J Neurophysiol. 2010;104:713–725. doi: 10.1152/jn.00282.2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Ryu JK, Franciosi S, Sattayaprasert P, Kim SU, McLarnon JG. Minocycline inhibits neuronal death and glial activation induced by β-amyloid peptide in rat hippocampus. Glia. 2004;48:85–90. doi: 10.1002/glia.20051. [DOI] [PubMed] [Google Scholar]
  32. Sivilotti L, Woolf CJ. The contribution of GABAA and glycine receptors to central sensitization: disinhibition and touch-evoked allodynia in the spinal cord. J Neurophysiol. 1994;72:169–179. doi: 10.1152/jn.1994.72.1.169. [DOI] [PubMed] [Google Scholar]
  33. Sung B, Lim G, Mao J. Altered expression and uptake activity of spinal glutamate transporters after nerve injury contribute to the pathogenesis of neuropathic pain rats. J Neurosci. 2003;23:2899–2910. doi: 10.1523/JNEUROSCI.23-07-02899.2003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Sung CS, Cherng CH, Wen ZH, Chang WK, Huang SY, Lin SL, Chan KH, Wong CS. Minocycline and fluorocitrate suppress spinal nociceptive signaling in intrathecal IL-1β-induced thermal hyperalgesic rats. Glia. 2012;60:2004–2017. doi: 10.1002/glia.22415. [DOI] [PubMed] [Google Scholar]
  35. Teng YD, Choi H, Onario RC, Zhu S, Desilets FC, Lan S, Woodard EJ, Snyder EY, Eichler ME, Friedlander RM. Minocycline inhibits contusion-triggered mitochondrial cytochrome c release and mitigates functional deficits after spinal cord injury. Proc Natl Acad Sci USA. 2004;101:3071–3076. doi: 10.1073/pnas.0306239101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Weng HR, Aravindan N, Cata JP, Chen JH, Shaw AD, Dougherty PM. Spinal glial glutamate transporters downregulate in rats with taxol-induced hyperalgesia. Neurosci Lett. 2005;386:18–22. doi: 10.1016/j.neulet.2005.05.049. [DOI] [PubMed] [Google Scholar]
  37. Xin WJ, Weng HR, Dougherty PM. Plasticity in expression of the glutamate transporters GLT-1 and GLAST in spinal dorsal horn glial cells following partial sciatic nerve ligation. Mol Pain. 2009;5:15. doi: 10.1186/1744-8069-5-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Yoon SY, Robinson CR, Zhang H, Dougherty PM. Spinal astrocyte gap junctions contribute to oxaliplatin-induced mechanical hypersensitivity. J Pain. 2013;14:205–214. doi: 10.1016/j.jpain.2012.11.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Yrjänheikki J, Tikka T, Keinänen R, Goldsteins G, Chan PH, Koistinaho J. A tetracycline derivative, minocycline, reduces inflammation and protects against focal cerebral ischemia with a wide therapeutic window. Proc Natl Acad Sci USA. 1999;96:13496–13500. doi: 10.1073/pnas.96.23.13496. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Zhang D, Hu X, Qian L, O'Callaghan JP, Hong JS. Astrogliosis in CNS pathologies: is there a role for microglia? Mol Neurobiol. 2010;41:232–241. doi: 10.1007/s12035-010-8098-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Zhang H, Xin W, Dougherty PM. Synaptically evoked glutamate transporter currents in Spinal Dorsal Horn Astrocytes. Mol Pain. 2009;5:36. doi: 10.1186/1744-8069-5-36. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Zhang H, Yoon SY, Zhang H, Dougherty PM. Evidence that spinal astrocytes but not microglia contribute to the pathogenesis of Paclitaxel-induced painful neuropathy. J Pain. 2012;13:293–303. doi: 10.1016/j.jpain.2011.12.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Zheng FY, Xiao WH, Bennett GJ. The response of spinal microglia to chemotherapy-evoked painful peripheral neuropathies is distinct from that evoked by traumatic nerve injuries. Neuroscience. 2011;176:447–454. doi: 10.1016/j.neuroscience.2010.12.052. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Zhuang ZY, Kawasaki Y, Tan PH, Wen YR, Huang J, Ji RR. Role of the CX3CR1/p38 MAPK pathway in spinal microglia for the development of neuropathic pain following nerve injury-induced cleavage of fractalkine. Brain Behav Immun. 2007;21:642–651. doi: 10.1016/j.bbi.2006.11.003. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES