Prevention of Paclitaxel-induced allodynia by Minocycline: Effect on loss of peripheral nerve fibers and infiltration of macrophages in rats

The administration of paclitaxel at a cumulative dose of 24 mg/kg (3 × 8 mg/kg, 3 days apart, i.p) caused a marked and prolonged mechanical allodynia as evidenced by 50% withdrawal threshold compared with day 0 (P < 0.05). On day 4 and 12 following initial paclitaxel treatment, the 50% withdrawal threshold significantly reduced to 8.87 ± 1.33 g and 3.03 ± 1.92 g respectively compared with the value (17.28 ± 1.85 g) on day 0 (Figure 1). Application of minocycline at a daily dosage of 30 mg/kg or 50 mg/kg, but not 5 mg/kg, initiated one day before paclitaxel and continued for 8 days significantly attenuated mechanical allodynia on day 4, 8 (P < 0.05) and 12 (P < 0.01) compared with the paclitaxel group. Furthermore, 50% withdrawal threshold had no significantly difference between the 30 mg/kg minocycline/paclitaxel group and the 50 mg/kg minocycline/paclitaxel group or the vehicle group. Continuous injection of minocycline at a dose of 30 mg/kg alone had no effect on mechanical hypersensivity in rats (Figure 1), so the dose of 30 mg/kg of minocycline was applied in subsequent experiments.

Consistent with previous studies [5], PGP9.5-labeled IENF emerged from cutaneous nerves and traveled vertically into the epidermis where they branched into terminal (Figure 2A). Following application of paclitaxel, there was a gradual decrease in the number of IENF (Figure 2H). The number of IENF per sight decreased from 12.71 ± 1 in vehicle rats to 11.92 ± 0.88, 6.2 ± 0.86, 4.5 ± 0.82 and 2.7 ± 0.53 on day 2, 4, 8 and 12 after paclitaxel treatment, respectively (Figure 2B-E). However, application of minocycline significantly inhibited the loss of IENF induced by paclitaxel on day 4 (10.69 ± 0.64) and 12 (8.58 ± 0.66) compared with that of paclitaxel group respectively (P < 0.01) (Figure 2F and 2G). Vehicle or minocycline alone did not affect the density of IENF compared with normal rats (data not shown).

To assess whether the loss of IENF is accompanied by the sensory cell injury. We examined the levels of ATF3 in DRG at various time points. The result showed that paclitaxel treatment significantly increased the number of ATF3-immunoreactivity (IR) positive cells in L4 DRG (Figure 3B-E and 3H) compared with vehicle-treated rats in that ATF3-IR positive cells were hardly found (Figure 3A). Meanwhile, double immunofluorescence staining showed that ATF3 was co-localized with NF200-labeled cells (A fiber neuronal marker), but not with IB4-labeled cells (C fiber neuronal marker) or GFAP-labeled cells (satellite cell marker) (Figure 4). Furthermore, pretreatment with minocycline inhibited the increase of ATF3 (0.44 ± 0.18 and 1.33 ± 0.50 in minocycline/paclitaxel group versus 12.44 ± 1.68 and 24.33 ± 2.12 in paclitaxel group on day 4 and 12, respectively) (Figure 3F and 3G).

Few macrophages stained by ED1 were detected in L4 DRG of vehicle treated rats (Figure 5A). However, following paclitaxel treatment, a significant increase in the number of macrophages was observed in DRG on day 2 (16.83 ± 1.81), 4 (37.33 ± 1.30), 8 (50.66 ± 2.66) and 12 (77.66 ± 1.62) (Figure 5B-E). Furthermore, the infiltration of macrophages induced by paclitaxel was significantly inhibited by minocycline treatment on day 4 (Figure 5F) and 12 (Figure 5G). Because recent study showed that the infiltration of macrophages in the spinal cord played a vital role in neuropathic pain [14], we also examined whether paclitaxel could induce the macrophage to infiltrate into the spinal cord. However, in our experiment, no obvious ED1-positive cells were detected in the spinal dorsal horn in either paclitaxel group or vehicle group (data not shown).

It is well established that paclitaxel treatment could elicit peripheral sensory neuropathy. Degeneration of nervous fibers has been currently suggested as the possible mechanism underlying the paclitaxel induced mechanical allodynia. It has been reported that application of low-dose paclitaxel (2 mg/kg) induced the loss of IENF [5, 15]. In vitro study also showed that paclitaxel directly applied to the axonal resulted in degeneration of axons [16]. In our present study, moderate-dose paclitaxel (8 mg/kg) also significantly decreased the number of IENF. There are several explanations to such degeneration of fibers. For example, Nogales et al indicated that paclitaxel impaired axoplasmic transport by binding toβ-tubulin which has been thought as the cause resulting in degeneration of IENF [17]. While evidence against this hypothesis implied that paclitaxel directly impaired mitochondria function which might lead to degeneration of the fiber terminals [2].

We also found that paclitaxel induced the expression of ATF3 in DRG. It has been shown that ATF3 might have a survival/regenerative function in sensory neurons [18]. Evidence has shown that lack of target derived growth factors secondary to nerve injury resulted in the ATF3 up-regulation [19, 20]. Therefore, in our study, it appeared that the decreased availability of target-derived growth factors due to the degeneration of IENF following paclitaxel administration induced expression of ATF3. In addition, double-staining showed that expression of ATF3 mainly focused on the NF200-positive cells, the result indicated that paclitaxel mainly induced the injury of Aβ-fiber neurons and is consistent with Dougherty's report that paclitaxel treatment in cancer patients impairs the Aβ fiber function [21].

In the present study, we also observed marked hyperplasia of macrophage in the DRG following application of paclitaxel. The result is consistent with Peter's report that intravenous infusion of high-dose paclitaxel induced hypertrophy and hyperplasia of macrophage in DRG and sciatic nerve [22]. The increased macrophages observed in the current study may be due to infiltration of macrophages into the DRG. This hypothesis is supported by our latest observation that application of moderate-dose paclitaxel elevated the level of chemotatic factor in DRG (unpublished data). Functionally, the activated macrophage may help remove degeneration neuronal debris and myelin following the peripheral nerve injury [23]; on the other hand, it may also contribute to the pathological pain through the release of proinflammatory cytokines which is capable of sensitizing primary afferent neurons [24, 25].

Cata's study showed that minocycline, an inhibitor of microglia/macrophage activation, ameliorated taxol-induced hyperalgesia. It has been hypothesized that the immunomodulatory activity of minocycline underlies its protective effect on taxol-induced neuropathic pain. However, the exact mechanism is still unclear. In our present study, minocycline attenuated the loss of IENF, which was parallel with the reduced allodynia. It has been shown that minocycline decreased recruitment and activation of macrophage thereby slowing Wallerian degeneration [26]. In addition, minocycline treatment reduces oligodendrocyte death and attenuates axonal dieback after spinal cord injury [13]. Moreover, mitochondrial impairment, which has been suggested to contribute to degeneration of nerve fibers, could be prevented by minocycline. Therefore, it is possible that, by protecting the integrity of IENF, minocycline attenuated the loss of IENF induced by paclitaxel. This hypothesis was also supported by our present finding that minocycline decreased ATF3 up-regulation in DRG neurons.

Furthermore, we observed that paclitaxel-induced macrophages infiltration into DRG was obviously prevented in minocycline treated rats. Several lines of evidence proved that minocycline could inhibit the activation and migration of macrophages and reduce production of macrophage proinflammatory factors [27, 28] which mediated peripheral nerve degeneration [28]. Furthermore, inhibition of macrophage responses might prevent nerve fiber degeneration by prohibiting the phagocytosis of axon ends [13, 28]. Although in the present study, activation of macrophages around the peripheral nerve fibers was not examined, its destructive effect on the IENF could not be excluded. Therefore, inhibition of macrophage responses might contribute to minocycline preventive effect on IENF loss induced by paclitaxel. However, additional studies are needed to elucidate the role of minocycline in protecting IENF from paclitaxel-induced injury.

Male Sprague-Dawley rats weighting 220-280 g were purchased from the laboratory animal center of Sun Yat-Sen University, the permit number for this study is SCXK(Guangdong)2009-0011. All animal experimental procedures were approved by the Sun Yat-Sen University Animal Care and Use Committee and were carried out in accordance with the guideline of the National Institutes of Health on animal care and the ethical guidelines. The minimum number of animals was used in each experiment, and in all cases every effort was made to minimize any pain and suffering in the subject animals.

Paclitaxel (Taxol, Bristol-Myers Squibb, 6 mg/ml) was diluted with saline (1:3) and injected i.p (cumulative dose of 24 mg/kg) on 3 alternate days (days 1, 4 and 7)[29]. Minocycline (Sigma) was administered (i.p) daily at a dosage of 30 mg/kg and begin prior one day to the paclitaxel [30]. Rats received minocycline always 30 min before application of paclitaxel and continued up to eight days. Minocycline was dissolved in saline and stored in solution at 200 mg/ml and then diluted in saline before the administration.

The rats were accommodated to the testing environment by placement within testing chambers for 15-20 minutes on the three separate days just prior to drug administration. Mechanical sensitivity was assessed using von Frey hairs as described previously [12]. Briefly, rats were placed under three different transparent Plexiglas chambers positioned on a wire mesh floor. Fifteen minutes were allowed for habituation. Each stimulus consisted of a 2-3s application of the von Frey hair to the middle of plantar surface of the foot with 5 min interval between stimuli. Brisk withdrawal or licking of the paw following the stimulus was considered a positive response. The experimenter who conducted the behavioral tests was blinded to all treatments.

All rats used in the immunohistochemistry experiments had confirmed to have a characteristic of behavior. Rats were deeply anesthetized with urethane (1.5 mg/kg, i.p.) at different time points, the chest was opened, and then quickly perfused through the ascending aorta with a warm heparinized saline, followed by 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.2-7.4, 4°C. The glabrous skin of hind paw and the L4 DRG was excised, post-fixed overnight and cryoprotected for 24 h in 30% sucrose in PB. Cryostat sections (16 μm) were cut and processed for immunohistochemical staining as previously described [31, 32]. Sections were blocked with 3% donkey serum in 0.3% Triton X-100 for 1 hour at the room temperature, and then incubated overnight at 4°C with rabbit anti-protein gene product 9.5 primary antibody (PGP9.5, 1:2000, Chemicon) for skin or rabbit anti-ATF3 antibody (1:200, Santa Cruz) or mouse anti-Macrophage antibody (ED-1, 1:200, Chemicon) for DRG. For double staining, the sections were incubated with rabbit anti-ATF3 antibody (1:200, Santa Cruz) and mouse anti-neurofilament-200, an A-fiber neuronal marker (NF200, 1:300, Chemicon), mouse anti-isolectin B4, a C-fiber neuronal marker (IB4, 1:200, Sigma) or glial fibrillary acidic protein, a satellite cell marker (mouse anti-GFAP, 1:2000, Chemicon). After rinsing three times with PBS, sections were incubated in donkey anti-rabbit IgG secondary antibody labeled with Cy3 (1:500, Jackson) or a mixture of IgG secondary antibody labeled with Cy3 and FITC respectively (1:500, Jackson) for 1 h at a room temperature. Five rats were included for each group for immunohistochemistry quantification. Three DRG tissue sections per animal are randomly selected, the number of ATF3 or ED-1 positive cells was examined with a Leica (Leica, Germany) fluorescence microscope and images were captured with a Leica DFC350 FX camera. For IENF quantification, we selected five plantar skin sections per animal and chose three sights for each section randomly. Images of immunohistochemical results were obtained using an Zeiss LSM710 confocal microscope and analyzed with a Bitplane Imaris V6.4. All ascending nerve fibers that were seen to cross into the epidermis were counted, no minimum length was required and fibers that branched within the epidermis were counted as one. The number of IENF per sight was counted. To confirm the specificity of the primary antibody, control sections were incubated without primary antiserum.

Blinded evaluator analyzed all images. The number of the fibers, ATF3 or ED1 positive cells every sight was expressed as mean ± SEM. Data were compared with student's t-test, P < 0.05 was considered significant. For behavioral experiments, one-way ANOVA followed by post hoc test was used and P < 0.05 was considered significant.