Studies of peripheral sensory nerves in paclitaxel-induced painful peripheral neuropathy: evidence for mitochondrial dysfunction

Abstract

Paclitaxel chemotherapy frequently induces neuropathic pain during and often persisting after therapy. The mechanisms responsible for this pain are unknown. Using a rat model of paclitaxel-induced painful peripheral neuropathy, we have performed studies to search for peripheral nerve pathology. Paclitaxel-induced mechano-allodynia and mechano-hyperalgesia were evident after a short delay, peaked at day 27 and finally resolved on day 155. Paclitaxel- and vehicle-treated rats were perfused on days 7, 27 and 160. Portions of saphenous nerves were processed for electron microscopy. There was no evidence of paclitaxel-induced degeneration or regeneration as myelin structure was normal and the number/density of myelinated axons and C-fibres was unaltered by paclitaxel treatment at any time point. In addition, the prevalence of ATF3-positive dorsal root ganglia cells was normal in paclitaxel-treated animals. With one exception, at day 160 in myelinated axons, total microtubule densities were also unaffected by paclitaxel both in C-fibres and myelinated axons. C-fibres were significantly swollen following paclitaxel at days 7 and 27 compared to vehicle. The most striking finding was significant increases in the prevalence of atypical (swollen and vacuolated) mitochondria in both C-fibres (1.6- to 2.3-fold) and myelinated axons (2.4- to 2.6-fold) of paclitaxel-treated nerves at days 7 and 27. Comparable to the pain behaviour, these mitochondrial changes had resolved by day 160. Our data do not support a causal role for axonal degeneration or dysfunction of axonal microtubules in paclitaxel-induced pain. Instead, our data suggest that a paclitaxel-induced abnormality in axonal mitochondria of sensory nerves contributes to paclitaxel-induced pain.

Fig. 1

Behavioural time-course of mechano-allodynia and mechano-hyperalgesia induced by paclitaxel treatment. Graph shows the mean ± SEM of the response frequency to mechanical stimulation by (A) von Frey 4 g, (B) von Frey 8 g and (C) von Frey 15 g. n = 8–14 for vehicle treatment, n = 10–17 for paclitaxel treatment. *p < 0.05, ^#p < 0.01, one-way repeated-measures ANOVA with Dunnett’s post hoc analysis compared to baseline (BL).

Fig. 2

Saphenous nerve sections from (A) vehicle-treated and (B) paclitaxel-treated rats at day 27 post-treatment initiation, i.e., at the time of peak of paclitaxel-induced pain behaviour. Examination at this magnification (725×) shows that the overall morphology of peripheral sensory nerves is unaltered by paclitaxel treatment. Note: the fine black lines seen in (A) are small folds in the section. Graphs show the mean ± SEM of the number of (C) myelinated axons and (D) C-fibres in complete saphenous nerve sections of vehicle-treated and paclitaxel-treated rats, n = 3 per group (C) and n = 2 per group (D). There is no significant difference in the number of myelinated axons or C-fibres following paclitaxel treatment, at any time point, compared to vehicle treatment (two-tailed unpaired t-tests, Welch correction applied as appropriate).

Fig. 3

ATF3-immuno-reactivity in nuclei of L4/5 DRG cells. (A) Three days following complete sciatic nerve transection, the majority of cells are ATF3-positive. (B) Day 27 post-paclitaxel initiation, virtually no ATF3-positive cells. Micrographs taken at 10× magnification. Scale bar, 100 μm.

Fig. 4

Effect of paclitaxel on cross-sectional area of C-fibres. Graph shows the mean ± SEM of the C-fibre area in vehicle-treated and paclitaxel-treated nerves at days 7, 27 and 160 post-initiation of treatment. At each time point, 120 C-fibres were measured from two vehicle-treated rats and 180 C-fibres from three paclitaxel-treated rats. Vehicle n = 120, paclitaxel n = 180: ^#p < 0.01, two-tailed unpaired t-tests, Welch correction applied as appropriate.

Fig. 5

Effect of paclitaxel on total microtubule densities in myelinated axons and C-fibres. Graphs show the mean ± SEM of the total microtubule density in (A) myelinated axons and (B) C-fibres of vehicle-treated and paclitaxel-treated nerves at days 7, 27 and 160 postinitiation of treatment. At each time point, microtubules were counted in 120 myelinated axons/C-fibres randomly sampled from two vehicle-treated rats and 180 myelinated axons/C-fibres randomly sampled from three paclitaxel-treated rats. Vehicle n = 120, paclitaxel n = 180: ^#p < 0.01, two-tailed unpaired t-tests.

Fig. 6

Circular and oblique microtubule profiles in a myelinated axon. Axoplasm of a myelinated axon in the saphenous nerve from a paclitaxel-treated rat at day 27 post-treatment initiation (44,400×). Circular (arrowheads) and oblique (arrows) microtubules. Note that circular microtubules occur with a much greater frequency than oblique microtubules.

Fig. 7

Effect of paclitaxel on circular and oblique microtubule populations in myelinated axons. Graphs show the mean ± SEM of the density of (A) circular microtubule profiles and (B) oblique microtubule profiles in myelinated axons of vehicle-treated and paclitaxel-treated nerves at days 7, 27 and 160 post-initiation of treatment. At each time point, circular and oblique microtubules were counted in 120 myelinated axons randomly sampled from two vehicle-treated rats and 180 myelinated axons randomly sampled from three paclitaxel-treated rats. Vehicle n = 120, paclitaxel n = 180: ^#p < 0.01, two-tailed unpaired t-tests, Welch correction applied as appropriate. Note: marked 16-fold change in y-axis range in (B) compared to (A).

Fig. 8

Effect of paclitaxel on microtubule populations in C-fibres. Graphs show the mean ± SEM of the number of (A) total microtubules, (B) circular microtubules and (C) oblique microtubules per C-fibre in vehicle-treated and paclitaxel-treated nerves at days 7, 27 and 160 post-initiation of treatment. At each time point, microtubules were counted and classified in 120 C-fibres randomly sampled from two vehicle-treated rats and 180 C-fibres randomly sampled from three paclitaxel-treated rats. Vehicle n = 120, paclitaxel n = 180: *p < 0.05, two-tailed unpaired t-tests, Welch correction applied as appropriate. Note: the marked 20-fold change in y-axis range in (C) compared to (A) and (B).

Fig. 9

Atypical and normal mitochondria in C-fibres of the saphenous nerve of paclitaxel-treated rats. Normal mitochondria (arrowheads) and atypical (swollen and vacuolated) mitochondria (arrows). (A) Remak bundle with Schwann cell nucleus from a paclitaxel-treated rat at day 27 post-treatment initiation. (B, C) C-fibres from nucleated Remak bundle shown in (A) of paclitaxel-treated rat at day 27 post-treatment initiation. (D) C-fibres from a paclitaxel-treated rat at day 7 post-treatment initiation. Magnifications: (A) 13,800×, (B–D) 44,400×.

Fig. 10

Atypical and normal mitochondria in myelinated axons of the saphenous nerve of paclitaxel-treated rats. Normal mitochondria (arrowheads) and atypical (swollen and vacuolated) mitochondria (arrows). (A) Small, thinly myelinated Aδ-fibre and (B) large myelinated Aβ-fibre of a paclitaxel-treated rat at day 7 post-treatment initiation. (C) Small, thinly myelinated Aδ-fibre and (D) large myelinated Aβ-fibre of a paclitaxel-treated rat at day 27 post-treatment initiation. Magnification: (A–D) 44,400×.

Fig. 11

Effect of paclitaxel on the prevalence of atypical mitochondria in C-fibres and myelinated axons. Graphs show the mean ± SEM of the percentage of (A) C-fibres and (B) myelinated axons that contained atypical mitochondria in vehicle-treated and paclitaxel-treated nerves at days 7, 27 and 160 post-initiation of treatment. Sixty myelinated axons/C-fibres were randomly sampled per animal and the number of myelinated axons/C-fibres that contained atypical mitochondria counted and expressed as a percentage of the total C-fibres/myelinated axons sampled. At each time point, n = 2 for vehicle treatment and n = 3 for paclitaxel treatment. ^#p < 0.01, Fisher’s exact test.