Introduction

Chronic pain conditions can develop through several phases: initiation of pain, pain maintenance, and finally, either resolution or persistency of hypersensitivity/pain1. Bulk RNA-sequencing (RNA-seq) experiments show that pain conditions trigger substantial gene expression plasticity in nociceptive pathway tissues, including the DRG and spinal cord2,3,4,5,6,7,8,9. The cellular mechanisms that cause transitioning from acute to chronic pain are usually decided during the initiation phase. Several studies have suggested that this initiation phase can last as short as 6–48 h for both male and female mice10,11. The maintenance phase of chronic pain is usually accompanied by consistently lowered mechanical and thermal nociceptive thresholds. It is not clear whether cell phenotypes at the peripheral, DRG, or CNS portions of the pain pathways undergo changes that ultimately decide whether an animal will transition to this chronic pain state. If chronic pain conditions undergo resolution, it has been suggested that several (1–4 days) days prior to nociceptive hypersensitivity starting to decline represents a critical phase called the pre-resolution phase1,12. Cell plasticity at this time point for different pain conditions in females and males is also largely unknown. Finally, full pain resolution leads to the post-resolution phase. The post-resolution phase is also associated with cell plasticity compared to naïve animals13.

Despite the progress made in understanding the different phases of pain conditions, there are still several gaps in knowledge1,13. First, gene expression plasticity in peripheral tissues, such as the paw and muscles, during different phases of pain has not been well addressed. Second, there are reports showing substantial distinctions between the mechanisms governing transition to chronic pain and pain resolution12,14. Third, the higher prevalence of pain conditions in women suggests that there are likely mechanistic sex differences in chronic pain that are also not well undestood15,16,17,18,19. Although there is a consensus that gene expression plasticity in the peripheral and central portions of nociceptive pathways can be sex-dependent20, potential sex differences in gene expression occurring at different phases of neuropathic pain development and maintenance have been understudied.

Based on the aforementioned gaps in knowledge, the aims of this study were to investigate the time course of gene expression plasticity in the hind paws of male and female mice and the DRG during the progression of persistent painful neuropathy. The rationale for these aims was to understand whether gene expression plasticity undergoes changes during distinct time points after paclitaxel (PTX) treatment; and whether these changes are sex-dependent. We used a PTX-induced peripheral neuropathy model to produce persistent pain in male and female mice, because this model enabled the generation of both persistent and resolving chronic pain conditions. Transcriptomic profiling was used to measure gene expression changes at different post-PTX time points within the initiation and maintenance phases to elucidate the biological processes underlying pain persistency development in males and females.

Results

Persistent painful chemotherapy-induced peripheral neuropathy in male and female mice

To define biological processes associated with persistent pain in males and females, we titrated systemic paclitaxel (PTX) doses to achieve long-lasting (> 30 days) persistent pain in mice. To do so, we initially used two (one day apart) i.p. PTX treatment with a 2-mg/kg dosage. This treatment produced 14–18 days lasting pain conditions in both males and females. Increasing dosage to 3 mg/kg prolonged the pain condition to 28–30 days. An increase in the cumulative PTX dosage to 18 mg/kg applied three times at individual dose of 6-mg/kg i.p. PTX given over 7 days (Fig. 1A) led to a gradual increase in chronicity and persistent mechanical hypersensitivity in male and female hind paws for at least 47 days (Fig. 1B). PTX-induced persistent mechanical hypersensitivity was almost identical in males and females (2-way ANOVA; F (13, 182) = 1.278; p = 0.23; n = 8). An additional i.p. injection of 6 mg/kg PTX did not increase pain level or change chronicity over 47 days. An extra increase in the overall PTX dosage to 40 mg/kg led to increased mortality and unpredictable development of pain trajectories (data not shown). To determine the trajectory of the development of PTX-induced painful neuropathies, we evaluated cold and heat hypersensitivity in this PTX model (Fig. 1A). Spontaneous pain measurements with condition place preference (CPP) were not evaluated because mice loose habituation over time, which makes it challenging to evaluate the trajectory for pain development. We also omitted the Randall-Selitto paw-pressure test because it shows data similar to those of the von Frey test21. PTX did not induce heat hypersensitivity in males (2-way ANOVA; F (8, 72) = 0.2659; p = 0.97; n = 5; Suppl Fig. 1A), as reported previously21,22. In contrast, PTX produced persistent cold hypersensitivity lasting more than 49 days (2-way ANOVA; for males F (9, 80) = 1.943; p = 0.0473; n = 5; for females F (9, 80) = 1.782; p = 0.0845; n = 5; Suppl Fig. 1B). Altogether, 3 i.p. injections of 6 mg/kg PTX every third day produced long-lasting (> 47 days) and persistent mechanical and cold, but not heat hypersensitivity in both female and male mice (Fig. 1B; Suppl Fig. 1A,B).

Figure 1
figure 1

A Paclitaxel (PTX) treatment paradigm for persistent painful chemotherapy-induced mechanical hypersensitivity in males and females. (A) Schematic for systemic PTX treatment with 6 mg/kg dosages at − 6, − 3 and 0 days; and time points for measurements of mechanical hypersensitivity and hind paw and DRG sample collections. (B) Time course for the development of mechanical hypersensitivity after systemic PTX treatments of male and female mice. Statistic 2-way ANOVA post-hoc Bonferroni test (non-significant; p > 0.05; n = 8).

Full size image

Chronic pain conditions develop through several phases1. As presented in Fig. 1A, the PTX-induced painful neuropathy model could be divided into the following putative stages: an initiation phase from 0–3 days post-PTX; an early maintenance phase from 3–25 days post-PTX; and a late maintenance phase from > 25 days post-PTX, for males and females (Fig. 1A). For tissue collection, we selected one time point in each phase. Thus, for the initiation phase, male and female tissues for bulk RNA-seq were collected 1 day after PTX; for the early maintenance phase, male and female skin and DRG tissues for bulk RNA-seq were collected 16 days after PTX; and for the later maintenance phase, male and female skin and DRG tissues for bulk RNA-seq were collected 31 days after PTX.

Transcriptomic changes in male and female hind paws and DRG 1 day after PTX

RNA-seq data have been deposited in GEO accession number GSE240708. Using criteria of RPKM > 1, FC > 1.5 and Padj < 0.05 for determination of DEGs, at 1 day post-PTX, representing the initiation phase of chronic pain conditions, 47 such genes were upregulated and 411 downregulated in female DRGs, whereas 676 DEGs were upregulated and 852 downregulated in male DRGs (Suppl Fig. 2A,B). At the same time point in hindpaws, 243 DEGs were upregulated and 469 DEGs were downregulated in females, and 916 DEGs were upregulated and 1447 DEGs were downregulated in males (Suppl Fig. 2C,D). In addition to the substantially higher number of regulated DEGs in males compared with females, PTX treatment led to the regulation of drastically different sets of DEGs in males compared with females for both the DRG and hindpaw 1d post-PTX (Suppl Fig. 1). Venny diagram analysis (URL: https://bioinfogp.cnb.csic.es/tools/venny/) showed that these differences were so substantial that only a few overlaps in DEGs were detected (Suppl Fig. 1). Such sex-dependent regulation levels are in a sharp contrast compared to the naïve condition, where gene expression differences between male and female hind paws and DRGs (all cells) were minimal (25 and 8 DEGs, respectively)20.

Biological processes at this PTX post-treatment time point were evaluated using the statistical overrepresentation test using PANTHER analysis (URL: http://www.pantherdb.org/), which associates changes with specific tissues and cell types. The data in Figs. 2 and 3 show that regulated DEGs at 1 day post-PTX were almost exclusively linked to increased or decreased developmental processes in a variety of tissues and cell types. At 1 day post-PTX in female DRGs, upregulation of developmental processes was marginal (Fig. 2A), whereas downregulation of DEGs participating in development was extensive (Fig. 2B). These suppressions of developmental processes affected epithelial cells, the vascular system, and axonogenesis (Fig. 2B). Downregulation of developmental processes was accompanied by reductions in the expression of DEGs related to cell migration, communication, and extracellular matrix remodeling (Fig. 2B). In males, PTX treatment led to processes similar to those in female DRGs 1 day post-PTX (Fig. 2C,D). However, increases in developmental processes were more robust compared with those in females, and downregulation was not as widespread (Fig. 2C,D).

Figure 2
figure 2

Up- and down-regulated biological processes at 1 d post-PTX in female and male DRG. Biological processes for up-regulated (A) and down-regulated (B) DEGs in female DRG at 1 d post-PTX. Biological processes for up-regulated (C) and down-regulated (D) DEGs in male DRG at 1d post-PTX. The X-axis on the panels (A–D) represents numbers of DEGs. The Y-axis notes biological processes. N = 3.

Full size image
Figure 3
figure 3

Up- and down-regulated biological processes at 1d post-PTX in female and male hind paws. Biological processes for up-regulated (A) and down-regulated (B) DEGs in female hind paws at 1d post-PTX. Biological processes for up-regulated (C) and down-regulated (D) DEGs in male hind-paws at 1d post-PTX. The X-axis on the panels (A–D) represents numbers of DEGs. The Y-axis notes biological processes. N = 3.

Full size image

In the hindpaw, similar processes were detected, and they were again distinct for females and males (Fig. 3). Specifically, upregulation of developmental processes associated with muscles, the vascular system, and neuronal tissues was observed, while downregulation of developmental processes was minimal in female hindpaws at 1 day post-PTX (Fig. 3A,B). In males, the picture was the opposite; downregulation of these processes was found with evidence for genes associated with muscles, epithelium, the vascular system, and neuronal tissues, including a reduction in myelination and axonogenesis (Fig. 3C,D). This down-regulation of developmental processes was accompanied by attenuation of extracellular matrix remodeling, cytoskeleton organization, and ATP metabolic processes for cellular energy (Fig. 3D). Interestingly, regulation of DEGs related to the immune system was relatively minimal. Nevertheless, up-regulation of developmental processes was correlated with a slight increase in immune system activity in female paws (Fig. 3A), while down-regulation of these processes was complemented by a decrease in expression of immune system DEGs in male paws (Fig. 3D). Altogether, according to PANTHER analysis (URL: http://www.pantherdb.org/), at the initial stage of post-PTX treatment, regulated DEGs and biological processes were almost exclusively related to upregulation or downregulation of DEGs related to developmental processes in muscles, epithelium, the vascular system, and neuronal tissues, including myelination, neurogenesis, and axonogenesis. Interestingly, regulation of immune system-related genes was minimal. Furthermore, regulated DEGs, but to a lesser extent biological processes, were substantially different between females and males in paw and DRG tissues.

Transcriptomic changes at 16 days post-PTX in the hind paws and DRG of male and female mice

Bulk RNA-seq of the hind paw and DRG tissues of males and females showed that at 16 days post-PTX, 51 DEGs were upregulated and 31 were downregulated in the female DRG, and 108 were upregulated and 81 were downregulated in the hind paws (Suppl Fig. 3). The numbers of regulated DEGs in males were higher in the DRG and similar numbers were observed in paws, with 657 up and 388 down in the DRG and 147 up and 98 down in the paw (Suppl Fig. 3). According to venny diagram analysis (URL: https://bioinfogp.cnb.csic.es/tools/venny/), similar to the initiation stage, 16-day post-PTX treatment resulted in distinct outcomes in the identities of regulated DEGs in male and female paws and DRGs (Suppl Fig. 3).

Biological processes at the initiation (1 d post-PTX) and maintenance (16d post-PTX) phases were similar in the DRGs and paws, but there was a reduction in the scope of biological processes and the numbers of regulated DEGs (Fig. 2 vs Fig. 4; and Fig. 3 vs Fig. 5). In female DRGs, up- and downregulation of DEGs related to developmental processes became minimal, and few biological processes were identified using PANTHER analysis (Fig. 4A,B). In male DRGs, upregulation of the developmental processes of epithelial cells, muscle, and neuronal tissues increased (Fig. 4C), whereas downregulation of DEGs representing developmental processes remained at a similar level (Fig. 4D). In female paws, the upregulation of developmental processes was drastically reduced (Fig. 3A vs Fig. 5A), and downregulation remained at approximately the same levels (Fig. 3B vs Fig. 5B). In male paws, the upregulation of developmental processes was stable compared with the 1-day post-PTX (Fig. 3C vs Fig. 5C), whereas downregulation was reduced (Fig. 3D vs Fig. 5D). In the male hindpaw, reduced neurogenesis, axonogenesis, and myelination were still observed (Fig. 5D). Overall, according to the PANTHER analysis (URL: http://www.pantherdb.org/), at 16 d post-PTX treatment, DEGs representing the developmental processes of a variety of tissues were regulated, but the numbers of DEGs were substantially reduced compared with those at 1 d post-PTX with one notable exception: upregulation of developmental processes in male DRGs, which was substantially increased (Fig. 2C vs Fig. 4C).

Figure 4
figure 4

Up- and down-regulated biological processes at 16d post-PTX in female and male DRG. Biological processes for up-regulated (A) and down-regulated (B) DEGs in female DRG at 16d post-PTX. Selected DEGs showed statistical difference when Pval < 0.05. For other panels in Figs. 2, 3, 4, 5, 6, 7, selection criteria for DEGs was Padj < 0.05. Biological processes for up-regulated (C) and down-regulated (D) DEGs in male DRG at 16d post-PTX. The X-axis on the panels (A–D) represents numbers of DEGs. The Y-axis notes biological processes. N = 3.

Full size image
Figure 5
figure 5

Up- and down-regulated biological processes at 16d post-PTX in female and male hind paws. Biological processes for up-regulated (A) and down-regulated (B) DEGs in female hind paws at 16d post-PTX. Biological processes for up-regulated (C) and down-regulated (D) DEGs in male hind paws at 16d post-PTX. The X-axis on the panels (A–D) represents numbers of DEGs. The Y-axis notes biological processes. N = 3.

Full size image

Transcriptomic changes at 31 days post-PTX in the DRG and hind paws of male and female mice

At the 31 days post-PTX time point, 7 DEGs were up- and 137 down-regulated in female DRGs (Suppl Fig. 4A,B), and only 31 up- and 22 down-regulated were found in the hindpaw (Suppl Fig. 4C,D). The numbers of regulated DEGs in males were far higher, with 376 up- and 314 down-regulated in DRGs (Suppl Fig. 4A,B), and 159 up- and 323 down-regulated in hindpaws (Suppl Fig. 4C,D). The overlaps between regulated DEGs in the female and male DRG and hind paws were low, similar to those at 1 and 16 days post-PTX (Suppl Fig. 4). Upregulated DEGs in the female DRG remained the same as at 16 days post-PTX (Fig. 4A vs Fig. 6A). However, downregulated DEG counts increased (Fig. 4B vs Fig. 6B). In contrast, up- and downregulated developmental processes in male DRGs were similar at 16 and 31 days post-PTX (Fig. 4C vs Fig. 6C; and Fig. 4D vs Fig. 6D). In the hindpaw, processes were similar at 16 vs. 31 d post-PTX for females (Fig. 5A vs Fig. 7A; and Fig. 5B vs Fig. 7B) as well as males (Fig. 5C vs Fig. 7C; and Fig. 5D vs Fig. 7D). Notably, upregulation of DEGs related to the immune system was yet again related to an increase in developmental processes, whereas downregulation of immune processes was associated with attenuation of developmental processes (Figs. 5B,C, 7B). In summary, at 31 days post-PTX treatment, the upregulation and downregulation of DEGs representing developmental processes underwent only slight changes from 16 to 31 days post-PTX in both female and male DRGs as well as hindpaws (Figs. 6, 7).

Figure 6
figure 6

Up- and down-regulated biological processes at 31d post-PTX in female and male DRG. Biological processes for up-regulated (A) and down-regulated (B) DEGs in female DRG at 31d post-PTX. Selected DEGs showed statistical difference when Pval < 0.05. Biological processes for up-regulated (C) and down-regulated (D) DEGs in male DRG at 31d post-PTX. The X-axis on the panels (A–D) represents numbers of DEGs. The Y-axis notes biological processes. N = 3.

Full size image
Figure 7
figure 7

Up- and down-regulated biological processes at 31d post-PTX in female and male hind paws. Biological processes for up-regulated (A) and down-regulated (B) DEGs in female hind paws at 31d post-PTX. Selected DEGs showed statistical difference when Pval < 0.05. Biological processes for up-regulated (C) and down-regulated (D) DEGs in male hind paws at 31d post-PTX. The X-axis on the panels (A–D) represents numbers of DEGs. The Y-axis notes biological processes. N = 3.

Full size image

PTX-induced immune cell profile changes in the male and female DRGs and hind paw

In males, immune system DEGs were slightly upregulated in the DRG at 16 and 31 days post-PTX (Figs. 4C, 6C) and in the hindpaw at 16 days post-PTX (Fig. 5C); while immune system DEGs were downregulated at 1 day post-PTX in the hindpaw (Fig. 3D). In females, the immune system DEGs were mainly downregulated in the DRG at 16 and 31 days post-PTX (Figs. 4B, 6B) and in the hindpaw at 31 days post-PTX (Fig. 7B); the exception was minor upregulation at 1 day post-PTX in the hindpaw (Fig. 3A). Changes in immune system-related gene expression may be due to immune cell infiltration/proliferation in tissues and/or the activation of resident immune (or potentially non-immune) cells in these tissues.

We investigated whether the regulation of immune-related genes correlated with changes in immune cell profiles in the DRG and hind paw of male and female mice. Flow cytometry and FlowJo analysis sftware (URL: https://www.flowjo.com/solutions/flowjo/downloads/previous-versions) were used to examine immune cell profiles in male and female hindpaw skin and DRG tissues at different time points after PTX. To evaluate immune cell types, the following gating strategy was used. Live/singlets/CD45+ cells were gated using the markers listed below to define specific cell populations: neutrophils (Nph, CD11b+/Ly6G+); macrophages (Mph, CD11b+/MHCIIhi/CD64+/CD24lo/CD11c−); inflammatory monocytes (iMo, CD11b+/MHCIIlo/SSClo/CD64+/Ly6Chi); CD11b+ dendritic cells; (DCs; CD11b+/CD64−/CD24hi/MHCIIhi/CD11c+); CD11b− DCs (CD11b−/CD64−/CD24hi/MHCIIhi/CD11c+); natural killer cells (NK; NK1.1+/TCRβ−); B cells (B, B220+/CD11b−/CD11c−); and T cells (T, CD3+/CD11b−/CD11c−). All immune cell counts were normalized to the numbers of live “singlet” cells. Overall changes in immune cell (CD45+ cells) counts after PTX treatments were minimal and statistically insignificant in male hindpaw (1-way ANOVA; F (3, 12) = 0.9914; p = 0.4298; n = 3–4; Suppl Fig. 5A), male DRG (1-way ANOVA; F (3, 12) = 4.798; p = 0.0202; n = 3–4; Suppl Fig. 5B), female hindpaw (1-way ANOVA; F (3, 12) = 2.118; p = 0.1512; n = 3–4; Suppl Fig. 5C), and female DRG (1-way ANOVA; F (3, 12) = 1.883; p = 0.1862; n = 3–4; Suppl Fig. 5D).

Changes in individual immune cell type profiles are presented as normalized total counts (Figs. 8A,B, 9A,B) and as counts among CD45+ cells (Figs. 8C,D, 9C,D). In accordance with published single-cell RNA-sequencing data, DRG of naïve males have had substantial numbers of macrophages (Mph) and neutrophils (Nph)23,24. Moreover, it was reported that Nph is mainly located in the dura mater surrounding sensory ganglia24. Flow cytometry of DRG samples, which included the dura mater, showed that vehicle-treated male (Fig. 8B,D) and female DRGs (Fig. 9B,D) were dominated by Mph and Nph.

Figure 8
figure 8

Immune cell profiles in DRG and hind paws of PTX-treated male mice. (A) Normalized (by live cells) counts of immune cells in the male DRG at 1, 16 and 31d post-PTX systemic treatments. (B) Normalized counts of immune cells in male hind paws at 1, 16 and 31d post-PTX systemic treatments. (C) Immune cell counts per 104 CD45+ cells in the DRG at 1, 16 and 31d post-PTX systemic treatments. (D) Immune cell counts per 104 CD45+ cells in the hind paws at 1, 16 and 31d post-PTX systemic treatments. BL is immune cell counts in male DRG and hind paws at 1d post vehicle-treatment. iMo inflammatory monocytes, Mph macrophages, Nph neutrophils, DCs dendritic cells, NK natural killer cells, T T-cells, B B-cells. Statistic is 2-way ANOVA (*p < 0.05; **p < 0.01; ***p < 0.001; #p < 0.0001; n = 3–4).

Full size image
Figure 9
figure 9

Immune cell profiles in DRG and hind paws of PTX-treated female mice. (A) Normalized (by live cells) counts of immune cells in the female DRG at 1, 16 and 31d post-PTX systemic treatments. (B) Normalized counts of immune cells in female hind paws at 1, 16 and 31d post-PTX systemic treatments. (C) Immune cell counts per 104 CD45+ cells in the DRG at 1, 16 and 31d post-PTX systemic treatments. (D) Immune cell counts per 104 CD45+ cells in the hind paws at 1, 16 and 31d post-PTX systemic treatments. BL is immune cell counts in female DRG and hind paws at 1d post vehicle-treatment. iMo inflammatory monocytes, Mph macrophages, Nph neutrophils, DCs dendritic cells, NK natural killer cells, T T-cells, B B-cells. Statistic is 2-way ANOVA (#p < 0.0001; n = 3–4).

Full size image

PTX treatment slightly altered the immune cell profiles of Mph and Nph, but not inflammatory monocytes (iMo), dendritic (DCs), natural killer (NK), T- and B-cells in male DRGs (2-way ANOVA variables are cell types and post-PTX time points; F (18, 84) = 9.760, p < 0.001; n = 3–4; Fig. 8B,D). Specifically, Mph was downregulated at 16 days post-PTX and slightly upregulated at 31 days post-PTX (Fig. 9C), whereas Nph was upregulated only at 31 days post-PTX in male DRGs (Fig. 8B,D). This may correspond to minor changes in the immune system-related DEG expressions reflected in Figs. 4C and 6C. PTX treatment did not significantly change Mph, Nph, as well as iMo, DCs, NK, T, and B cells in female DRGs (2-way ANOVA variables are cell types and post-PTX time points; Fig. 9B,D), despite slight downregulation of immune-related DEGs in DRGs at 16 and 31 days post-PTX (Figs. 4B, 6B).

Hind paws of naïve males and females had detectable numbers of Mph, DCs, and T cells (Figs. 8A,C, 9A,C). PTX treatment altered immune cell profiles of only T cells, but not Mph, Nph, DCs, NK, and B cells in male hindpaws (2-way ANOVA variables are cell types and post-PTX time points; F (15, 72) = 2.017, p = 0.025; n = 3–4; Fig. 8A,C). Specifically, male hind paw T cells were upregulated at 16 days post-PTX (Fig. 8C). This may correspond to (only 23 DEGs) changes in immune system-related DEG expression, as shown in Fig. 5C. Overall, the profiles of the majority of immune cell types did not significantly change in female hindpaws after PTX treatment (Fig. 9A,C). However, Mph cell numbers were reduced in the female hind paw at 31 days post-PTX (2-way ANOVA variables are cell types and post-PTX time points; F (15, 72) = 2.248; p = 0.0118; n = 3–4; Fig. 9A,C). Overall, our data indicate that PTX induces slight infiltration and/or proliferation of immune cells predominantly in male DRGs and the hind paws. These changes in immune cell numbers at different post-PTX time points correspond to the detected small changes in gene expression plasticity in immune system-related DEGs (Figs. 3A,D, 4B,C, 5C, 6B,C, 7B).

Discussion

Gene expression plasticity observed at different phases after PTX treatment was dramatically different between males and females. PANTHER analysis predominantly attributed these changes to developmental processes for a variety of non-neuronal and neuronal cell types and/or tissues in both males and females. One way to interpret these data is that the upregulation of DEGs related to developmental processes could represent regenerative processes, whereas the downregulation of DEGs associated with tissue development could characterize degenerative processes. An increase in the expression of genes representing cytoskeleton organization, extracellular matrix remodeling, and oxidative phosphorylation could contribute to tissue and nerve regeneration, whereas a reduction in these processes could reflect degeneration. Thus, muscle and vascular degeneration was associated with the downregulation of DEGs related to muscle development, mitochondrial function, and cytoskeleton organization25,26,27,28. Amyotrophic lateral sclerosis (ALS) as a neurodegenerative disease was associated with reduced myelination and neurogenesis, synapse organization, synaptic transmission, and mitochondrial function29. In contrast, regenerative processes in tissues and nerves were linked to extracellular matrix formation and remodeling, increased axonogenesis, improved tissue development and differentiation, and increased mitochondrial function30,31,32,33,34.

Based on the outlined interpretation framework, we schematically presented the time courses of regenerative and degenerative processes in the hindpaw and DRG of males and females in Fig. 10. Levels of regeneration and degeneration in tissues were estimated according to the number of regulated DEGs and biological processes associated with tissue development. Figure 8A,B also show that the time course of regenerative and degenerative processes distinctly depends on tissue type (DRG vs paw) and sex. Thus, in DRG, degenerative processes persisted through 31 days post-PTX in female (Fig. 10A). In contrast, in male DRGs, regenerative processes were dominant, but no balance between these processes was achieved even after 31 days of PTX treatment (Fig. 10A). In the hindpaw, processes were also sex-dependent. Regenerative processes at 1 day post-PTX were switched to degenerative processes at 16 and 31 days post-PTX treatment in females (Fig. 10B). In males, a high level of degenerative-associated gene expression was observed throughout the 31 days post-PTX (Fig. 10B). Overall, our data indicated that pain persistence after PTX treatment was associated with a variety of continuously ongoing tissue-damaging and repairing processes in the hindpaw and DRG of both males and females. However, individual gene expression plasticity after PTX treatment was substantially sex-dependent. Degenerative events in the skin and DRG after PTX treatment are not surprising, as they have been reported in multiple previous studies35,36,37. Nevertheless, more detailed studies on the differences in these processes between males and females could be conducted in follow-up studies.

Figure 10
figure 10

Post-PTX time courses for changes in degenerative and regenerative processes in DRG and hind paw of males and females. (A) Schematic representation of degenerative and regenerative processes in DRG of males and females at 1d, 16d and 31d post-PTX. (B) Schematic representation of degenerative and regenerative processes in hind paws of males and females at 1d, 16d and 31d post-PTX. The X-axis on the panels (A) and (B) represents post-PTX days. The Y-axis notes putative levels of degenerative (blue box) or regenerative (orange box) processes.

Full size image

Some PTX-induced processes, such as neurogenesis and axonogenesis, were related to neuronal development. It is understandable that such processes occur in DRGs. But it could be surprising to find them in the hindpaw. One possible explanation for this phenomenon is that the skin is densely innervated by sensory neurites/afferents, which transport substantial amounts of mRNA from cell bodies38,39,40,41. The transported mRNA is used for local translation, which is critical for the initiation and maintenance of pain chronicity in a sex-dependent manner11,39,42. To test this possibility, spatial sequencing technologies could be used to localize transcripts to areas of axonal inputs to the hindpaw.

Somewhat surprisingly, DEGs related to proinflammatory processes were lower in numbers for both sexes. However, they were more evident in the hindpaw and DRG of males than in females, and this was confirmed by flow cytometry measurements. Previous reports have indicated that the immune system could be a major contributor to pain persistency (see review43). Many reports also implied that the immune system is implicated in each pain development phase, including the chronic pain maintenance phase1,16,17,41,43,44,45,46. There is also strong evidence for immune involvement in pain resolution. For instance, activation of the anti-inflammatory IL-13/IL-10 T cell/macrophage axis in the DRG resolves chemotherapy-induced painful neuropathies (CIPN) in both males and females47. Reported data also indicate that immune cell-mediated mechanisms are sexually dimorphic in chronic pain conditions, leading to the sex-dependent promotion of pain chronicity48,49,50. For instance, despite similar progressions of persistent neuropathic pain in males and females, microglia-linked mechanisms contribute to neuropathic and inflammatory pain in male mice but not in female mice51,52. Stimulation of TLR-9 results in CIPN pain initiation only in male mice, whereas an active IL-23/Il-17A/TRPV1 axis promotes pain persistence in females45. The general theme emerging from this body of evidence is that males may be more sensitive to the effects of monocyte-lineage cells than, whereas females are more sensitive to T cells52. One commonly held notion is that pain will persist as long as inflammation is present in the DRG and/or spinal cord43. Consistent with this, the elimination of M1 macrophages or activation of M2 macrophages in DRGs reverses mechanical hypersensitivity in several neuropathic pain models43,44,45,49,53,54,55,56. This conclusion is also supported by the fact that most proinflammatory mediators can trigger nociceptive responses in naïve animals by either directly or indirectly sensitizing sensory neurons by changing the microenvironment in the vicinity of peripheral and/or central terminals43.

Recently, different mechanisms for pain resolution have been proposed, implying that elevation of proinflammatory processes may be required for pain resolution14. Prolonged intraplantar treatment with prolactin rapidly elevates proinflammatory processes in males but not in females and consequently supports pain resolution only in males12. Elevated neutrophil counts in the blood and CD11c+ microglia in the spinal cord are associated with chronic pain resolution in patients and result in pain resolution rodent pain models, respectively14,57. Our data are broadly consistent with this mechanism and indicate that the association between proinflammatory processes and pain persistency is minimal in both males and females. Despite mechanical hypersensitivity from day 1 onward, the number of macrophages in DRGs was not different from baseline (2 d), significantly reduced from baseline (16 d), or moderately increased from baseline (35 d) (Fig. 8A–D), and transcriptomic signatures did not reveal major shifts in genes associated with macrophages. We do not rule out the possibility that macrophage populations may play different roles at different time points; however, this inconsistent correlation with pain behavior suggests that macrophage populations are not the only drivers of pain behavior.

Our data indicate that pain persistence in male and female mice treated with PTX is correlated with either degenerative or regenerative processes in the DRG and hindpaw skin. Remarkably, in our observations, degeneration was accompanied by a reduction in immune processes, and regeneration was complemented by an increase in immune system activity. This observation suggests an alternative hypothesis that proinflammatory macrophages react to global changes in tissue physiology after PTX treatment and drive degenerative and regenerative processes toward tissue homeostasis without a clear role in driving pain behaviors.

In conclusion, despite PTX-induced persistent neuropathy having the same pain progression trajectories in males and females, the condition triggers dramatic sex-dependency in gene expression plasticity and biological processes at every post-PTX time point. Dominant biological processes in the hindpaw and DRG of females and males during the initiation and maintenance phases of PTX-induced persistent mechanical and cold hypersensitivity were consistently related to alterations in the development of many types of tissues and cells at every post-PTX time point. These developmental processes have resemblances to degeneration or regeneration processes that could affect non-neuronal cells (angiogenesis, cartilage, bone, muscle and epithelial cell development), sensory neurons (neurogenesis), and sensory neurites (myelination, axonogenesis) in skin. As expected, these events were accompanied by cytoskeletal and extracellular matrix remodeling, oxidative phosphorylation, and cell energy production. In contrast, despite the commonly held notion that pain will persist as long as inflammation is present in the skin and DRG, we found a minimal increase in pro-inflammatory responses during the maintenance phase of the pain condition. The points presented here require functional experiments to support these findings. Perhaps, some of these functional experiments could be performed on resolving chronic pain models. Nevertheless, the results presented herein add new depth to our understanding of the biological processes linked to chronic pain in males and females.

Materials and methods

Ethical approval

The reporting in this manuscript follows the recommendations in the ARRIVE guidelines (PLoS Bio 8(6), e1000412, 2010). We also followed the guidelines issued by the National Institutes of Health (NIH) and the Society for Neuroscience (SfN) to minimize the number of animals used and their suffering. All animal experiments conformed to protocols approved by the University Texas Health Science Center at San Antonio (UTHSCSA) Institutional Animal Care and Use Committee (IACUC). The protocol numbers are 20180001AR and 20200011AR.

Reagents and mouse lines

Paclitaxel (PTX) was purchased from Millipore-Sigma (Cat: T7402; St. Louis, MO). The PTX powder was replaced every 4 months to maintain the reagent’s activity at the same level. The PTX stock (40 mg/ml) was prepared in 100% ethanol, which was kept in − 80 °C for no more than 2 weeks. Stock was diluted to working concentrations using a mixture of 650 ml saline/160 ml Kolliphor-EL (Millipore-Sigma; Cat: C5135; St. Louis, MO, USA). Remnant diluted stock was discarded after every injection.

Experiments were performed on 10–18-week-old C57BL/6 wild-type male and female mice, which were originally purchased from Jackson Laboratory (Bar Harbor, ME).

Paclitaxel (PTX)-induced pain model and hypersensitivity measurement

PTX was injected systemically (i.p.). An appropriate sample size for each group was computed using power analysis (see below). PTX was titrated (see “Results”) for achieving persistent (> 47 days) painful neuropathies in chemotherapy-induced models. PTX was injected three times at − 6, − 3, and 0 days at a dosage of 6 mg/kg for every male and female to generate persistent mechanical hypersensitivity for at least 47 days in both right and left hindpaw (see schematic; Fig. 1A,B). Heat, cold, and mechanical hypersensitivity were measured in the right hindpaw.

Mechanical hypersensitivity, which was reflected as the mechanical withdrawal threshold at the hindpaw, was assessed using the up-down von Frey filament as previously described42. Mice were habituated for 45–60 min before measurements. Baseline (BL) readings were obtained on the right hindpaw before PTX injection. No readings were performed on day 0 after the last PTX injection.

Heat nociception was assessed using Hargreaves’ apparatus (Ugo Basile) as previously described58. In brief, mice were placed on a glass surface with temperature held constant at 20–22 °C. Following habituation, thermal withdrawal latencies to a radiant heat beam were recorded at each time point (3× measurements at each time point, averaged to obtain the data value used in analyses). To prevent tissue damage, the stimulus was terminated after ≈ 20 s if the animal did not withdraw its hind paw.

Cold nociception in mice was measured as described59. In brief, mice were placed in acrylic boxes with glass bottoms. They were habituated to the boxes for 45–60 min. Crushed dry ice was packed into heavy-gauge aluminum foil and applied to glass underneath the hindpaw. Thermal withdrawal latencies to packed dry ice were recorded (3× measurements at each time point, averaged to obtain the data value used in analyses). One mouse at a time was used to measure cold nociception.

Mechanical, heat, and cold hypersensitivity development were monitored at the right hindpaw at the post-last PTX injection time points specified in the text and shown in the figures. The behavior experiments were blinded so that the experimenter was not aware of the treatment conditions. We also used randomized designs for behavioral experiments in which animals were randomly assigned to various experimental groups. Additionally, experiments were performed in several trials with small numbers of mice. Data from several trials with small groups were pooled together without normalization.

Biopsy and RNA isolation

We used different mouse groups for tissue collection for RNA sequencing and flow cytometry. The development of mechanical hypersensitivity was monitored in mice used to obtain biopsies for RNA sequencing and flow cytometry. To eliminate the contribution of immune cells from blood, all animals for biopsy were perfused with cold PBS before tissue dissection. The time points for tissue collections were measured from the last PTX or vehicle injection. Moreover, at each time point, mechanical thresholds were measured, and then biopsies for RNA sequencing were collected 4–6 h after the last measurement of mechanical hypersensitivity. Biopsies for flow cytometry were obtained 12–16 h after the last measurement of mechanical hypersensitivity. DRG dissection was performed as described20,60. Briefly, the spinal canal was exposed, and left and right L3–L5 DRGs from each animal were removed using fine dissection and pooled together into RNA-easy or Hank’s buffer. For hindpaw glabrous skin, a 3-mm punch biopsy of the plantar surfaces of the left and right hindpaw was performed, and epidermal, dermal, and subcutaneous connective tissue. Subcutaneous connective tissue was removed using forceps, and the epidermis/dermis was finely minced before placing into RNA-easy or Hank’s buffer14.

RNA from relatively soft tissues, such as DRGs, was isolated using RNA-easy solution (Qiagen) and a Bead Mill Homogenizer (Omni International, Kennesaw, GA) as previously described20,61. The isolation of skin RNA required careful slicing/dissing of tissues before the homogenization step, which is detailed in62. RNA quality and integrity were checked using an Agilent 2100 Bioanalyer RNA 6000 Nanochip (Agilent Technologies, Santa Clara, CA). The RNA quality selected for sequencing had a RIN score in the range of 7–9.

Bulk RNA sequencing and analysis

Approximately 500 ng of total RNA was used for cDNA library preparation with oligo-dT primers according to the Illumina TruSeq preparation guide (Illumina, San Diego, CA)20. cDNA libraries were quantified, pooled for cBot amplification, and sequenced using a 50-bp single read run on an Illumina HiSeq 3000 platform. The depth of reads was 30–50 × 106 bp. RNA-seq cDNA libraries were prepared using oligodT according to the SMART-seq-2 protocol63,64 with previously described modifications20.

RNA-sequencing data analysis was previously presented in detail20. In summary, demultiplexing to generate FastQ files was performed using CASAVA. The RNA-seq data were aligned to the mm9/UCSC hg19 mouse genome using the TopHat2 default settings. The alignment BAM files were processed using HTSeq-count to obtain counts per gene for all samples. DESeq-2 was used to identify differentially expressed genes (DEGs) after median normalization. Quality control of the outliers, intergroup variability, distribution levels, PCA, and hierarchical clustering analysis were performed to statistically validate the experimental data. Samples that did not pass these quality control steps were not used in the final analysis. Multiple test correction was performed using the Benjamini–Hochberg procedure, and an adjusted p value (Padj) was generated. If not specified in the text selection criteria for DEGs, reads per kilobase of transcript, per million mapped reads (RPKM) > 1, fold-change (FC) > 1.5 and statistically significant DEGs with Padj < 0.05. Venn diagrams were generated using (URL: https://bioinfogp.cnb.csic.es/tools/venny/). Genes were clustered according to biological processes using PANTHER software (URL: http://www.pantherdb.org/)20.

Flow cytometry

Flow cytometry was used to assess immune cell profiles in tissue biopsies. Single-cell suspensions from tissues were created by treating them for 60–90 min (60 min for DRG and 90 min for skin) at 37 °C with 250 μg/ml Liberase (Millipore-Sigma; St. Louis, MO) and 100 μg/ml dispase I (Millipore-Sigma; St. Louis, MO), washing with Dulbecco’s modified eagle medium (DMEM) containing 5% fetal calf serum (FCS), triturating with a Pasteur pipette, and then filtering cell suspensions through 70 μm strainer.

Cell suspensions were first stained for viability using Zombie NIR™ Fixable Viability Kit (BioLegend; San Diego, CA) for 20 min at room temperature in phosphate buffer solution (pH 7.2S) combined with FcR blocking antibody (1 μg, clone 2.4G2, BioXCell; Lebanon, NH) to block nonspecific binding. Cells were then washed with 2% FBS/PBS and stained with antibodies against surface antigens for 30 min on ice. Fluorochrome-conjugated antibodies against mouse CD45 (clone 30-F11), CD3 (145-2C11), CD24 (M1/69), B220 (RA3-6B2), CD11b (M1/70), CD64 (X54-5/7.1), CD11c (N418), NK1.1 (PK136), TCRβ (H57-597), MHC-II (M5/114.15.2), Ly-6G (1A8), and Ly-6C (KH1.4) were purchased from BioLegend (San Diego, CA), eBioscience (San Diego, CA), or BD Biosciences (San Jose, CA). Flow cytometry was performed using a Celesta or LSRII cytometer (BD Biosciences; San Jose, CA). Data were analyzed using the FlowJo LLC v10.6.1 software (URL: https://www.flowjo.com/solutions/flowjo/downloads/previous-versions).

The gating strategy for selecting immune populations in the skin and DRG was modified from a previously described approach65. Live/singlets/CD45+ cells were gated using the markers listed below to define specific cell populations: neutrophils (Nph, CD11b+/Ly6G+); macrophages (Mph, CD11b+/MHCIIhi/CD64+/CD24lo/CD11c−); inflammatory monocytes (iMo, CD11b+/MHCIIlo/SSClo/CD64+/Ly6Chi); CD11b+ dendritic cells; (DCs; CD11b+/CD64−/CD24hi/MHCIIhi/CD11c+); CD11b− DCs (CD11b−/CD64−/CD24hi/MHCIIhi/CD11c+); natural killer cells (NK; NK1.1+/TCRβ−); B cells (B, B220+/CD11b−/CD11c−); and T cells (T, CD3+/CD11b−/CD11c−).

Statistical analysis

Power analysis considered our experience with next-generation RNA sequencing from similar tissues and treatments and behavioral measurement of nociception modalities. To reduce variability, an inbred mouse line was used, and samples that did not pass quality controls during pre- or post-sequencing procedures were not included. Power analysis for RNA-seq experiments on tissue biopsy samples identified that a sample size of 3 for each group of mice attains > 80% power for each test to detect at least a 1.5-fold significant difference (FC) in gene expression with an estimated standard deviation of 0.6 (considering low mouse to mouse variation and rejection of certain samples after quality controls as described above) with a false discovery rate (FDR) of 0.05 using a Benjamini–Hochberg test assuming that the actual distribution is normal. Power analysis for behavioral experiments showed that a sample size of 6 for each group of mice achieved > 80% power for each test to detect a significant difference.

GraphPad Prism 8.0 (GraphPad, La Jolla, CA) was used for all statistical analyses. Data in the figures are mean ± standard error of the mean (SEM), with “n” referring to the number of animals per group. Differences between groups with two variables were assessed using two-way ANOVA and Bonferroni post hoc tests. A difference was accepted as statistically significant when p < 0.05. The interaction F ratios and associated p values are reported.

Ethical approval and informed consent

The reporting in the manuscript follows the recommendations in the ARRIVE guidelines 2.0 (PLoS July 14, 2020; https://doi.org/10.1371/journal.pbio.3000410). All experimental protocols were approved by the UTHSCSA IACUC. The protocol numbers are 20180001AR and 20200011AR.