A transient inflammatory response contributes to oxaliplatin neurotoxicity in mice

Abstract

Objectives: Peripheral neuropathy is a relevant dose-limiting adverse event that can affect up to 90% of oncologic patients with colorectal cancer receiving oxaliplatin treatment. The severity of neurotoxicity often leads to dose reduction or even premature cessation of chemotherapy. Unfortunately, the limited knowledge about the molecular mechanisms related to oxaliplatin neurotoxicity leads to a lack of effective treatments to prevent the development of this clinical condition. In this context, the present work aimed to determine the exact molecular mechanisms involved in the development of oxaliplatin neurotoxicity in a murine model to try to find new therapeutical targets.

Methods: By single-cell RNA sequencing (scRNA-seq), we studied the transcriptomic profile of sensory neurons and satellite glial cells (SGC) of the Dorsal Root Ganglia (DRG) from a well-characterized mouse model of oxaliplatin neurotoxicity.

Results: Analysis of scRNA-seq data pointed to modulation of inflammatory processes in response to oxaliplatin treatment. In this line, we observed increased levels of NF-kB p65 protein, pro-inflammatory cytokines, and immune cell infiltration in DRGs and peripheral nerves of oxaliplatin-treated mice, which was accompanied by mechanical allodynia and decrease in sensory nerve amplitudes.

Interpretation: Our data show that, in addition to the well-described DNA damage, oxaliplatin neurotoxicity is related to an exacerbated pro-inflammatory response in DRG and peripheral nerves, and open new insights in the development of anti-inflammatory strategies as a treatment for preventing peripheral neuropathy induced by oxaliplatin.

Figure 1

Electrophysiological, behavioral, and morphological characterization of the OIPN mouse model. (A) Oxaliplatin was administered i.v. twice a week (2 × w) for 8 weeks (Induction time). After this time, animals were further evaluated to assess the coasting effect for 2 additional weeks (coasting‐effect time). (B–D) Nerve conduction studies (NCS): SNAP (B), CNAP (C), and CMAP (D) recordings. Data are expressed as change vs basal. n = 15 Control mice and 35 Oxaliplatin‐treated mice. (E) Von Frey test of control and oxaliplatin‐treated mice. Withdrawal threshold (force, G) to mechanical stimuli is represented. n = 9 Control mice and 19 Oxaliplatin‐treated mice. (F) Cold plate test of control and oxaliplatin‐treated mice. Time (seconds) of first hind paw lift to −4°C noxious stimuli is represented. n = 10 Control mice and 13 Oxaliplatin‐treated mice. (G) Plantar test of control and oxaliplatin‐treated mice. Time (seconds) of paw lift to heat stimuli is represented. n = 8 Control mice and 10 Oxaliplatin‐treated mice. (H, I) Quantification of the number of PGP+ (H), and CGRP+ (I) IENF. n = 5 mice/group. (J, K) Quantification of the number of myelinated axons in the sciatic (J) and the tibial (K) nerves of control and oxaliplatin‐treated mice at 8 and 10 weeks. n = 4–5 mice/group in sciatic nerve; n = 7 mice/group in tibial nerve. *p < 0.05 **p < 0.01 vs Control;^$ p < 0.05^$$ p > 0.01^$$$ p < 0.001 veruss Basal. 2‐Way RM ANOVA with Bonferroni post hoc test was used for the analysis of NCS and algesimetry tests. One‐way ANOVA with Bonferroni post hoc test was used for the analysis of histological studies. All data are represented as group mean ± SD.

Figure 2

scRNA‐seq analysis of DRG cell population after oxaliplatin treatment. (A) Single‐cell sorting strategy used to isolate DRG cells. Only cells that were negative for propidium iodide (PI) and positive for TRKs labeling were selected and sorted into the 96‐well plates. All other events were discarded. (B) t‐SNE plot showing the distribution of different cell populations obtained in the analysis. (C) Neuronal populations overexpress the specific neuronal markers Tubb3 (left) and Eno2 (right). Color key represents normalized gene expression with the highest expression marked red and the lowest marked gray. (D, E) Volcano plots of DEGs (p < 0.05) between control and oxaliplatin‐treated mice in sensory neurons (D) and SGCs (E). Each point represents a single DEG. The negative log of p‐val‐adj (base 10) is plotted on the Y‐axis, and the log of the FC (base 2) is plotted on the X‐axis. Only DEGs above the dashed line have a p‐val‐adj <0.05. Green genes are up‐regulated and red genes are down‐regulated in oxaliplatin‐treated mice. Vertical lines indicate Log₂(FC) of 1 or −1. Horizontal line indicates the point in which p‐val‐adj <0.05 (−log₁₀(p‐val‐adj) = 1.3). (F, G) Inflammation process‐related GO annotations enriched in neurons (F) and SGCs (G) of oxaliplatin‐treated mice. Vertical lines indicate the point in which p < 0.05. The numbers right of the bars correspond to the “number of genes associated with the given GO term and found in the oxaliplatin vs control comparison”/“number of genes associated with the given GO term in the reference genome”.

Figure 3

Oxaliplatin treatment induces a pro‐inflammatory response to mice peripheral nervous system. (A) Representative western blots of NF‐kB p65 protein levels in DRGs of control and oxaliplatin‐treated mice at 8 and 10 weeks of the study. Graph below shows the corresponding quantification. n = 4–6 mice/group. Two‐way ANOVA with Bonferroni post hoc test. (B) Bar graphs showing the fold change of cytokine protein levels in DRGs (left) and sciatic nerves (right) of oxaliplatin‐treated mice at 8 and 10 weeks. Horizontal lines indicate fold increase equal 1 corresponding to control animals. n = 3–6 mice/group. Two‐way ANOVA with Bonferroni post hoc test. (C) Representative images of Iba1 staining (red) in the DRG of control and oxaliplatin‐treated mice at 8 and 10 weeks. β‐III‐tubulin (green) was used to label neurons. Scale bar: 20 μm. Graph in the right shows the quantification of Iba1 intensity (Integrated density). n = 3–4 mice/group. One‐way ANOVA with Bonferroni post hoc test. *p < 0.05 vs control **p < 0.01 versus control ***p < 0.001 versus control. All data are represented as group mean ± SD.

Figure 4

ICD is not activated in the DRG as a response to oxaliplatin treatment. (A) Left: quantification of the protein levels of ICD hallmarks (calreticulin, peIF2a, Hsp90, and Hsp70) in DRGs of control and oxaliplatin‐treated mice at 8 and 10 weeks. Right: representative western blots of each protein analyzed. n = 4–6 mice/group. Two‐way ANOVA with Bonferroni post hoc test. Data are expressed as fold change vs the average of the control group for each time point (group mean ± SD). (B) Representative images of calreticulin staining (CALR, red) in DRG of control and oxaliplatin‐treated mice at 8 and 10 weeks of the study. β‐III‐tubulin (green) was used to label neurons and nuclei were counterstained with Dapi (blue). White boxes are amplified in the images below (magnified view) for a better visualization of calreticulin staining. White arrows indicate calreticulin staining in the cell surface. Scale bar: 100 μm; Scale bar of magnified views: 20 μm. Graph in the right shows the quantification of the percentage of neurons that express calreticulin in their cell surface. n = 7–9 mice/group. One‐way ANOVA with Bonferroni post hoc test. All data are represented as group mean ± SD.

Figure 5

Oxaliplatin treatment induces a transient DNA damage in mice DRGs. Left: representative western blots of γH2AX and p21 protein levels in DRGs of control and oxaliplatin‐treated mice at 8 and 10 weeks of the study. Right: graphs with the corresponding quantification. n = 4–5 mice/group. Two‐way ANOVA with Bonferroni post hoc test. *p < 0.05. Data are represented as group mean ± SD.