Cisplatin Neurotoxicity Targets Specific Subpopulations and K+ Channels in Tyrosine-Hydroxylase Positive Dorsal Root Ganglia Neurons

Abstract

Among the features of cisplatin chemotherapy-induced peripheral neuropathy are chronic pain and innocuous mechanical hypersensitivity. The complete etiology of the latter remains unknown. Here, we show that cisplatin targets a heterogeneous population of tyrosine hydroxylase-positive (TH⁺) primary afferent dorsal root ganglion neurons (DRGNs) in mice, determined using single-cell transcriptome and electrophysiological analyses. TH⁺ DRGNs regulate innocuous mechanical sensation through C-low threshold mechanoreceptors. A differential assessment of wild-type and vitamin E deficient TH⁺ DRGNs revealed heterogeneity and specific functional phenotypes. The TH⁺ DRGNs comprise; fast-adapting eliciting one action potential (AP; 1-AP), moderately-adapting (≥2-APs), in responses to square-pulse current injection, and spontaneously active (SA). Cisplatin increased the input resistance and AP frequency but reduced the temporal coding feature of 1-AP and ≥2-APs neurons. By contrast, cisplatin has no measurable effect on the SA neurons. Vitamin E reduced the cisplatin-mediated increased excitability but did not improve the TH⁺ neuron temporal coding properties. Cisplatin mediates its effect by targeting outward K⁺ current, likely carried through K2P18.1 (Kcnk18), discovered through the differential transcriptome studies and heterologous expression. Studies show a potential new cellular target for chemotherapy-induced peripheral neuropathy and implicate the possible neuroprotective effects of vitamin E in cisplatin chemotherapy.

Keywords: cancer; chemotherapy; neuropathy; sensory; vitamin E.

FIGURE 1

TH1 and TH2 are distinct subpopulations of TH⁺ DRGNs of adult mice. (A) Using single-cell RNA-sequencing data from our previously published study (Finno et al., 2019), a heat map of the top five unique genes defining each DRGN subpopulation was generated to clearly define the differences in transcriptional profiles of the two TH⁺ DRGN subpopulations (TH1 and TH2; red). (B) TSNE plots demonstrating overall expression of Slc17a8 (VGluT3) and Fam19a4 (TAFA4). (C) Differential transcript expression of K⁺ two-pore domain subfamily K channels (Kcnk2, Kcnk3, Kcnk12, and Kcnk18) across the DRGN subpopulations, with Kcnk18, differentially expressed across five subpopulations [peptidergic 1 (PEP1), peptidergic 2 (PEP2), TH1, TH2, unknown (UNK)]. (D) TSNE plots demonstrating overall expression of Pla2g7 (TH2 > TH1), Kcna4 (TH1 > TH2), and Kcnd3 (TH1 > TH2) from single-cell RNA-sequencing (Finno et al., 2019). DE, differentially expressed; NF, neurofilament; NP, non-peptidergic; PEP, peptidergic; TH, tyrosine hydroxylase; UN, unknown cluster. n = 2 mice per group with ∼3,600 cells/mouse profiled. (E) Single-molecule fluorescence in situ hybridization (smFISH) with RNAscope from TH-EGFP+ mice, demonstrating expression of Kcna4, Kcnd3, and Pla2g7 in TH⁺ DRGNs. Green: TH⁺ DRGN, Blue: DAPI nuclear stain, Red: Kcna4 (first row), Kcnd3 (second row), and Pla2g7 (third row) mRNA. Scale bars represent 5 μm. (F) Quantification of mRNA using RNAscope for Kcna4, Kcnd3, and Pla2g7. Within TH⁺ DRGNs, Kcnd3 was the most highly expressed. Mean ± SD, N = 11 neurons per experimental group (3 mice in each group), one-way ANOVA, or Kruskal-Wallis. Compared to GAPDH (not shown) one-way ANOVA, *p < 0.05, ^****p < 0.0001.

FIGURE 2

Functional differences in response properties of TH⁺ DRGNs. (A–C) Representative AP traces were recorded from TH-EGFP+ transgenic mice DRGNs. (A) Fast-adapting TH⁺ DRGNs elicit 1-AP upon current injection (0.2 nA for the example shown). (B) Moderately-adapting TH⁺ DRGNs, eliciting ≥2-APs (0.2 nA). (C) Spontaneously active (SA) TH⁺ DRGNs. (D–F) Whole-cell K⁺ currents in 1-AP (D), ≥2-APs (E), and SF (F) TH⁺ DRGNs. Voltage-clamp recordings of TH⁺ DRGNs following current-clamp assessment. TH⁺ DRGNs were held at −90, and −40 mV stepped from −100 to 40 mV using 10-mV increments. To differentiate between ion channels that have different activation potentials, the insets show a difference-current trace using 40-mV stepped potential between neurons held at −90 and −40 mV. (G,H) Summary data for the current-voltage (I/V) relationship showing differences in the current densities between the three classes of TH⁺ DRGNs. TH⁺ DRGNs with 1-AP (in black symbols), ≥2-APs (in red symbols), and SF (in blue symbols). (*p < 0.05; **p < 0.01; n = 15 neurons from four mice).

FIGURE 3

Electrical stimulation of TH⁺ DRGNs with oscillatory current injections. (A) Upper panel: polar plots for TH⁺ DRGNs with a 1-AP-elicited square pulse. Below are representative membrane responses of TH⁺ DRGNs to oscillatory current injection (0.2 nA) at 0.4, 5, and 10 Hz. (B) Row of similar polar plots derived from TH⁺ DRGNs responding with ≥2-APs after square pulse injection (0.2 nA) to 0.4, 5, and 10 Hz oscillatory currents. (C,D) Summary data of the two sets of TH⁺ DRGNs, showing the relationship between oscillatory current injection (0.4, 5, and 10 Hz, 0.2 nA) and spike frequency (open and solid symbols represent data from neurons with 1 AP, and ≥2 AP response features, respectively). (D) Data were obtained from the same neurons as in (C), with computed vector strength (VS). 1-AP TH⁺ DRGNs consistently showed higher VS, indicating better temporal coding. Each symbol represents a different TH⁺ DRGNs (n = 9 for 1-AP, n = 6 for ≥2-APs TH⁺ DRGNs).

FIGURE 4

Cisplatin differentially increased membrane excitability of TH⁺ DRGNs but reduced their temporal coding properties. (A) Time-dependent alterations of the input resistance of TH⁺ DRGNs after application of 2 μM cisplatin. Black and blue symbols represent TH⁺ DRGNs that elicit 1-AP (n = 5) and ≥2-APs (n = 5) upon current injection. In purple are spontaneously active neurons (n = 4). (B) Exemplary effects of cisplatin on rectangular-shaped current-injected APs on two distinct TH⁺ DRGNs [1-AP (upper, shown in black traces, and ≥2-APs (lower, shown in blue traces, panels)]. (C) Whole-cell currents from 1-AP (black) and ≥2-APs TH⁺ DRGNs (blue) show cisplatin-mediated outward currents reduction. The left lower panel shows the corresponding current density (pA/pF) and voltage relation (n = 9). For outward currents elicited from −70 to 0 mV, cisplatin reduced the current by ∼23% for 1-AP TH⁺ DRGNs and ∼30% for ≥2-APs TH⁺ DRGNs. The middle-lower panel shows the difference in current, or the cisplatin-sensitive current, which appears to exhibit time-independent properties. The inset shows current from 1-AP TH⁺ DRGNs (black), and spontaneously active TH⁺ DRGNs (purple). Summary data from currents generated from a holding potential of −70 mV to step potential of 40 mV. Mean for 1-AP = 52.1 ± 6.23 and spontaneously active neuron = 32.8 ± 6.09 (****p = 5.6 × 10^–6; n = 9 neurons from three mice). (D) Representative response properties of 1-AP TH⁺ DRGNs generated by simultaneous sinusoidal current injection (0.2 nA, 20 Hz). The amplitude criterion of a valid AP was 0 mV overshoot (dashed line). The lower panels summarize spike frequencies and vector strength (VS) changes before and after cisplatin (1 mM) application. Data were collected using 5 Hz (n = 8) and 20 Hz (n = 7) sinusoidal current (***p < 0.001).

FIGURE 5

Vitamin E reduces cisplatin-mediated increased excitability of TH⁺ DRGNs but does not improve the coding properties. (A) Summary of the effects of cisplatin on the input resistance of TH⁺ DRGNs after application of 1.5 μM cisplatin and the subsequent effects of the application of vitamin E (vitE, 100 μM) and 1.5 μM cisplatin. Differences between control and the treatment were tested using one-way ANOVA. ^#p = 3.75 × 10^–4 and *p = 2.32 × 10^–2 (n = 12). (B) Dose-response relation on cisplatin-mediated increased spike frequency and applied cisplatin concentrations. A Hill coefficient of 2 and EC50 of 1.9 ± 0.3 μM (n = 4) was estimated. The inset on the right shows the effects of 0.5 μM cisplatin. (C) Effects of cisplatin (0.5 μM) and vitE (100 μM) on the response properties of TH⁺ DRGNs generated by injecting sinusoidal current (0.2 nA, 10 Hz). (D) The relation between spike-frequency in control, cisplatin and after cisplatin/vitE application (n = 6). Data were assessed from TH⁺ DRGNs that generated 1-AP in response to a square pulse. (E) The corresponding computed VS, cisplatin, and vitamin E reduce the VS (n = 6).

FIGURE 6

Cisplatin reduced induced membrane potential depolarization and reduced action potential threshold by suppressing a resting current. (A) Left panel. Using a 5-ms current injection of different amplitudes, membrane depolarizations and action potentials were generated. The strategy was used to determine the action potential threshold before and after applying 1 μM cisplatin (right panel). The dashed line indicates the 0-mV level, and the dotted line represents the resting membrane potentials. Note that cisplatin mediated ∼5–8-mV membrane depolarization. (B) Average and raw data of the effective rheobase evaluated from (A) plotted for control and after application of cisplatin (1 μM). p-values for statistical comparison are shown, and statistical significance is indicated with an asterisk (***p < 0.001, n = 14 neurons from three mice). (C) Exemplary action potentials were generated using 150 pA current injection for control (in black) and after cisplatin application (in blue). (D) Phase plot (dV/dt) of the action potentials in C comparing controls (in black) and cisplatin effect (in blue). The membrane voltage threshold is determined from the phase plots for control and after cisplatin application. Cisplatin reduced the threshold voltage significantly (***p < 0.001) as summarized in (E) (n = 13 from three mice). (F) Current traces showing generated with a voltage-ramp from −70 mV holding voltage to −30 mV for control (black) and in the presence of cisplatin (1 μM, blue) followed by application of solution containing 1 μM cisplatin and 100 μM vitE (red). Cisplatin reduced the holding current, which was reversed with vitE application. Summary of the measured holding current at −70 mV holding voltage in control, after cisplatin and vitE (***p < 0.001, n = 16 from four mice).

FIGURE 7

Effects of vitE pre-incubation on cisplatin-mediated alterations on expressed K_2P18.1 and K_v1.4 in HEK 293 cells. Outward K⁺ currents from HEK 293 cells transfected with the mouse K_2P18.1 and K_v1.4 plasmids. (A) K_2P18.1 channel expression; the recordings were obtained from a holding potential of −80 mV and stepped to −100 mV with a ramp to 60 mV within 500 ms (in black) or a square pulse ranging from −70 to 60 mV (ΔV = 10 mV). The insets show square-pulse elicited current. The red- square and ramp traces depict the effects of cisplatin application (10 μm). Cisplatin (10 μM) blocked ∼40% of the current. (B) In contrast, pre-incubation (∼2 h) in 10-μM vitE sufficed to abolish the effects of cisplatin on K_2P18.1 current (blue). I-V relationship in response to a ramp stimulus. (C) Summary data of effects of cisplatin and after vitE pre-incubation obtained from n = 14 cells. (D–F) Outward K⁺ currents were elicited after K_v1.4 transfection in HEK 293 cells. Cisplatin produced ∼45% reduction of the current density (E). However, pre-incubation of vitE did not reverse the cisplatin-mediated reduction of the K_v1.4-mediated outward current (F; n = 12 cells) (***p < 0.001).