Prolactin receptor expression in mouse dorsal root ganglia neuronal subtypes is sex-dependent

Abstract

Sensory neurons exhibit sex-dependent responsiveness to prolactin (PRL). This could contribute to sexual dimorphism in pathological pain conditions. The aim of this study is to elucidate mechanisms underlying sex-dependent PRL sensitivity in sensory neurons. Quantitative RT-PCR show that prolactin receptor (Prlr) long and short isoform mRNAs are expressed at comparable levels in female and male mouse dorsal root ganglia (DRG). In Prlr^cre/+;Rosa26^{LSL-tDTomato/+} reporter mice, percentages of Prlr⁺ sensory neurons in female and male DRG are also similar. Characterization of Prlr⁺ DRG neurons using immunohistochemistry and electrophysiology revealed that Prlr⁺ DRG neurons are mainly peptidergic nociceptors in females and males. However, sensory neuron type-dependent expression of Prlr is sex dimorphic. Thus, Prlr⁺ populations fell into three small- and two medium-large-sized sensory neuronal groups. Prlr⁺ DRG neurons are predominantly medium-large sized in males and are proportionally more comprised of small-sized sensory neurons in females. Specifically, Prlr⁺/IB4⁺/CGRP⁺ neurons are 4–5-fold higher in numbers in female DRG. In contrast, Prlr⁺/IB4⁻/CGRP⁺/5HT3a⁺/NPYR2⁻ are predominant in male DRG. Prlr⁺/IB4⁻/CGRP⁻, Prlr⁺/IB4⁻/CGRP⁺ and Prlr⁺/IB4⁻/CGRP⁺/NPYR2⁺ neurons are evenly encountered in female and male DRG. These differences were confirmed using an independently generated single-cell sequencing dataset. Overall, we propose a novel mechanism whereby sensory neuron type-dependent expression of Prlr could explain the unique sex dimorphism in responsiveness of nociceptors to PRL.

Introduction

Responsiveness to PRL in a variety of cell types depends on sex, estrous phase, pregnancy status and lactation (1–5). PRL signaling in cells is mediated by its cognate receptor, Prlr (6, 7). Prlr is the product of a single gene (8) that undergoes splicing in a region of the cytoplasmic domain generating short (Prlr-S) and long (Prlr-L) isoforms in rodents and humans (9). PRL-induced sensitization of transient receptor potential (TRP) and other ligand-gated channels in sensory neurons is sex-dependent (10, 11). It appears that PRL responsiveness is predominantly detected in female DRG and trigeminal ganglia neurons, and is linked to estrogen status (5, 10). This sex-dependency in sensitivity to PRL could be a driving force behind sexual dimorphism in pathological conditions. For example, global ablation of PRL or Prlr leads to a reversal in thermal (i.e. heat and cold) inflammatory and postoperative hypersensitivity/pain only in females via unknown mechanisms (10, 12). Hence, understanding mechanisms controlling sex-dependency of PRL sensitivity in neurons, including sensory neurons, are vitally important.

Overexpression of the Prlr-S isoform in male DRG neurons establishes their responsiveness to PRL to the level observed in females (4). This indicates that naïve male DRG do not express enough levels of functionally active Prlr to show modulation of neuronal activities by PRL. Therefore, the aim of the present study is to evaluate whether there is a sex-dependency in expression of Prlr mRNA in DRG neurons. Overall, our work strongly links sex dimorphism in PRL responsiveness in DRG neurons to differences in sensory neuron type-dependent expression of Prlr in females versus males. This paradigm may be critical for understanding sexual dimorphisms in pain pathophysiology.

Materials and Methods

Ethical Approval

All animal experiments conformed to APS’s Guiding Principles in the Care and Use of Vertebrate Animals in Research and Training, and to protocols approved by the University Texas Health Science Center at San Antonio (UTHSCSA) Animal Care and Use Committee (IACUC). We followed 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.

Mouse lines

Eight-to-twelve-week-old female and male mice were used for all described experiments. The Rosa26^{LSL-tDTomato/+} mouse line on B6.129 background was obtained from the Jackson Laboratory (Bar Harbor, ME). 5HT3a-GFP and TRPV1-GFP mouse lines were purchased from the GENSAT program (MMRRC services; UNC, NC and UC Davis, CA). The NPY2R-tdTomato mouse line was kindly provided by Dr. Xinzhong Dong (John Hopkins University Medical School, Baltimore, MD). The CGRP^cre/+-ER mouse line was kindly provided by Dr. Pao-Tien Chuang (UC San Francisco, San Francisco, CA). The Prlr^cre/+ mouse line (Prlr-cre) was made by Dr. Ulrich Boehm (University of Saarland School of Medicine, Homberg, Germany) using homologous recombination in mouse embryonic stem (ES) cells (13). The targeting construct was designed to insert an internal ribosome entry site (IRES) and the sequence for Cre recombinase immediately after exon 10 in the Prlr gene, with an FRT-flanked neomycin-resistance cassette for selection of clones.

Real-Time PCR

Real-Time RT-PCR was performed as previously described (10, 14). L3-L5 DRG total RNA was extracted using RNA Later (Qiagen, Valencia, CA, USA), the QIAzol lysis reagent and the RNAeasy Mini Kit (Qiagen), all were used per manufacturer’s instructions. cDNA was prepared using Superscript III First Strand Synthesis kit (Invitrogen, CA, California, USA). Amplification of target sequences was detected by a sequence detector ABI 7500 Fast RT-PCR system (Applied Biosystems, Foster City, CA) utilizing TaqMan Fast Universal PCR Master Mix and Prlr-S (Assay# Mm02017047_s1; amplicon size is 109bp within a single exon; Applied Biosystems) or Prlr-L isoform specific primers (Assay# Mm00619170_s1; amplicon size is 135bp within a single exon; Applied Biosystems) for mice tissues. The reactions were run in triplicates, including the endogenous control, mouse GAPDH (Assay# Mm03302249_g1; amplicon size is 70bp within several exons; Applied Biosystems), for each individual gene expression assay. For quantitative analysis, comparative delta–delta Ct (ddCt) was utilized to normalize the data based on the endogenous control (i.e. GAPDH), and to express it as the relative fold change, after the exclusion criteria were verified by comparing primer efficiencies.

Primary DRG neuronal culture

DRG from male and estrous phase females were used. Estrous phase was determine by vaginal gavage as described (15). Prlr^cre/+;Rosa26^{LSL-tDTomato/+} mice were deeply anaesthetized with isoflurane (0.3 ml in 1 liter administered for 60–90 sec) and sacrificed by cervical dislocation. L3-L5 DRG was quickly removed, and sensory neurons cultured as previously described (4). DRG neurons were dissociated by treatment with a 1mg/ml collagenase-dispase (Roche, Indianapolis, IN) solution. Cells were maintained in DMEM supplemented with 2% fetal bovine serum (FBS), 2mM L-glutamine, 100U/ml penicillin and 100μg/ml streptomycin. The experiments were performed within 6–24 h after DRG neuron plating.

Electrophysiology: Recording

Recordings were made in patch clamp whole-cell voltage (holding potential (V_h) of –60mV) or current clamp configurations at 22–24^oC as described (16, 17). Briefly, data were acquired and analyzed using an Axopatch 200B amplifier and pCLAMP10.6 software (Molecular Devices, Sunnyvale, CA). Recorded data were filtered at 0.5–5 kHz and sampled at 2–20 kHz depending on current kinetics. Access resistance (R_s) was compensated (40–80%) when appropriate up to the value of 6–8 MΩ. Data were rejected when R_s changed >20% during recording, leak currents were >200pA, or input resistance was < 100 MΩ. Currents were considered positive when their amplitudes were 5-fold bigger than displayed noise (in root mean square). Standard external solution (SES) contained (in mM): 140 NaCl, 5 KCl, 2 CaCl₂, 1 MgCl₂, 10 D-glucose and 10 HEPES, pH 7.4. The standard pipette (internal) solution (SIS) contained (in mM): 140 KCl, 1 MgCl₂, 1 CaCl₂, 10 EGTA, 10 D-glucose, 10 HEPES, pH 7.3, 2.5 ATP and 0.2 GTP. Drugs were applied by a fast, pressure-driven and computer controlled 4-channel system (ValveLink8; AutoMate Scientific, San Francisco, CA) with quartz application pipettes.

Electrophysiology: Protocols and Data Analysis

Electrophysiology protocols have been described in detail previously (16). Before patch clamp recording, DRG cells were stained for 0.5–4h with IB4 Alexa Fluor-488 or IB4 Alexa Fluor-594 (1:1000; Thermo-Fisher Scientific, Waltham, MA). Cells were selected according to Prlr/TdTomato or sensory neuronal marker expression (red or green), capacitance (pF) and IB4 staining (Fig 1A). On the selected DRG neuron, using SES and SIS solution, we used the following sequence of recording protocols: (1) a single AP in current clamp configuration was generated with a 0.5ms and 1nA (2nA for >40pF cells) current step (Fig 1B); (2, 3) after switching to voltage clamp configuration (V_h=−60mV), ATP (30μM) and then 5-HT (30μM) was sequentially applied for 5 sec (each) to evaluate I_ATP and I_5HT, respectively; (4) the next protocol for a current (18) was generated by step down from a V_h to −40mV kept for 500ms, and then 200-ms depolarizing command steps (20 mV) were applied to −40mV with a final potential of +20 mV (Fig 1C); and (5) finally, capsaicin (CAP; 100 nM) was applied for 30 sec for recording I_CAP. In some sets of experiments CAP was substituted with mustard oil (25μM; MO) to evaluate I_MO (or Ca²⁺ influx).

Figure 1: Electrophysiology protocols.

A: Female Prlr⁺ DRG neurons (red; blue arrow) selected for recording were classified by size (capacitance in pF) and positive or negative staining with IB4 (green; yellow arrow). B: Stimulus waveform (1,000 pA, 0.5 msec) generates single AP in female Prlr⁺ neuron which is indicated below trace. AP and AHP parameters (resting membrane potential – Vm; rise time – RT; fall time - FT; duration at base – dB and the time required for the AHP (measured in mV) to decay by 80% - (AHP₈₀) was measured as indicated. Characteristic “hump” on the falling phase of AP is indicated by black arrow. C: A current was generated in standard extracellular (SES) and intracellular solutions (SIS) by indicated below traces waveforms. The decay constant τ was derived from single exponential fits between points indicated by arrows to the final outward current trace (+20 mV). Typical “spike-like” peak observed in a majority of medium-large-sized Prlr⁺, CGRP-cre⁺, NPY2R-TdT⁺ and 5HT3a-GFP⁺ neurons by applied waveforms is marked with blue arrow on trace generated by stepping to +0 mV.

Data was accumulated from 6–7 independent mouse DRG neuronal cultures. Each culture was generated from one male or female. 20–24 neurons were recorded from each mouse. Chi-square analysis with Fisher’s exact test was performed to assess P values on the numbers of cells from each of the sensory neuronal groups between males and females (see Table 2). Diameter of cells was estimated from capacitance values using d=5x√(Cm/$4\pi )$; where Cm is capacitance in pF, d is diameter in μm. AP duration at the base (dB) as well as 80% recovery of AHP to baseline (AHP₈₀) were measured from data generated by protocol-1 (Fig 1B). Protocol-2, −3 and −5 reveal DRG neurons that respond to ATP (I_ATP), CAP (I_CAP), 5-HT (I_5HT) or MO as well as I_ATP characteristics. I_CAP, I_5HT and MO-induced response characteristics were not analyzed, since sequential recordings could have desensitized and changed kinetics and magnitude of these currents/responses. From the current protocol-4, the trace evoked by +20 mV was fit with a standard (i.e single) exponential function. Fitting and decay tau (τ; ms) calculation was performed using pCLAMP10.6 software (Fig 1C). Presence or absence of a “spike-like” feature on a current produced by steps to 0 and +20 mV was an important clustering variable (16, 18).

Table 2.

Properties of Prlr-cre⁺ sensory neuronal groups in females and males

Group	Sex	N+	P	Vm	Size (pF)	dB (ms)	AHP80 (ms)	τ	IATP (pA)	ICAP	IB4	CGRP	V1	I5HT3a	IMO
S2	F	12	0.6500 NS	−49.3±1.6NS	28.4±1.8***	6.0±0.3**	148.6±10.2****	5.9±1.0*	188±57	+	−	+	+	−	−
	M	9		−51.4±1.4	30.0±1.9	6.6±0.2	158.6±20.8	7.9±1.2	167±67
S3	F	29	0.0007***	−49.9±0.9NS	18.4±1.7NS	6.9±0.3****	123.6±13.9**	7.6±0.9*	−	+	+/−	+	+	+/−	−
	M	7		−48.7±1.0	16.2±1.4	6.7±0.4	133.1±16.6	4.9±0.8	−
S6	F	14	0.7000 NS	−49.8±1.6	13.1±0.9	4.5±0.3	40.6±9.6	19.4±2.8	−	+	−	−	+	−	+
	M	16		−49.3±2.0	13.4±1.3	4.1±0.2	30.3±5.9	19.1±2.0	−
S6a	F	12	0.0004***	−47.5±2.8NS	15.6±1.9NS	4.1±0.2NS	28.3±6.7NS	21.5±3.9NS	581±128	+	−	−	+	−	−
	M	0
ML1	F	13	0.0066**	−52.3±1.6NS	47.8±4.6****	5.1±0.3NS	107.5±10.2*	19.4±3.1NS	−	−	−	+	−	+	−
	M	29		−51.9±1.1	50.8±2.6	4.9±0.1	111.1±8.9	22.4±3.5	−
ML1a	F	6	0.0411*	−54.0±0.6NS	46.3±2.9****	4.5±0.5NS	78.0±11.3NS	36.1±13.6NS	316±74	−	−	+	−	+	−
	M	15		−52.3±0.9	49.0±6.0	4.6±0.3	67.0±5.2	25.7±5.7	513±20
ML3	F	17	0.3862 NS	−55.5±1.5NS	55.7±4.4****	3.8±0.2NS	158.4±20.7****	33.8±5.3*	−	−	−	+	−	+/−	−
	M	22		−55.8±0.7	59.4±2.1	3.6±0.1	168.2±16.4	34.0±6.8	−
ML3a	F	13	0.8405NS	−56.9±0.6*	66.8±4.2****	3.4±0.2NS	114.7±9.4*	31.2±6.6NS	483±112	−	−	+	−	+/−	−
	M	14		−56.7±1.1	65.8±7.1	3.4±0.3	88.46±23.5	24.4±4.2	398±82

⁺ − 15 male and 9 female Prlr⁺ neurons did not fit any group outline in Table 1.

Sign “+/−“ for IB4 indicates weak expression.

Sign “+/−“ for 5HT3a indicates weak (<400 pA) responses to 5-HT.

N⁺ is numbers of neurons; Vm is resting membrane potential; IB4 is staining to IB4; CGRP is expression of CGRP/TdTomato; V1 is expression of TRPV1-GFP.

Statistics:

P values on the numbers of cells are assessed by chi-square analysis with Fisher’s exact test for each of the functional groups between males and females.

Statistic for female data on Vm, Size, dB, AHP80 and tau is 1-way ANOVA comparing mean of each column with control column (S1), post-hoc test Bonferroni, NS p>0.05

^*p<0.05

^**p<0.01

^***p<0.001

^****p<0.0001

Statistic between specific groups is presented in the “Result” section.

Ca²⁺ imaging in DRG neurons

Fluorescent imaging was performed as previously described (17). Fluorescence was detected by a Nikon TE 2000U microscope fitted with a 20x/0.75 NA S Fluor objective. Data were collected and analyzed with NIS-elements software (Nikon Instruments, Melville, NY). The experiments were performed in SES solution (see Electrophysiology section) using calcium-sensitive dye Fura-2 AM (2 μM; Molecular Probes, Carlsbad, CA). The net changes in Ca⁺² influx were calculated by subtracting the basal [Ca⁺²]_i (mean value collected for 60 s prior to agonist addition) from the peak [Ca⁺²]_i value achieved after exposure to the agonists. Ca²⁺ imaging experiments were sometimes combined with patch-clamp recording to measure MO-activated responses.

Immunohistochemistry

DRG (L3-L5) from perfused female and male Prlr^cre/+;Rosa26^{LSL-tDTomato/+} mice were fixed again with 4% paraformaldehyde, cryoprotected with 30% sucrose in phosphate buffer, embedded in Neg 50 (Richard Allan Scientific, Kalamazoo, MI); and 30 μm cryo-sections were generated. Immunohistochemistry (IHC) was carried out as previously described (4). The following previously characterized primary antibodies were used: anti-TRPV1 guinea pig polyclonal (Neuromics; Bloomington, MN; catalogue GP14100; 1:700) (19); anti-CGRP rabbit polyclonal (Sigma; C8198; 1:300) (20); anti-tyrosine hydroxylase (TH) rabbit polyclonal (Pel-Freez; Rogers, AR; P40101; 1:400) (21); anti-mrgD rabbit polyclonal (Alamone Lab; ASR-031; 1:200) (22); anti-NPY2R rabbit polyclonal (Sigma; SAB4502029; 1:50); anti-5HT3a rabbit polyclonal (Invitrogen; PA1–41033; 1:100); anti-trkC goat polyclonal (R&D systems; Minneapolis, MN; AF1404; 1:200) (23); anti-trkB goat polyclonal (R&D systems; AF1494; 1:200) (23); rabbit anti-parvalbumin (Swant, PV25, 1:500) (23); and rabbit anti-Calbindin D28k (Swant, CB-38a, 1:500) (23). Sections were incubated with species appropriate Alexa Fluor secondary antibodies (1:200; Molecular Probes, Eugene, OR). Images were acquired using a Nikon Eclipse 90i microscope (Melville, NY, USA) equipped with a C1si laser scanning confocal imaging system. Images were processed with NIS-elements software (Nikon Instruments, Melville, NY).

Control IHC was performed on tissue sections processed as described but either lacking primary antibodies or lacking primary and secondary antibodies. IHC images were obtained from 3 independent tissue sections from 2–4 animals. IHC images for cell counting were from Z-stacks. Cell counts were performed using Image J software. Percentages of marker-positive cells from total or Prlr-cre⁺ numbers of cells were counted on each coverslip or section. Mean values from 3–4 sections or 9–12 coverslips generated from an animal represented data for this animal (i.e. on biological replicate). Numbers of animals are considered as the biological replicates. Intensity of immunoreactivity or TdTomato labeling was also calculated with Image J software; subtractions of background intensity from signal levels were applied.

Single Cell Sequencing Analysis

Single cell sequencing data was downloaded as expression values and metadata per cell (L5_all.loom) from mousebrain.org (Linnarsson Lab). Cells marked as coming from a single sex were picked for analysis. Among a total of 160796 cells, 46302 cells were marked as male, and 52996 cells were marked as female. Sex specific genes (Xist, Uty) expression level were checked for each group cells to confirm the accuracy of the sex label. Once these cells were sorted, dorsal root ganglion (DRG) sensory neurons were picked from these male and female sorted cells. Cluster IDs from the dataset were used to select sensory neurons. Cells labeled with PSNF (neurofilament), PSNP (non-peptidergic), PSPEP (peptidergic) were picked and separated to male and female groups for further analysis. Both male and female DRG neurons sequencing raw counts were imported and clustered in R using Seurat package v2.3.4 (24). Cluster results were compared with original clustering from Linnarsson lab to check for consistency. All genes counts were normalized to TPM (transcripts per million) for every cell. For every gene, average expression level, as well as the percentage of cells expressing the gene, were calculated for both males and females, for every cluster of cells, based on the output from Seurat package.

Statistics

GraphPad Prism 7.0 (GraphPad, La Jolla, CA) was used for all statistical analyses of data. Data in the figures are mean ± standard error of the mean (SEM), with “n” referring to the number of cells, the number of animals per group. Some of the experiments were performed in duplicate. Differences between groups were assessed by chi-square analysis with Fisher’s exact test, unpaired t-test, or 2-way ANOVA with Bonferroni’s post-hoc test. A difference is accepted as statistically significant when p<0.05. Interaction F ratios, and the associated p values are reported.

Results

Prlr mRNA is expressed at the same level in female and male mouse DRG

We have previously shown that female TRPV1⁺ and TRPA1⁺ sensory neurons are approximately 40-fold more sensitive to PRL than male neurons (4). Here, we evaluated transcription levels of the Prlr-L and Prlr-S isoforms in female and male mouse DRG tissues. Despite the dramatic difference in sensitivity to PRL, the data show statistically insignificant differences for Prlr-L (un-paired t-test; t=0.3971 df=9; P=0.7005; n=5–6; Fig 2A) and Prlr-S (un-paired t-test; t=0.7651 df=10; P= 0.4619; n=6; Fig 2B) mRNA levels in female versus male DRG. Mean thresholds for amplification of Prlr-L were 27 cycles and for Prlr-S 29 cycles, while the internal control (i.e. GAPDH) was 17 cycles.

Figure 2: Prlr mRNA expression levels and patterns in female and male DRG neurons.

Normalized mRNA expressions of Prlr-L (A) and -S (B) isoforms in female and male mouse DRG were assessed by quantitative RT-PCR (non-significant p>0.05; unpaired t-test; n=3–4). Prlr⁺ DRG neurons in Prlr^cre/+;Rosa26^{LSL-tDTomato/+} female (C) and male (D) mice. Yellow arrows mark small (<30μm) DRG neurons, blue arrows indicate medium-large (>30μm) sensory neurons, and green arrows non-neuronal DRG cells. E: Percentage of small and medium-large Prlr-cre⁺ DRG neurons in female and male mice (counts from 9–12 coverslips per mouse; n=3 mice per group; NS - non-significant P>0.05; *p<0.05; ***p<0.001; **** p<0.0001; 2-way ANOVA; sex and cell size as variables; cell size is “within subject factor”; Bonferroni’s post-hoc test).

To further characterize mouse female and male DRG sensory neurons expressing Prlr, we crossed a Prlr^cre/+ line with Rosa26^{LSL-tDTomato/+}. The resulting Prlr^cre/+;Rosa26^{LSL-tDTomato/+} mouse line shows Prlr⁺ cells in red in female (Fig 2C) and male (Fig 2D) DRG. Approximately equal numbers of male (10.92±0.43%) and female (11.47±0.37%) DRG neurons are Prlr⁺ (non-significant; unpaired t-test; n=27–36; 9–12 sections from three animals per group; t=0.5713 df=93; p=0.5692). However, separation of Prlr-cre⁺ neurons into two groups composed of small-sized <30μm and medium-large-sized >30μm neurons (Figs 2C, 2D) revealed that Prlr-cre⁺ male DRG neurons are predominantly medium-large, while female Prlr-cre⁺ neurons were evenly distributed among small and medium-large neurons (2-way ANOVA; F (1, 4) = 808.5; P<0.0001; Fig 2E). Some non-neuronal DRG cells also contain Prlr (Figs 2C, 2D). Culturing DRG cells from Prlr^cre/+;Rosa26^{LSL-tDTomato/+} mice showed that Prlr-cre⁺ non-neuronal cells exhibited fibroblast-like and macrophage-like morphology (data not shown). Overall, Prlr mRNA levels were not significantly different in female versus male DRG. However, our data suggest that Prlr mRNA could be expressed by different subsets of DRG neurons in female and male mice.

Female and male DRG sensory neuronal groups expressing Prlr mRNA

Female-selective in vitro (25) sensory neuronal responsiveness to PRL are not explained by differences in Prlr mRNA levels in female versus male sensory neurons (Figs 2A, 2B). However, differences in cell-type selective Prlr expression could be one of the mechanisms by which PRL exerts its sex-dependent behavioral effects. Accordingly, in the next set of experiments, we identified sensory neuronal groups expressing Prlr mRNA in female and male DRG neurons. To precisely identify Prlr-cre⁺ neuronal sub-types, electrophysiological profiling was used as recently described (16). Sensory neurons were clustered according to their size, IB4 labeling, neuronal marker expression in reporter mice, action potential (AP) and after hyperpolarization (AHP) parameters, responses to ATP, CAP, 5-HT and MO, and current generated by protocol-4 features (Fig 1C) (16).

Prlr⁺ DRG neurons were visualized using Prlr^cre/+;Rosa26^{LSL-tDTomato/+} reporter mice. 125 male (selected in the ratio 1 : 3 = small : medium-large) and 128 female (selected in ratio 1 : 1 = small : medium-large, based on distributions shown in Fig 2E) Prlr-cre⁺ DRG neurons were recorded with sequential protocols as described in the “Material and Methods” section (see also (16)). Approximately 95% of Prlr-cre⁺ neurons could be divided into 5 main groups/clusters (in bold) and two sub-clusters (in italic; Table 1). Prlr-cre⁺ neurons were assigned to one of these groups based on the criteria that every parameter, with the possibility of one exception, should fit the characteristics summarized in Table 1. Detailed electrophysiological profiles of Prlr-cre⁺ female and male DRG neuron groups are summarized in Table 2. Marking groups of small, medium and large-sized neurons (as S2, S3, ML1 etc.) was adopted from a previous publication (16).

Table 1:

Prlr-cre⁺ sensory neuronal group clusters

Group	Size (pF)	dB (ms)	FT:RT	AP shape	AHP80 (ms)	IB4	ICAP	I5HT (pA)	IATP	I; τ (ms)
S2	20–40	>6.5	>2:1	hump	>100	−	+	<200	<300pA Act: 2–5 s; Inact: none	−; <15
S3	10–40	>6	>2:1	hump	>50	+/−	+	<200	−	−; <15
S6	<20	2.5–5.5	>1:1	−	10–60	−	+	−	−	−; 10–30
ML1	>35	4–6.5	>2:1	hump	>60	−	−	>600	−	+; 10–40
ML1a	>35	4–6.5	>2:1	hump	>60	−	−	>600	<500pA Act: <1.5 s; Inact: 15–30 s	+; 10–40
ML3	>50	3–4.5	>1:1	deflection	>80	−	−	<400	−	+; 10–40
ML3a	>50	3–4.5	>1:1	deflection	>60	−	−	<400	<500pA Act: <1.5 s; Inact: 15–30 s	+; 10–40

Sign “+/−“ for IB4 indicates weak expression.

Groups with clearly detected “spike-like” feature for outward portion of current (I) is marked “+”; groups with no “spike-like” feature are marked “-“. The τ obtained after fitting with single or double exponential equation is noted (see “Material and Methods”).

Characteristic features of AP waveform on its return to baseline, such as “hump” and “deflection” (a.k.a. lesser pronounced “hump”) are noted, and are shown on Supplementary Figs 1B, 4B.

I_ATP - current size and kinetic parameters are noted. “Act” is activation time to reach 95% of peak. “Inact” is inactivation time for 50% decline during ATP delivery.

Small-sized (<35pF) DRG neurons

IB4-negative S2 Prlr-cre⁺ DRG neurons were almost equally represented in males and females (Table 2). S2 neurons have a slow AHP and broad AP with a pronounced “hump” on the falling phase (Table 2; Figs 3, 4A), a slow and small I_ATP and respond to CAP, but not MO (Fig 3). The S3 group among Prlr-cre⁺ DRG neurons was mostly represented in female but not male samples. S3 neurons have almost identical AP to S2, similar AHP, protocol-4 current parameters, and responded to CAP, but not MO; however, in contrast to S2, the S3 population did not respond to ATP or 5-HT application (Table 2; Figs 3, 4D). These S3 neurons are also weakly stained with IB4 (Table 1). Comparisons with previously characterized DRG neuronal clusters indicate that S2 and S3 are peptidergic small neurons (16), which could likely be assigned to the PEP1 group revealed by single-cell sequencing (26).

Figure 3: Electrophysiological profiles for sub-classes of Prlr-cre⁺ DRG neurons.

For each Prlr-cre⁺ sensory neuronal group recorded from female mouse DRG, AP, responses to ATP (30μM), capsaicin (100nM; CAP), 5-HT (30μM) and mustard oil (10μM; MO) are presented from left to right. The AP time scale (horizontal bar) is 5 msec for each panel. I_CAP time scale is indicated, and magnitude scale (vertical bar) is 100pA for each panel. I_5-HT time and magnitude scales are indicated. MO responses were measured by Ca²⁺ imaging. Prlr-cre⁺ sensory neuronal groups are indicated. Drug application times are illustrated by horizontal bar above traces. Additional information on subgroups is presented in Table 2.

Figure 4: Electrophysiology profile differences between Prlr-cre⁺ groups.

(A) Comparison of single AP in S3 and S6 Prlr-cre⁺ neuronal groups. (B) Comparison of single AP in ML1 and ML3 Prlr-cre⁺ neuronal groups. (C) Protocol-4 current (I) produced from Prlr-cre⁺ S6 group neurons. (D) I produced from the Prlr-cre⁺ S3 group neurons. (E) I produced from Prlr-cre⁺ ML1 group neurons. (F) I produced from the Prlr-cre⁺ ML3 group neurons. Time scales are indicated as horizontal bars.

Prlr-cre⁺ DRG neurons classified as S6 were almost equally represented in males and females (Table 2). S6 neurons were small-sized, IB4⁻ and had unusually fast AP and AHP compared to other small Prlr-cre⁺ DRG neurons (Table 2; Figs 3; 4A). S6 neurons mostly did not respond to ATP, but a sub-group of S6 neurons (termed S6a), which were found exclusively in females, produced a fast ATP-gated current (Table 2; Fig 3). S6 and S6a responded to CAP and MO, but not 5-HT (Fig 3). S6/S6a protocol-4 current (I) had a longer τ compare to other Prlr-cre⁺ small neurons (Table 2; Fig 4C). Characterization of marker-expression showed that the S6 neuron group is non-peptidergic and TRPV1⁺ (16), and likely corresponds to the NP3 cluster identified by single cell sequencing (26). In summary, Prlr-cre⁺ neurons belonged mostly to the S2 peptidergic and S6 non-peptidergic C-fiber groups and were similarly represented in males and females. In contrast, the S3 and S6a Prlr-cre⁺ neurons were predominantly present in female DRG. S1, S4 and S5 DRG neuronal groups, characterized previously (16), were not found in the Prlr-cre⁺ population in males or females.

Medium-large-sized (>35pF) DRG neurons

Prlr-cre⁺ ML1 and ML1a DRG neurons were twice as common in males compared to females (Table 2). ML1a’s only distinction from ML1 is that ML1a is responsive to ATP (Fig 3). Otherwise, ML1/ML1a neurons are medium-sized, are IB4⁻, insensitive to CAP and MO, and have different AP shape than S2-S3 neurons (1-way ANOVA; F (3, 35) = 2.242; P=0.1006) and slow AHP (Table 2; Fig 3). ML1/ML1a neurons have a clearly detectable “hump” on the AP falling phase (Figs 3; 4B). ML1/ML1a neurons have an obvious “spike-like” feature on protocol-4 current traces (Fig 4E), and τ is substantially longer than for S2/S3 neurons (1-way ANOVA; F (3, 43) = 15.31; P<0.0001; Table 2). A distinctive feature of ML1/ML1a neurons is their high sensitivity to 5-HT (I_5HT=1–4 nA; Fig 3). Classification of marker-expressing neuronal subsets suggests that ML1/ML1a neurons are probably medium-sized peptidergic neurons clustered as the PEP-2 group by single cell RNA sequencing (16, 26).

Prlr-cre⁺ ML3 and ML3a DRG neurons were almost equally present in females and males (Table 2). ML3 is distinct from ML3a in the same way as ML1 from ML1a: its responsiveness to ATP (Fig 3). Prlr-cre⁺ ML3/ML3a had many comparable characteristics to ML1/ML1a: no IB4 staining, slow AHP, and no CAP and MO responsiveness (Fig 3). But, unlike ML1/ML1a, a majority of Prlr-cre⁺ ML3/ML3a neurons did not respond to 5-HT (Fig 3). The most distinct feature of ML3/ML3a neurons was their significantly faster AP compare to S2, S3 and ML1/ML1a with no “hump”, but a slight deflection of the falling phase of the AP (1-way ANOVA; F (4, 51) = 15.25; P<0.0001; Figs 3, 4B). ML3/ML3a is likely a subgroup of medium-sized peptidergic PEP-2 neurons that uniquely express the Npy2r gene (16, 23, 26). Altogether, ML1 and ML3 Prlr-cre⁺ neurons could be classified as peptidergic, myelinated A-fiber sensory neurons. Among Prlr-cre⁺ neurons, while ML3/ML3a was equally represented in female and male cells, ML1/ML1a was encountered twice as often in male than female neurons (Table 2).

Expression of sensory neuronal markers in female and male Prlr-cre⁺ DRG neurons

To further corroborate our electrophysiology data, we used IHC to examine co-expression of sub-population markers with Prlr-cre⁺ neurons in female and male Prlr^cre/+;Rosa26^{LSL-tDTomato/+} L3-L5 DRG sections. Consistent with our electrophysiology analysis, the majority of Prlr-cre⁺ DRG neurons were peptidergic as CGRP expression was observed in ≈ 80% of Prlr-cre⁺ neurons in female and male DRG (2-way ANOVA; F (3, 24) = 5.049; P=0.0075; Figs 5A; 5C). The CGRP⁻/TRPV1⁺ population that represents the S6/NP3 group (16, 26) comprised ≈ 12–15% of female and male Prlr-cre⁺ DRG neurons (Figs 5A; 5C). Prlr-cre⁺/TRPV1⁺ neurons were twice as abundant in female compared to male DRG (2-way ANOVA; p= 0.0032; Figs 5A; 5C). In concordance with our electrophysiological data, female DRG had 3-fold more CGRP⁺/TRPV1⁺/Prlr-cre⁺ neurons than male DRG (mean values are 31.585 vs 9.332%; 2-way ANOVA; p= 0.0002; Table 2; Fig 5). These neurons account for the greater number of S2 and S3 groups found in female Prlr-cre⁺ mice. NPY2R⁺/Prlr-cre⁺ neurons were almost equal in female and male DRG (mean values are ≈ 26 vs 31%; Fig 5B). 5HT3a, which is expressed in both S3 and ML1/ML1a neurons (Fig 3; (16, 26), was expressed equally (≈25–27%) in Prlr-cre⁺ neurons of female and male DRG (Figs 6A–6C). This is consistent with electrophysiology recording showing that summed numbers of S3 and ML1/ML1a Prlr-cre⁺ neurons expressing 5HT3a are approximately equal in female vs male even though they fall in different sub-populations between the sexes (Table 2).

Figure 5: CGRP, TRPV1 and NPY2R expression in female and male Prlr-cre⁺ DRG neurons.

(A) Immunohistochemistry (IHC) of L3-L5 DRGs from female Prlr^cre/+;Rosa26^{LSL-tDTomato/+} mice to examine co-expression of Prlr-cre⁺ neurons (red) in CGRP⁺ peptidergic (green) and TRPV1⁺ (blue) nociceptors. Examples of Prlr-cre⁺/CGRP⁺ cells are marked with yellow arrows; Prlr-cre⁺/TRPV1⁺/CGRP⁻ cells are marked with white arrows and Prlr-cre⁺/TRPV1⁺/CGRP⁺ cells are marked with a sapphire arrow. (B) Immunostaining of L3-L5 DRGs from female mice to examine co-expression of Prlr-cre⁺ neurons (red) in NPY2R⁺ (green) neurons. White arrows show Prlr-cre⁺/NPY2R⁺ cells and sapphire arrows mark Prlr⁺/NPY2R⁻ cells. (C) Percentage of Prlr-cre⁺ DRG neurons that co-express markers (i.e. CGRP, TRPV1, TRPV1/CGRP⁻ and NPY2R) for the defined subpopulation of sensory neurons in female and male mice. Two-way ANOVA, sex and cell type as variables; cell type is “between subject factor”; Bonferroni’s post-hoc test; **p<0.01; counts are from 3 sections per mouse; n = 4 mice per group).

Figure 6: 5HT3a expression in Prlr-cre⁺ DRG neurons.

(A) IHC of L3-L5 DRGs from female Prlr^cre/+;Rosa26^{LSL-tDTomato/+} mice to examine co-expression of Prlr-cre⁺ neurons (red) in 5HT3a⁺ (B; green) nociceptors. Examples of Prlr-cre⁺/5HT3a⁺ cells are marked with blue arrows; Prlr-cre⁺/5HT3a⁻ cells are marked with white arrows. (C) Percentage of Prlr-cre⁺ DRG neurons that co-express 5-HT⁺ in female and male mice. Statistic is unpaired t-test, t=0.6261 df=8; P=0.5487; n=3–4 mice with 2–3 section per mouse.

We next evaluated whether Prlr is co-expressed with other markers for nociceptive and non-nociceptive (i.e. low threshold mechano-receptors; LTMRs) sensory neuronal groups. IHC analysis showed that Prlr-cre⁺ neurons that fall into the C-fiber class are exclusively C-fiber nociceptors, as they do not co-express tyrosine hydroxylase (TH; Fig 7A), a marker for C-LTMR (27). We also found that Prlr-cre⁺ neurons do not co-express MrgD (Fig 7B), which is a marker for non-peptidergic C-fiber nociceptors (28). Parvalbumin (PV), a proprioceptor marker, and TrkC, which is expressed in some cutaneous and muscle Aβ-LTMRs and proprioceptors (29), was also not detected in Prlr-cre⁺ neurons (Figs 7C; 7F). In electrophysiology experiments, 9 medium-to-large-sized Prlr-cre⁺ neurons in males and 3 in females did not correspond to any sensory neuronal groups described in Tables 1 and 2. These neurons had very fast APs and fast AHPs. This implies that these neurons may belong to the LTMR class of neurons (30). Indeed, using IHC, a few Prlr-cre⁺ neurons co-expressed calbindin, a marker of rapidly-adapting (RA) Aβ-LTMRs (Fig 7D) (31), and trkB, a marker for Aδ-LTMRs (Fig 7E) (32).

Figure 7: Co-expression of other neuronal markers with Prlr-cre⁺ DRG neurons.

IHC of L3-L5 DRGs from female Prlr^cre/+;Rosa26^{LSL-tDTomato/+} mice to examine expression of Prlr-cre⁺ (red) in (A) TH⁺ marker for C-LTMRs; (B) mrgD⁺ marker for non-peptidergic C-fiber nociceptors; (C) PV⁺ marker for proprioceptors; (D) Calb⁺ marker for Aβ-RA-LTMRs; (E) trkB⁺ marker for Aδ-LTMR; and (F) trkC⁺ marker for Aβ-LTMRs and proprioceptors. An example of Prlr-cre⁺/Calb⁺ and Prlr-cre⁺/trkB⁺ cells are marked with blue arrows (panels D and E).

Finally, we used existing single cell sequencing data from mouse DRG neurons to assess expression of Prlr mRNA in different clusters of sensory neurons. First, we clustered male and female sensory neurons and made tSNE plots (Fig 8A). With this approach we found 12 clusters that roughly match those described previously (33). We then examine Prlr expression in each of these sub-clusters. We found 4 major clusters expressing Prlr, and 3 of these showed some level of sexual dimorphic expression. Cluster 9, which matches characteristics of the ML1a subgroup, showed higher proportions of expressing neurons in males, matching the electrophysiology data (Fig 8B). Cluster 7, matching the S6 and S6a groups, showed substantially higher expression in female neurons, again matching electrophysiology findings for the S6a cluster. There was also higher Prlr expression in the 4^th cluster, this one matching group S3. Again, this matched electrophysiology findings. We identified a final cluster with Prlr expression (cluster 2) where an equal proportion of neurons expressed Prlr although overall expression was higher in females. This 2^nd cluster shows some, but not all, characteristics of S2, where we did not find differences in Prlr expression in physiology studies. Collectively, this re-analysis of single cell RNA sequencing data, stratified by sex, supports conclusions using electrophysiology and genetic tracing techniques.

Figure 8: Single cell RNA sequencing analysis demonstrates sex-specific expression of Prlr mRNA.

(A) t-Distribution Stochastic Neighbor Embedding (t-SNE) plots are shown for male and female DRG neuron single cell clustering from data at mousebrain.org. Clustering on the left shows cell types divided by sex in 11 clusters and clustering in the right panel labels the clusters by number (matching cluster ID (Clst.ID) in panel B). (B) Split dot plot of marker genes and Prlr across the 11 clusters with expression level and number of cells represented. Clusters identified by electrophysiological measures are matched (S2, S3, S6a and ML1a) and show differential expression of Prlr in males and females in corresponding clusters.

Overall, our electrophysiology, single-cell database analysis and IHC results support the conclusion that female and male Prlr-cre⁺ neurons are predominantly C- and A-fiber nociceptors, a majority of which are peptidergic. There is an important sexual dimorphism amongst these populations wherein the highest percentages of Prlr-cre⁺ neurons in female DRG are from the C-fiber nociceptive S3 group (CGRP⁺/TRPV1⁺) while in male DRG they are from the A-fiber nociceptive ML1/ML1a group (5HT3a⁺/NPY2R⁻/CGRP⁺). Since function of C and A-fiber neurons in nociception substantially differ (34, 35), such sex differences in sensory neuron type-dependent distribution of Prlr could affect differential responsiveness to PRL in female versus male sensory neurons and nociception.

Discussion

Multiple studies in humans and animals have demonstrated that Prlr-mediated PRL effects are sex-dependent in many tissues and cell types (4, 6, 7, 12). Sex-dependent regulation of Prlr expression has been studied primarily in non-neuronal cells, such as liver and mammary gland epithelial cells (36–39). Understanding sex-specific regulation of Prlr in neurons is essential in explaining the role of PRL in neurological diseases and pathological conditions, including nociceptive processing (5, 10).

We and others have reported that PRL responsiveness is at least 10 fold higher in females compared to male DRG and TG sensory neurons (4, 10, 11). Surprisingly, Prlr-L and Prlr-S mRNA expression patterns and levels in DRG neurons were not sex-dependent (Fig 2). Prlr has several promoters, which do not have classical gonadal-response elements (40). Nevertheless, Prlr mRNA is regulated by E2 in non-neuronal cells by utilizing other transcription binding sites, such as C/EBP, Sp3 and/or Sp1A (41, 42). There is no consensus on whether neuronal Prlr mRNA is regulated by gonadal hormones and during estrous phases. Our data show that despite female-selective PRL sensitivity in DRG neurons (4, 43), Prlr-L and Prlr-S mRNA expression is not sex-dependent. It was proposed that neuronal Prlr sex dimorphic function could be regulated via mechanisms involving cell type-dependent utilization of multiple Prlr promoters (44, 45). Thus, the Prlr gene has multiple 5’-untranslated regions (UTRs) and corresponding promoters (36). Cell type-dependent expression of Prlr mRNA may, then, be defined by cell type-specific utilization of Prlr gene promoters (36, 37). Moreover, selection of different Prlr promoters could be controlled by gonadal hormones. This mechanism in regulation of Prlr expression by estrogen was reported in the liver (36, 39). If such a mechanism is also applied to the regulation of sensory neuronal Prlr mRNA, then sex-specific operation of multiple Prlr promoters could occur in particular types of sensory neurons. Indeed, we showed that Prlr expressing subsets in DRG differ in female versus males (Table 2, Figs 3–7). This indicates that Prlr mRNA expression could be sex-dependently controlled via post-transcriptional mechanisms in sensory neurons.

Sex-dependent expression of Prlr in unique functional subsets of sensory neurons is of considerable interest. These neuronal subsets could have different roles in pain and nociceptive transmission during a variety of pathological conditions (46). We showed, using 3 independent methods, electrophysiology, IHC and single cell RNA sequencing, that Prlr expression is different in subsets of mouse DRG neurons in females versus males. However, this difference was only seen in females where Prlr expression was 4-fold more abundant in the S3 and S6a groups (i.e. C-nociceptors), while male Prlr⁺ neurons were more prominent in the ML1/ML1a group (i.e. Aδ-HTMRs). Other Prlr⁺ groups were almost equally represented in female versus male mouse DRG. Thus, reported female-selective regulation of C-fiber neuron-expressing TRPV1 and TRPA1 channels by PRL could be explained by the predominant presence of Prlr mRNA in these neurons (Table 2) (25, 43).

In conclusion, based on the findings that PRL responses in sensory neurons are totally blocked in Prlr KO (4), that PRL only selectively sensitizes TRP channels in female sensory neurons (25) and the data presented here, we favor the hypothesis that sex-dependency of PRL responsiveness in sensory neurons can be explained by sex dimorphic sensory neuron type-dependent expression of Prlr in C- and A-fiber peptidergic nociceptors. These results advance our understanding of sex dimorphisms in pain signaling and provide evidence for the critical role that transcriptional regulation via alternative promoters could play in setting nociceptor excitability in response to a broad variety of important physiological stimuli.