Uterine Cervical Afferents in Thoracolumbar Dorsal Root Ganglia Express Transient Receptor Potential Vanilloid Type 1 Channel and Calcitonin Gene–related Peptide, but Not P2X3 Receptor and Somatostatin

The study was approved by the Animal Care and Use Committee of Wake Forest University School of Medicine (Winston-Salem, North Carolina), and experiments followed the animal research ethical guidelines of the International Association for the Study of Pain. Female Sprague-Dawley rats were housed two per cage in our central animal facility, with free access to food and tap water ad libitum  and with a 12 h:12 h light–dark cycle.

Anesthesia was induced with 5% and maintained with 2% halothane in oxygen with spontaneous ventilation. A laparotomy was performed via  a midline incision in the lower abdomen. A bilateral ovariectomy was performed, and then a pellet containing either estrogen or placebo was implanted subcutaneously. Intact animals served as a control group; estrous stage was not determined at the time of surgery. The effectiveness of estrogen replacement was confirmed by visual inspection of the uterine cervical tissue at the time of death. Typically, ovariectomized animals showed a small, dense, and pale cervix, whereas estrogen supplemented rats showed a red, soft, and enlarged cervix.

A fluorescent retrograde neuronal tracer (4% fluorogold) was injected into the dorsal surface of the uterine cervix and the lower uterine segment at multiple sites (2 μl each, maximum 8–10 μl) using a 10-μl Hamilton syringe and a 30-gauge needle. Overflow of tracer solution was rinsed with sterile saline, and the injection sites were dried with a cotton swab immediately and then coated with Dermabond (ETHICON, Somerville, NJ) to prevent spread of dye. The incision was closed using 4-0 silk suture. In addition, two rats received a bilateral hypogastric neurectomy before fluorogold labeling. At the end of surgery, 0.5% bupivacaine was infiltrated in the incision for postoperative pain control. Postoperatively, animals showing any signs of infection or hematoma or impairment of water and food intake resulting in significant weight loss were killed and not included in the study.

Twenty-one days after surgery, animals were anesthetized with sodium pentobarbital (100 mg/kg, intraperitoneal) and perfused transcardially with 250 ml phosphate-buffered solution, 0.01 m, containing 1% sodium nitrite and then with 500 ml of 4% paraformaldehyde in 0.1 m phosphate-buffered solution. Bilateral DRGs at the T12–L2 levels were removed and postfixed in the same fixative for 4 h and then transferred into 30% sucrose in phosphate-buffered solution for cryoprotection for 48 h. A pair of DRGs from each level was mounted in Tissue-tek O.C.T. (Sakura Finetechnical Co., Torrance, CA) embedding medium and cryostat sectioned at 16 μm. Sections were thaw-mounted onto poly-l-lysine–coated slides (Fisher Scientific, Pittsburgh, PA) and stored in a −80°C freezer until use.

After thawing, slides were washed with 0.001 m phosphate-buffered solution–triton (pH 7.4) four times and blocked with 1.5% normal donkey serum for 1 h. Slides were then incubated with one of the following specific antibodies: TRPV1 (guinea pig, 1:2,500; Neuromics, Northfield, MN), P2X3 (guinea pig, 1:1,000; Neuromics), CGRP (rabbit, 1:1,000; Bachem, King of Prussia, PA), or somatostatin (rabbit, 1:500; Santa Cruz, Santa Cruz, CA) overnight at 4°C in a humidified chamber. After washing with phosphate-buffered solution–triton (2 × 5 min), the slides were incubated with donkey species-specific tetramethyl rhodamine–labeled secondary antibodies (Jackson, West Grove, PA) diluted 1:200 in 1.5% normal donkey serum for 1 h. The slides were then coverslipped with Vectashield Mounting Medium (Vector, Burlingame, CA). Two control experiments were performed to validate the specific staining for each antibody: (1) incubation with specific blocking peptide (obtained from the same source company, except somatostatin, which was not performed) and (2) omitting the primary antibody.

Only sections with fluorogold-labeled cells were selected for quantification. Every third section was used to avoid double counting, and effort was made to obtain three sections from each spinal level. Only cells with intact structure and observable nucleus were counted.

Images were captured with Q-image under epifluorescence microscopy (Nikon E600; Tokyo, Japan) and with proper filters (UV-2A for fluorogold and G-2E/C for tetramethyl rhodamine). Images were later analyzed using Sigma ScanPro5.0 (SPSS, Chicago, IL) by a person blinded to animal treatment. Calibrations were performed for each staining experiment to determine an optical density threshold for each specific positive staining on a 0–255 gray scale. This was done by comparing the level of fluorescence of cells incubated with primary antibodies to cells on slides that had been processed without primary antibodies. Cell diameter was determined by averaging the measured long and short axis distance through the nucleus.

Halothane was obtained from Halocarbon Laboratories (River Edge, NJ); 17β-estradiol (1.5 mg/pellet, 21-day sustained release) and placebo pellets were obtained from Innovative Research of America (Sarasota, FL). Fluorogold-hydroxystilbamidine methanesulfonate was obtained from Molecular Probes (Eugene, OR). Bupivacaine HCl was obtained from AstraZeneca (Wilmington, DE).

Dorsal root ganglion sections from T12–L2 were quantified for TRPV1 and P2X3 immunoreactivity. Because the fluorogold-labeled neurons were found mostly in L1 and the frequency of colabeling of TRPV1 with fluorogold-traced neurons did not vary with spinal levels, only the L1 DRG sections were used for CGRP and somatostatin immunostaining. An average of 7–12 sections for TRPV1 and P2X3, and 3 sections for CGRP and somatostatin from each animal was used for data analysis. Data are expressed as mean ± SEM. The percentage of specific staining for fluorogold-traced and nontraced neurons within a group was compared by chi-square, whereas the difference among the groups was determined using one-way analysis of variance. P < 0.05 was considered statistically significant.

A total of 26 rats were studied (ovariectomy = 9, estrogen supplement = 8, intact = 7, bilateral hypogastric neurectomy = 2), and 727 fluorogold-labeled cells were identified. Injection of fluorogold into the dorsal surface of the uterine cervix and lower uterine segment labeled neurons bilaterally in DRGs at T12–L2 spinal levels. This labeling was via  retrograde transport through the hypogastric nerve, because fluorogold-positive neurons were not found in animals that received bilateral hypogastric neurectomy at the time of tracer injection. Fluorogold-labeled cells appeared solid white to light blue in color with fluorescent staining in the cytoplasm (fig. 1), and those cells with calibrated density above 180 were counted as positively stained. Overall, the population of fluorogold-labeled neurons was 1–2% in studied DRG sections. In staining for TRPV1 and P2X3 receptors, a total of 412 fluorogold-positive cells were identified, and 315 fluorogold-positive cells were identified for the staining of CGRP and somatostatin. Most fluorogold-positive cells were found at L1, followed by L2 and T13, with few fluorogold-labeled cells at T12 (fig. 2). The average diameter of fluorogold-labeled neurons was 24 ± 6.1 μm (average ± SD), which was significantly larger than neurons stained with TRPV1 (20 ± 6.0 μm), P2X3 (22 ± 5.6 μm), CGRP (20 ± 5.2 μm), or somatostatin (18 ± 3.7 μm) (P < 0.001; fig. 3).

Cells staining positively for TRPV1 were small- to medium-diameter neurons (figs. 1 and 3). TRPV1 staining was numerically more common in fluorogold-labeled neurons than unlabeled neurons in all groups, although the difference was only significant in estrogen-treated animals (P < 0.001; fig. 4).

Cells staining positively for P2X3 receptors were also small to medium size (figs. 1 and 3). Interestingly, not a single fluorogold-labeled uterine cervical afferent neuron colabeled with the P2X3 receptor in any group. In contrast, DRG neurons that were not traced from the uterine cervix expressed the P2X3 receptor with no differences among treatment groups (19 ± 5.4% in ovariectomy, 21 ± 3.7% in estrogen replacement, and 27 ± 3.6% in control animals; P > 0.05).

Calcitonin gene–related peptide-immunostained neurons were also of small to medium size (figs. 1 and 3). Within each group, the proportion of neurons with CGRP staining was significantly greater in fluorogold-labeled neurons (P < 0.001; fig. 5). There was no effect of ovariectomy or estrogen replacement on the proportion of uterine cervical afferent or control neurons expressing CGRP (fig. 5).

Very few DRG neurons expressed somatostatin (fig. 1), with no differences among groups (4.2 ± 1.0% in ovariectomy, 3.9 ± 1.6% in estrogen replacement, and 6.5 ± 0.7% in control animals; P > 0.05). As with the P2X3 receptor, there were no fluorogold-labeled neurons expressing somatostatin.

Stimulation of the uterine cervix, whether from chronic inflammation, cancer, or acute dilatation, is a common cause of pelvic pain in women, yet its anatomic and neurophysiologic bases have received little previous attention. Berkley et al.  2,3characterized hypogastric afferents that innervate the uterine body of the rat and their responses to distension, heat, and bradykinin, but the relevance of these afferents to acute nociception is unclear, because conscious animals display remarkably little response to uterine body distension in the absence of inflammation.28Papka et al.  36–39described the presence of some neurotransmitters and enzymes in nerve fibers in the rat cervix as well as in L6–S1 DRGs and the influence of pregnancy and estrogen treatment on their expression. These studies, however, focused primarily on pelvic afferents which terminate in low lumbar and upper sacral cord and did not trace uterine cervical afferents in the DRG, so the studies could not determine whether estrogen- or pregnancy-related changes occurred specifically in these afferents. Therefore, the current study fills an important gap in our knowledge regarding the sensory phenotype of hypogastric nerve afferents innervating the uterine cervix and lower uterine segment. Because all identified uterine and uterine cervical afferents in electrophysiologic studies conduct at C-fiber velocity,2,6,8we focused our initial characterization on neurotransmitters and ion channels that are commonly expressed in C fibers and considered important in nociception. The results indicate a remarkably different phenotypic profile of uterine cervical hypogastric nerve afferents compared with other visceral or somatic C fibers.

Transient receptor potential vanilloid type 1 is a nonselective ion channel expressed in normal conditions by C fibers that responds to noxious heat, low pH, and vanilloid compounds, including capsaicin.10Its expression is regulated by nerve growth factor13and increases in uninjured neurons in models of neuropathic pain.40Expression of TRPV1 in bladder urothelium in animals11and pain and hyperalgesia from topical application of capsaicin to gastrointestinal endothelium in humans41suggests that TRPV1 signaling is important to pain from these viscera. Importantly, TRPV1 channel activation can, including colonic visceral afferents, transduce responses to mechanical stimuli and sensitization to distension after visceral inflammation.9,16We previously demonstrated that some uterine cervical afferents respond to heat,8consistent with expression of this channel. The current study suggests that approximately one third of uterine cervical hypogastric nerve afferents express TRPV1. Female sex hormones, at least at levels seen in the nonpregnant state, do not seem to regulate expression of this channel, as evidenced by no effect of ovariectomy or estrogen treatment. However, there was a difference in proportion of uterine cervical afferent neurons expressing TRPV1 compared with untraced neurons in the presence of estrogen. Therefore, these results indicate it is possible that increased TRPV1 expression could explain estrogen-mediated sensitization of uterine cervical afferents.8 

Of course, immunocytochemistry studies only indicate expression of the antigen to which the antibody was raised at a level great enough to provide a signal above background and do not necessarily indicate the presence of functional protein or its quantitative level of expression. Other techniques, including single-cell proteomics or neurophysiologic recording, could examine more subtle changes in level of expression or in function. We used a stringent, population-based approach to determine whether a cell would be considered immunostained, and the proportion of cells meeting this criterion was somewhat less for TRPV1 and the other markers in the current study than in previous publications.42–44 

P2X3 is one of a family of adenosine triphosphate–gated receptors on C fibers that is regulated by glial-derived nerve growth factor14and is important in sensitization of somatic responses after inflammation and nerve injury.15Unlike TRPV1, P2X3 receptor expression is unchanged in undamaged afferents after peripheral nerve injury.17P2X3 receptors are present and functional in urinary bladder^12,18and gastrointestinal afferents19and are believed to play a key role in visceral nociception. Thus, it is surprising that we observed a total lack of P2X3 receptor immunostaining in uterine cervical afferents in the current study. As noted above, we used a stringent criterion for indicating positive immunoreactivity and could have missed weakly stained cells. We consider this unlikely, because the proportion of control DRG neurons meeting the criterion for P2X3 receptor staining, 20–25%, is similar to that previously observed.45Lack of P2X3 receptors on uterine cervical afferents of hypogastric nerve in the nonpregnant state represents an important distinction between this and other viscera and suggests that targeting this receptor would be unlikely to yield analgesia from acute uterine cervical nociception.

Calcitonin gene–related peptide is an important pain neurotransmitter expressed by a subset of C fibers and that increases in expression in states of sensitization, including inflammation,46nerve injury,20or chronic opioid exposure.21CGRP also plays an important efferent function and is released in areas of inflammation, including the gastrointestinal mucosa,22to result in vasodilatation and recruitment of leukocytes. The uterus36and uterine cervix37contain neuronal fibers that express CGRP, and some have speculated that CGRP plays an important efferent role in the cervical ripening process preceding labor.23,47As previously reported,4uterine cervical afferent neurons were more likely than untraced DRG neurons to express CGRP in the current study, consistent with an important role of this neurotransmitter in sensation and possibly peripheral signaling in the uterine cervix. Lack of effect of ovariectomy or estrogen suggests that changes in expression of this neurotransmitter are unlikely to explain estrogen-mediated sensitization of uterine cervical afferents, although the effects of pregnancy are worthy of exploration.

Somatostatin is present in a small subset of C fibers that express c-ret and, paradoxically, inhibit dorsal horn,48and peripheral26responses of somatic nociceptors, in part by modulating TRPV1 channel activity.27Inflammation of the gastrointestinal mucosa leads to a reduction in somatostatin content in colonic afferents,49consistent with its presence and regulation in afferents to this visceral structure. The current study did not identify any afferents that innervate the uterine cervix and express somatostatin, suggesting targeting somatostatin receptors might not reduce responses to nociception from this organ.

In summary, despite the importance of the uterine cervix in acute and chronic pain states, little is known regarding the hypogastric afferents that innervate this structure. The current study demonstrates a unique phenotype of these afferents. The common expression of CGRP likely participates in both sensory and efferent functions in this tissue, and expression of TRPV1 may play a role in mechanosensitivity and, particularly, sensitization from inflammation or estrogen exposure. Interestingly, uterine cervical afferents in the hypogastric nerve do not express P2X3 receptor, which is thought to play a key signaling role in pain from the urinary bladder, nor do they express the inhibitory neuropeptide, somatostatin. Female sex hormones, at levels present in the nonpregnant state, do not seem to regulate expression levels of these neurotransmitters and channels, with the possible exception of estrogen dependence of TRPV1. These results provide testable hypotheses regarding novel strategies to treat pain from the uterine cervix and lay the groundwork for future studies examining the effect of pregnancy and labor on the neurophysiology of these afferents.