Special Reviews

The Pains of Endometriosis

See allHide authors and affiliations

Science  10 Jun 2005:
Vol. 308, Issue 5728, pp. 1587-1589
DOI: 10.1126/science.1111445

Abstract

Endometriosis is a disease defined by the presence of endometrial tissue outside of the uterus. Severe pelvic pain is often associated with endometriosis, and this pain can be diminished with therapies that suppress estrogen production. Many women with endometriosis also suffer from other chronic pain conditions. Recent studies suggest that mechanisms underlying these pains and sensitivity to estrogen involve the growth into the ectopic endometrial tissue of a nerve supply, which could have a varied and widespread influence on the activity of neurons throughout the central nervous system.

Endometriosis is a common disorder that occurs mainly in women of reproductive age. Because ectopic endometrial implants respond to natural or induced decreases in estrogen levels, the disorder is considered “estrogen dependent” (1). Symptoms of endometriosis include reduced fertility and several types of pain such as severe dysmenorrhea (excessive menstrual pain), deep dyspareunia (pelvic pain with coitus), dyschezia (pelvic pain with defecation), and chronic pelvic pain. In some women, pain can be exacerbated by the co-occurrence of other severe chronic pain conditions such as irritable bowel syndrome, interstitial cystitis, repetitive kidney stones, vulvodynia, temporomandibular syndrome, migraine, and fibromyalgia (24). Little is known about the association between the ectopic implants and pain; however, recent studies of women and animal models are beginning to provide clues.

The most common pharmacological treatment for endometriosis uses a class of drugs called gonadotropin-releasing hormone (GnRH) agonists. Because these drugs down-regulate GnRH receptors, they suppress pituitary gonadotropin secretion and sex steroid production, thereby producing a systemic hypoestrogenic state. This treatment results in the elimination or reduction in size of the implants in women (5) as well as in a rat model of endometriosis (6). In addition, optimized treatment with GnRH agonists is effective in reducing endometriosis-related pain symptoms in women (5).

Because GnRH agonists reduce both implant size and the pains associated with endometriosis, these pains may be due to the presence of the abnormal implants. Numerous studies, however, have failed to find a correlation among pain scores, types of pain, and various aspects of the anatomy and biochemistry of the implants (7). In addition, although surgical removal of the ectopic implants alleviates pain symptoms in many women, the surgery can fail to alleviate the pain and/or pain may recur even without evidence of residual or recurrent disease or any other identifiable visceral or somatic pathology (8).

On the other hand, correlations have been found between pain severity and both the depth of “infiltration” into peritoneum or pelvic organs and the proinflammatory cytokines, prostaglandins, chemokines, and other substances released by the implants or neighboring tissues into peritoneal fluid (1, 9). Of relevance here are findings that the percentage of patients reporting pain is greater in women with deeply infiltrating implants in highly innervated areas (such as the uterosacral region) than in women with other types of implants, and that the former implants are more likely to infiltrate nerves (10). Others have found that the nerve fibers are closer to the implants in women with pelvic pain than in women without pain (11). These results implicate the nervous system in the various pains of endometriosis. Indeed, recent results drawn from a rat model of endometriosis support this idea. This model (Fig. 1A) involves autotransplantation of parts removed from one uterine horn onto abdominal blood vessels, where the transplants then grow into cysts.

Fig. 1.

(A) Rat model of endometriosis. Abdominal and pelvic organs are on the right; at left is the pattern of their input/output neural connections with different segments of the spinal cord and brain (via the vagus nerve). T1, T13, L1, L6, and S1 designate spinal segments. Also shown is the location of ectopic endometrial cysts (red dots) that develop after autotransplantation of pieces of uterus taken from one uterine horn (X) in the rat model of endometriosis discussed in the text (1215). (B) Section through a rat endometrial cyst and lumen (lum) showing immunofluorescence for the pan-neuronal marker PGP9.5 (protein gene product 9.5); note bundle of nerves (arrow) entering the hilus. The box indicates an area from which images (D) to (F) were taken. (C) Section through a human cyst and lumen stained with hematoxylin and eosin. The box indicates an area from which images (G) to (I) were taken. (D) Hilus with nerves entering the cyst (arrow) and coursing toward the luminal epithelium (epi). (E) Hilus immunostained with markers for efferent nerves [vesicular monoamine transporter (VMAT), sympathetic fibers] in rat cysts. (F) Hilus immunostained with markers for sensory nerves [calcitonin gene–related peptide (CGRP)] in rat cysts. Neurites (arrows) are evident in the myometrial stroma (myo) and near the luminal epithelium. In (F), sensory fibers (upper arrow) enter the luminal epithelium. (G to I) Sections through human cysts were also immunostained for PGP9.5 (G), VMAT (H), and CGRP (I). (G) PGP9.5 fibers (arrows) in myometrial stroma and near the epithelium. (H) VMAT-sympathetic fibers (arrows) in myometrial stroma (myo) and along a capillary (cap; thin arrows). (I) CGRP-sensory fibers near and in the luminal epithelium. Scale bars, 250 μm (B), 100 μm (C), 25 μm [(D) to (I)]. [(A) modified from (14); (B), (D), (E), and (F) modified from (15)]

This rat model appears to be valid for studying both the signs of endometriosis (ectopic implants) and its symptoms (subfertility and pain). The implants in rats and women respond similarly to hormonal treatment and show similar alterations in protein production (12). Rats with endometriosis (ENDO rats), like some women with endometriosis, are subfertile (12). Although ENDO rats do not exhibit spontaneous pain behaviors (13), they develop an increased pain sensitivity (hyperalgesia) of the vagina, the severity of which correlates positively with estradiol levels during their ovarian cycle (14). Thus, ENDO rats suffer from a painful condition that is estrogen sensitive and is similar to dyspareunia in women. Furthermore, urinary bladder capacity is reduced in ENDO rats (14). In interstitial cystitis in women, the most salient symptom, other than pain, is excessively frequent urination. Endometriosis in rats also exacerbates pain behaviors associated with an artificial stone implanted in the ureter, and the stone in turn induces new pain behaviors like those that occur when the uterus is treated with an inflammatory agent (13). Thus, ENDO rats appear to develop pain symptoms similar to those associated with conditions that co-occur with endometriosis in women, specifically interstitial cystitis, uterine pain, and kidney stones (4).

The ectopic implants develop a sensory and sympathetic nerve supply both in rats (Fig. 1, B and D to F) and in women (Fig. 1, C and G to I), similar to that of the healthy rat uterus (15). In the rat model, this supply connects the implants directly with the central nervous system via the splanchnic and vagus nerves (14, 15). Input to the spinal cord from the implants arrives at the same spinal segments as those receiving input from the ureter (Fig. 1A, green shading), but rostral to segments receiving input from the vaginal canal (blue shading) or bladder (pink and blue shading). Thus, vaginal hyperalgesia and ureteral pain in this rat model likely involve central neural mechanisms [i.e., intersegmental spinal communication (14, 15)], whereas the effects of endometriosis on bladder function could be via peripheral interactions and/or in the caudal spinal cord.

Two other factors may help to explain the variable types and severity of endometriosis-associated pains, their co-occurrence with other painful disorders, and their amelioration by a hypoestrogenic state: (i) central sensitization, and (ii) divergent and convergent connectivity in the central nervous system. First, it is commonly recognized that sensory fibers of the type observed in both the rat and human ectopic implants (Fig. 1, F and I, respectively) are activated and sensitized by many inflammatory agents in them (1, 9, 16). By means of several molecular processes reviewed elsewhere (16), sensitization of the sensory fibers would in turn produce central sensitization, which is a long-lasting hyperexcitability of neurons in the central nervous system that can continue long after the originally sensitized input is reduced or eliminated (e.g., by surgery).

Second, sensory input arriving at the spinal cord from individual internal organs diverges within the cord. Thus, although information from different organs is delivered most densely to spinal neurons within the entry segments, it is also delivered, less densely, to widespread spinal regions extending for many segments rostrally and caudally (14). The anatomical divergence gives rise to considerable convergence of information on central neurons. This “viscero-viscero-somatic convergence” (13, 14) produces a situation in which the activity of somato-visceral neurons in the spinal cord and brain is dominated by information from individual peripheral structures but can be augmented, particularly in sensitized neurons, by events occurring elsewhere. Such convergence thereby provides a substrate by which sensitized input from ectopic implants augmenting that from healthy organs can have widespread influences on the activity of neurons normally associated with input from different individual organs and tissues, a situation that has been demonstrated in women (17).

These results suggest that inconsistency in the various pains associated with the ectopic implants could reflect variability in a number of factors associated with the implants' nerve supply. These factors include the types of nerves that innervate the implants, agents that activate or sensitize them, sites in the central nervous system where the nerves deliver information, and how that information is modulated by estradiol—both peripherally (18) and centrally (19)—as well as by other central dynamic processes (14, 17, 20).

Much remains to be learned about how endometriosis comes to be associated so variably with pain symptoms and how those symptoms are ameliorated by a hypoestrogenic state. One promising area of research concerns the implants' sensory and autonomic nerve supply and its potentially estradiol-modulated influence on activity within the central nervous system.

References and Notes

View Abstract

Navigate This Article