The effects of Escherichia coli STa (heat stable) toxin on the contractility of isolated human myometrium in vitro.

OBJECTIVE: The purpose of the study was to assess the effects of Escherichia coli STa (heat stable) toxin on isolated human myometrial response to oxytocin. METHODS: One hundred and sixteen muscle strips were obtained from the lower uterine segment of 42 women undergoing cesarean section at term. Amniotic membranes and decidua were excluded. Uterine contractility in response to cumulative doses of E. coli STa toxin was recorded, as well as uterine response to cumulative doses of oxytocin before and after incubation with STa toxin or vehicle. The 50th percentile effective oxytocin concentration (EC50) of muscle strips with and without spontaneous activity before and after the incubation with STa toxin or vehicle was calculated. A paired t test was used for comparison. RESULTS: Muscle strips with and without spontaneous activity responded to cumulative doses of oxytocin before and after the incubation with STa toxin or vehicle. No differences in contraction force, duration, or frequency were noted between the groups (P > 0.05). Furthermore, this toxin was not able to induce uterine contractility when tested alone. CONCLUSIONS: The inability of this toxin to affect myometrial response to oxytocin in this study may be due to the absence of amnion cells, chorion, or decidua. Other possible explanations for the lack of response are discussed.

he incidence of urogenital infections during pregnancy has been well established, 1,2 and their association with preterm delivery reported. 3,4 Several authors have suggested a relationship between maternal infection, preterm rupture of membranes, and prematurity. [5][6][7] Some microorganisms and bacterial products isolated from women with premature rupture of membranes and chorioamnionitis have been shown to stimulate prostaglandin synthesis.  These substances may induce or potentiate contractility of the pregnant uterus as well as other smooth muscles. Romero et al. found that several lyophilized bacteria, including two strains of Escherichia coli, were able to induce the synthesis of prostaglandins by human amnion cells. 11,1z Nevertheless, we have not found reports of the effects of either E. coli or its products (e.g., heat stable [STa] toxin) on isolated human myometrial con- tractility. The present study was conducted to assess the possible effects of E. coli STa toxin on oxytocin-stimulated human myometrium.

SUBJECTS AND METHODS
Human myometrial samples were collected during elective cesarean section at term of nonlaboring patients whose membranes were intact. These women were informed of the study, which was approved by the institutional review board. Only those who gave written consent were included. A single 2.0-cm long full-thickness strip of myometrium was removed from the upper margin of the incision after delivery. The tissue was immediately immersed in ice-cold salt solution. After separating the amnion or serosa if present, two to three strips of circular muscle (3 x 3 x 12 mm) were carefully cut from the middle portion of the sample under magnification (12x). The muscle strips were tied at each end with 4.0 silk thread and placed in a 10-mL organ bath vertically within one hour of sur- washed twice every 15 min. As soon as the baseline force restabilizcd, a single dose of potassium chloride was added to the bath, elevating its concentration to 40 raM, and 5 min. of muscle response were recorded at the beginning and at the end of each experiment. Only muscle strips that responded to this challenge were included. After a 30-rain. period of registered contractility, the strips were separated into two groups according to the presence or absence of spontaneous uterine activity.
In the first series of experiments, cumulative dose-response curves to oxytocin and E. coli STa toxin (Sigma Co., St. Louis, MO) were assessed. 13 Nine cumulative doses of STa toxin ranging from 0.1 to 19.2 nM were tested. A different group of strips without spontaneous activity was challenged with seven cumulative doses of oxytocin ranging from 0.1 to 10.0 ng.
In the experimental protocol, three cumulative doses of oxytocin (X 2.38 10 -l M, Y 1.4 10 -9 M, Z 6.56 10 -9 M) that corresponded approximately to the initial (25%), middle (50%), and final (90-100%) part of the first dose-response curves, respectively, were added before and after the incubation of the muscle strips with 1.0 ng of STa toxin 14 for 30 min. (Fig. 1). The tissue was washed twice, and the same dose of STa toxin (1.0 ng) or vehicle was added back to the bath. The selection of vehicle or STa toxin was made ran- domly (computerized distribution using Latin squared method). The second and third oxytocin doses were applied until no further increase in tension was observed. The volume of the drugs did not exceed 0.2 mL. The average contraction force that developed within 10 rain. of the maximum contractile response was measured and divided by tissue weight. Frequency was recorded as the number of contractions, and duration was registered as the time from the beginning to the end of each contraction in this same period of time. A change in the resting tone excluded the strip from experimental protocols. Several samples exposed to these conditions were sent for histological analysis, and their measurements were not included.

Statistical Analysis
Each strip was considered its own control. The oxytocin concentration producing half maximal response (EC50) was used for comparison. A twotailed paired t test (alfa 0.05) was used.

RESULTS
The clinical profiles, of the patients who donated tissue for the study revealed a mean gestational age of 38.4 + 1.5 weeks; all patients were nonlaboring patients with intact membranes. Indications for cesarean delivery included elective repeat cesarean in 28 patients and malpresentation in 14. tn the elective repeat group, the primary cesarean section was indicated because of cephalopelvic disproportion in eight, malpresentation in nine, macrosomia in six, chorioamnionitis in two, and fetal distress in three. Nine of 14 patients with malpresentation were nul-liparous. Of a total of 116 myometrial strips obtained from these patients, 103 responded to the potassium chloride challenge. Sixty five percent of the strips developed spontaneous activity. No changes in resting tone were registered in any group. The ECs0 of oxytocin in tissue with and without spontaneous activity was calculated as 6.67 1.1 x 10 -1 M (n 8) and 7.25 _ _ _ 0.8 x 10 -9 M (n 8), respectively. Cumulative doses of STa toxin were administered to several strips, and no change in uterine contractility was noted (n 10).
No histopathologic changes of the myometrial strips exposed to this toxin or vehicle were ob-[,,] ox roc,.

DISCUSSION
Clinical profiles of patients who donated tissue for this study, including those who did not respond to high potassium, were quite uniform. An estimated 88% of the myometrial strips obtained .from these patients were considered viable judging from their ability to contract with a high-potassium stimulus.
As reported previously, approximately half of the strips used in these experiments developed spontaneous activity, and the maximal mean tension developed was our greatest source of variability. This has been attributed to the differences in total smooth muscle content of the lower uterine segment strips obtained through this technique, is Therefore, we considered each muscle strip to be its own control.
Cumulative dose-response curves to E. coli STa toxin revealed that the toxin was not able to induce contractility in strips without spontaneous activity, and no changes were seen when spontaneous activity was present. A wide range of toxin doses above and below the dose that produces symptoms in humans was tested (1.0 ng per 10 mL). 16 '7 In order to exclude whether the lack of response was because not enough time was given to the toxin to act or because the sample was obtained from nonlaboring patients, oxytocin-induced uterine contractions before and after incubation with toxin were registered. In this series of experiments, cumulative dose-response curves to oxytocin were obtained and a ECso was calculated for strips with and without spontaneous activity. In the experimental protocols, both groups of myometrial strips were incubated with toxin or vehicle, and the mean of maximal tension developed in response to three cumulative doses of oxytocin after incubation did not differ from controls before incubation (Figs. and 2). Analysis of frequency and duration of the contractions revealed no differences between the groups either. The experiments carried out with vehicle assured that no changes in response to oxytocin were present during the experimental protocol. Damage to the tissue and presence of amnion, chorion, or decidua was excluded histologically.
The results from these experiments suggest that the incubation of this tissue with STa toxin did not affect the ability of isolated human myometrium strips to contract in response to oxytocin under these conditions. Furthermore, the toxin was not able induce contractility in strips without spontaneous activity over a wide range of doses tested.
The inability of this tissue to respond to STa toxin in the present study may be explained by the absence of amnion, chorion, or decidual cells. As demonstrated by Romero et al., lyophilized bacteria can produce mediators (e.g., prostaglandins, interleukin-1, and tumor necrosis factor) from surrounding contiguous tissues that can trigger myometrial activity. Therefore, the use of lyophilized E. coli or STa toxin could affect uterine contractility when amnion or decidua are present.
Other possibilities for the lack of effect of this toxin seen in these experiments may include the absence of STa receptors in the tissue. Schulz and coworkers have designated guanylyl cyclase as a heat stable enterotoxin receptor. 18 These enterotoxin receptors have been described in small bowel smooth muscle 19 and liver z but not in uterine myometrium. Studies that could answer this question were not done, because without ahy apparent effect of this toxin on isolated human myometrium, isolation of these receptors appeared not to be indicated.
These results confirm the importance of contiguous tissues (amnion, chorion, and decidua) in mediating an increase of uterine response and their possible role in the initiation of labor in patients with concomitant infection. In order to better understand how these events happen, it appears necessary to perform in vitro studies that include myometrium with decidua or amniotic membranes to determine if the lack of effect is solely due to the absence of these surrounding tissues or because this toxin has no inherent effect on the contractility of human pregnant uterus.