Superoxide anion increases intracellular free calcium in human myometrial cells.

We investigated the effects of superoxide anion on the intracellular free calcium concentration ([Ca2+]i) in human cultured myometrial cells using a calcium-sensitive fluorescent dye, indo-1, and a digital imaging fluorescence microscopic system. Hypoxanthine (HX) plus xanthine oxidase induced a rise in [Ca2+]i in a manner dose-dependent on xanthine oxidase. The increase in [Ca2+]i in the absence of extracellular calcium ([Ca2+]ex) was 10% of that in the presence of [Ca2+]ex. Nifedipine, which blocks voltage-sensitive calcium channels, also reduced the increase in [Ca2+]i induced by HX-xanthine oxidase. Superoxide dismutase or superoxide dismutase plus catalase, which metabolizes superoxide anion, inhibited the effect of HX-xanthine oxidase on [Ca2+]i. The desensitization of the effect of superoxide anion on [Ca2+]i was investigated by pulsatile administration of HX and xanthine oxidase. Desensitization was observed on pulsatile administration of HX-xanthine oxidase at 2-min intervals. These data suggest that superoxide production may participate in uterine contraction via [Ca2+]i increase.

Recent studies have suggested that intrauterine bacterial infection may be one of the main causes of premature labor (1). Prostaglandin biosynthesis, which is stimulated by bacterial or host signals, causes uterine contractions (2). The uterine muscle is one of the most sensitive organs to prostaglandin Fza, E1, and E,. Phagocytes also generate various kinds of monokines (e.g. interleukin 1, tumor necrosis factor, platelet activating factor) which stimulate prostaglandin production (3-5). Phagocytes that accumulate at infectious focuses also produce superoxide anion (6). But, the role of this superoxide anion is still unknown.
On the other hand, the role of increase in the intracellular free calcium concentration ([Ca"],) in muscle contraction is well established (7,8 22533 cium concentrations used to prepare the standard curve were determined using Ca'+-EGTA solution (16). Two-dimensional fluorescence images were made by moving the specimen over the focal point of the laser, making fluorescence measurements, and correlating them with the coordinates of movement. For administration of drugs and their removal by washing, we used the perfusion system. Medium was perfused around the cells at a rate of 0.6 ml/min at 37 "C. The effects of each treatment on the [Ca'+], in 10 cells were measured. Solution, of which Ca*+ concentration was controlled in the range of 50-100 nM by addition of EGTA, was prepared as a Ca'+-free solution.
Indirect Zmmunofluorescence-On days 2,4,6, and 8 after isolation, cultured cells were fixed with methanol at -20 "C for 20 min. The cells were washed with 0.1 M phosphate-buffered saline and stained with monoclonal antibodies against desmin (muscle-specific intermediate filament) or myosin (contractile protein) for 40 min. After washing three times, fluorescein isothiocyanate-labeled goat IgG antimouse immunoglobulins were applied for 40 min. Then, the cells were washed and observed with a fluorescence microscope.
Analysis of Data-Data are shown as mean values plus or minus the standard error of the mean (mean + S.E.) of multiple determinations in at least three replicate experiments. The significance of the differences was assessed by analysis of variance, and a p value of less than 0.01 was considered significant.

RESULTS
Superoxide was produced by adding HX and xanthine oxidase (HX-xanthine oxidase). At first, we investigated the effect of superoxide anion on [Ca"+li in suspension of freshly isolated cells. HX-xanthine oxidase significantly induced increases of [Cat+]; (Fig. 1). Subsequently, we examined by using single myometrial cells to clarify that the effect was induced especially in myometrium.
The Graphics show the distribution of [Ca*+k before (A) and 15 s (B), 30 s (C), and 120 s (D) after the stimulation.
The mean [Ca'+], in the resting condition was 95 nM, but the concentration of [Ca'+], was higher in the cytoplasm than in the nucleus ( Fig. 2A). After the addition of HX (1 mM)xanthine oxidase (10 milliunits/ml), the [Ca'+], increased to 180 nM in 15 s (Fig. 2B), reached a peak of 211 nM after 30 s (Fig. 2C), and then returned to nearly the basal level (103 nM) after 120 s (Fig. 20) oxidase (10 milliunits/ml) was 10 times greater in the presence of [Ca2+lex more than in its absence (Fig. 3). These results indicate that an increase of [Ca"], by HX-xanthine oxidase is mainly caused by influx of extracellular Ca2+ through the membrane. To exclude the possibility of a nonspecific effect of HX-xanthine oxidase, i.e. cell damage, the basal [Ca'+], and the response of [Ca2+]l to oxytocin were checked 20 min before and 20 min after the addition of HX-xanthine oxidase (Fig. 4). The basal [Ca2+lr levels and the responses of [Ca'+], to oxytocin before and after treatment with HX-xanthine oxidase were not significantly different. Thus, the effect of HX-xanthine oxidase on [Ca2+lr seemed to be reversible and physiological.
Details of the change in [Ca2+lr in these cells induced by oxytocin will be reported in a separate paper. For confirmation of the specificity of the effect of superoxide anion, superoxide dismutase (10 units/ml) and superoxide dismutase (10 units/ml) + catalase (100 pg/ml) were added 5 min before HX-xanthine oxidase stimulation.
Increase of [Ca"], induced by HXxanthine oxidase was significantly inhibited by either superoxide dismutase or superoxide dismutase + catalase (Fig. 5). Therefore, this increase was concluded to be due to a specific effect of superoxide.
In fact, hydrogen peroxide in various concentrations did not affect [Ca'+], (not shown). Nifedipine (1 PM) also significantly reduced the [Ca2+], increase induced by HX-xanthine oxidase (Fig. 5). These data indicated that not hydrogen peroxide but superoxide anion induced an increase of [Ca"+], in myometrial cells, partly by activation of with superoxide anion and oxytocin are compared in Fig. 6. The response of [Ca"+], to superoxide anion had a shorter unresponsive period than that to oxytocin. Pulsatile stimulation with superoxide anion caused desensitization of [Ca"], increases, but less than the desensitization on pulsatile stimulation with oxytocin. Using indirect immunofluorescent technique, cultured cells were stained with monoclonal antibodies against desmin and fluorescein isothiocyanate-labeled IgG. At least 90% of cultured cells on day 2 showed positive labeling of desmin. Nearly all of oxytocin-responded cells were stained with similar high intensity on days 2,4, and 6. The intensity started to decrease on day 8. Oxytocin-nonresponsive cells started to be weakly stained on days 4 or 6. On the other hand, nearly all of the cultured cells on days 2, 4, and 6 were stained with similar high intensity with monoclonal antibodies against myosin. On day 8, 70% of the cells was weakly stained, and other cells were strongly stained (Fig. 7). These results show that myometrial cells cultured for 2 days, especially oxytocin-responded cells, could not modulate from contractile to synthetic phenotype.

DISCUSSION
In the present study, we measured the two-dimensional distribution of [Ca2+J1 in cultured human myometrial cells and found that superoxide anion induced increases in [Ca'+],. HX-xanthine oxidase is known to generate superoxide anion, which is converted first to hydrogen peroxide and oxygen by the scavenger enzyme superoxide dismutase, and then to water and oxygen by catalase. Superoxide anion is reported to inactivate the endoplasmic reticular Ca2+ pump in coronary artery (22) and to activate Na+-Ca'+ exchange in heart cells (23). We found that the increase of [Ca'+], in myometrial cells was mainly caused by activation of L-type calcium channels. But the roles of superoxide anion on T-type calcium channels, Ca*+-Na2+ exchange, and release of intracellular calcium stores were not clear. Anwer and Sanborn (24) reported that oxytocin and carbachol increased [Ca*+]i in rat myometrial cells in suspension, and that the increase was due to both mobilization of intracellular stores and influx of extracellular calcium. We showed that the increase of [Ca"+]i in human myometrial cells induced by oxytocin was dependent in part on intracellular stores but mainly on extracellular Can+.' In this point, superoxide anion and oxytocin may have similar effects. We found that resting cells have a high level of [Ca"']; near the nucleus, probably due to calcium stores in sarcoplasmic reticulum in this region. The increase in calcium near the cell membrane was heterogeneous, suggesting that this distribution of calcium channels in the membrane is also heterogeneous.
Desensitization of the response of [Ca'+]; to superoxide anion is less than that of the response to oxytocin. Since the existence of a receptor for superoxide anion has not been demonstrated, the mechanism of the increase of [Ca'+]; induced by superoxide anion is not clear. Superoxide anion may induce the increase of [Ca2+li in the mechanism different from that of the oxytocin-activating calcium channel via receptor. We studied the measurement of [Ca'+]; with myometrial cells cultured for 24-48 h. On the other hand, smooth muscle cells, including myometrial cells, in primary culture have been demonstrated to modulate from contractile to synthetic phenotype, paralleling a decrease of contractile protein (myosin, actin, and caldesmon) (25). Palmberg and Thyberg (15) reported that the process of modulation in cultured myometrial cells involved loss of desmin, and the proliferative potential of myometrial cells was markedly lower than that of arterial smooth muscle cells. Our examination of immunofluorescence showed that contractile phenotype remained in primary cultured myometrial cells, especially oxytocin-responded cells, for 24-48 min.
The increase of [Ca2+lL might be a trigger of uterine contraction. Superoxide is supposed to induce uterine contraction. Superoxide can be generated in vivo in the following ways: 1) stimulation of phagocytes in infectious focuses, 2) ischemia-reoxygenation, and 3) activation of arachidonate metabolism.
There is much suggestive evidence that intrauterine infection is associated with the onset of premature labor (1). Prostaglandins have been considered to play an important role in uterine contraction due to infection. In fact, bacterial products stimulate prostaglandin production in am-' K. Tasaka, N. Matsumoto, A. Miyake, and 0. Tanizawa, manuscript in preparation. niotic membrane.
Interleukin-1 and tumor necrosis factor produced by mononuclear cells and decidua also stimulate prostaglandin production in response to bacterial products (3, 4). Cherouny et al. (26) reported that hydrogen peroxide induced prostaglandin production and contractions of pregnant rat uterus. But, the relationship between prostaglandin production and superoxide generation is still unknown. Our data showed that superoxide anion itself produced by phagocytes activated in infectious focuses could induce increases in [ Ca2+li and uterine contraction.