MECHANICALLY-INDUCED HYDROSALPINX: LONG-TERM OVIDUCTAL DILATATION DOES NOT IMPAIR CILIARY TRANSPORT FUNCTION*

Transcription

1 FERTILITY AND STERILITY Copyright c 1981 The American Fertility Society Vol. 36, No. 6, December 1981 Printed in U.S A. MECHANICALLY-INDUCED HYDROSALPINX: LONG-TERM OVIDUCTAL DILATATION DOES NOT IMPAIR CILIARY TRANSPORT FUNCTION* DOROTHY L. PATTON, PH.D.t SHERIDAN A. HALBERT, PH.D. Department of Biological Structure, University of Washington, Seattle, Washington Long-term hydrosalpinges were mechanically induced in rabbits by ligation of the fimbriated end of the oviduct. The structure and function ofthese model hydrosalpinges were studied 28 to 52 weeks following ligation. This procedure caused a nonpurulent, clear serous fluid to accumulate within the lumen of the ampulla, which resulted in dilatation of the ampulla and thinning of the tubal wall. The intraluminal mucosal folds were atten(j-ated in the expanded regions; however, normal ciliation was predominant throughout the endosalpinx. Two rabbits, one with a single distal tubal ligation and the other with double ampullary ligation, showed abnormal epithelium with distinct patches of flattened polygon-shaped, nonciliated cells. In vivo and in vitro observations of luminal transport of surrogate ova in cumulus showed that cilwry transport was not affected by the long -term gross distension of the ampullae. The authors conclude that long-term tubal dilatation by itself does not alter tubal morphology sufficiently to impair ovum transport function in the oviductal ampulla. Fertil Steril 36:808, 1981 In an earlier study,! the effects of distal tubal occlusion on oviductal structure and egg transport function in rabbits and monkeys were investigated. Marked dilatation of the oviductal ampulla occurred when the oviduct was ligated at the base of the fimbria, and the intraluminal mucosal surface of the stretched walls became flattened; however, ciliary transport on the ampullacy endosalpinx was not impeded even after a mechanically-induced hydrosalpinx of 20 weeks' duration, suggesting that tubal dilatation alone does not cause significant functional damage to the endosalpinx and thus does not account for the persistent infertility in patients following surgical repair of distal tubal occlusion. However, one of the most notable observations in the clinical evaluation of hydrosalpinx in women is the extreme variability in the extent of tubal disease. Received June 10, 1981; revised and accepted July 31, *Supported by National Institutes of Health Grant ND treprint requests: Dorothy L. Patton, Ph.D., Department of Biological Structure, University of Washington School of Medicine, Seattle, Washington Presumably both the degree of distension and the loss of function of the tube's structural components may increase with time. To learn more about the morphologic and functional changes that occur in hydrosalpinx, experiments were conducted toinvestigate the long-term effects of dilatation following distal tubal occlusion, with the hypothesis that such prolonged distension may result in a greater degree of structural change than a shorter-term dilatation would, perhaps including deciliation of the luminal epithelium,2 which could interfere with gamete transport function. MATERIALS AND METHODS To mechanically induce a hydrosalpinx, we ligated the fimbriated end of the left oviduct in five New Zealand White rabbits. The right oviduct served as a control. At intervals from 28 to 52 weeks after tubal occlusion, ciliary transport function of the endosalpinx of the distended ampulla was assessed both in vivo and in vitro. Every 3 to 4 weeks laparotomies were done to assess hydrosalpinx development. 808

2 Vol. 36, No.6 MECHANICALLY-INDUCED HYDROSALPINX 809 scanning and transmission electron-microscopic examination as described previously.5 FIG. 1. Scanning electron micrograph of the large mucosal folds that have become attenuated after 38 weeks of induced hydrosalpinx. Note the intricate folding pattern between the large folds (magnification x 30). In preparation for the transport observations, the rabbits were given an injection of 100 f.lg luteinizing hormone to induce ovulation. Ten to 12 hours later, at the time of ovulation, the animals were anesthetized with halothane. The oviducts were exposed with a midline abdominal incision, and an observation chamber was inserted into the body wall according to the procedure outlined by Blandau and Boling. 3 4 Next, the experimental oviduct was gently placed into the observation dish and submerged with warmed Hanks' balanced salt solution. Direct observations of ampullary transport were subsequently made in this chamber. A surrogate cumulus oophorus made of chicken egg white containing 15 f.l black microspheres to sharpen surrogate visibility was injected transmurally into the ampullary portion of the distended oviduct, as described previously. 1 The injected bolus of equal 10 IJ.l volumes of albumen and saline was deposited within 3 to 5 mm from the tip of the injection needle, which was glued into place in the ampulla to prevent fluid leakage. Progress of the surrogates was observed through a stereomicroscope and permanently recorded with 16 mm cine film; we determined the transport rate by measuring the surrogate's position in sequential frames. Tissue removed for in vitro observations was bathed in warmed Hanks' balanced salt solution ' slit open, and pinned out fiat to its in situ dimensions in a paraffin dish. Surrogates were carefully applied to the exposed endosalpingeal surface and their rates of transport recorded with a stopwatch. Then the entire ampulla was prepared for RESULTS The diameter of the ligated ampulla ranged from 4.5 to 8 mm, i.e., two to four times normal size after 28 to 52 weeks' ligation. Only the ampullary portion of the oviduct was involved in each rabbit. The distended walls of the ampulla were thin and translucent, and the lumen was filled with a nonpurulent, clear serous fluid. The overall diameter and extent of the hydrosalpinx did not change appreciably after the initial laparotomy check. In one animal in which minimal distension was noted at 4- and 8-week laparotomy checks, a second suture was added about 12 mm proximal to the initial suture. Three weeks later the diameter of the tube between the ligatures had doubled. The dilatation progressed to three times normal size 10 weeks later, after which it remained constant. When the distended ampulla was slit open in vitro prior to transport observations, it was immediately obvious that the normal longitudinal, closely packed mucosal fold configuration had become greatly attenuated. Only remnants of the large mucosal folds persisted. Low-power scanning electron micrographs (SEMs) showed that intersecting small ridges formed a weblike pattern on the luminal surface between the larger folds (Fig. 1). Occasionally small "pillars" or fingerlike projections densely covered with cilia were noted between the attenuated folds (Fig. 2). The mucosal surface consisted mostly of appar- FIG. 2. On higher power, the remnants of the mucosal folds were typically densely covered with cilia. In between the larger folds, small pillars were often found that were also densely ciliated (magnification x 290).

3 ~ i'. ~ ; r 810 PATrON AND HALBERT December 1981 t FIG. 3. Scanning electron micrograph of a specimen of ampullary mucosa after 29 weeks of mechanical hydrosalpinx, showing an unusual example of abnormal epithelial cells. In this extensive field of flattened, polygon-shaped deciliated cells, note the clumps of densely ciliated cells. On the surfaces of the flattened cells, long, stringy single cilia are abundant. Discrete patches of abnormal cells such as these were isolated in regions of predominantly normal cells (magnification x570). ently normal secretory and ciliated epithelial cell types. The dome-shaped secretory cells usually protruded above the tips of the long, slender cilia densely clustered on the surface of the ciliated cells. In two animals, one of which had been doubly ligated, discrete patches of flat polygon-shaped, nonciliated epithelial cells (Fig. 3) were observed among a predominance of relatively normal cells. The nonciliated cells were covered by microvilli and frequently contained only a single long, FIG. 5. Scanning electron micrograph of ampullary mucosa 35 weeks after ligation. In this particular view, one of the remaining small mucosal folds has a large patch of atypically flattened secretory-like cells running along its apex. These cells are covered by microvillous projections. Occasional clumps of cilia are seen between these flattened cells. Views such as this were seen only occasionally within essentially normal tissue (magnification x 250). stringy cilium (Fig. 3). In other regions, clumps of atypical thin filaments were interspersed between the predominantly normal secretory and ciliated epithelial cells (Fig. 4). Occasionally small patches of flattened, amorphous nonciliated cells were observed on the crest of the remaining mucosal folds (Fig. 5). In some instances, blebs of apparent secretory product were noted. In the doubly ligated ampulla, the tubal segment between the ligatures developed into a cystlike structure with opalescent walls in which mu- FIG. 4. Scanning electron micrograph of a specimen from the doubly ligated ampulla, 53 weeks after ligation. Most of the mucosa of the luminal surface showed areas of normal secretory and ciliated cell populations. Thin filamentous strands were often found sparsely lacing the surface of the otherwise normal cells (magnification x 1060). FIG. 6. This is a low-power scanning electron micrograph of a sample from the endosalpinx of the doubly ligated segment of ampulla after 53 weeks. Complete attenuation of the mucosal folds was observed in this specimen (magnification x 12).

4 Vol. 36, No.6 MECHANICALLY-INDUCED HYDROSALPINX 811 FIG. 7. Transmission electron micrograph of the doubly ligated ampulla after 53 weeks. The ciliated cell is filled with numerous mitochondria and rough endoplasmic reticulum. The apical portion of the cell contains several ciliary basal bodies. Typically the epithelial cells of all the experimental oviducts were ultrastructurally normal except for a tendency toward a more cuboidal rather than a columnar shape and a decrease in the number of secretory granules in the secretory cell (magnification x 4250). cosal folds were nearly nonexistent (Fig. 6). Variation and irregularity of the epithelium, including apparently flattened, nonciliated cells with numerous microvilli on their surfaces, were pronounced in this segment, as cited above; yet at least half of the surface in this region consisted of normal ciliated and secretory cells. Scanning electron-microscopic examination of the remainder of this ampulla showed normal ciliated and secretory epithelial cells. Transmission electron microscopy revealed a thinning of the muscular oviductal wall. In some areas, the tall columnar cells of the endosalpinx had become cuboidal in shape (Fig. 7). Both ciliated and secretory epithelial cells were present, and the majority were quite normal in appearance. Within the secretory cells, the granules appeared to be fewer in number. Occasionally large, pale, round nuclei were noted within the more cuboidal-shaped epithelial cells. In only a few instances, an almost complete vacuolization of the epithelial cell internal components was seen (Fig. 8). Of the five experimental oviducts, three were suitable for in vivo ovum transport studies. In the other two animals, excessive fat deposition about the oviduct or adhesions to the body wall made clear visualization of the hydrosalpinx in the abdominal dish preparation impossible. In one, an attempt was made to observe surrogate transport; after injection, the surrogate disappeared from sight preventing assessment of transport. In two of the three animals in which transport was observed directly in situ, the cumulus oophorus was clearly seen to move in a slow, regular progression on the inner surface of the tubal wall toward the ampullo-isthmic junction. No contractile activity of the thinned tubal wall was detected that might have influenced the motion of the surrogates. In the third animal, the film lacked distinctive definition, and accurate measurement of transport rate was not possible. However, transport of the surrogate at least to midampulla was confirmed when the oviduct was slit open in vitro. Only unobstructed cine records of transport were analyzed to determine the in vivo transport rate. Three definitive measurements of in vivo transport rate averaged 0.15 mm/sec. In all five of the experimental oviducts, transport was observed in vitro. Measurements of surrogate transport rate-39 in all-were made over various regions of the endosalpinx, notably the tops of the ridges of the large folds, the smaller intertwining folds, and the completely flattened valleys between those folds. The average of all measurements was 0.13 mm/sec. Even in the doubly ligated ampulla, in vitro transport occurred and was found to be no different from that in the rest of the oviducts. Although it is not possible to compare in vivo and in vitro transport rates directly, because of the different number of observations, there is no evidence that the rates are appreciably different. FIG. 8. Transmission electron micrograph of a doubly ligl!\ted ampulla after 53 weeks. The tight junctions between tljle cells are intact and unimpaired. Small intracytoplasmic vacuoles (V), (arrows) are found within one ciliated cell (C). On the left of these ciliated cells is a portion of a single secretory cell (S) whose cytoplasm~ contains numerous large vacuoles and no definitive secretory granules. These examples of intracellular abnormalities are rare exceptions to the predominantly normal tissue seen in most sections (magnification x 3360).

5 812 PATrON AND HALBERT December 1981 DISCUSSION In both the earlier short-term study of 8 to 20 weeks and in the present long-term study of 28 to 52 weeks, similar changes were seen to occur in all animals, i.e., distension and thinning of the ampullary wall. Maximal dilatation reached as much as 8 mm in both the short- and long-term mechanically-induced hydrosalpinges; prolongation of the occlusion did not increase the extent of dilatation. Flattening of the mucosal folds was accompanied by largely normal populations of ciliated and secretory cells in both experiments. In the current study, two instances of abnormal epithelium were noted; yet in these cases as well, most cells were essentially normal. In all cases of experimentally induced hydrosalpinx-including those with cellular abnormalities-ciliary transport was unimpaired. This suggests that prolonged dilatation does not significantly alter the structure and function of the epithelial cells of the ampullary lumen. More important, these results indicate that ovum transport can proceed despite some deciliation. The extent to which this holds true cannot be assessed from the current studies, however. The cell surface pattern noted in the nonciliated cells of two rabbits resembled the deciliated epithelium in the ampulla of the female pig-tailed macaque monkey in the late luteal phase of its menstrual cycle. 6 It is particularly noteworthy that flattened, nonciliated cells in the experimental oviducts were observed atop the remaining mucosal folds as well as in the fiat regions. Certainly the cells on the crest of a mucosal fold or ridge would not have been subjected to the same forces that were acting to stretch the cells in flattened regions, suggesting that the deciliation in the distended tubes may be unrelated to the process of stretching. In a recent, similar study ofexperimental hydrosalpinx in rabbits by Vasquez et al., 7 metal clips were used to occlude both the proximal and distal ends of the oviduct. When only a single clip was applied at the distal end, marginal dilatation occurred with only subtle cellular changes after 8 weeks. However, when clips were placed on both the proximal and distal ends of the tube, more dramatic changes occurred by 4 weeks, such as marked flattening of the luminal mucosa and modest to extensive deciliation; by 8 to 16 weeks deciliation had increased, mucosal folds were replaced by mere ridges, and the entire epithelial surface was deciliated. The different results in these experiments are probably due to differences in procedure. The metal clip is perhaps a less effective occlusive device, since a single application did not produce dilatation, whereas a single ligature did. It may be postulated that the added clip at the proximal oviduct increased dilatation by preventing loss of fluid through the uterine horn. However, the additional clip is not normally necessary to produce oviductal dilatation in rabbits, which, if maintained in estrus, consistently have a very high luminal resistance to fluid flow in the proximal isthmus. Support of this fact comes not only from the results of these single-ligation experiments but also from pilot studies preceding this initial investigation of experimental hydrosalpinx, in which additional ligatures placed on the proximal isthmus failed to increase ampullary dilatation (Halbert and Patton, unpublished results). In the long-tailed monkey, it was also found that double ligation equaled a single distal ligation in inducing a hydrosalpinx. 1 The single example of double-ligation longterm hydrosalpinx in this current series showed the most severe morphologic changes of any of the ligature experiments, yet the extent of deciliation and cellular damage was mild compared to that in the double-clipped tubes of Vasquez et al. 7 Although it is possible that the doubly ligated tube had undergone spontaneous cellular regeneration after 1 year, in spite of a persistent, severe dilatation, it is difficult to believe that the pathologic differences in the tubes in these two studies can be explained simply by differing occlusion methods. To a certain extent, it may be expected that excess luminal fluid would be removed by the vascular and lymphatic systems as hydrostatic pressure in the oviductal wall increased. Blockage of the lymphatic or vascular systems along the oviduct could possibly impair these alternate escape routes, thereby causing greater fluid accumulation, increased intraluminal pressure, and more severe cellular damage. Since the vasculature supply was specifically avoided by Vasquez et al. 7 when clipping the oviducts, the above argument is unlikely to apply. In fact, since the oviductal diameters recorded in the two studies were comparable, it may well be that the differences in cellular damage are not due to the differences in degree of dilatation. The intracellular changes of the epithelium following the double clipping, i.e., pale nuclei with sparse chromatin and intracytoplasmic vacuolization, which were virtually never seen in this study, are remarkably similar to transmission electron-microscopic observations in rabbits with experimental salpin-

6 Vol. 36, No.6 MECHANICALLY-INDUCED HYDROSALPINX 813 gitis. 8 Such salpingitis was found to be rapidly self-limiting but was accompanied by damage to the luminal epithelium that persisted for many weeks. Thus an infection induced at the time of experimental occlusion could cause damage such as that found by Vasquez et al. 7 at 4 weeks; yet no symptoms of an infection would be present at that time. On the other hand, consideration must be given to the fact that the rabbit is very resistant to infection under normal circumstances. Since the results were apparently uniform throughout the double-clipped group, infection would not seem to be a prime suspect. Thus the discrepancies in these two studies appear to go unexplained. The results of this study suggest that cellular damage on the endosalpinx in human hydrosalpinges cannot be blamed simply on chronic tubal distension. While rapid and extensive stretching of the very compliant walls of the rabbit ampulla could possibly result in damage to the epithelium, it is not as likely that this would happen in human tubes, which have much thicker and less easily distensible walls. Also, when only the fimbria! end is occluded, a passageway remains for excess fluid to flow into the uterus; indeed, there is no evidence in human hydrosalpinges of the existence of extreme intraluminal pressures that could rapidly stretch and damage the ampullary mucosa. In addition, fimbriectomy sterilization, which often results in a sterile hydrosalpinx, has been found to be reversible in 38% to 44% of cases, 9 10 an indication that in such instances the dilated ampulla does contain healthy mucosa. It is thus concluded, as before, 1 that persistent infertility following surgical reversal of hydrosalpinx is more likely to be related to abnormalities caused by a previous tubal infection than to damage resulting from tubal distension. REFERENCES 1. Halbert SA, Patton DL: Hydrosalpinx: effect of oviductal dilatation on egg transport. Fertil Steril 35:69, Woodruff JD, Pauerstein CJ: The Fallopian Tube. Baltimore, Williams and Wilkins Co, 1969, p Blandau RJ: Observing ovulation and egg transport. In Methods in Mammalian Embryology, Edited by JC Daniel. San Francisco, W. H. Freeman, 1971, p 1 4. Blandau RJ, Boling JL: An experimental approach to the study of egg transport through the oviducts of mammals. In The Regulation of Mammalian Reproduction, Edited by SJ Segal, P Crozier, PA Corfman, PC Condliffe. Springfield, lll, Charles C Thomas, 1973, p Patton DL, Halbert SA: Electron-microscopic examination of the rabbit oviduct ampulla following microsurgical end-to-end anastomosis. Fertil Steril 32:691, Rumery RE, Gaddum-Rosse P, Blandau RJ, Odor DL: Cyclic changes in ciliation of the oviductal epithelium in the pig-tailed macaque (Macaca nemestrina). Am J Anat 153:234, Vasquez G, Oberti C, Winston RML, Brosens la: The evolution of experimentally induced hydrosalpinges in rabbits. Fertil Steril 35:342, Patton DL, Halbert SA, Wang SP: Experimental salpingitis in rabbits provoked by Chlamydia trachomatis. In preparation 9. Gomel V, McComb P: Surgical aspects of infertility. In Gynecology and Obstetrics, Edited by JJ Sciarra. Hagerstown, Harper & Row, 1981, p Novy MJ: Reversal of Kroener fimbriectomy sterilization. Am J Obstet Gynecol137:198, 1980