- Email: [email protected]
Ultrastructural aspects of preeclampsia
Ultrastructural aspects of preeclampsia I. Placental bed and uterine boundary vessels Douglas R. Shanklin, MD, and Baha M. Sibai, MD Memphis, Tennessee Biopsy specimens were obtained under direct vision at the time of cesarean section from 42 patients (33 preeclamptic and nine normotensive patients) and from three hysterectomies (all in normotensive patients). Mean gestational age was 32.8 ± 0.9 weeks (range, 26 to 40 weeks) in women with preeclampsia and 36.1 ± 1.1 weeks in normotensive women (range, 32 to 41 weeks). Tissues were obtained from the central placental bed and from non placental sites. Specimens were examined histologically and by electron microscopy. Ultrastructural changes in small vessels, primarily venules, were compared with preeclamptic blood pressure. Extensive ultrastructural endothelial injury was found consistently in both site and nonsite areas in all of the specimens from women with preeclampsia but not in normotensive women (p < 0.0001). There was no apparent correlation between the type or degree of endothelial damage and maternal hypertension. The same types and relative severity of specific vascular injury were present in both placental and nonplacental sites. Endothelial and derivative vascular injury occurs more or less uniformly in the uterus in preeclampsia, especially along the boundary zone between maternal and fetal tissues. (AM J OBSTET GVNECOL 1989;161 :735-41.)
Key words: Placental site biopsy. preeclampsia, vascular injury, ultrastructural changes, uterine boundary zone The continuing search for both the cause and pathogenesis of preeclamptic toxemia (gestosis humanis) over the past 20 years has turned to the integrity of uterine vessels and changes in coagulation, which may relate to endothelial injury as shown by histologic and electron microscopic methods!·18 Histologic,I.3. 9. 10. 13 transmission ultrastructurap·7. 12. 13. 14. 18. 19 and scanning electron microscopic" studies have considered uterine arteries at several levels of the arborization, as shown especially clearly by Sheppard and Bonnar. 18. 19 Interest has been directed primarily at spiral arteries* and more general features of the placental bed. I. 3. 5·7.12·14.17·19 These largely European studies have identified a number of lesions: acute atherosis, I. 7. 10. 13 hyperplastic arteriosclerosis, 13 and alterations I. 9. 13. 18. 19 in the expected physiologic restructuring of the spiral arteries during pregnancy.3.5. 6.13.14.17·19 The cited articles account for at least 426 placental bed biopsies and samples of placental site tissue from uteri studied after cesarean section hysterectomy or pregnancy termination by earlier hysterec-
From the Departments of Pathology and Obstetrics and Gynecology, University of Tennessee, MemphIS. Presented at the Ninth Annual Meetmg of the Society of Pennatal Obstetricians, New Orleans, Louisiana, February 2-4,1989. Reprint requests: Douglas R. Shanklin, MD, Department of Pathology, University of Tennessee, MemphIS, 800 Madison Ave., Memphis, TN 38163. *References 1, 3, 5-7, 9, 12-14, 17-19. tReferences 1, J, 5, 6, 9, 12-14, 17-19. 616113891
tomy. Two large series were identified only as "over 100"13 and "several hundred"3 biopsies, comprising a possible total of perhaps 750 individual cases studied to date. This material was derived from clinical backgrounds of normal pregnancy,t preeclampsia or other hypertensive states;' 7. 9. 10. 13. 18. 19 and intrauterine growth retardation. I. 10. 18. 19 Some studies included multiple categories of clinical condition of either mother or fetus. I. 9.10.18.19 Several studies also examined placentas;' 7. 10. 13. 18. 19 including two in which the placenta remained in situ after hysterectomy!' 13 The basic techniques for obtaining placental bed biopsies and an overview of the information obtained up to 1986 were summarized in a critical review by Robertson et al. 11 The observation that the placental bed is a heterogeneous tissue field, making sampling difficult, has been emphasized in a number of reports. 3. 13. 14 Less often, but no less emphatically, several comparative studies have indicated that no distinction can be made between normotensive and hypertensive cases by histologic or ultrastructural examination of spiral arteries,18. 19 despite the finding by Brosens et al.I that acute atherosis was found principally in hypertension combined with fetal growth retardation and not in normotensive pregnancies, a finding later disputed presumptively by study of 39 placentas but not placental beds.1O In light of these studies and the ongoing need to elucidate primary vascular events in early preeclampsia, we began in mid-1987 to collect central placental bed biopsy material at time of cesarean section delivery with
735
736
Shanklin and Sibai
Fig. 1. Normal small venule in placental bed region of the uterine boundary zone, preeclampsia. Platelets and erythrocytes indicate a diameter of the vessel at IS to 20 !Lm. Note the excellent preservation of platelets. Bundles of collagen (C) are outside the vascular wall (arrows), which is composed of flat endothelium with some overlapping of cells and tight junctions. Occasional normal venules are found in the boundary zone of women with preeclampsia. Figs. 1 to 3 show the range of change from normal to severe endothelial injury. (Original electron microscopic magnification x3150.)
simultaneous biopsy of non placental site uterine boundary tissue, with emphasis on smaller blood vessels. Histologic and transmission electron microscopic studies to date are the subject of this report. Extensive ultrastructural endothelial injury was found consistently in both placental site and uterine non placental site boundary myometrium in all of the specimens from the preeclamptic patients but not in specimens from normotensive women (p < 0.0001). Material and methods
Informed consent for placental bed and uterine boundary biopsies was obtained from various patients in the prenatal clinics of the E. H. Crump Hospital who were scheduled for cesarean section for obstetric or perinatal indications. A total of 42 patients, 33 with preeclampsia (systolic blood presure > 140 mm Hg, or diastolic blood pressure >90 torr or both, with proteinuria >300 mg/24 hours), and nine normotensive
September 1989 Am J Obstet Gynecol
Fig. 2. Segmentally injured large venule in placental bed region of the uterine boundary zone, preeclampsia. Vessel diameter is approximately 20 to 30 !Lm. Enlarged endothelial nuclei (N) and segmentally swollen cytoplasm (arrows). Note also cavitated endothelial mitochondria (M). Similar change in smaller venules or somewhat greater change in vessels of this size would show reduction in lumen. (Original electron microscopic magnification x 1600.)
women had biopsies. Three cesarean section hysterectomies from normotensive patients were also studied; thus a total of 45 patients (12 normotensive and 33 hypertensive) was studied. At the time of cesarean section, biopsy specimens approximately 0.8 X 0.6 X 1.0 em in size were excised by sharp dissection and were immediately placed in the appropriate fixative. Placental site samples were taken from the center of the attachment site just after placental removal, and comparable nonsite samples were taken from the opposite uterine wall. On receipt in the pathology laboratory, appropriate accession numbers were assigned and the samples were trimmed into pieces for paraffin or plastic embedding according to ordinary techniques, beginning with 10% neutral buffered aqueous formalin for histology and net 1.0% glutaraldehyde buffered in Sorenson's solution for ultrastructural study. Paraffin sections were cut at 5 /-Lm and were stained by hematoxylin and eosin for light microscopy. Histologic preparations were available for all 45 patients. Plastic sections for electron microscopy were pre-
Volume 161 l\umber 3
Fig. 3. Marked reduction of lumen of large venule, placental bed regIOn of uterine boundary zone, preeclampsia. One intact erythrocyte (Ei is pressed by greatly swollen endothelial cells including herniated endothelial cytoplasm from a different plane (e). Venular basement membrane is irregular but intact (arrows). (Original electron microscopic magnification x 8000.)
pared in the usual manner with lead acetate staining. Electron micrographs were taken at X 1600 to x 2500 original magnification in a Zeiss EM-IOC and were printed at an average further enlargement of x 2.7. Electron micrographs were available for 27 of 33 preeclamptic women and for 9 of 12 normotensive women. Results
Vascular and ultrastructural changes were compared between the two groups and with preeclamptic maternal blood pressure. Clinical aspects. The average gestational age in preeclamptic pregnancies was 32.8 ± 0.9 weeks (range, 26 to 40 weeks) and 36.1 ± 1.1 weeks in normotensive pregnancies (range, 32 to 41 weeks; the difference was not significant, t = 0.50, P > 0.5). Maximum recorded blood pressures fulfilled criteria used in the primary diagnosis of preeclampsia (systolic blood pressure> 140 mm Hg, diastolic blood pressure >90 torr, or both, with proteinuria >300 mg/24 hr). The variability of blood pressure in each patient was noted, but since no cor-
Ultrastructural aspects of preeclampsia 737
Fig. 4. Partial surface erosion, swollen endothelium, large venule, non placental site region of uterine boundary zone, preeclampsia. Considerable swelling is marked by zonal and linear erosion (arrows). Closely proximate adventitial collagen lies next to hypertrophic uterine muscle (M). (Original electron microscopic magnification x 4000.)
relation with average or maximum values and tissue findings could be demonstrated, further correlation with the variability was not undertaken. There were no maternal deaths in this series. Perinatal mortality was low; there were no stillbirths and only one neonatal death in the hypertensive group, which yielded a group rate of 3.0% and 2.2% overall. None of these patients had eclamptic seizures. The differences in blood pressures were strongly significant (systolic: t = 10.5, P < 0.001; diastolic: t = 10.8, P< 0.001). The more than 3-week difference in mean gestational age suggested some significance by Student's t test (t = 2.32,0.05 > P > 0.02). The F value for variation of population, however, was 5.38 (not significant), which devalued the t test result. We anticipate that enlargement of the normotensive portion of the series will further reduce this limited difference without impairing the distinction with regard to vascular injury. Pathologic aspects. Histological findings were helpful in assessing the ultrastructural changes and will be described first.
738
Shanklin and Sibai
Fig. 5. Linear fibrin deposition (arrows) on eroded endothelium, large venule, nonplacental site region, uterine boundary zone, preeclampsia. Note intermingling of cellular debris and strands of fibrin. (Original electron microscopic magnification X 4000.)
Histologic changes. Histologic changes that were readily observed included some apparent medial hypertrophy of smaller arteries and arterioles, primarily but not exclusively in the placental site, dilated veins and venules with and without thrombosis, occasional placental site trophoblastic giant cells, focal hemorrhage, some interstitial edema, and uterine smooth muscle reflecting the expected gestational hypertrophy. The adaptive changes in spiral arteries were seen in most specimens, but no distinction with regard to degree or character of the change was observed between preeclamptic and normal subjects. Decidua and regressed or atrophic endometrial glands were observed less often. Occasional lymphocytes were found in the junctional tissues and occasional circulating blood elements, such as polymorphonuclear leukocytes, macrophages, and platelets were found in addition to many erythrocytes. Whereas overt thrombosis was uncommon, small foci of fibrin deposit were found more often, especially in relation to tears and erosions of the endothelium, and in overt tears into the perivascular space and myometrium. None of the changes were
September 1989 Am J Obstet Gynecol
Fig. 6. Compacted heavy fibrin deposition (arrows) and loose luminal fibrin replacing extensively eroded endothelium, nonplacental site region, uterine boundary zone, preeclampsia. (Original electron microscopic magnification X 2500.)
present such that unequivocal, consistent identification of placental site as opposed to non placental site sampling was possible. Placental site trophoblast, when present, would identify the site, but only one third of these biopsies contained trophoblast. The possibility that trophoblastic site giant cells may be found beyond the placental bed was enhanced by one case in which both biopsies contained them. More importantly, an overall view of the boundary between the uterus and the placental sac is possible from a composite of these observations. Vascular injury and tears into uterine soft tissues, with hemorrhage, are commonplace; essentially these processes occur both in the placental site and elsewhere throughout the boundary zone either equally randomly or with comparable frequency. In fact, such observations make it possible to define the boundary zone as the combination of the atrophic or regressive endometrium, which is displaced and compressed by the pregnancy sac, plus the interdigitating connective tissues of the inner myometrium including blood vessels. The ultimate maternal vascularization of the placental site is therefore a modification of this region, with
Ultrastructural aspects of preeclampsia 739
Volume 161 Number 3
Fig. 7. Rupturing bleb (arrow) of endothelium, large venule, non placental site region of uterine boundary zone, preeclampsia. This is one possible precursor to overt paraluminal erosion of endothelium. Note a similar structure (B) near lower edge of photograph, protruding from a different level in the block. (Original electron microscopic magnification
Fig. 8. Microhemorrhage into connective tissue of nonplacental site region of uterine boundary zone, preeclampsia. Note fibrin deposition onto surfaces of the dissection planes of the hemorrhage and the absence of venular media or basement membrane. (Original electron microscopic magnification x 2500.)
x3150.)
the added change of hypertrophy and dilation in the spiral arteries with additional changes in the deeper arterial arcades of the myometrium. Since the pathologic changes seemed to us to be more or less uniform throughout the boundary zone, the larger vessels beneath the placental site itself would appear to be less significant. Our subsequent studies of the ultrastructural details of these regions have confirmed the unitary and integral nature of the boundary zone. Ultrastructural changes. There was no apparent correlation between the type or degree of endothelial damage and the level of maternal hypertension. Endothelial injury was pervasive in the preeclamptic women, but the same types and relative severity of specific changes were present in samples from both placental and nonplacental boundary sites. We concentrated on smaller vessels, particularly the venules, as the most likely site for injury (Figs. 1 and 2). The restricted size of electron microscopic samples signified that all kinds and degrees of change were not present in all cases. Nevertheless,
a coherent spectrum of change was obvious. The endothelium was both swollen (Fig. 3) and eroded (Fig. 4). Even mild swelling was associated with enlargement of endothelial nuclei (Fig. 2); the result was apparent reduction of lumen (which was sometimes very extreme), with marked cytoplasmic swelling of endothelium and a reduced number of nuclei per crosssectional field (Fig. 3). Luminal fibrin polymerization (Fig. 5) was observed in the case of both swollen and eroded endothelium, but was more obvious and more common with erosion (Fig. 6). In some cases erosion was incomplete with a light deposit of fibrin, but in others complete erosion was related to a heavy deposit (Fig. 6). Endothelial cytoplasmic tears short of erosion of the whole thickness were found, and as the sample became larger, at least one source for these focal tears became apparent: the endothelium occasionally showed vacuoles that enlarged into blebs bulging from the surface (Fig. 7). These blebs would then rupture, exposing the endothelial cytosol to the vascular lumen. Subendothelial edema was seen less often. Platelets
740
Shanklin and Sibai
were found but not as a regular feature of these early lesions. Instead, linear or circumferential fibrin deposit seemed to be the principal means by which the earliest repair of the injury was attempted. Hemorrhages into the deeper boundary tissues, although not common, were almost always associated with fibrin deposit at that level (Fig. 8). The vacuoles have been identified as endothelial mitochondria, which initially show swelling with loss of cristae. A general expansion of cytoplasm, probably from electrolyte shifts, was a conspicuous feature. Platelet adhesion was observed, but was not a consistent feature. Altered permeability was seen occasionally with subendothelial plasma and fluid accumulation. Atherotic changes were not seen by electron microscopy; in two cases mucinous degeneration of larger vessels was seen on light microscopy. Comment
As has been noted, prior studies have concentrated heavily on spiral arteries* in the placental bed. 1. 3. 5-7. 12. 13. 14. 17-20 Critical comparative studies have shown no effective distinction in spiral arteries in the placental bed between hypertensive and normotensive pregnancies, lB. 19 findings not explained by procedural and technical difficulties in obtaining biopsies. 14 On a priori grounds, we considered that the effects of changes in coagulation factors seen in preeclamptic toxemia4. 15. 16 and other aspects of the vascular lesion might be more easily identified in smaller vessels of the placental-maternal tissue boundary. The spiral arteries are subject to such profound physiologic adaptation,' and pathologic changes especially in late pregnancy, 13 that subtle lesions might not be easily distinguished. Moreover, recent challenges lO • lB. 19 to the putative specificity of acute atherosis and other changes in hypertensive pregnancies 7. 13 have led researchers to consider pathogenetic effects at differing zones or segments of the arterial tree. 1. 9.12 In fact, the scope of sampling may need to be extended in three dimensions, since significant differences in abnormal uterine waveform velocity have now been described between the fundal and the low arcuate segment vascular fields." Our observations extend this work with emphasis on the proximal venous side of the uterine boundary circulation. U nderstanding of the phenomenon of altered coagulation has also changed, with more precise and sensitive methods now available. I6 Some changes are present very early in pregnancy, such as decreased fibrinolytic activity (present by 12 weeks), and decreased fibrinolytic capacity (seen by 20 weeks)! Saleh et al. I6 examined clotting factors by a variety of techniques. They found levels of serum fibronectin to be increased and
September 1989 Am J Obstet Gynecol
anti plasmin to be decreased in preeclampsia. Both levels returned toward normal post partum, with a greater relative change in patients with severe preeclampsia. 16 It is tempting to consider that the increased fibronectin, a substance known to be produced by both liver and endothelium,I6 originates in the injured smaller vessels we have examined. Obviously, further study is necessary to establish the relationship of these observations. In particular, correlation with early changes in the coagulation mechanism must await the availability of appropriate uterine tissue from early second trimester, if not before. Earlier tissue samples might also provide information on the vascular adaptations that result, near term, in a lack of significant differences in vascular resistance and other circulatory parameters between preeclamptic pregnancy and the nonpregnant state, whereas normal pregnancy differs from both. 20 Meanwhile we are attempting to describe the mature vascular lesions in the third trimester. The existence of a circulating factor toxic to endothelium has been suggested by recent evidence. I5 It is unclear whether this is a primary agent or a factor involved in further pathogenesis promoting extension of the lesions. The ultrastructural changes are those of endothelial injury, probably to cytoplasmic organelles and very possibly to mitochondria. Both diffuse and organelle swelling occur, and either leads to erosion of endothelium. Our observations suggest that this erosion begins with loss of the cortical cytoplasm rather than slough of the entire endothelial cell. Fibrin is deposited during the erosive process just as would occur in most forms of endothelial i~ury. Occasionally fluid appears to collect behind the endothelium. This can occur either by transudation or as dissection from another point along the vessel. Of interest are the recent studies of Glawanakowa and Popowa. B They examined peripheral blood for circulating desquamated endothelial cells and found a progressive rise during pregnancy, with a further increase in patients with edema-proteinuriahypertension gestosis (preeclampsia). Once erosion has begun, whole-cell desquamation can proceed from that focus. Our later-stage lesions (Figs. 7 and 8) can represent either whole-cell desquamation or multistage erosion. Confirmation of the results of Glawanakowa and Popowa will contribute to a better understanding of the vascular effects of pregnancy as modified by preeclampsia. A major source for circulating endothelial cells might be the venules we have examined, since they are numerous. Our experience has shown that small vessels, particularly venules, are injured in preeclampsia irrespective of their location in the boundary zone. It remains to be determined whether comparable vascular injury occurs elsewhere in the uterus or in
Ultrastructural aspects of preeclampsia
Volume 161 Number 3
other body sites. Changes in endothelial organelles such as mitochondria suggest a metabolic link that merits further exploration. We thank Deborah Hollis for preparation of tissues for electron microscopy and Frank Moretta for the electron micrographs. REFERENCES 1. Brosens I, Dixon HG, Robertson WB. Fetal growth retardation and the arteries of the placental bed. Br] Obstet Gynecol 1977;84:656-63. 2. Brosens I, Renaer M. On the pathogenesis of placental infarcts in pre-eclampsia.] Obstet Gynaecol Br Commonw 1972;79:794-9. 3. Brosens I, Robertson WB, Dixon HG. The physiological response of the vessels of the placental bed to normal pregnancy.] Pathol Bacteriol 1967;93:569-80. 4. Condie RG, Ogston D. Sequential studies on components of the haemostatic mechanism in pregnancy with particular reference to the development of pre-eclampsia. Br] Obstet Gynaecol 1976;83:938-42. 5. De Wolfe F, De Wolf-Peeters C, Brosens l. Ultrastructure of the spiral arteries in the human placental bed at the end of normal pregnancy. AM] OBSTET GYNECOL 1973; 117:833-48. 6. De Wolf F, De Wolf-Peeters C, Brosens I, Robertson WB. The human placental bed: electron microscopic study of trophoblastic invasion of spiral arteries. AM] OBSTET GyNECOL 1980;137:58-70. 7. De Wolf F, Robertson WB, Brosens I. The ultrastructure of acute atherosis in hypertensive pregnancy. AM] OBSTET GYNECOL 1975;123:164-74. 8. Glawanakowa W, Popowa G. Untersuchungen tiber die desquamation des gefassendothels bei EPH-gestosen. Cor Vasa 1988;30:140-5. 9. Khong TV, De Wolf F, Robertson WB, Brosens l. Inadequate maternal vascular response to placentation in pregnancies complicated by pre-eclampsia and by small-for-
10.
11. 12. 13. 14.
15. 16. 17.
18.
19.
20.
741
gestational age infants. Br ] Obstet Gynaecol 1986; 93:1049-59. Khong TV, Pearce ]M, Robertson WB. Acute atherosis in preeclampsia: maternal determinants and fetal outcome in the presence of the lesion. AM ] OBSTET GYNECOL 1987;157:360-3. Leiberman]R, Meizner I, Mazor M, Insler V. Blood supply to the uterus in preeclampsia. Eur] Obstet Gynecol Reprod BioI 1988;28:23-7. Marais WD. Human decidual spiral arteries; ultrastructural of the intima in normal vessels. ] Obstet Gynaecol Br Commonw 1968;75:552-67. Robertson WB, Brosens I, Dixon HG. The pathological response of the vessels of the placental bed to hypertensive pregnancy.] Pathol Bacteriol 1967;93:581-91. Robertson WB, Khong TV, Brosens I, De Wolf F, Sheppard BL, Bonnar]. The placental bed biopsy: review from three European centers. AM] OBSTET GYNECOL 1986; 155:401-12. Rodgers GM, Taylor RN, Roberts ]M. Preeclampsia is associated with a serum factor cytotoxic to human endothelial cells. AM] OBSTET GYNECOL 1988; 159:908-14. Saleh AA, Bottoms SF, Weich RA, Ali AM, Mariona FG, Mammen RF. Preeclampsia, delivery, and the hemostatic system. AM] OBSTET GYNECOL 1987; 157:331-6. Sheppard BL, Bonnar J. Scanning electron microscopy of the human placenta and decidual spiral arteries in normal pregnancy.] Obstet Gynaecol Br Commonw 1974; 81 :20-9. Sheppard BL, Bonnar]. The ultrastructure of the arterial supply of the human placenta in pregnancy complicated by fetal growth retardation. Br] Obstet Gynaecol 1976; 83:948-59. Sheppard BL, Bonnar J. An ultrastructural study of uteroplacental spiral arteries in hypertensive and normotensive pregnancy and fetal growth retardation. Br ] Obstet Gynaecol 1981 ;88:695-705. Sohn Ch, Fendel H. Arterielle renale und uterine Durchblutung in normalen und gestotischen Schwangerschaften. Z Geburtshilfe Perinatol 1988; 192:43-8.
Key words: Placental site biopsy. preeclampsia, vascular injury, ultrastructural changes, uterine boundary zone The continuing search for both the cause and pathogenesis of preeclamptic toxemia (gestosis humanis) over the past 20 years has turned to the integrity of uterine vessels and changes in coagulation, which may relate to endothelial injury as shown by histologic and electron microscopic methods!·18 Histologic,I.3. 9. 10. 13 transmission ultrastructurap·7. 12. 13. 14. 18. 19 and scanning electron microscopic" studies have considered uterine arteries at several levels of the arborization, as shown especially clearly by Sheppard and Bonnar. 18. 19 Interest has been directed primarily at spiral arteries* and more general features of the placental bed. I. 3. 5·7.12·14.17·19 These largely European studies have identified a number of lesions: acute atherosis, I. 7. 10. 13 hyperplastic arteriosclerosis, 13 and alterations I. 9. 13. 18. 19 in the expected physiologic restructuring of the spiral arteries during pregnancy.3.5. 6.13.14.17·19 The cited articles account for at least 426 placental bed biopsies and samples of placental site tissue from uteri studied after cesarean section hysterectomy or pregnancy termination by earlier hysterec-
From the Departments of Pathology and Obstetrics and Gynecology, University of Tennessee, MemphIS. Presented at the Ninth Annual Meetmg of the Society of Pennatal Obstetricians, New Orleans, Louisiana, February 2-4,1989. Reprint requests: Douglas R. Shanklin, MD, Department of Pathology, University of Tennessee, MemphIS, 800 Madison Ave., Memphis, TN 38163. *References 1, 3, 5-7, 9, 12-14, 17-19. tReferences 1, J, 5, 6, 9, 12-14, 17-19. 616113891
tomy. Two large series were identified only as "over 100"13 and "several hundred"3 biopsies, comprising a possible total of perhaps 750 individual cases studied to date. This material was derived from clinical backgrounds of normal pregnancy,t preeclampsia or other hypertensive states;' 7. 9. 10. 13. 18. 19 and intrauterine growth retardation. I. 10. 18. 19 Some studies included multiple categories of clinical condition of either mother or fetus. I. 9.10.18.19 Several studies also examined placentas;' 7. 10. 13. 18. 19 including two in which the placenta remained in situ after hysterectomy!' 13 The basic techniques for obtaining placental bed biopsies and an overview of the information obtained up to 1986 were summarized in a critical review by Robertson et al. 11 The observation that the placental bed is a heterogeneous tissue field, making sampling difficult, has been emphasized in a number of reports. 3. 13. 14 Less often, but no less emphatically, several comparative studies have indicated that no distinction can be made between normotensive and hypertensive cases by histologic or ultrastructural examination of spiral arteries,18. 19 despite the finding by Brosens et al.I that acute atherosis was found principally in hypertension combined with fetal growth retardation and not in normotensive pregnancies, a finding later disputed presumptively by study of 39 placentas but not placental beds.1O In light of these studies and the ongoing need to elucidate primary vascular events in early preeclampsia, we began in mid-1987 to collect central placental bed biopsy material at time of cesarean section delivery with
735
736
Shanklin and Sibai
Fig. 1. Normal small venule in placental bed region of the uterine boundary zone, preeclampsia. Platelets and erythrocytes indicate a diameter of the vessel at IS to 20 !Lm. Note the excellent preservation of platelets. Bundles of collagen (C) are outside the vascular wall (arrows), which is composed of flat endothelium with some overlapping of cells and tight junctions. Occasional normal venules are found in the boundary zone of women with preeclampsia. Figs. 1 to 3 show the range of change from normal to severe endothelial injury. (Original electron microscopic magnification x3150.)
simultaneous biopsy of non placental site uterine boundary tissue, with emphasis on smaller blood vessels. Histologic and transmission electron microscopic studies to date are the subject of this report. Extensive ultrastructural endothelial injury was found consistently in both placental site and uterine non placental site boundary myometrium in all of the specimens from the preeclamptic patients but not in specimens from normotensive women (p < 0.0001). Material and methods
Informed consent for placental bed and uterine boundary biopsies was obtained from various patients in the prenatal clinics of the E. H. Crump Hospital who were scheduled for cesarean section for obstetric or perinatal indications. A total of 42 patients, 33 with preeclampsia (systolic blood presure > 140 mm Hg, or diastolic blood pressure >90 torr or both, with proteinuria >300 mg/24 hours), and nine normotensive
September 1989 Am J Obstet Gynecol
Fig. 2. Segmentally injured large venule in placental bed region of the uterine boundary zone, preeclampsia. Vessel diameter is approximately 20 to 30 !Lm. Enlarged endothelial nuclei (N) and segmentally swollen cytoplasm (arrows). Note also cavitated endothelial mitochondria (M). Similar change in smaller venules or somewhat greater change in vessels of this size would show reduction in lumen. (Original electron microscopic magnification x 1600.)
women had biopsies. Three cesarean section hysterectomies from normotensive patients were also studied; thus a total of 45 patients (12 normotensive and 33 hypertensive) was studied. At the time of cesarean section, biopsy specimens approximately 0.8 X 0.6 X 1.0 em in size were excised by sharp dissection and were immediately placed in the appropriate fixative. Placental site samples were taken from the center of the attachment site just after placental removal, and comparable nonsite samples were taken from the opposite uterine wall. On receipt in the pathology laboratory, appropriate accession numbers were assigned and the samples were trimmed into pieces for paraffin or plastic embedding according to ordinary techniques, beginning with 10% neutral buffered aqueous formalin for histology and net 1.0% glutaraldehyde buffered in Sorenson's solution for ultrastructural study. Paraffin sections were cut at 5 /-Lm and were stained by hematoxylin and eosin for light microscopy. Histologic preparations were available for all 45 patients. Plastic sections for electron microscopy were pre-
Volume 161 l\umber 3
Fig. 3. Marked reduction of lumen of large venule, placental bed regIOn of uterine boundary zone, preeclampsia. One intact erythrocyte (Ei is pressed by greatly swollen endothelial cells including herniated endothelial cytoplasm from a different plane (e). Venular basement membrane is irregular but intact (arrows). (Original electron microscopic magnification x 8000.)
pared in the usual manner with lead acetate staining. Electron micrographs were taken at X 1600 to x 2500 original magnification in a Zeiss EM-IOC and were printed at an average further enlargement of x 2.7. Electron micrographs were available for 27 of 33 preeclamptic women and for 9 of 12 normotensive women. Results
Vascular and ultrastructural changes were compared between the two groups and with preeclamptic maternal blood pressure. Clinical aspects. The average gestational age in preeclamptic pregnancies was 32.8 ± 0.9 weeks (range, 26 to 40 weeks) and 36.1 ± 1.1 weeks in normotensive pregnancies (range, 32 to 41 weeks; the difference was not significant, t = 0.50, P > 0.5). Maximum recorded blood pressures fulfilled criteria used in the primary diagnosis of preeclampsia (systolic blood pressure> 140 mm Hg, diastolic blood pressure >90 torr, or both, with proteinuria >300 mg/24 hr). The variability of blood pressure in each patient was noted, but since no cor-
Ultrastructural aspects of preeclampsia 737
Fig. 4. Partial surface erosion, swollen endothelium, large venule, non placental site region of uterine boundary zone, preeclampsia. Considerable swelling is marked by zonal and linear erosion (arrows). Closely proximate adventitial collagen lies next to hypertrophic uterine muscle (M). (Original electron microscopic magnification x 4000.)
relation with average or maximum values and tissue findings could be demonstrated, further correlation with the variability was not undertaken. There were no maternal deaths in this series. Perinatal mortality was low; there were no stillbirths and only one neonatal death in the hypertensive group, which yielded a group rate of 3.0% and 2.2% overall. None of these patients had eclamptic seizures. The differences in blood pressures were strongly significant (systolic: t = 10.5, P < 0.001; diastolic: t = 10.8, P< 0.001). The more than 3-week difference in mean gestational age suggested some significance by Student's t test (t = 2.32,0.05 > P > 0.02). The F value for variation of population, however, was 5.38 (not significant), which devalued the t test result. We anticipate that enlargement of the normotensive portion of the series will further reduce this limited difference without impairing the distinction with regard to vascular injury. Pathologic aspects. Histological findings were helpful in assessing the ultrastructural changes and will be described first.
738
Shanklin and Sibai
Fig. 5. Linear fibrin deposition (arrows) on eroded endothelium, large venule, nonplacental site region, uterine boundary zone, preeclampsia. Note intermingling of cellular debris and strands of fibrin. (Original electron microscopic magnification X 4000.)
Histologic changes. Histologic changes that were readily observed included some apparent medial hypertrophy of smaller arteries and arterioles, primarily but not exclusively in the placental site, dilated veins and venules with and without thrombosis, occasional placental site trophoblastic giant cells, focal hemorrhage, some interstitial edema, and uterine smooth muscle reflecting the expected gestational hypertrophy. The adaptive changes in spiral arteries were seen in most specimens, but no distinction with regard to degree or character of the change was observed between preeclamptic and normal subjects. Decidua and regressed or atrophic endometrial glands were observed less often. Occasional lymphocytes were found in the junctional tissues and occasional circulating blood elements, such as polymorphonuclear leukocytes, macrophages, and platelets were found in addition to many erythrocytes. Whereas overt thrombosis was uncommon, small foci of fibrin deposit were found more often, especially in relation to tears and erosions of the endothelium, and in overt tears into the perivascular space and myometrium. None of the changes were
September 1989 Am J Obstet Gynecol
Fig. 6. Compacted heavy fibrin deposition (arrows) and loose luminal fibrin replacing extensively eroded endothelium, nonplacental site region, uterine boundary zone, preeclampsia. (Original electron microscopic magnification X 2500.)
present such that unequivocal, consistent identification of placental site as opposed to non placental site sampling was possible. Placental site trophoblast, when present, would identify the site, but only one third of these biopsies contained trophoblast. The possibility that trophoblastic site giant cells may be found beyond the placental bed was enhanced by one case in which both biopsies contained them. More importantly, an overall view of the boundary between the uterus and the placental sac is possible from a composite of these observations. Vascular injury and tears into uterine soft tissues, with hemorrhage, are commonplace; essentially these processes occur both in the placental site and elsewhere throughout the boundary zone either equally randomly or with comparable frequency. In fact, such observations make it possible to define the boundary zone as the combination of the atrophic or regressive endometrium, which is displaced and compressed by the pregnancy sac, plus the interdigitating connective tissues of the inner myometrium including blood vessels. The ultimate maternal vascularization of the placental site is therefore a modification of this region, with
Ultrastructural aspects of preeclampsia 739
Volume 161 Number 3
Fig. 7. Rupturing bleb (arrow) of endothelium, large venule, non placental site region of uterine boundary zone, preeclampsia. This is one possible precursor to overt paraluminal erosion of endothelium. Note a similar structure (B) near lower edge of photograph, protruding from a different level in the block. (Original electron microscopic magnification
Fig. 8. Microhemorrhage into connective tissue of nonplacental site region of uterine boundary zone, preeclampsia. Note fibrin deposition onto surfaces of the dissection planes of the hemorrhage and the absence of venular media or basement membrane. (Original electron microscopic magnification x 2500.)
x3150.)
the added change of hypertrophy and dilation in the spiral arteries with additional changes in the deeper arterial arcades of the myometrium. Since the pathologic changes seemed to us to be more or less uniform throughout the boundary zone, the larger vessels beneath the placental site itself would appear to be less significant. Our subsequent studies of the ultrastructural details of these regions have confirmed the unitary and integral nature of the boundary zone. Ultrastructural changes. There was no apparent correlation between the type or degree of endothelial damage and the level of maternal hypertension. Endothelial injury was pervasive in the preeclamptic women, but the same types and relative severity of specific changes were present in samples from both placental and nonplacental boundary sites. We concentrated on smaller vessels, particularly the venules, as the most likely site for injury (Figs. 1 and 2). The restricted size of electron microscopic samples signified that all kinds and degrees of change were not present in all cases. Nevertheless,
a coherent spectrum of change was obvious. The endothelium was both swollen (Fig. 3) and eroded (Fig. 4). Even mild swelling was associated with enlargement of endothelial nuclei (Fig. 2); the result was apparent reduction of lumen (which was sometimes very extreme), with marked cytoplasmic swelling of endothelium and a reduced number of nuclei per crosssectional field (Fig. 3). Luminal fibrin polymerization (Fig. 5) was observed in the case of both swollen and eroded endothelium, but was more obvious and more common with erosion (Fig. 6). In some cases erosion was incomplete with a light deposit of fibrin, but in others complete erosion was related to a heavy deposit (Fig. 6). Endothelial cytoplasmic tears short of erosion of the whole thickness were found, and as the sample became larger, at least one source for these focal tears became apparent: the endothelium occasionally showed vacuoles that enlarged into blebs bulging from the surface (Fig. 7). These blebs would then rupture, exposing the endothelial cytosol to the vascular lumen. Subendothelial edema was seen less often. Platelets
740
Shanklin and Sibai
were found but not as a regular feature of these early lesions. Instead, linear or circumferential fibrin deposit seemed to be the principal means by which the earliest repair of the injury was attempted. Hemorrhages into the deeper boundary tissues, although not common, were almost always associated with fibrin deposit at that level (Fig. 8). The vacuoles have been identified as endothelial mitochondria, which initially show swelling with loss of cristae. A general expansion of cytoplasm, probably from electrolyte shifts, was a conspicuous feature. Platelet adhesion was observed, but was not a consistent feature. Altered permeability was seen occasionally with subendothelial plasma and fluid accumulation. Atherotic changes were not seen by electron microscopy; in two cases mucinous degeneration of larger vessels was seen on light microscopy. Comment
As has been noted, prior studies have concentrated heavily on spiral arteries* in the placental bed. 1. 3. 5-7. 12. 13. 14. 17-20 Critical comparative studies have shown no effective distinction in spiral arteries in the placental bed between hypertensive and normotensive pregnancies, lB. 19 findings not explained by procedural and technical difficulties in obtaining biopsies. 14 On a priori grounds, we considered that the effects of changes in coagulation factors seen in preeclamptic toxemia4. 15. 16 and other aspects of the vascular lesion might be more easily identified in smaller vessels of the placental-maternal tissue boundary. The spiral arteries are subject to such profound physiologic adaptation,' and pathologic changes especially in late pregnancy, 13 that subtle lesions might not be easily distinguished. Moreover, recent challenges lO • lB. 19 to the putative specificity of acute atherosis and other changes in hypertensive pregnancies 7. 13 have led researchers to consider pathogenetic effects at differing zones or segments of the arterial tree. 1. 9.12 In fact, the scope of sampling may need to be extended in three dimensions, since significant differences in abnormal uterine waveform velocity have now been described between the fundal and the low arcuate segment vascular fields." Our observations extend this work with emphasis on the proximal venous side of the uterine boundary circulation. U nderstanding of the phenomenon of altered coagulation has also changed, with more precise and sensitive methods now available. I6 Some changes are present very early in pregnancy, such as decreased fibrinolytic activity (present by 12 weeks), and decreased fibrinolytic capacity (seen by 20 weeks)! Saleh et al. I6 examined clotting factors by a variety of techniques. They found levels of serum fibronectin to be increased and
September 1989 Am J Obstet Gynecol
anti plasmin to be decreased in preeclampsia. Both levels returned toward normal post partum, with a greater relative change in patients with severe preeclampsia. 16 It is tempting to consider that the increased fibronectin, a substance known to be produced by both liver and endothelium,I6 originates in the injured smaller vessels we have examined. Obviously, further study is necessary to establish the relationship of these observations. In particular, correlation with early changes in the coagulation mechanism must await the availability of appropriate uterine tissue from early second trimester, if not before. Earlier tissue samples might also provide information on the vascular adaptations that result, near term, in a lack of significant differences in vascular resistance and other circulatory parameters between preeclamptic pregnancy and the nonpregnant state, whereas normal pregnancy differs from both. 20 Meanwhile we are attempting to describe the mature vascular lesions in the third trimester. The existence of a circulating factor toxic to endothelium has been suggested by recent evidence. I5 It is unclear whether this is a primary agent or a factor involved in further pathogenesis promoting extension of the lesions. The ultrastructural changes are those of endothelial injury, probably to cytoplasmic organelles and very possibly to mitochondria. Both diffuse and organelle swelling occur, and either leads to erosion of endothelium. Our observations suggest that this erosion begins with loss of the cortical cytoplasm rather than slough of the entire endothelial cell. Fibrin is deposited during the erosive process just as would occur in most forms of endothelial i~ury. Occasionally fluid appears to collect behind the endothelium. This can occur either by transudation or as dissection from another point along the vessel. Of interest are the recent studies of Glawanakowa and Popowa. B They examined peripheral blood for circulating desquamated endothelial cells and found a progressive rise during pregnancy, with a further increase in patients with edema-proteinuriahypertension gestosis (preeclampsia). Once erosion has begun, whole-cell desquamation can proceed from that focus. Our later-stage lesions (Figs. 7 and 8) can represent either whole-cell desquamation or multistage erosion. Confirmation of the results of Glawanakowa and Popowa will contribute to a better understanding of the vascular effects of pregnancy as modified by preeclampsia. A major source for circulating endothelial cells might be the venules we have examined, since they are numerous. Our experience has shown that small vessels, particularly venules, are injured in preeclampsia irrespective of their location in the boundary zone. It remains to be determined whether comparable vascular injury occurs elsewhere in the uterus or in
Ultrastructural aspects of preeclampsia
Volume 161 Number 3
other body sites. Changes in endothelial organelles such as mitochondria suggest a metabolic link that merits further exploration. We thank Deborah Hollis for preparation of tissues for electron microscopy and Frank Moretta for the electron micrographs. REFERENCES 1. Brosens I, Dixon HG, Robertson WB. Fetal growth retardation and the arteries of the placental bed. Br] Obstet Gynecol 1977;84:656-63. 2. Brosens I, Renaer M. On the pathogenesis of placental infarcts in pre-eclampsia.] Obstet Gynaecol Br Commonw 1972;79:794-9. 3. Brosens I, Robertson WB, Dixon HG. The physiological response of the vessels of the placental bed to normal pregnancy.] Pathol Bacteriol 1967;93:569-80. 4. Condie RG, Ogston D. Sequential studies on components of the haemostatic mechanism in pregnancy with particular reference to the development of pre-eclampsia. Br] Obstet Gynaecol 1976;83:938-42. 5. De Wolfe F, De Wolf-Peeters C, Brosens l. Ultrastructure of the spiral arteries in the human placental bed at the end of normal pregnancy. AM] OBSTET GYNECOL 1973; 117:833-48. 6. De Wolf F, De Wolf-Peeters C, Brosens I, Robertson WB. The human placental bed: electron microscopic study of trophoblastic invasion of spiral arteries. AM] OBSTET GyNECOL 1980;137:58-70. 7. De Wolf F, Robertson WB, Brosens I. The ultrastructure of acute atherosis in hypertensive pregnancy. AM] OBSTET GYNECOL 1975;123:164-74. 8. Glawanakowa W, Popowa G. Untersuchungen tiber die desquamation des gefassendothels bei EPH-gestosen. Cor Vasa 1988;30:140-5. 9. Khong TV, De Wolf F, Robertson WB, Brosens l. Inadequate maternal vascular response to placentation in pregnancies complicated by pre-eclampsia and by small-for-
10.
11. 12. 13. 14.
15. 16. 17.
18.
19.
20.
741
gestational age infants. Br ] Obstet Gynaecol 1986; 93:1049-59. Khong TV, Pearce ]M, Robertson WB. Acute atherosis in preeclampsia: maternal determinants and fetal outcome in the presence of the lesion. AM ] OBSTET GYNECOL 1987;157:360-3. Leiberman]R, Meizner I, Mazor M, Insler V. Blood supply to the uterus in preeclampsia. Eur] Obstet Gynecol Reprod BioI 1988;28:23-7. Marais WD. Human decidual spiral arteries; ultrastructural of the intima in normal vessels. ] Obstet Gynaecol Br Commonw 1968;75:552-67. Robertson WB, Brosens I, Dixon HG. The pathological response of the vessels of the placental bed to hypertensive pregnancy.] Pathol Bacteriol 1967;93:581-91. Robertson WB, Khong TV, Brosens I, De Wolf F, Sheppard BL, Bonnar]. The placental bed biopsy: review from three European centers. AM] OBSTET GYNECOL 1986; 155:401-12. Rodgers GM, Taylor RN, Roberts ]M. Preeclampsia is associated with a serum factor cytotoxic to human endothelial cells. AM] OBSTET GYNECOL 1988; 159:908-14. Saleh AA, Bottoms SF, Weich RA, Ali AM, Mariona FG, Mammen RF. Preeclampsia, delivery, and the hemostatic system. AM] OBSTET GYNECOL 1987; 157:331-6. Sheppard BL, Bonnar J. Scanning electron microscopy of the human placenta and decidual spiral arteries in normal pregnancy.] Obstet Gynaecol Br Commonw 1974; 81 :20-9. Sheppard BL, Bonnar]. The ultrastructure of the arterial supply of the human placenta in pregnancy complicated by fetal growth retardation. Br] Obstet Gynaecol 1976; 83:948-59. Sheppard BL, Bonnar J. An ultrastructural study of uteroplacental spiral arteries in hypertensive and normotensive pregnancy and fetal growth retardation. Br ] Obstet Gynaecol 1981 ;88:695-705. Sohn Ch, Fendel H. Arterielle renale und uterine Durchblutung in normalen und gestotischen Schwangerschaften. Z Geburtshilfe Perinatol 1988; 192:43-8.