- Email: [email protected]
Picornavirus Infection in Early Murine Gestation: Significance of Maternal Illness
Placenta (2000), 21, 840–846 doi:10.1053/plac.2000.0577, available online at http://www.idealibrary.com on
Picornavirus Infection in Early Murine Gestation: Significance of Maternal Illness M. J. Abzuga and R. W. Tyson Departments of Pediatrics (Infectious Diseases) and Pathology, University of Colorado School of Medicine and The Children’s Hospital, Denver, Colorado, USA Paper accepted 1 June 2000
To evaluate whether maternal illness following picornavirus infection during pregnancy adversely affects placental and fetal health, mice were inoculated with the GDVII strain of Theiler’s murine encephalomyelitis virus or control cell lysate during days 4–7 of gestation. Gross appearance, histopathology and viral culture, and in situ hybridization positivity of placentae and fetuses from ill GDVII-infected, healthy GDVII-infected and control mice were compared. Twenty of 34 (59 per cent) GDVII-infected dams became clinically ill. More placenta–fetus pairs from ill mice were grossly abnormal (68 per cent) than from well GDVII-infected (51 per cent; P<0.01) or control mice (9 per cent; P<0.001). Virus was detected by in situ hybridization in 73 per cent of placentae and 29 per cent of fetuses from sick GDVII-infected dams, and in 85 per cent of placentae and 19 per cent of fetuses from healthy GDVII-infected mice (differences not significant). Histological abnormalities consisting of necrosis or an increase in hyaline tissue in the vascular labyrinth layer were similarly frequent in placentae from ill and well GDVII-infected mice (58 per cent versus 67 per cent, P=0.5). Viral RNA, inflammation and necrosis were evident in the heart, great vessels, brain and spinal cord of GDVII-infected fetuses. Infection with GDVII in early pregnancy produces a high rate of gross placental and fetal abnormalities. The rate of gross abnormalities exceeds the incidence of fetal infection and more closely parallels the rates of infection and histopathology in the placenta, suggesting that much of the damage to placenta–fetus pairs is a consequence of placental infection. In addition, the occurrence of viral-induced maternal illness is associated with additive risk to placental and fetal health not explained by an increased rate of placental or fetal infection. 2000 Harcourt Publishers Ltd Placenta (2000), 21, 840–846
INTRODUCTION Maternal infections during pregnancy with polioviruses and non-poliovirus enteroviruses are associated with fetal infection and adverse outcomes, such as spontaneous abortion, stillbirth, congenital defects and perinatal illness (Brown and Karunas, 1971; Cherry, 1990; Axelsson et al., 1993; Abzug, 1995). To investigate the pathogenesis of gestational enterovirus infections, we have developed an animal model using a murine picornavirus, Theiler’s murine encephalomyelitis virus (TMEV), in its natural host (Abzug, Rotbart and Levin, 1989; Abzug et al., 1991; Abzug, 1993; Abzug, 1997; Palmer et al., 1997). TMEV is a member of the cardiovirus genus in the picornavirus family, whose behaviour in mice is similar to that of the closely related enterovirus genus, also of the picornavirus family, in humans. We have shown in this model that infection with either the DA strain, a member of the less acutely neurovirulent group II strains of TMEV, or the more a To whom correspondence should be addressed at: The Children’s Hospital, Box B055, 1056 E. 19th Avenue, Denver, CO 80218, USA. Fax: +1 303-837-2707; E-mail: [email protected]
0143–4004/00/080840+07 $35.00/0
acutely neurovirulent group I strain, GDVII, in early gestation (days 4–7 of a 19 day gestation) is associated with high rates of placental and fetal infection, whereas infection in late pregnancy (days 12–13) is associated with a high rate of placental but not fetal infection (Abzug, Rotbart and Levin, 1989; Abzug et al., 1991; Abzug and Tyson, 1999). Maternal health factors, such as compromised uteroplacental blood flow and concomitant bacterial infection, may increase the risk of transplacental TMEV infection (Abzug, 1997). Infection with DA virus in early pregnancy can produce gross and histological abnormalities of placentae such as small placentae with fibrin deposition, dilated vascular lakes and infarction, and of fetuses, e.g. degenerated fetal tissue containing necrotic cellular debris; specific fetal organ defects are not apparent, and maternal health is not altered (Palmer et al., 1997). Because GDVII virus administered parenterally causes acute paralytic encephalomyelitis in some adult mice, whereas infection with DA virus does not cause short-term clinical changes (Abzug, personal observation), in this study we used the GDVII strain of TMEV to compare placental and fetal health of GDVIIinfected, clinically ill dams and GDVII-infected, well dams. 2000 Harcourt Publishers Ltd
Abzug and Tyson: Maternal Illness in Gestational Picornavirus Infection
MATERIALS AND METHODS Virus preparation Tissue culture-adapted GDVII virus was grown on baby hamster kidney (BHK) 21 cells and titered by TCID50 assay (Abzug, Rotbart and Levin, 1989).
841
of homogenates, approximately 150 l of each supernatant was cultured on BHK 21 cells for 14 days. Positive cultures were identified by detection of typical cytopathic effect. Maternal sera obtained at sacrifice were thawed and approximately 50 l of each was cultured to detect viraemia at the time of tissue harvesting (Abzug, Rotbart and Levin, 1989).
Animal experiments The investigations using pregnant mice were performed according to the US Department of Health and Human Services and the University of Colorado School of Medicine guidelines. Outbred ICR mice, seronegative for TMEV (Sasco Laboratories, Omaha, NE, USA), were bred. The first day of gestation was counted as the day following visualization of a vaginal plug. Implantation occurs on days 4–5 of pregnancy; the normal length of gestation is approx 19 days. After mating, females in which vaginal plugs were visualized were splenectomized [to increase the incidence of placental infection (Abzug, Rotbart and Levin, 1989)] and inoculated intravenously with either 2.5107 TCID50 or 2.5106 TCID50 GDVII, during days 4–7 of gestation. Uninfected control animals were inoculated on days 4–7 with BHK 21 cell lysate (Abzug, Rotbart and Levin, 1989). Mice were phlebotomized by cardiac puncture and sacrificed at varying times following viral inoculation. Dams that developed signs of TMEVassociated encephalomyelitis, i.e. hunched appearance, ruffled fur, tachypnea and limb paralysis, were sacrificed on the first day they exhibited clinical signs. Control dams and GDVIIinfected dams that remained clinically healthy were sacrificed between days 12 and 18 of gestation. Placentae and whole fetuses were harvested, washed in phosphate buffered saline (PBS), and either frozen in PBS at 70C for viral culture or fixed sequentially in 4 per cent paraformaldehyde followed by 70 per cent ethanol prior to paraffin embedding for in situ hybridization (ISH) and histopathological examination. Maternal sera were frozen at 70C. At sacrifice, the number of placental–fetal pairs in each dam and their gross morphology were recorded. Placental–fetal pairs were defined as abnormal if placentae had an unusually small surface area or if they were unusually pale; if fetuses were incompletely formed, markedly stunted, autolysed, or absent despite the presence of a placenta; or if tissue present was grossly unidentifiable as placental or fetal. Placental–fetal pairs from one dam inoculated with 2.5107 TCID50 GDVII that became ill were excluded from this analysis because a short incubation period of 3 days required sacrifice at a gestational age of 7 days, when an accurate placental–fetal count and assessment of morphology were not possible. Placenta–fetal pairs from one dam in the BHK Lysate cohort were similarly excluded because delivery occurred prior to sacrifice. Virus cultures Frozen placentae and whole fetuses were thawed and homogenized in dounce tissue homogenizers. After centrifugation
In situ hybridization (ISH) A minimum of three placental–fetal pairs per animal were studied by ISH as previously described (Abzug, 1997). Two slides, each containing at least two 5–7 m thick serial sections of placentae and whole fetuses, were evaluated per placental– fetal pair. Tissues were deparaffinized and pretreated in 0.2 N HCL and 5 g/ml Proteinase K (Boehringer Mannheim; Mannheim, Germany). Slides were then treated with a prehybridization solution containing 50 per cent de-ionized formamide and Denhardt’s solution, dextran sulphate, yeast t-RNA (Boehringer Mannheim; Mannheim, Germany) and standard saline citrate (SSC) for 1 h. Hybridization was performed overnight at 42C in the same solution containing approximately 0.05 g digoxigenin-labelled RNA probe/slide. Probe was made by linearizing a PGEM-2 plasmid vector (Promega; Madison, WI, USA) into which a cDNA fragment of the 3 end of TMEV strain BeAn 8386 described previously had been cloned (Rotbart, Abzug and Levin, 1988); transcribing in the presence of digoxigenin-labelled UTP (Boehringer Mannheim; Mannheim, Germany); and hydrolyzing the product in a carbonate buffer, pH 10.2, at 60C for 20 min. After hybridization, slides underwent serial washes in SSC, including one wash at 37C containing 20 g/ml RNase A (Boehringer Mannheim; Mannheim, Germany). The digoxigenin label on hybridized probe was detected by overnight incubation with a 1 : 3000 dilution of polyclonal sheep antidigoxigenin antibody conjugated to alkaline phosphatase, followed by colour development with nitroblue tetrazolium and X-phosphate (Boehringer Mannheim; Mannheim, Germany). Slides were counterstained with eosin. ISH of placentae was characterized by overall positivity and by histological identification of the tissues containing signal: decidua, spongiotrophoblast, or the labyrinth closest to the fetus. Fetuses were characterized according to the presence or absence of hybridization signal and the specific organs containing signal. At least one negative control of placenta and fetus from a BHK lysate-inoculated mouse and one positive control of placenta shown to be ISH-positive in prior experiments were included in every ISH run (Abzug, 1997; Abzug and Tyson, 1999). Hybridization slides were read blind, without prior knowledge of which experimental group dams had been assigned to. For both placentae and fetuses, ISH was considered positive only if hybridization signal was detected in the same location on more than one tissue section.
Placenta (2000), Vol. 21
842
Table 1. Pregnancy outcome in mice inoculated with GDVII virus or BHK lysate on days 4–7 of gestation GDVII Virus/Ill dams 2.510 TCID50 Mice with placental–fetal pairs at sacrifice/no. inoculated Mean no. of placental–fetal pairs/ pregnant dam (s.d.) Placental–fetal pairs with abnormal gross exam/no. harvested (per cent)
7
2.510 TCID50
6
GDVII/Well dams Total GDVII/ Ill dams
2.5107 TCID50
2.5106 TCID50
Total GDVII/ Well dams
BHK lysate
7/9a
7/11b
14/20
4/4c
7/10d
11/14
7/12e
9.3 (5.6)
11.8 (3.2)
10.7 (4.5)
12.5 (7.8)
13.8 (1.5)
13.4 (4.5)
13.7 (1.9)
47/56 (84)f
48/83 (58)
95/139 (68)f
27/50 (54)
48/97 (49)
75/147 (51)f
7/82 (9)f
a
Four received 2.5107 TCID50 GDVII intravenously on gestational day 4, 4 on day 5, 1 on day 7. Three received 2.5106 TCID50 GDVII intravenously on gestational day 4, 3 on day 5, 3 on day 6, 2 on day 7. c One received 2.5107 TCID50 GDVII intravenously on gestational day 5, 2 on day 6, 1 on day 7. d Four received 2.5106 TCID50 GDVII intravenously on gestational day 4, 2 on day 5, 3 on day 6, 1 on day 7. e Three received BHK lysate supernatant on day 4, 4 on day 5, 2 on day 6, 3 on day 7. f P<0.01, abnormal gross placenta–fetus morphology in total GDVII/Ill dam group versus total GDVII/Well dam group and between 2.5107 TCID50 GDVII/Ill dam subgroup and 2.5106 TCID50 GDVII/Ill dam, 2.5107 TCID50 GDVII/Well dam and 2.5106 TCID50 GDVII/Well dam subgroups. P<0.001, abnormal gross placenta–fetus morphology in total GDVII/Ill dam or total GDVII/Well dam groups (and in each dose subgroup) versus BHK lysate group. b
Histopathology
RESULTS
Consecutive sections of tissues studied by ISH were stained with haematoxylin and eosin. Placentae were classified as normal; containing increased hyaline tissue among blood vessels in the vascular labyrinth layer without cellular degeneration; or having evidence of cellular degeneration with loss of staining and/or cell dropout and loss of normal architecture within the vascular labyrinth, regardless of presence or absence of increased hyaline tissue (Palmer et al., 1997). Whole mount sections of fetuses were examined for lesions such as necrosis and inflammation within individual organs. However, within grossly abnormal fetuses that consisted of autolysed tissue, individual organs could not be recognized. Placentae and fetuses were evaluated for abnormalities blind to experimental group assignment.
Nine of 13 (69 per cent) mice initially inoculated with 2.5107 TCID50 GDVII virus became ill. In order to provide a sufficient number of clinically well GDVII-infected mice to compare with ill GDVII-infected mice, a 10-fold lower inoculum was subsequently administered to 21 mice, of which 11 (52 per cent) became ill. The incubation period of GDVIIassociated encephalomyelitis in the 20 ill mice ranged from 3–11 days; these mice were sacrificed between days 7 and 18 of gestation, with a mean interval from inoculation to sacrifice of 6 days in the 2.5107 TCID50 GDVII/Ill Dam subgroup and 8 days in the 2.5106 TCID50 GDVII/Ill Dam subgroup. The 14 clinically healthy GDVII-inoculated mice were sacrificed between days 12 and 18 of gestation, with a mean interval from inoculation to sacrifice of 6 days in the 2.5107 TCID50 GDVII/Well Dam subgroup and 12 days in the 2.5106 TCID50 GDVII/Well Dam subgroup. Twelve control mice inoculated with BHK lysate remained clinically well and were sacrificed between days 15 and 18 of gestation. Seventy per cent of ill GDVII-inoculated mice, 78 per cent of healthy GDVII-inoculated mice, and 58 per cent of control BHK lysate-inoculated mice had placentae and fetuses at sacrifice (Table 1). Of the dams containing placentae and fetuses at sacrifice, there was a trend for mean litter size to be lower among ill GDVII-infected dams when compared to well GDVII-infected and control mice (10.7 versus 13.4 and 13.7 placenta–fetus pairs per dam, respectively; P=0.2). Of the 14 ill GDVII-inoculated dams that were pregnant at sacrifice, five (36 per cent) had <10 placenta–fetus pairs, compared with one of the 11 (9 per cent) pregnant, well GDVII-inoculated dams (P=0.2) and none of the seven pregnant control dams (P=0.1). Ninety-five of 139 (68 per cent) placenta–fetus pairs from GDVII-infected ill dams were grossly abnormal compared with 75/147 (51 per cent) recovered from infected well dams
Statistical analysis Dams that developed clinical encephalomyelitis following GDVII virus inoculation (higher or lower dose) were compared with virus-inoculated mice that remained healthy and with dams that received BHK lysate. Comparisons among the higher dose GDVII/ill dam, lower dose GDVII/ill dam, higher dose GDVII/well dam, and lower dose GDVII/well dam subgroups were also performed. Statistical analysis was performed with Epi Info version 5.0 (Centers for Disease Control Epidemiology Program Office; Atlanta, GA, USA). Tests used for comparisons of proportions were chi-squared with Yates correction and, where indicated, Fisher exact test (two-tailed). Comparisons of means was done with the Kruskal–Wallis and Wilcoxon tests for non-parametric data. Significance was defined as P<0.05.
Abzug and Tyson: Maternal Illness in Gestational Picornavirus Infection
843
Table 2. Virological and histological evaluation of placentae from mice inoculated with GDVII virus or BHK lysate on days 4–7 of gestation GDVII Virus/Ill dams 2.510 TCID50 Culture+placentae/no. cultured (per cent) In situ hybridization+placentae/ no. examined (per cent) Placentae with cellular degeneration/ no. examined (per cent) Placentae with increased hyaline in labyrinth/no. examined (per cent)
7
2.510 TCID50
6
GDVII/Well dams Total GDVII/ Ill dams
2.5107 TCID50
2.5106 TCID50 4/28 (14)a
Total GDVII/ Well dams
BHK lysate
20/44 (45)a
1/22 (8)
7/12 (58)a
16/19 (84)
23/31 (74)a
16/16 (100)a
5/13 (38)b
25/28 (89)
30/41 (73)
15/15 (100)
26/33 (79)
41/48 (85)
0/24 (0)
7/17 (41)
18/31 (58)
25/48 (52)c
4/21 (19)c
18/34 (53)
22/55 (40)c
4/24 (17)c
2/17 (12)
1/31 (3)
3/48 (6)d
11/21 (52)d
4/34 (12)
15/55 (27)d
4/24 (17)
a
P=0.02, placenta culture positivity in total GDVII/Ill dam group versus total GDVII/Well dam group. Pc0.01, placenta culture positivity in 2.5107 TCID50 GDVII/Ill dam subgroup versus 2.5107 TCID50 GDVII/Well dam subgroup, and placenta culture positivity in 2.5106 TCID50 GDVII/Well dam subgroup versus 2.5107 TCID50 GDVII/Ill dam, 2.5106 TCID50 GDVII/Ill dam, and 2.5107 TCID50 GDVII/Well dam subgroups. b Pc0.01, placenta in situ hybridization positivity in 2.5107 TCID50 GDVII/Ill dam subgroup versus 2.5106 TCID50 GDVII/Ill dam, 2.5107 TCID50 GDVII/Well dam, and 2.5106 TCID50 GDVII/Well dam subgroups. c P<0.05, placentae with cellular degeneration in total GDVII/Ill dam or total GDVII/Well dam groups versus BHK lysate group, and placentae with cellular degeneration in 2.5107 TCID50 GDVII/Well dam subgroup versus 2.5106 TCID50 GDVII/Ill dam and 2.5106 TCID50 GDVII/Well dam subgroups. d P<0.05, placentae with increased hyaline in labyrinth in total GDVII/Well dam group versus total GDVII/Ill dam group, and placentae with increased hyaline in labyrinth in 2.5107 TCID50 GDVII/Well dam subgroup versus 2.5107 TCID50 GDVII/Ill dam, 2.5106 TCID50 GDVII/Ill dam, and 2.5106 TCID50 GDVII/Well dam subgroups.
(P<0.01) (Table 1). In particular, the incidence of abnormal placenta-fetus pairs in ill dams inoculated with 2.5107 TCID50 GDVII significantly exceeded the incidence of abnormal placenta-fetus pairs in dams inoculated with 2.5107 TCID50 GDVII that remained well (47/56, 84 per cent, versus 27/50, 54 per cent, P<0.01), whereas the difference between ill and well mice inoculated with 2.5106 TCID50 GDVII was not statistically significant (Table 1). Overall, the percentage of placenta–fetus pairs from dams inoculated with 2.5107 TCID50 GDVII virus that were abnormal was higher than from dams inoculated with 2.5106 TCID50 GDVII (74/106, 70 per cent, versus 96/180, 53 per cent, P<0.01) and, among ill dams, abnormal tissues were more frequent in those that received the higher dose of GDVII. The incidence of grossly abnormal placenta–fetus pairs from both ill and well GDVIIinoculated mice was markedly greater than that from BHK control-inoculated mice (Table 1). Twenty-three of 31 (74 per cent) placentae from GDVIIinoculated mice that became ill were culture-positive compared with 20/44 (45 per cent) of those from virus-inoculated dams that remained well (P=0.02) (Table 2). This difference was due to a very low culture positivity rate (14 per cent) among tissues from healthy dams that received the lower inoculum of GDVII. All maternal sera obtained at sacrifice were culturenegative. One placenta from a BHK-inoculated dam was falsely culture-positive. In contrast to culture, the overall rates of ISH positivity of placentae from ill and well GDVII-infected dams (and of placentae from the subgroup of healthy dams that received the lower inoculum of GDVII virus in particular) were similarly high (73–85 per cent) (Table 2). Although the 38 per cent rate of ISH positivity in the 2.5107 TCID50 GDVII/Ill dam
subgroup was lower than in the other subgroups, there was no overall difference between tissues from mice that received the higher inoculum of GDVII and those from dams that received the lower inoculum (20/28, 71 per cent, versus 51/61, 84 per cent, P=0.3). Most ISH-positive placentae contained hybridization signal in the decidua and spongiotrophoblast layers, and approximately half were also positive in the labyrinth layer. No placentae from BHK control dams were ISH-positive. Histopathological evidence of degenerative changes was found in 47/103 (46 per cent) placentae from GDVIIinoculated dams compared to 4/24 (17 per cent) placentae from BHK control dams (P<0.05), but there was no difference in the number of placentae with cellular degeneration between sick and healthy GDVII-infected mice (Table 2). Increased hyaline tissue in the vascular labyrinth in the absence of cellular degeneration was evident in more placentae from well GDVII-inoculated dams than in placentae from ill dams, but there were no differences in comparison to BHK control dams. Overall, placentae from GDVII-infected dams had either of these two histological abnormalities (degeneration or increased hyaline tissue) more often than did placentae from BHKinoculated dams (65/103, 63 per cent, versus 8/24, 33 per cent, P=0.02), but there was no difference between placentae from sick GDVII-infected mice and those from healthy GDVII-infected mice (28/48, 58 per cent, versus 37/55, 67 per cent, P=0.5), between placentae from mice that received the higher GDVII inoculum and those that received the lower inoculum (24/38, 63 per cent, versus 41/65, 63 per cent, P=0.8), or among subgroups (Pd0.4). Approximately one quarter of fetuses examined from GDVII-inoculated mice were culture-positive, with no overall difference between ill and well mothers (Table 3). No fetuses
Placenta (2000), Vol. 21
844
Table 3. Virological evaluation of fetuses from mice inoculated with GDVII virus or BHK lysate on days 4–7 of gestation GDVII Virus/Ill dams 7
Culture+fetuses/no. cultured (per cent) In situ hybridization+fetuses/ no. examined (per cent)
GDVII/Well dams Total GDVII/ Ill dams
2.5107 TCID50
2.5106 TCID50
Total GDVII/ Well dams
BHK lysate
4/17 (24)a
6/25 (24)
11/16 (69)a
0/25 (0)a
11/41 (27)
0/22 (0)
7/19 (37)
7/24 (29)
1/14 (7)
6/23 (26)
7/37 (19)
0/24 (0)
2.510 TCID50
2.510 TCID50
2/8 (25) 0/5 (0)
6
a
P=0.02, fetus culture positivity in 2.5106 TCID50 GDVII/Ill dam subgroup versus 2.5107 TCID50 GDVII/Well dam and 2.5106 TCID50 GDVII/Well dam subgroups. P<0.001, fetus culture positivity in 2.5107 TCID50 GDVII/Well dam subgroup versus 2.5106 TCID50 GDVII/Well dam subgroup.
Figure 1. Fetal heart from a dam inoculated with GDVII virus (2.5106 TCID50) on day 7 of gestation. In situ hybridization (a, 100; bar=45 m) showed viral RNA (dark stain; arrows) in myocardium (M). Haematoxylin and eosin stain of the myocardium (b, 500; bar=9 m) demonstrated focal myocarditis, with cellular necrosis and debris (large arrows) and scattered lymphocytes (small arrows).
from BHK-inoculated mice were culture-positive. Rates of fetal ISH positivity were similar for the sick and healthy GDVII-infected cohorts, and no fetuses from BHK-inoculated controls were ISH-positive (Table 3). The subgroup of healthy dams that received the lower inoculum of GDVII had no culture-positive fetuses but 6/23 (26 per cent) fetuses that were ISH-positive. Hybridization signal in GDVII-infected fetuses was located in heart (n=4) and great vessels (n=1), brain (n=2) and spinal cord (n=2), intestine (n=3), connective tissue (n=3), and unidentifiable degenerative tissue (n=8). One ISH-positive heart presented a lymphocytic myocarditis and another focal necrosis (Figure 1). Focal necrosis was observed in the brain tissue of three fetuses from GDVII-infected mice and in spinal cord from one fetus (Figure 2). Several fetuses from GDVIIinoculated dams had an increased number of apparent lymphocytes in brain tissue detected by routine light microscopy, as compared to fetuses from BHK-inoculated dams. Other, grossly abnormal, fetuses from GDVII-inoculated dams had generalized tissue autolysis histologically, but no specific organ anomalies were noted. In additional cases, no fetal tissue could be identified histologically in association with placental tissue.
DISCUSSION Infection by the GDVII strain of TMEV during early pregnancy caused high rates of maternal illness, which we have not observed with the DA strain (Abzug, Rotbart and Levin, 1989; Abzug et al., 1991; Palmer et al., 1997; Abzug and Tyson, 1999), accompanied by very high rates of gross placental and fetal abnormalities. Decreasing the inoculum of GDVII 10-fold only modestly reduced the incidence of maternal clinical illness and the rate of placental and fetal abnormalities. Comparison of placentae and fetuses from infected dams that became ill and those that remained asymptomatic revealed an increased incidence of gross placental and fetal abnormalities in sick dams, particularly among mice that received the higher viral inoculum. Rates of gross fetal abnormalities in both ill and well mothers exceeded the incidence of documented fetal infection but paralleled the rates of infection and histological abnormalities of placentae, suggesting that much of the fetal degeneration we observed was secondary to damaging effects of virus on the placenta. However, the incremental increase in gross abnormalities of placental–fetal pairs from infected ill mothers, as compared to infected healthy dams, was not explained by consistently higher rates of placental or fetal
Abzug and Tyson: Maternal Illness in Gestational Picornavirus Infection
Figure 2. Central nervous system tissue from fetuses harvested from dams inoculated with GDVII virus (2.5106 TCID50) on day 7 of gestation. Haematoxylin and eosin stain in the region of the brainstem (a, 125; bar=36 m; b, 250; bar=18 m) showed areas of necrosis (large arrows) and lymphocytic inflammation (small arrrows). Haematoxylin and eosin stain of the spinal cord of another fetus from the same dam (c, 500; bar=9 m) demonstrated a region of cellular necrosis (large arrow) and lymphocytes (small arrows).
infection or of placental histological abnormalities. This suggests that adverse effects of virus on maternal health may confer an additive risk to placental–fetal outcome, above and beyond direct viral infection of the placenta or fetus. Alternatively, it is possible that disease in mothers is a marker for
845
increased viral replication and greater quantity of virus in infected gestational tissues. It is unlikely that the more favourable pregnancy outcome in infected dams that remained well (as compared to virus-inoculated dams that became ill) was due to pre-existing immunity of these animals to GDVII, as the mice used in these experiments were from a known TMEVseronegative colony and no consistent differences in placental and fetal infection rates were evident between healthy and sick dams. However, it is possible that differences in the immune response among these outbred mice to viral challenge had an impact on the development of both maternal illness and adverse pregnancy outcome. Two other findings deserve comment. Firstly, placentae from healthy GDVII-inoculated mice were less frequently culture-positive than were the same tissues from sick GDVIIinfected dams, due to the very low rate of culture positivity in the subgroup of healthy mice inoculated with the lower dose of GDVII. In contrast, ISH positivity of placentae from the entire healthy GDVII group, and the healthy, lower dose GDVII subgroup in particular, exceeded culture positivity and was similar to the rate of ISH positivity among placentae from the sick GDVII cohort. A similar discrepancy between culture and ISH was observed with respect to fetal tissues in the healthy, lower dose GDVII subgroup. These discrepancies are most likely explained by the longer mean interval from inoculation to sacrifice for the lower dose subgroup of well GDVII-infected dams than for the ill GDVII-infected group. We have previously observed that with longer intervals from inoculation to tissue harvesting, viral culture positivity decreases but ISH yields persist and are a more accurate indicator of placental and fetal infection rates (Palmer et al., 1997). The most likely explanation for this finding is that as time increases from inoculation, host immune responses limit viral replication, leaving only non-cultivable virus in tissues, detectable by ISH but not tissue culture. Secondly, GDVII was identified primarily in fetal heart and great vessels and in brain and spinal cord, with inflammation and necrosis present in both cardiac and central nervous system tissues. This is in contrast to DA-infected fetuses, in which infection and histological changes were more often limited to the heart and great vessels (Abzug et al., 1991; Abzug and Tyson, 1999). This latter difference is consistent with the enhanced neurovirulence of the GDVII group of TMEV strains. The results presented parallel epidemiological observations of human enterovirus infections (Abzug, 1995). Maternal poliomyelitis is associated with fetal loss, stillbirth, intrauterine growth retardation and an increased incidence of premature delivery. These sequelae are more often due to effects on maternal health than due to infection of the fetus, and fetal loss is more frequent when maternal illness is severe (Abzug, 1995). Neonatal poliomyelitis, in some cases reflecting in utero infection, may follow maternal poliovirus infection in the peripartum period, but is uncommon in the absence of clinical disease in the mother. Maternal infections during pregnancy with non-poliovirus enteroviruses frequently occur, with many of the infections being asymptomatic. However, among
846
neonates with enterovirus disease, high rates of viral illness are reported among their mothers, and, in some series, maternal illness has been shown to be a risk factor for severe neonatal disease (Abzug, 1995). Thus, for humans infected with entero-
Placenta (2000), Vol. 21
viruses, as for mice infected with GDVII virus, symptomatic viral infection in the mother appears to be a marker for increased risk to the fetus or newborn.
ACKNOWLEDGEMENTS The authors would like to thank Linda Sanders for her technical expertise; Pat Lenhart and Kate Fasth for their assistance; and Harley Rotbart, MD, for his support, valuable suggestions, and generosity of time. Tissue culture-adapted strains of GDVII virus were provided by R. Murray, Swedish Medical Center, Englewood, CO, USA. This study was supported by a grant from the National Institutes of Health (5 R29 HD 27692).
REFERENCES Abzug MJ (1993) Identification of trophoblastic giant cells as the initial principal target of early gestational murine enterovirus infection. Placenta, 14, 137–148. Abzug MJ (1995) Perinatal enterovirus infections. In Human Enterovirus Infections (Ed.) Rotbart HA, pp. 221–231. Washington, DC: American Society for Microbiology. Abzug MJ (1997) Maternal factors affecting the integrity of the late gestation placental barrier to murine enterovirus infection. J Infect Dis, 176, 41–49. Abzug MJ, Tyson RW (2000) Protection of murine gestational tissues from picornavirus infection in the preimplantation period. Placenta, 21, 422–426. Abzug MJ, Rotbart HA & Levin MJ (1989) Demonstration of a barrier to transplacental passage of murine enteroviruses in late gestation. J Infect Dis, 159, 761–765. Abzug MJ, Rotbart HA, Magliato SA & Levin MJ (1991) Evolution of the placental barrier to fetal infection by murine enteroviruses. J Infect Dis, 163, 1336–1341.
Axelsson C, Bondesatam K, Frisk G, Bergstrom S & Diderholm H (1993) Coxsackie B virus infections in women with miscarriage. J Med Virol, 39, 282–285. Brown GC & Karunas RS (1971) Relationship of congenital anomalies and maternal infection with selected enteroviruses. Am J Epidemiol, 95, 207–217. Cherry JD (1990) Enteroviruses. In Infectious Diseases of the Fetus and Newborn Infant. 3rd ed (Eds) Remington JS, Klein JO, pp. 325–366. Philadelphia: WB Saunders. Palmer AL, Rotbart HA, Tyson RW & Abzug MJ (1997) Adverse effects of maternal enterovirus infection on the fetus and placenta. J Infect Dis, 176, 1437–1444. Rotbart HA, Abzug MJ & Levin MJ (1988) Development and application of RNA probes for the study of picornaviruses. Mol Cell Probes, 2, 65–73.
Picornavirus Infection in Early Murine Gestation: Significance of Maternal Illness M. J. Abzuga and R. W. Tyson Departments of Pediatrics (Infectious Diseases) and Pathology, University of Colorado School of Medicine and The Children’s Hospital, Denver, Colorado, USA Paper accepted 1 June 2000
To evaluate whether maternal illness following picornavirus infection during pregnancy adversely affects placental and fetal health, mice were inoculated with the GDVII strain of Theiler’s murine encephalomyelitis virus or control cell lysate during days 4–7 of gestation. Gross appearance, histopathology and viral culture, and in situ hybridization positivity of placentae and fetuses from ill GDVII-infected, healthy GDVII-infected and control mice were compared. Twenty of 34 (59 per cent) GDVII-infected dams became clinically ill. More placenta–fetus pairs from ill mice were grossly abnormal (68 per cent) than from well GDVII-infected (51 per cent; P<0.01) or control mice (9 per cent; P<0.001). Virus was detected by in situ hybridization in 73 per cent of placentae and 29 per cent of fetuses from sick GDVII-infected dams, and in 85 per cent of placentae and 19 per cent of fetuses from healthy GDVII-infected mice (differences not significant). Histological abnormalities consisting of necrosis or an increase in hyaline tissue in the vascular labyrinth layer were similarly frequent in placentae from ill and well GDVII-infected mice (58 per cent versus 67 per cent, P=0.5). Viral RNA, inflammation and necrosis were evident in the heart, great vessels, brain and spinal cord of GDVII-infected fetuses. Infection with GDVII in early pregnancy produces a high rate of gross placental and fetal abnormalities. The rate of gross abnormalities exceeds the incidence of fetal infection and more closely parallels the rates of infection and histopathology in the placenta, suggesting that much of the damage to placenta–fetus pairs is a consequence of placental infection. In addition, the occurrence of viral-induced maternal illness is associated with additive risk to placental and fetal health not explained by an increased rate of placental or fetal infection. 2000 Harcourt Publishers Ltd Placenta (2000), 21, 840–846
INTRODUCTION Maternal infections during pregnancy with polioviruses and non-poliovirus enteroviruses are associated with fetal infection and adverse outcomes, such as spontaneous abortion, stillbirth, congenital defects and perinatal illness (Brown and Karunas, 1971; Cherry, 1990; Axelsson et al., 1993; Abzug, 1995). To investigate the pathogenesis of gestational enterovirus infections, we have developed an animal model using a murine picornavirus, Theiler’s murine encephalomyelitis virus (TMEV), in its natural host (Abzug, Rotbart and Levin, 1989; Abzug et al., 1991; Abzug, 1993; Abzug, 1997; Palmer et al., 1997). TMEV is a member of the cardiovirus genus in the picornavirus family, whose behaviour in mice is similar to that of the closely related enterovirus genus, also of the picornavirus family, in humans. We have shown in this model that infection with either the DA strain, a member of the less acutely neurovirulent group II strains of TMEV, or the more a To whom correspondence should be addressed at: The Children’s Hospital, Box B055, 1056 E. 19th Avenue, Denver, CO 80218, USA. Fax: +1 303-837-2707; E-mail: [email protected]
0143–4004/00/080840+07 $35.00/0
acutely neurovirulent group I strain, GDVII, in early gestation (days 4–7 of a 19 day gestation) is associated with high rates of placental and fetal infection, whereas infection in late pregnancy (days 12–13) is associated with a high rate of placental but not fetal infection (Abzug, Rotbart and Levin, 1989; Abzug et al., 1991; Abzug and Tyson, 1999). Maternal health factors, such as compromised uteroplacental blood flow and concomitant bacterial infection, may increase the risk of transplacental TMEV infection (Abzug, 1997). Infection with DA virus in early pregnancy can produce gross and histological abnormalities of placentae such as small placentae with fibrin deposition, dilated vascular lakes and infarction, and of fetuses, e.g. degenerated fetal tissue containing necrotic cellular debris; specific fetal organ defects are not apparent, and maternal health is not altered (Palmer et al., 1997). Because GDVII virus administered parenterally causes acute paralytic encephalomyelitis in some adult mice, whereas infection with DA virus does not cause short-term clinical changes (Abzug, personal observation), in this study we used the GDVII strain of TMEV to compare placental and fetal health of GDVIIinfected, clinically ill dams and GDVII-infected, well dams. 2000 Harcourt Publishers Ltd
Abzug and Tyson: Maternal Illness in Gestational Picornavirus Infection
MATERIALS AND METHODS Virus preparation Tissue culture-adapted GDVII virus was grown on baby hamster kidney (BHK) 21 cells and titered by TCID50 assay (Abzug, Rotbart and Levin, 1989).
841
of homogenates, approximately 150 l of each supernatant was cultured on BHK 21 cells for 14 days. Positive cultures were identified by detection of typical cytopathic effect. Maternal sera obtained at sacrifice were thawed and approximately 50 l of each was cultured to detect viraemia at the time of tissue harvesting (Abzug, Rotbart and Levin, 1989).
Animal experiments The investigations using pregnant mice were performed according to the US Department of Health and Human Services and the University of Colorado School of Medicine guidelines. Outbred ICR mice, seronegative for TMEV (Sasco Laboratories, Omaha, NE, USA), were bred. The first day of gestation was counted as the day following visualization of a vaginal plug. Implantation occurs on days 4–5 of pregnancy; the normal length of gestation is approx 19 days. After mating, females in which vaginal plugs were visualized were splenectomized [to increase the incidence of placental infection (Abzug, Rotbart and Levin, 1989)] and inoculated intravenously with either 2.5107 TCID50 or 2.5106 TCID50 GDVII, during days 4–7 of gestation. Uninfected control animals were inoculated on days 4–7 with BHK 21 cell lysate (Abzug, Rotbart and Levin, 1989). Mice were phlebotomized by cardiac puncture and sacrificed at varying times following viral inoculation. Dams that developed signs of TMEVassociated encephalomyelitis, i.e. hunched appearance, ruffled fur, tachypnea and limb paralysis, were sacrificed on the first day they exhibited clinical signs. Control dams and GDVIIinfected dams that remained clinically healthy were sacrificed between days 12 and 18 of gestation. Placentae and whole fetuses were harvested, washed in phosphate buffered saline (PBS), and either frozen in PBS at 70C for viral culture or fixed sequentially in 4 per cent paraformaldehyde followed by 70 per cent ethanol prior to paraffin embedding for in situ hybridization (ISH) and histopathological examination. Maternal sera were frozen at 70C. At sacrifice, the number of placental–fetal pairs in each dam and their gross morphology were recorded. Placental–fetal pairs were defined as abnormal if placentae had an unusually small surface area or if they were unusually pale; if fetuses were incompletely formed, markedly stunted, autolysed, or absent despite the presence of a placenta; or if tissue present was grossly unidentifiable as placental or fetal. Placental–fetal pairs from one dam inoculated with 2.5107 TCID50 GDVII that became ill were excluded from this analysis because a short incubation period of 3 days required sacrifice at a gestational age of 7 days, when an accurate placental–fetal count and assessment of morphology were not possible. Placenta–fetal pairs from one dam in the BHK Lysate cohort were similarly excluded because delivery occurred prior to sacrifice. Virus cultures Frozen placentae and whole fetuses were thawed and homogenized in dounce tissue homogenizers. After centrifugation
In situ hybridization (ISH) A minimum of three placental–fetal pairs per animal were studied by ISH as previously described (Abzug, 1997). Two slides, each containing at least two 5–7 m thick serial sections of placentae and whole fetuses, were evaluated per placental– fetal pair. Tissues were deparaffinized and pretreated in 0.2 N HCL and 5 g/ml Proteinase K (Boehringer Mannheim; Mannheim, Germany). Slides were then treated with a prehybridization solution containing 50 per cent de-ionized formamide and Denhardt’s solution, dextran sulphate, yeast t-RNA (Boehringer Mannheim; Mannheim, Germany) and standard saline citrate (SSC) for 1 h. Hybridization was performed overnight at 42C in the same solution containing approximately 0.05 g digoxigenin-labelled RNA probe/slide. Probe was made by linearizing a PGEM-2 plasmid vector (Promega; Madison, WI, USA) into which a cDNA fragment of the 3 end of TMEV strain BeAn 8386 described previously had been cloned (Rotbart, Abzug and Levin, 1988); transcribing in the presence of digoxigenin-labelled UTP (Boehringer Mannheim; Mannheim, Germany); and hydrolyzing the product in a carbonate buffer, pH 10.2, at 60C for 20 min. After hybridization, slides underwent serial washes in SSC, including one wash at 37C containing 20 g/ml RNase A (Boehringer Mannheim; Mannheim, Germany). The digoxigenin label on hybridized probe was detected by overnight incubation with a 1 : 3000 dilution of polyclonal sheep antidigoxigenin antibody conjugated to alkaline phosphatase, followed by colour development with nitroblue tetrazolium and X-phosphate (Boehringer Mannheim; Mannheim, Germany). Slides were counterstained with eosin. ISH of placentae was characterized by overall positivity and by histological identification of the tissues containing signal: decidua, spongiotrophoblast, or the labyrinth closest to the fetus. Fetuses were characterized according to the presence or absence of hybridization signal and the specific organs containing signal. At least one negative control of placenta and fetus from a BHK lysate-inoculated mouse and one positive control of placenta shown to be ISH-positive in prior experiments were included in every ISH run (Abzug, 1997; Abzug and Tyson, 1999). Hybridization slides were read blind, without prior knowledge of which experimental group dams had been assigned to. For both placentae and fetuses, ISH was considered positive only if hybridization signal was detected in the same location on more than one tissue section.
Placenta (2000), Vol. 21
842
Table 1. Pregnancy outcome in mice inoculated with GDVII virus or BHK lysate on days 4–7 of gestation GDVII Virus/Ill dams 2.510 TCID50 Mice with placental–fetal pairs at sacrifice/no. inoculated Mean no. of placental–fetal pairs/ pregnant dam (s.d.) Placental–fetal pairs with abnormal gross exam/no. harvested (per cent)
7
2.510 TCID50
6
GDVII/Well dams Total GDVII/ Ill dams
2.5107 TCID50
2.5106 TCID50
Total GDVII/ Well dams
BHK lysate
7/9a
7/11b
14/20
4/4c
7/10d
11/14
7/12e
9.3 (5.6)
11.8 (3.2)
10.7 (4.5)
12.5 (7.8)
13.8 (1.5)
13.4 (4.5)
13.7 (1.9)
47/56 (84)f
48/83 (58)
95/139 (68)f
27/50 (54)
48/97 (49)
75/147 (51)f
7/82 (9)f
a
Four received 2.5107 TCID50 GDVII intravenously on gestational day 4, 4 on day 5, 1 on day 7. Three received 2.5106 TCID50 GDVII intravenously on gestational day 4, 3 on day 5, 3 on day 6, 2 on day 7. c One received 2.5107 TCID50 GDVII intravenously on gestational day 5, 2 on day 6, 1 on day 7. d Four received 2.5106 TCID50 GDVII intravenously on gestational day 4, 2 on day 5, 3 on day 6, 1 on day 7. e Three received BHK lysate supernatant on day 4, 4 on day 5, 2 on day 6, 3 on day 7. f P<0.01, abnormal gross placenta–fetus morphology in total GDVII/Ill dam group versus total GDVII/Well dam group and between 2.5107 TCID50 GDVII/Ill dam subgroup and 2.5106 TCID50 GDVII/Ill dam, 2.5107 TCID50 GDVII/Well dam and 2.5106 TCID50 GDVII/Well dam subgroups. P<0.001, abnormal gross placenta–fetus morphology in total GDVII/Ill dam or total GDVII/Well dam groups (and in each dose subgroup) versus BHK lysate group. b
Histopathology
RESULTS
Consecutive sections of tissues studied by ISH were stained with haematoxylin and eosin. Placentae were classified as normal; containing increased hyaline tissue among blood vessels in the vascular labyrinth layer without cellular degeneration; or having evidence of cellular degeneration with loss of staining and/or cell dropout and loss of normal architecture within the vascular labyrinth, regardless of presence or absence of increased hyaline tissue (Palmer et al., 1997). Whole mount sections of fetuses were examined for lesions such as necrosis and inflammation within individual organs. However, within grossly abnormal fetuses that consisted of autolysed tissue, individual organs could not be recognized. Placentae and fetuses were evaluated for abnormalities blind to experimental group assignment.
Nine of 13 (69 per cent) mice initially inoculated with 2.5107 TCID50 GDVII virus became ill. In order to provide a sufficient number of clinically well GDVII-infected mice to compare with ill GDVII-infected mice, a 10-fold lower inoculum was subsequently administered to 21 mice, of which 11 (52 per cent) became ill. The incubation period of GDVIIassociated encephalomyelitis in the 20 ill mice ranged from 3–11 days; these mice were sacrificed between days 7 and 18 of gestation, with a mean interval from inoculation to sacrifice of 6 days in the 2.5107 TCID50 GDVII/Ill Dam subgroup and 8 days in the 2.5106 TCID50 GDVII/Ill Dam subgroup. The 14 clinically healthy GDVII-inoculated mice were sacrificed between days 12 and 18 of gestation, with a mean interval from inoculation to sacrifice of 6 days in the 2.5107 TCID50 GDVII/Well Dam subgroup and 12 days in the 2.5106 TCID50 GDVII/Well Dam subgroup. Twelve control mice inoculated with BHK lysate remained clinically well and were sacrificed between days 15 and 18 of gestation. Seventy per cent of ill GDVII-inoculated mice, 78 per cent of healthy GDVII-inoculated mice, and 58 per cent of control BHK lysate-inoculated mice had placentae and fetuses at sacrifice (Table 1). Of the dams containing placentae and fetuses at sacrifice, there was a trend for mean litter size to be lower among ill GDVII-infected dams when compared to well GDVII-infected and control mice (10.7 versus 13.4 and 13.7 placenta–fetus pairs per dam, respectively; P=0.2). Of the 14 ill GDVII-inoculated dams that were pregnant at sacrifice, five (36 per cent) had <10 placenta–fetus pairs, compared with one of the 11 (9 per cent) pregnant, well GDVII-inoculated dams (P=0.2) and none of the seven pregnant control dams (P=0.1). Ninety-five of 139 (68 per cent) placenta–fetus pairs from GDVII-infected ill dams were grossly abnormal compared with 75/147 (51 per cent) recovered from infected well dams
Statistical analysis Dams that developed clinical encephalomyelitis following GDVII virus inoculation (higher or lower dose) were compared with virus-inoculated mice that remained healthy and with dams that received BHK lysate. Comparisons among the higher dose GDVII/ill dam, lower dose GDVII/ill dam, higher dose GDVII/well dam, and lower dose GDVII/well dam subgroups were also performed. Statistical analysis was performed with Epi Info version 5.0 (Centers for Disease Control Epidemiology Program Office; Atlanta, GA, USA). Tests used for comparisons of proportions were chi-squared with Yates correction and, where indicated, Fisher exact test (two-tailed). Comparisons of means was done with the Kruskal–Wallis and Wilcoxon tests for non-parametric data. Significance was defined as P<0.05.
Abzug and Tyson: Maternal Illness in Gestational Picornavirus Infection
843
Table 2. Virological and histological evaluation of placentae from mice inoculated with GDVII virus or BHK lysate on days 4–7 of gestation GDVII Virus/Ill dams 2.510 TCID50 Culture+placentae/no. cultured (per cent) In situ hybridization+placentae/ no. examined (per cent) Placentae with cellular degeneration/ no. examined (per cent) Placentae with increased hyaline in labyrinth/no. examined (per cent)
7
2.510 TCID50
6
GDVII/Well dams Total GDVII/ Ill dams
2.5107 TCID50
2.5106 TCID50 4/28 (14)a
Total GDVII/ Well dams
BHK lysate
20/44 (45)a
1/22 (8)
7/12 (58)a
16/19 (84)
23/31 (74)a
16/16 (100)a
5/13 (38)b
25/28 (89)
30/41 (73)
15/15 (100)
26/33 (79)
41/48 (85)
0/24 (0)
7/17 (41)
18/31 (58)
25/48 (52)c
4/21 (19)c
18/34 (53)
22/55 (40)c
4/24 (17)c
2/17 (12)
1/31 (3)
3/48 (6)d
11/21 (52)d
4/34 (12)
15/55 (27)d
4/24 (17)
a
P=0.02, placenta culture positivity in total GDVII/Ill dam group versus total GDVII/Well dam group. Pc0.01, placenta culture positivity in 2.5107 TCID50 GDVII/Ill dam subgroup versus 2.5107 TCID50 GDVII/Well dam subgroup, and placenta culture positivity in 2.5106 TCID50 GDVII/Well dam subgroup versus 2.5107 TCID50 GDVII/Ill dam, 2.5106 TCID50 GDVII/Ill dam, and 2.5107 TCID50 GDVII/Well dam subgroups. b Pc0.01, placenta in situ hybridization positivity in 2.5107 TCID50 GDVII/Ill dam subgroup versus 2.5106 TCID50 GDVII/Ill dam, 2.5107 TCID50 GDVII/Well dam, and 2.5106 TCID50 GDVII/Well dam subgroups. c P<0.05, placentae with cellular degeneration in total GDVII/Ill dam or total GDVII/Well dam groups versus BHK lysate group, and placentae with cellular degeneration in 2.5107 TCID50 GDVII/Well dam subgroup versus 2.5106 TCID50 GDVII/Ill dam and 2.5106 TCID50 GDVII/Well dam subgroups. d P<0.05, placentae with increased hyaline in labyrinth in total GDVII/Well dam group versus total GDVII/Ill dam group, and placentae with increased hyaline in labyrinth in 2.5107 TCID50 GDVII/Well dam subgroup versus 2.5107 TCID50 GDVII/Ill dam, 2.5106 TCID50 GDVII/Ill dam, and 2.5106 TCID50 GDVII/Well dam subgroups.
(P<0.01) (Table 1). In particular, the incidence of abnormal placenta-fetus pairs in ill dams inoculated with 2.5107 TCID50 GDVII significantly exceeded the incidence of abnormal placenta-fetus pairs in dams inoculated with 2.5107 TCID50 GDVII that remained well (47/56, 84 per cent, versus 27/50, 54 per cent, P<0.01), whereas the difference between ill and well mice inoculated with 2.5106 TCID50 GDVII was not statistically significant (Table 1). Overall, the percentage of placenta–fetus pairs from dams inoculated with 2.5107 TCID50 GDVII virus that were abnormal was higher than from dams inoculated with 2.5106 TCID50 GDVII (74/106, 70 per cent, versus 96/180, 53 per cent, P<0.01) and, among ill dams, abnormal tissues were more frequent in those that received the higher dose of GDVII. The incidence of grossly abnormal placenta–fetus pairs from both ill and well GDVIIinoculated mice was markedly greater than that from BHK control-inoculated mice (Table 1). Twenty-three of 31 (74 per cent) placentae from GDVIIinoculated mice that became ill were culture-positive compared with 20/44 (45 per cent) of those from virus-inoculated dams that remained well (P=0.02) (Table 2). This difference was due to a very low culture positivity rate (14 per cent) among tissues from healthy dams that received the lower inoculum of GDVII. All maternal sera obtained at sacrifice were culturenegative. One placenta from a BHK-inoculated dam was falsely culture-positive. In contrast to culture, the overall rates of ISH positivity of placentae from ill and well GDVII-infected dams (and of placentae from the subgroup of healthy dams that received the lower inoculum of GDVII virus in particular) were similarly high (73–85 per cent) (Table 2). Although the 38 per cent rate of ISH positivity in the 2.5107 TCID50 GDVII/Ill dam
subgroup was lower than in the other subgroups, there was no overall difference between tissues from mice that received the higher inoculum of GDVII and those from dams that received the lower inoculum (20/28, 71 per cent, versus 51/61, 84 per cent, P=0.3). Most ISH-positive placentae contained hybridization signal in the decidua and spongiotrophoblast layers, and approximately half were also positive in the labyrinth layer. No placentae from BHK control dams were ISH-positive. Histopathological evidence of degenerative changes was found in 47/103 (46 per cent) placentae from GDVIIinoculated dams compared to 4/24 (17 per cent) placentae from BHK control dams (P<0.05), but there was no difference in the number of placentae with cellular degeneration between sick and healthy GDVII-infected mice (Table 2). Increased hyaline tissue in the vascular labyrinth in the absence of cellular degeneration was evident in more placentae from well GDVII-inoculated dams than in placentae from ill dams, but there were no differences in comparison to BHK control dams. Overall, placentae from GDVII-infected dams had either of these two histological abnormalities (degeneration or increased hyaline tissue) more often than did placentae from BHKinoculated dams (65/103, 63 per cent, versus 8/24, 33 per cent, P=0.02), but there was no difference between placentae from sick GDVII-infected mice and those from healthy GDVII-infected mice (28/48, 58 per cent, versus 37/55, 67 per cent, P=0.5), between placentae from mice that received the higher GDVII inoculum and those that received the lower inoculum (24/38, 63 per cent, versus 41/65, 63 per cent, P=0.8), or among subgroups (Pd0.4). Approximately one quarter of fetuses examined from GDVII-inoculated mice were culture-positive, with no overall difference between ill and well mothers (Table 3). No fetuses
Placenta (2000), Vol. 21
844
Table 3. Virological evaluation of fetuses from mice inoculated with GDVII virus or BHK lysate on days 4–7 of gestation GDVII Virus/Ill dams 7
Culture+fetuses/no. cultured (per cent) In situ hybridization+fetuses/ no. examined (per cent)
GDVII/Well dams Total GDVII/ Ill dams
2.5107 TCID50
2.5106 TCID50
Total GDVII/ Well dams
BHK lysate
4/17 (24)a
6/25 (24)
11/16 (69)a
0/25 (0)a
11/41 (27)
0/22 (0)
7/19 (37)
7/24 (29)
1/14 (7)
6/23 (26)
7/37 (19)
0/24 (0)
2.510 TCID50
2.510 TCID50
2/8 (25) 0/5 (0)
6
a
P=0.02, fetus culture positivity in 2.5106 TCID50 GDVII/Ill dam subgroup versus 2.5107 TCID50 GDVII/Well dam and 2.5106 TCID50 GDVII/Well dam subgroups. P<0.001, fetus culture positivity in 2.5107 TCID50 GDVII/Well dam subgroup versus 2.5106 TCID50 GDVII/Well dam subgroup.
Figure 1. Fetal heart from a dam inoculated with GDVII virus (2.5106 TCID50) on day 7 of gestation. In situ hybridization (a, 100; bar=45 m) showed viral RNA (dark stain; arrows) in myocardium (M). Haematoxylin and eosin stain of the myocardium (b, 500; bar=9 m) demonstrated focal myocarditis, with cellular necrosis and debris (large arrows) and scattered lymphocytes (small arrows).
from BHK-inoculated mice were culture-positive. Rates of fetal ISH positivity were similar for the sick and healthy GDVII-infected cohorts, and no fetuses from BHK-inoculated controls were ISH-positive (Table 3). The subgroup of healthy dams that received the lower inoculum of GDVII had no culture-positive fetuses but 6/23 (26 per cent) fetuses that were ISH-positive. Hybridization signal in GDVII-infected fetuses was located in heart (n=4) and great vessels (n=1), brain (n=2) and spinal cord (n=2), intestine (n=3), connective tissue (n=3), and unidentifiable degenerative tissue (n=8). One ISH-positive heart presented a lymphocytic myocarditis and another focal necrosis (Figure 1). Focal necrosis was observed in the brain tissue of three fetuses from GDVII-infected mice and in spinal cord from one fetus (Figure 2). Several fetuses from GDVIIinoculated dams had an increased number of apparent lymphocytes in brain tissue detected by routine light microscopy, as compared to fetuses from BHK-inoculated dams. Other, grossly abnormal, fetuses from GDVII-inoculated dams had generalized tissue autolysis histologically, but no specific organ anomalies were noted. In additional cases, no fetal tissue could be identified histologically in association with placental tissue.
DISCUSSION Infection by the GDVII strain of TMEV during early pregnancy caused high rates of maternal illness, which we have not observed with the DA strain (Abzug, Rotbart and Levin, 1989; Abzug et al., 1991; Palmer et al., 1997; Abzug and Tyson, 1999), accompanied by very high rates of gross placental and fetal abnormalities. Decreasing the inoculum of GDVII 10-fold only modestly reduced the incidence of maternal clinical illness and the rate of placental and fetal abnormalities. Comparison of placentae and fetuses from infected dams that became ill and those that remained asymptomatic revealed an increased incidence of gross placental and fetal abnormalities in sick dams, particularly among mice that received the higher viral inoculum. Rates of gross fetal abnormalities in both ill and well mothers exceeded the incidence of documented fetal infection but paralleled the rates of infection and histological abnormalities of placentae, suggesting that much of the fetal degeneration we observed was secondary to damaging effects of virus on the placenta. However, the incremental increase in gross abnormalities of placental–fetal pairs from infected ill mothers, as compared to infected healthy dams, was not explained by consistently higher rates of placental or fetal
Abzug and Tyson: Maternal Illness in Gestational Picornavirus Infection
Figure 2. Central nervous system tissue from fetuses harvested from dams inoculated with GDVII virus (2.5106 TCID50) on day 7 of gestation. Haematoxylin and eosin stain in the region of the brainstem (a, 125; bar=36 m; b, 250; bar=18 m) showed areas of necrosis (large arrows) and lymphocytic inflammation (small arrrows). Haematoxylin and eosin stain of the spinal cord of another fetus from the same dam (c, 500; bar=9 m) demonstrated a region of cellular necrosis (large arrow) and lymphocytes (small arrows).
infection or of placental histological abnormalities. This suggests that adverse effects of virus on maternal health may confer an additive risk to placental–fetal outcome, above and beyond direct viral infection of the placenta or fetus. Alternatively, it is possible that disease in mothers is a marker for
845
increased viral replication and greater quantity of virus in infected gestational tissues. It is unlikely that the more favourable pregnancy outcome in infected dams that remained well (as compared to virus-inoculated dams that became ill) was due to pre-existing immunity of these animals to GDVII, as the mice used in these experiments were from a known TMEVseronegative colony and no consistent differences in placental and fetal infection rates were evident between healthy and sick dams. However, it is possible that differences in the immune response among these outbred mice to viral challenge had an impact on the development of both maternal illness and adverse pregnancy outcome. Two other findings deserve comment. Firstly, placentae from healthy GDVII-inoculated mice were less frequently culture-positive than were the same tissues from sick GDVIIinfected dams, due to the very low rate of culture positivity in the subgroup of healthy mice inoculated with the lower dose of GDVII. In contrast, ISH positivity of placentae from the entire healthy GDVII group, and the healthy, lower dose GDVII subgroup in particular, exceeded culture positivity and was similar to the rate of ISH positivity among placentae from the sick GDVII cohort. A similar discrepancy between culture and ISH was observed with respect to fetal tissues in the healthy, lower dose GDVII subgroup. These discrepancies are most likely explained by the longer mean interval from inoculation to sacrifice for the lower dose subgroup of well GDVII-infected dams than for the ill GDVII-infected group. We have previously observed that with longer intervals from inoculation to tissue harvesting, viral culture positivity decreases but ISH yields persist and are a more accurate indicator of placental and fetal infection rates (Palmer et al., 1997). The most likely explanation for this finding is that as time increases from inoculation, host immune responses limit viral replication, leaving only non-cultivable virus in tissues, detectable by ISH but not tissue culture. Secondly, GDVII was identified primarily in fetal heart and great vessels and in brain and spinal cord, with inflammation and necrosis present in both cardiac and central nervous system tissues. This is in contrast to DA-infected fetuses, in which infection and histological changes were more often limited to the heart and great vessels (Abzug et al., 1991; Abzug and Tyson, 1999). This latter difference is consistent with the enhanced neurovirulence of the GDVII group of TMEV strains. The results presented parallel epidemiological observations of human enterovirus infections (Abzug, 1995). Maternal poliomyelitis is associated with fetal loss, stillbirth, intrauterine growth retardation and an increased incidence of premature delivery. These sequelae are more often due to effects on maternal health than due to infection of the fetus, and fetal loss is more frequent when maternal illness is severe (Abzug, 1995). Neonatal poliomyelitis, in some cases reflecting in utero infection, may follow maternal poliovirus infection in the peripartum period, but is uncommon in the absence of clinical disease in the mother. Maternal infections during pregnancy with non-poliovirus enteroviruses frequently occur, with many of the infections being asymptomatic. However, among
846
neonates with enterovirus disease, high rates of viral illness are reported among their mothers, and, in some series, maternal illness has been shown to be a risk factor for severe neonatal disease (Abzug, 1995). Thus, for humans infected with entero-
Placenta (2000), Vol. 21
viruses, as for mice infected with GDVII virus, symptomatic viral infection in the mother appears to be a marker for increased risk to the fetus or newborn.
ACKNOWLEDGEMENTS The authors would like to thank Linda Sanders for her technical expertise; Pat Lenhart and Kate Fasth for their assistance; and Harley Rotbart, MD, for his support, valuable suggestions, and generosity of time. Tissue culture-adapted strains of GDVII virus were provided by R. Murray, Swedish Medical Center, Englewood, CO, USA. This study was supported by a grant from the National Institutes of Health (5 R29 HD 27692).
REFERENCES Abzug MJ (1993) Identification of trophoblastic giant cells as the initial principal target of early gestational murine enterovirus infection. Placenta, 14, 137–148. Abzug MJ (1995) Perinatal enterovirus infections. In Human Enterovirus Infections (Ed.) Rotbart HA, pp. 221–231. Washington, DC: American Society for Microbiology. Abzug MJ (1997) Maternal factors affecting the integrity of the late gestation placental barrier to murine enterovirus infection. J Infect Dis, 176, 41–49. Abzug MJ, Tyson RW (2000) Protection of murine gestational tissues from picornavirus infection in the preimplantation period. Placenta, 21, 422–426. Abzug MJ, Rotbart HA & Levin MJ (1989) Demonstration of a barrier to transplacental passage of murine enteroviruses in late gestation. J Infect Dis, 159, 761–765. Abzug MJ, Rotbart HA, Magliato SA & Levin MJ (1991) Evolution of the placental barrier to fetal infection by murine enteroviruses. J Infect Dis, 163, 1336–1341.
Axelsson C, Bondesatam K, Frisk G, Bergstrom S & Diderholm H (1993) Coxsackie B virus infections in women with miscarriage. J Med Virol, 39, 282–285. Brown GC & Karunas RS (1971) Relationship of congenital anomalies and maternal infection with selected enteroviruses. Am J Epidemiol, 95, 207–217. Cherry JD (1990) Enteroviruses. In Infectious Diseases of the Fetus and Newborn Infant. 3rd ed (Eds) Remington JS, Klein JO, pp. 325–366. Philadelphia: WB Saunders. Palmer AL, Rotbart HA, Tyson RW & Abzug MJ (1997) Adverse effects of maternal enterovirus infection on the fetus and placenta. J Infect Dis, 176, 1437–1444. Rotbart HA, Abzug MJ & Levin MJ (1988) Development and application of RNA probes for the study of picornaviruses. Mol Cell Probes, 2, 65–73.