Prenatal toxoplasmosis diagnosis from amniotic fluid by PCR Diagnóstico pré-natal da toxoplasmose no líquido amniótico através da técnica de PCR

Toxoplasmosis is one of the most common infections all over the world. Most cases are asymptomatic, except in immunosuppressed individuals and fetuses, which can be seriously damaged. Prenatal diagnosis should be made as soon as possible since treatment of the mother can minimize fetal sequelae. Our aim in this study was to test the polymerase chain reaction technique (PCR) in 86 samples of amniotic fluid from women who seroconverted during pregnancy. DNA was amplified using external primers and, in a second step, internal primers, in a nested PCR system. Samples were also inoculated into mice and the newborn were evaluated by T. gondii serology, skull x-ray, transfontanel ultrasound, fundoscopic examination, lumbar puncture and clinical examination. PCR was positive in seven cases and negative in 79. Among PCR-positive cases, two were negative by inoculation into mice and by clinical evaluation; among PCR-negative ones, three had clinical evidence of toxoplasmosis and one was positive after inoculation into mice. PCR showed values of sensitivity = 62.5% and specificity = 97.4%; the values of inoculation into mice where 42.9% and 100%, respectively. Although PCR should not be used alone for prenatal diagnosis of congenital toxoplasmosis, it is a promising method and deserves more studies to improve its efficacy. Key-words: Congenital toxoplasmosis. PCR. Amniotic fluid. Resumo A toxoplasmose é infecção freqüente em todo o mundo, mas na maioria dos casos não traz repercussões importantes para o paciente, exceto indivíduos imunodeprimidos e fetos, os quais podem apresentar graves seqüelas. O diagnóstico precoce durante a gravidez é altamente desejável, já que o tratamento da gestante reduz a freqüência e gravidade da infecção fetal. Neste estudo aplicou-se a técnica de PCR em 86 amostras de líquido amniótico de gestantes que apresentaram soroconversão durante a gravidez. O DNA foi amplificado usando-se iniciadores externos e internos, num sistema de nested PCR. As amostras foram também inoculadas em camundongos e os recém-nascidos acompanhados clinicamente através de sorologia, RX de crânio, ultrassom transfontanela, exame de fundo de olho e punção lombar. Pela PCR, sete casos foram positivos e 79 negativos. Entre os positivos, dois não se confirmaram pela inoculação em camundongo nem pela avaliação clínica da criança; dos negativos, três apresentaram clínica de toxoplasmose congênita, e em um deles o exame de inoculação em camundongo foi positivo. A sensibilidade e especificidade da PCR foram 62,5% e 97,4%, respectivamente; a inoculação em camundongos mostrou 42,9% de sensibilidade e 100% de especificidade. Embora a PCR não deva ser usada como único teste diagnóstico da toxoplasmose congênita, trata-se de método promissor e merece mais estudos para aumentar sua eficácia. Palavras-chaves: Toxoplasmose congênita. PCR. Líquido amniótico. 1. Departamento de Anatomia Patológica e Medicina Legal da Faculdade de Medicina da Universidade Federal de Minas Gerais (UFMG). 2. Departamento de Ginecologia e Obstetrícia da Faculdade de Medicina da UFMG. 3. Santa Casa de Misericórdia de Belo Horizonte. 4. Departamento de Parasitologia do Instituto de Ciências Biológicas da Universidade Federal de Minas Gerais, Belo Horizonte, MG. This study has been approved by Ethic Committee of Hospital das Clínicas – UFMG (process # 95/99). Address to: Prof. Geraldo Brasileiro Filho. Dept° de Anatomia Patológica e Medicina Legal/FM/UFMG. Av. Alfredo Balena 190, 30310-100 Belo Horizonte, MG, Brasil. Tel: 55 31 3248-9632; Fax: 55 31 3248-9664 e-mail: gbrasil@medicina.ufmg.br Recebido para publicação em 23/2/2001. Congenital toxoplasmosis is the result of transplacental transmission of Toxoplasma gondii from an acutely infected mother. The risk of fetal infection varies according to the gestational age at the time of maternal infection. When this occurs in the 11-14 weeks of gestation, the incidence of fetal infection is 7.2%, whereas in the 31-34 weeks period it is 67%; the overall rate of fetal infection is 7.4%. Congenitally infected infants can present, at birth, the classical signs and symptoms of congenital toxoplasmosis, such as hydrocephalus, chorioretinitis, cerebral calcifications and stroke, or they can be symptom-free and develop long-term sequelae during childhood or adult life. Toxoplasmosis can be diagnosed by: 1) serological profile, although this may be unreliable in those who

Vidigal PVT et al are immunodeficient, as in the fetus; 2) isolation of the parasite by animal inoculation or cell culture, which is time consuming; 3) demonstration of the parasite genome by polymerase chain reaction (PCR) technique, a promising method due to its reported good sensitivity and high specificity 12 .
Early anti-parasitic treatment of infected mothers has been shown to both decrease the transmission rate to the fetus and to reduce sequelae when intrauterine infection has already occurred 17 .Since the choice of anti-parasitic treatment depends on whether or not the fetus is infected, early detection of the parasite is very important 10 .
The aim of this study was to apply the PCR technique to amniotic fluid for the prenatal diagnosis of congenital toxoplasmosis and to compare these results to the postnatal clinical diagnosis.

PATIENTS AND METHODS
We studied 86 amniotic fluid samples from women in whom acquired T. gondii infection during pregnancy has occurred.The infection was detected during a routine serologic screening program performed in the Prenatal Unit of Hospital das Clínicas da Universidade Federal de Minas Gerais (HC-UFMG) and Santa Casa de Misericórdia de Belo Horizonte, Minas Gerais.Initially serology (ELISA, ELIFA, immunofluorescence) was performed at each gestation trimester.In the case of a positive result, it was then performed montly.For diagnosis of fetal infection, part of the amniotic fluid was submitted to PCR and the other part was inoculated into mice.After birth, the infants were followed periodically by the infectologist staff of HC-UFMG, and T. gondii serology, skull x-ray, transfontanel ultrasound, fundoscopic examination, lumbar puncture and clinical evaluation were performed.Congenital infection was confirmed (clinical signs and/or serology) or discarded (clinical signs absent and negative serology) after at least six months of follow-up.
DNA extraction.From the samples of amniotic fluid, 1mL was used for DNA extraction.This consisted of centrifugation at 5,000rpm for 10 minutes followed by discharge of 800µL of the supernatant.The 200µL left were incubated at 100 o C for 15 minutes; 450µL of ethanol and 20µL of ammonium acetate were then added and the tube was kept at -80 o C for 60 minutes or at -20 o C overnight.Then the sample was centrifugated at 15,000rpm at 4 o C during 30 minutes, the supernatant was discharged and the pellet was ressuspended in 50µL of deionized water.
T. gondii DNA amplification.PCR was performed using two pairs of primers that anneal to gene B1 2 of the Toxoplasma gondii, leading to amplification of 690 and 178 bp fragments (Table 1).The reaction mixture for the first amplification consisted of: 1) 10µL of the After the first amplification, 1µL (10x diluted) of the product was reamplified, now using inner primers 2 and 3.The amplification program was the same, except for the annealing temperature (60 o C).
All amplification reactions contained a positive (one microliter of DNA extracted from a T. gondii culture) and negative (10µL DDW) controls.Products were submitted to electrophoresis and analyzed in silver-stained 6% polyacrilamide gel.
Mouse inoculation.2-5mL of the amniotic fluid were centrifugated at 2,000rpm during 10 minutes; after discharging the supernatant, 1mL of PBS (phosphate buffer solution) was added to the pellet and the product was inoculated intraperitoneally into two mice.The animals were clinically evaluated daily; when death occurred, the peritoneal liquid was analyzed at the microscope (looking for tachyzoites).If after 30 days the animals were still alive, they were sacrificed; serology for T. gondii was performed and the brain was examined for tissue cysts.
Statistical analysis were performed in Epi-Info Software 5 Version 6.04, using Chi-square (comparison of frequencies), ANOVA (comparison of means) and Kruskal-Wallis (comparison of medians) tests, depending on the data profile and distribution.

RESULTS
From the 86 cases of acute gestational toxoplasmosis studied, maternal-fetal transmission occurred in 8 (9.3%).Clinical manifestations of the infected newborns were variable, as shown in Table 2. Mean mother's age in the group with congenital infection was 19.2 ± 3.0 (median = 19 years) and in the group without congenital toxoplasmosis was 23.9 ± 5.4 (median = 24 years) (p = 0.02).There were no difference in the mother's gestational history, gestational age (median = 37.5 weeks in the group with congenital infection and median = 38 weeks in the group without infection, p = 0.21) and newborn gender (p = 0.26).Mean birth weight was lower in the infected group (mean = 2,236.2± 922g; median = 2,470g) than in non-infected one (mean = 3,023 ± 495g; median = 3,045g) (p = 0.01).
One minute-Apgar score showed similar medians in the two groups (8 and 8; p = 0.49); however, five minute-Apgar score was statistically different (medians 8 and 9 in the infected and non-infected groups, respectively; p = 0.04) (Table 3).Except for Apgar scores, the missing data (which could not be obtained from the records available) were less than 10%, meaning good statistical confidence.
Prenatal exams.In five of eight cases of congenital infection, prenatal ultrasound (US) showed some alteration (four hydrocephalus, two ventricular dilatation, two hepatic calcifications and one intracerebral   7).Amniotic fluid PCR was positive in five of the eight cases of maternal-fetal transmission (three falsenegative cases).In the group without fetal transmission, PCR was positive in two (two false-positives) (Table 4).Sensitivity = 62.5% (CI 95% 25.9-89.8),specificity = 97.44%(CI 95% 90.2-99.6),positive predictive value = 71.4% (CI 95% 30.3-94.9),negative predictive value = 96.2%(CI 95% 88.6-99) (Table 7).In the positive PCR samples (five correct results and two false-positive ones), only in one case the result was achieved after the first amplification; in all the others the second step was necessary.This occurred when the same pair of primers (inner) was used for both amplifications or when the outer and inner (nested PCR) system was applied.
Table 8 shows a comparation between prenatal and postnatal exams and the final clinical diagnosis of congenital toxoplasmosis.

DISCUSSION
Congenital toxoplasmosis, an worldwide infection, can vary from asymptomatic to serious sequels.Prevention of neonate damage can be attained by information of the mothers, treatment of pregnant women already infected and early treatment of the newborns.Treatment is known to reduce the transmission rates and clinical manifestations 9 .The earlier the diagnosis is made, the better the outcome.
The maternal-fetal transmission rate found in this study was 9.3%, higher than 7.4% obtained by Hohlfeld et al 12 and lower than 38.7% and 29% from Foulon et al 8 and Dunn et al 6 , respectively.However, we could not determine the transmission rate according to the period of maternal seroconversion, which can vary widely.
There are many studies describing the use of PCR to diagnose toxoplasmosis, using P30 gene 2 , ribosomal DNA 3 or TGR1E 4 (a repetitive original DNA sequence) as target for amplification.In our study, using B1 gene as target 1 , we obtained 62.5% of sensitivity and 97.4% of specificity (two false-positive and three false-negative results), values lower than those achieved by Hohlfeld et al 12 , Cazenave et al 3 and Gratzl et al 11 .Similar results were the sensitivity of 76.2% of Gangneux et al 16 and 55% of Jenum et al 13 .
The two false-positive cases could be explained: 1) PCR can detect T. gondii DNA already damaged by drugs used in the treatment, which is unable to infect mouse, cell culture or the fetus; 2) although all measures were taken to prevent sample contamination, this can not be ruled out; 3) after birth, children were clinically followed for at least six months, when treatment was interrupted and patients were not seen anymore.Since there are cases of sequels appearing 15-20 years later, we can be misdiagnosing some cases 14 ; 4) a slight possibility is the contamination of the sample by maternal blood.
To improve sensitivity the number of amplification cycles can be increased, but this raises the risk of contamination.In the same way, to improve specificity amplification can be done in two steps, although this also increases contamination and, therefore, falsepositive results.
The three false-negatives results could be explained by: 1) short time between maternal infection and amniocentesis, fetal transmission occurring after amniocentesis; 2) inadequate storage of the samples and DNA degradation; 3) although B1 gene is repeated 35 times in the toxoplasma genome, PCR was usually performed in 1mL of amniotic fluid, which might not be representative; 4) the presence of cells, human DNA and especially blood can interfere in PCR reaction 17 .Aiming to rule out PCR inhibition, we submitted all samples to a phenol-chloroform extraction; the results, however, were not different (data not shown).
Amniotic fluid mouse inoculation gave 42.9% of sensitivity and 100% of specificity, results lower than those obtained by Foulon et al 8 (58% and 98%) and Gangneux et al 16 (52.9%and 100%).Mouse inoculation is a good test, but since the presence of viable parasites is necessary for mice infection, delay in inoculation, problems in storage or even treatment of the mother before amniocentesis can interfere in the results.
PCR and mouse inoculation are different methods that have different performances.In this study, the combination of both tests slightly improved sensitivity (75%) from 62.5% for PCR alone and 42.9% for mouse inoculation alone.
Taken together, our results showed that a negative prenatal diagnosis based on PCR alone does not always exclude the possibility of congenital infection, as well positive results do not always confirm it.Clinical evaluation of the newborns during the first year of life is obviously necessary.The final diagnosis of congenital toxoplasmosis should be based on clinical data, serological evaluation, PCR and mouse inoculation.

Table 1 -
Primers used in polymerase chain reaction.

Table 2 -
Clinical data in newborns with congenital toxoplasmosis.

Table 3 -
Mother's and newborn's profile in groups with and without congenital toxoplasmosis.

Table 4 -
PCR with amniotic fluid for diagnosis of congenital toxoplasmosis.

Table 5 -
Amniotic fluid inoculation into mouse for diagnosis of congenital toxoplasmosis.

Table 6 -
Combination of amniotic fluid PCR and mouse inoculation for diagnosis of congenital toxoplasmosis.

Table 7 -
Comparison between prenatal tests for diagnosis of congenital toxoplasmosis.

Table 8 -
Comparison between pre and postnatal exams and the final diagnosis of congenital toxoplasmosis.