Prenatal diagnosis of human parvovirus B19 in nonimmune hydrops fetalis by polymerase chain reaction

Prenatal diagnosis of human parvovirus B19 in nonimmune hydrops fetalis by polymerase chain reaction Bruce W. Kovacs, MD, Dru E. Carlson, MD, Bejan Shahbahrami, MS, and Lawrence D. Platt, MD Los Angeles, California OBJECTIVE: Nonimmune hydrops fetalis is a potentially lethal condition reflecting the clinical manifestation of several pathologic processes. Recently maternal infection by human parvovirus 819 has been reported to result in nonimmune fetal hydrops. We sought to develop a rapid and sensitive test to detect the presence of this agent in utero. STUDY DESIGN: Using a cloned isolate of the virus, we developed an assay based on enzymatic amplification of a segment of the human parvovirus 819 genome that allows direct detection of this agent in samples of fetal blood and amniotic fluid. RESULTS: The method detected as few as 100,000 genome equivalences and was specific for the viral genome alone. We used this assay to evaluate nine fetuses initially seen with nonimmune hydrops. Three cases were found to be positive for the human parvovirus 819 genome. CONCLUSION: The method is powerful in that it is rapid, sensitive, and simple. This assay may have general applicability in evaluation of nonimmune hydrops and in documentation of the natural history of fetal human parvovirus infections . (AM J OaSTET GVNECOL 1992;167:461-6.)

Key words: Prenatal diagnosis, parvovirus B-19, hydrops fetalis, polymerase chain reaction Hydrops fetalis is a potentially leth al condition that can be cau sed by immunologic or nonimmunologic factors. The advent of me asures to prevent rhesus factor isoimmunization has resulted in nonimmune fetal h ydrops becoming the most frequentl y encountered form of this condition in the United States. I 1t is estimated that th is cond ition com plicates 1 in every 3000 to 4000 pregnancies! Importantly, a variety of conditions have been associated with nonimmune fetal hydrops. Therefore depending on the cause the fetal outcome can be quite variable. In some cases the hydrops appears to resolve spontaneou sly, resulting in an apparently normal offspring. However, in other cases the prognosis for the fetus is grave.' When fetal h ydrops is diagnosed by ultrasonography a clinical dilemma arises with respect to the most appropriate counseling and management. Thus it is important th at as much information as possible be gathered relative to the cause so that the antenatal management of this disorder is optimal. 4 Recently maternal infection by human parvovirus B19 during gestation has been reported to be associated

From the Departments of Obstetrics-Gynecology and Pathology, University of S outhern Californ ia School of Medi cine. Received for publication Au gust 12 , 1991 ; revised Jamuu) 21 , 1992; accepted February 13, 1992 . Reprint requests: Bruce W. Kovacs, MD, M olecular and Cytogenetics Laboratory, 1240 N. Mis sion R oad, Room 1M-1 8, Los Angeles, CA 90033 .

6/1/37265

with nonimmune fetal hydrops.' Therefore in addition to other modalities of diagnosis a rapid test to detect the presence of human parvovirus B19 in utero would add to the clinical tools used in the evaluation of nonimmune hydrops and aid the clinician in management decisions. In this study we report on a sensitive method to detect the presence of the human parvovirus B19 deoxyribonucleic acid (DNA) in clinical specimens based on sequence-specific DNA amplification." This polymerase chain reaction detection method was used to identify a portion of the parvovirus B 19 genome in samples of fetal blood and amniotic fluid obtained from several pregnancies presenting with non immune hydrops fetal is. Material and methods A 20-base oligonucleotide probe complimentary to nucleotides 1730 through 1749 of a human parvovirus B19 isolate' was synthesized on an automated DNA synthesizer (model 380, Applied Biosystems, Foster City, Calif.) and purified by high-performance liquid chromatograph y and acrylamide gel electrophoresis. H ybridization of this probe to cloned plasmid DNA and DNA obtained from clinical samples was performed either in gel or with nylon membranes after radiolabeling the B19 oligonucleotide probe with phosphorus 32-labeled adenosine triphosphate T4 pol ynucleotide kinase." The conditions of hybridization for both the dried gels or nylon membranes was at 60° C for 8 to 12 hours in 6 x standard saline citrate buffer contain461

462

Kovacs at al.

August 1992

Am

J Obstet GynecoJ

Table I. Nonimmune fetal hydrops cases evaluated Case

Gestational age (wk)*

Fetal hemoglobin (gmldl)

Human parvovirus (DNA)

Fetal outcome

3476 9802 8217 5653 7721 2356 1121 5183 3325

29 21 35 32 32 22 30 37 36

14.9 12.1 13.5 15.2 9.1 2.5 15.0 12.6 17.1

Negative Negative Positive Negative Negative Positive Negative Positive Negative

Live birth Fetal death Infant death Neonatal death Neonatal death Termination Live birth Neonatal death Neonatal death

*At blood sampling.

tOf percutaneous umbilical blood sample.

ing 10 fLg/ml transfer ribonucleic acid and 2 x 10- 6 cpm/ml of the B19 oligonucleotide probe." After hybridizations the matrixes were washed in 6 x standard saline citrate buffer for 2 to 3 hours at room temperature and 30 minutes at 65° C. Gels and membranes were then autoradiographed with intensifying screens at - 70° C for 1 to 3 days . The primer set for enz ymatic amplification of the parvovirus B 19 genome by polymerase chain reaction was synthesized and purified in the same way as B 19 oligonucleotide probe. The 5' primer was identical to nucleotides 1667 through 1686 of the parvovirus genome; the 3' primer was the reverse compliment of nucleotides 1921 through 1940 of the viral genome.' The primers for the human genomic positive standard were composed of a pair of oligonucleotides that amplify a 700 base-pair segment previously described in the human actin gene. 9 The conditions for the polymerase chain reaction amplification were optimized for a 100 fLl reaction volume with 2.5 units of T aq polymerase (AmpliTaq, Perkin-Elmer Cetus, Norwalk, Conn.). Each reaction mixture contained 50 mmol/L potassium chloride, 10 mmol/L Tris pH 8.3, 2.5 mmol/L magnesium chloride, 1.0 mmol/L deoxynucleoside triphosphates, 40 pmol/L each primer, 4 mmol/L spermidine hydrochloric acid, and various amounts of template DNA as noted in results. Each reaction mixture was overlayed with mineral oil; after an initial incubation at 97° C for 4 minutes, 30 cycles of amplification were carried out, each consisting of I minute at 94° C and 3 minutes at 65° C. After the amplification 10 fLl aliquots of the reaction mixture were either subjected to electrophoresis in 2% agarose gels for 1 hour at 60 V or spotted onto nylon membranes with a dot blot device. Nonisotopic detection of the amplification product was done by staining gels with ethidium bromide for 15 minutes; the gels were then visualized and photographed with ultraviolet light. The known positive standard template DNA source was the plasmid pYTI04C (a gift from P. Tattersall) containing the entire coding sequence of a human par-

vovirus B19 isolat e.' Before hybridizations and polymerase chain reaction amplifications the plasmid DNA was linearized by Sal I digestion. Clinical samples consisted of 1.5 ml of amniotic fluid or 0.25 ml of fetal cord blood obtained antenatally by ultrasonographically guided amniocentesis or percutaneous umbilical blood sampling from gestations initially seen at various gestational ages with hydrops fetalis. Samples of maternal peripheral blood were also obtained at initial evaluation. The basic clinical aspects of the gestations evaluated are summarized in Table l. Blood and amniotic fluid samples were prepared for dot-blotting or polymerase chain reaction by a rapid freeze-thaw followed by centrifugation at 12,000 rpm in a microfuge for 5 minutes to obtain a supernatant free of cell debris. In both the dot blot and polymerase chain reaction experiments 20 fLl of supernatant was used without further treatment. Approval for use of the test on material from human subjects was obtained from the institutional research committee. Results

The sensitivity and specificity of the B 19 oligonucleotide probe and the optimal conditions for hybridization were determined by dot blots with serial dilutions of DNA isolated from plasmid pYT104C suspended in buffer. The quantity of DNA spotted onto the nylon membrane for this determination ranged from 10 pg to 1 fLg. In addition , 0.1 fLg of plasmid DNA was added to 0.5 mllysed whole blood from an uninfected individual and spotted onto the nylon membrane. An additional control of 0.5 ml lysed whole blood without pYT104C DNA was also used. The re sults of these hybridizations are shown in Fig. 1. This experiment demonstrated that , under our conditions, the B19 oligonucleotide probe detected as little as 10 pg of the parvovirus DNA, equivalent to about 2 x 10' viral particles. Parvovirus DNA was also detected in the presence of whole blood, although the signal strength was slightly less intense than that seen with an equivalent quantity of ta rget DNA suspended in buffer. Thus

Vnhn ne ](i7

Nu mber 2

in the presence of whole blood the lower limit of sensitivity was about fivefold less tha n the sensitivity in bu ffer, or about 10" vira l particles. No hybr idization signa l was detected at th e spo t conta ining only lysed who le blood , indicating absence of d etectable nonspecific hybridization to human DNA. We ne xt used th e pr imers B 19A/B with the line arized who le plasmid DNA for po lymerase cha in re action ' am plifications. T o o ptimize the reaction, various para meters of thermal cycling were tried . In these experiments we used a I /Jog qu antity of pYTI04C DNA as the tar get for amplification s. Negativ e contro ls consisted o f l/Jog of human DNA, a 30 /Jol sample of lysed whole blood (both obtained fro m a noninfected person), or an aliquot of sterile bu ffe r. Gel electrophoresis of th e amplification reaction pro ducts revealed that after an initi al denaturation ste p at 95° C for 1 minute th e 274-base-pair product was best amplified with a sing le te mpe ra tur e for an nea ling and extension. The negat ive contro ls did not yield any a mplification product. To d ocument that th e am plification product was d er ived from th e parvoviru s ge no me th e gels were h ybr id ized to radiolab eled B 19 oligon ucleo tid e probe. T he autora diograp hs from th is h ybr idization dem onstrated th at th e D NA fragm ent amplified was speci fic for th e seg me n t of th e par voviru s ge no me d ete cted by th e parvovirus-specific probe. We next d etermined th e minimum qu antity of parvovi ru s DNA necessary for am plificatio n. Setting the num ber of amplification cycles at 30, with 1 /Jog of eac h prime r 1 /Jo[ aliquot s o f serial dilution s o f pYTlO4C plasmid DNA ranging fro m 100 pg to 0.0 I pg were su bjected to amplification . After this 50 /Jol quantities of th e reac tion products were elect rophoresed for I hour, stai ne d with ethidium bromide, and photographed. As sho wn in Fig. 2 these expe rime n ts demonstrated th at as little as 1 pg or 2 x 10 ' vir al particles could be amplified and detected un d er our cond itions. Wh en th is experime nt was repeated with the same amou n ts of pYTI 04C DNA su spended in who le blood lysate a reduction in se nsitivity was agai n found. Thus under th ese conditio ns our lower limit of detectio n was 5 pg of p YTI 04 C DNA, eq uivale nt to 1 x 10' vira l particles. Aft er these preliminary stud ies clinical samples obtained at prenatal evalu at ion of fe tal hydrops were assayed by our pol ymerase cha in reaction method with th e optimized parameters d eveloped in ou r preliminar y studies. The results of th ese amplificatio ns are show n in Fig. 3. With the method descr ibed we were able to detect the presen ce o f parvoviru s B 19 in fetal blood sam ples in three o f nin e cases in which nonimmune hydrop s was detected by ultrason ogr aph y. Howe ver, only in case 2356 was th e vir us also d etected in amniotic fluid. In addition, as determined by visual inspection of the ethidium bromide-stained gel , the amount of

Prenatal detection of human parvovirus 819

463

Fig. 1. H ybridizat ion o f B 19 oligonucleotide probe to d ilut ion s of parvovir us DNA. The ra nge of conce ntration is fro m I tJ.g of DNA in the first dot to 10 pg in the fifth dot. Positio ns :1 and 6 co nta in equ ivalent amo unts of parvovir us DNA (0. 1 tJ.g) mixed with sterile buffer and whole blood lysate , re spect ively. Position 7 conta ins a negati ve cont ro l of who le blood lysate alone .

virus in th e fluid was mu ch less than in the fetal blood in tha t th e 274-base-pair am plification product was barel y d etectable. In all cases the success o f polymerase cha in react ion amplification was established by the presence of the 700-base pair actin seque nce on the gels. In no cases was parvovirus B19 d etectable in the samples o f maternal blood ob tain ed at the time of fetal sampling. In all ofthe positive cases follow-up polymerase chain reac tio n analysis was done on fetal samples. In case 82 17 th e clinical man ifestat ion s of h ydrops re solved and th e pat ient was deli vered of a male infant with out any a p paren t seq uelae . Test results for a cord blood sample obtai ned 4 wee ks after th e initial sample were still positive for parvoviru s B 19 in th e su bseque nt polymerase chai n rea ction ana lysis. Notably, in thi s fetus th e hydrops resol ved withi n 4 week s o f our evaluation , and th e pregnancy res ulted in a live-born male infant with ou t ap pa ren t problem s. T his infant died at home 53 days afte r deli ver y u nder circums tances presumed to be consiste nt with sud de n infant d eath syndrome . An a uto psy was not obta ine d on th is infant precluding further analysis. In case 2356 the h ydrop s persisted and worsened ove r a 7-day period. At 25 weeks' ge station ultrasonogr aphy revealed massive peri ca rdi al effusion, ascites, and no card iac motion in th e fetus. After documen-

464 Kovacs et al.

M

August 1992 Am J Obstet C ynecol

1

2

3

4

5

6

7

8

9 10

-274bp

Fig. 2. Gel electrophoresis of polymerase chain rea ctions with various amounts of parvovirus B19 DNA with whole blood lysate . Lanes 1,2, and 3 are negative controls with sterile buffer . I f.Lg purified human DNA, and whole blood lysate as the amplific ation template. Lan es 4 through 9 are amplifications with 0.1 pg, lpg, 5 pg , 50 pg , 100 pg , and 500 pg quantities of parvovi ru s DNA in whole blood lysate . Lan e 10 is a positive control of 500 pg of parvovirus DNA in sterile buffer . Fragment size is indicated in base pairs (bp) .

tation of an intrauterine fetal death the pregnancy was terminated . No autopsy was performed on this fetus ; however, after delivery a portion of the placenta and fetal blood from cardiac aspiration was obtained, and test results for both were positive for parvovirus B19 with our assay method . Case 5181 was delivered 36 hours after fetal blood sampling. The infant was delivered with marked hypotension and was admitted to the neonatal intensive care nursery, where he died 12 hours after delivery. Blood samples obtained in the nursery were sent for viral cultures, which were positive for parvovirus B19. In this case an autopsy revealed a 3280 gm male infant with generalized hydrops and edema. Hepatosplenomegaly was noted, along with an enlarged heart. Histopathologic examination of the heart demonstrated generalized inflammatory changes. The liver demonstrated perivascular mononuclear infiltrates. These findings were believed to be consistent with those reported in other cases of fetal infection with several viruses. including parvovirus B 19."Test results for a portion of fetal cardiac muscle and spleen tissue obtained at autopsy were positive for parvovirus B 19 with the polymerase chain reaction assay. The maternal histories obtained in the three positive cases were notable for several features. In case 8217 the mother gave a history of recent travel to Mexico to visit family. During that visit she came in contact with several children who had symptoms resembling "influenza." However, she never experienced any symptoms.

In case 2356 the mother gave a history of "flu" with fever approximately 4 weeks before evaluation. Case 5181 was associated with maternal exposure to a child with nausea, vomiting, fever, and a "red face" in her home about 5 weeks before evaluation. Parvovirus-specific immunoglobulin studies were not obtained in an y of these cases. Comment

The polymerase chain reaction is a rapid and relatively simple molecular genetic technique that enzy matically amplifies exceptionally small amounts of a specific DNA present in clinical samples. Since the first description of this technology the pol ymerase chain re action method has been applied to clinical specimens for diagnosis of a variety of neoplastic, genetic, and infectious diseases." Indeed , at the same time the studies reported here were underway another group of investigators also reported on the use of polymerase chain reaction to detect the presence of human parvovirus B 16 in formalin- fixed tissues obtained at autopsy from a case of an intrauterine fetal death. II In our study we used the pol ymerase chain reaction tech nique to amplify a different portion of the viral genome and demonstrated that the sens itivity of this method exceeds that which can be obtained with more traditional approaches, such as dot-blots. Importantly, both studies show that the detection of the viral genome was specific in that known parvovirus-negative samples did not yield amplification product. However, as has been

Prenatal detection of human parvovirus 819

Volu me 167 N umbe r ~

bp

1

2 3 4 5 6 7 8

465

9 10 11

700-..

274~

Fig . 3. Gel elect ropho resis of the am plification products o btai ned from fetal blood sam ples in the nine cases in itially see n with hydrops fet alis are see n in lanes I th rou gh 9. Lane I () is a negat ive cont ro l, and lan e II is a positive co nt ro l. The 274 -base-pai r (hpj frag ment is th e par vo vir us am p lificat ion produ ct, and th e 700-base-p air frag me nt co rresponds 10 the human act in sta ndard .

not ed by others, contamin ation of samples with exogeno us target DNA template can be a poten tial probl em in polymerase chain re action -based assays. III Therefore ap p ro priate measures must be tak en to avo id thi s problem , and negative contro ls sho u ld always be used in am plification protocols. I. Moreover, analysis of paired maternal samples will preclude th e possibility of falsepositive results resulting from cros s-contamination with viral particles in maternal blood . We also found that th e sens itivity of the assay decreased in the presence of lysed who le blood. This is not ent ire ly unexpected and is most likely the result of in hibi tion of the Taq polymerase enzyme by heme presen t in th e sam ple. Alternatively, it could also be cause d by some nonspecific binding of th e oligo nucleo tide primer sequences to human ge no mic DNA possessing some hom ologic seq ue nce becau se th e effect was also not ed in the do t blot s. In th e polyme rase chain reactio n amplification process non specific binding would result in a reduction of primer Item plate ratio , which would in turn affect amplificat ion efficiency." However , the se nsitivity of th e assay, even in th e presen ce of blood , is at th e level of about 100,000 vira l particles. Moreover , the use o f radioactive nu cleotides in the polymerase cha in rea ction would result in even greater sensitivity, as would increases in th e n u mber of amplificatio n cycles, without affectin g spec ificity. In our study we have d em on strated that parvovirus B IY DNA can be detected in amniotic fluid and fetal blood samples obtained in a p ren atal diagnostic setting

with en zymatic DNA amplification. T his technique represe nts an ex tremely sen sitive detec tion method that is es pecially useful when sma ll sa mple size limits dia gnostic techniques. It is also more ra pid th an viral culture for diagnos is, req uiring o nly a few hours to perform. Moreve r, in appl ying this techn ique to a group of clin ic pat ients we found th at some gestations with signs of nonimmune fetal hydrops have d efinitive evidence of parv ovirus B 19 infection. T he basic pathogenesis of fetal hydrops is thought to be th at of cardiovascular failu re as precipitated by anemia , myocardial dysfunction , vascular shunts, and ot her alte rations in card iovascular dynamics! In the case of parvovirus B 19 infection , ane mia or direct cyto pa thic effe cts on fe tal orga ns, including the myocard ium , have both been postul ated to be re sponsible for the fda I hydrops." In th e cases rep orted here anemia was noted in the fet al blood samples, a finding that is cons istent with other repurts o f fe tal infection with parvovirus B19 but is not an in variable findin g in all cases. " On th e basis of the limited num ber of cases in thi s se ries we cannot speculate on th e true inciden ce of anemia associated with human parvoviru s B 19 in fetuses initia lly see n with nonimmune hydrop s. Neither can we be certa in that the pathogen esis o f the hydrops was in fact ca used onl y by the ane mia . T he presence of detecta ble parvovirus in th e myocardium of one case combine d with relatively mild ane mia in another suggests that more than erythroid cell d estruction may contribute to th e fetal hydrops ; thi s has also been suggested

466 Kovacs et al.

in other studies of fetal parvovirus B19 infections. 13 Most important, however, the variable outcomes of the pregnancies seen in our study indicate that many additional questions regarding the association of human parvovirus B 19 infection and nonimmune hydrops remain to be answered. In our series we noted a live-born infant without obvious sequelae after an in utero infection. Thus our case of a documented fetal viremia and hydrops that spontaneously resolved adds to those cases reported by others of apparently normal outcomes in some fetuses after in utero parvovirus B 19 infections. 13 However, in our case a sudden infant death without a definitive cause of death is disturbing and might be related to early and apparently persistent chronic infection. This case emphasizes that much is yet to be learned about the pathologic make-up of parvovirus infection and its long-term effects on fetuses that have viremia in utero. In sum, on the basis of the data presented here we believe that the simplicity of this diagnostic method presents clinicians with an important tool for evaluation of nonimmune fetal hydrops. When applied to larger clinical studies this method will be an important tool in understanding the natural history of parvovirus B 19 infection occurring in pregnancy. REFERENCES 1. Machin GA. Differential diagnosis of hydrops fetal is. Am J Med Genet 1981;9:341-50. 2. Castillo RA, Devoe D, Hadi HA, Martin S, Geist D. Non-

August 1992 Am J Obstet Gynecol

3. 4. 5. 6. 7.

8. 9.

10.

11.

12. 13.

immune hydrops fetalis: clinical experience and factors related to poor outcome. AM J OBSTET GYNECOL 1986; 155:812-6. Watson J, Campbell S. Antenatal evaluation and management of nonimmune hydrops fetalis. Obstet Gynecol 1986;67:589-92. Romero R, Pilu G, Jeanty P, Ghidini A, Hobbins JC. Prenatal diagnosis of congenital anomalies. Norwalk, Conn.: Appleton & Lange, 1988:414-26. Anand A, Gray ES, Brown T, Clewley JP, Cohen BS. Human parvovirus infection in pregnancy and hydrops fetalis. N EnglJ Med 1987;316:183-6. Saiki RK, Gelfand DH, Stoffel S, et al. Primer directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 1988;239:487-91. Shade RO, Blundell MC, Cotemore SF, Tattersall P, Astell CR. Nucleotide sequence and genome organization ofhuman parvovirus B19 isolated from the serum of a child during aplastic crisis. J Virol 1986;58:921-36. Kovacs BM, Shahbahrami B, Medearis A, Comings DE. Prenatal determination of paternity by molecular genetic "fingerprinting." Obstet Gynecol 1989;75:1-6. Nakajima-lijima S, Hamada H, Reddy P, Kakunaga T. Molecular structure of the human cytoplasmic B-actin gene: interspecies homology of sequences in the introns. Proc Nat! Acad Sci USA 1985;82:6133-6. Wright PA, Wynford-Thomas D. The polymerase chain reaction: miracle or mirage? A critical review of its uses and limitations in diagnosis and reserach. J Pathol 1990; 162:99-117. Salimans MMM, Van de Rijke FM, Raap AK, ElsackerNiele AMW. Detection of parvovirus B19 DNA in fetal tissues by in situ hybridization and polymerase chain reaction. J Clin Pathol 1989;42:525-30. Sarkar G, Sommer SS. Shedding light on PCR contamination. Nature 1990;343:27. Brown KE. What threat is human parvovirus B-19 to the fetus? Br J Obstet Gynaecol 1989;96:764-7.

Bound volumes available to subscribers Bound volumes of the AMERICAN JOURNAL OFOBSTETRICS AND GYNECOLOGY are available to subscribers (only) for the 1992 issues from the Publisher, at a cost of $68.00 for domestic, $95.76 for Canada, and $91.00 for international for Vol. 166 (January-June) and Vol. 167 (July-December). Shipping charges are included. Each bound volume contains a subject and author index and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable buckram with the JOURNAL name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact Mosby- Year Book, Inc., Subscription Services, 11830 Westline Industrial Drive, St. Louis, MO 63146-3318, USA; phone (800) 325-4177, ext. 4351, or (314) 453-4351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular JOURNAL subscription.