|
|
| |||||||||||||||||||||||||||||||||||||||
Background and Methods We describe a child who apparently acquired human immunodeficiency virus type 1 (HIV-1) infection in the home setting. The suspected source of infection was a child with the acquired immunodeficiency syndrome who had received zidovudine and whose virus contained a mutation associated with in vitro zidovudine resistance. The children were born to different HIV-1-infected mothers, but they lived in the same home between the ages of two and five years. Child 1 was infected perinatally; Child 2 was not and was repeatedly found to be seronegative. Child 2 was examined because of acute lymphadenopathy and had seroconverted to HIV-1 positivity. HIV-1 proviral DNA was amplified from peripheral-blood mononuclear cells and subjected to sequence analysis. Sequences from Child 2 were compared with those from Child 2's mother, Child 1, and local HIV-1-infected control children.
Results HIV-1 nucleotide sequences from the third hypervariable region (V3) of the env gene from Child 2 were much more similar to those of Child 1 (with a difference of 1.3 percent) than to those of Child 2's mother (a difference of 9.9 percent) or those of four local, epidemiologically unrelated children (differences of 10.1 to 16.3 percent). A zidovudine-resistance mutation at codon 215 of the reverse transcriptase gene (Thr-to-Tyr) was found in Children 1 and 2, but not in Child 2's mother. Although the children had no documented exposure to each other's blood, there had been numerous opportunities, including nosebleeds, bleeding gums, and a laceration in Child 1.
Conclusions In the case we describe, HIV-1 with a mutation associated with zidovudine resistance was transmitted from one young child to another, apparently in the home and probably through unrecognized exposure to blood.
Case Reports
Child 1
Child 1 was born to an HIV-1-infected mother and had recurrent pulmonary infections as an infant. The acquired immunodeficiency syndrome (AIDS) was diagnosed at the age of 1 1/2 years, when a lung biopsy revealed lymphoid interstitial pneumonitis. At that time the child had a CD4+ lymphocyte count of 846 per cubic millimeter and a CD8+ lymphocyte count of 3934 per cubic millimeter. Culture and diagnostic polymerase chain reaction (PCR) were positive for HIV-1. Later, because of failure to thrive, zidovudine therapy was begun at a dose of 180 mg per square meter of body-surface area every six hours.
Child 2
Child 2 was born to a different HIV-1-infected mother, who had never received zidovudine. At six weeks of age, Child 2 was noted to have anemia of unknown origin, for which a blood transfusion was given. A urine culture was positive for cytomegalovirus, but there were no stigmata of congenital cytomegalovirus infection. Over the next several months, recurrent episodes of otitis media and sinusitis were treated with antibiotics, and for several months the child's growth slowed. The serum IgG2 concentration was found to be low (14 mg per deciliter), and monthly intravenous infusions of immune globulin were begun. Thereafter, the frequency of infections decreased and the growth rate returned to normal.
Evaluation for HIV-1 infection included cultures for HIV-1 at 3 weeks and 1 year of age, diagnostic PCR11 with two sets of primers and probes12 at 1 year of age, and testing for HIV-1 antibody by enzyme immunoassay (EIA) (Abbott Laboratories, Abbott Park, Ill.) and Western blotting (Bio-Rad Laboratories, Hercules, Calif.) on three occasions during the first 1 1/2 years of life. All were negative. CD4+ and CD8+ lymphocyte counts at 15 months of age were normal (3524 and 1436 cells per cubic millimeter, respectively). A fourth HIV-1-antibody test was again negative.
Twelve months after this last negative antibody test, disturbed sleep, easy fatigability, and tender swelling in the cervical and axillary regions developed progressively over a period of three to four weeks. Tender cervical and axillary lymphadenopathy was confirmed on examination. Laboratory evaluation revealed previous exposure to Epstein-Barr virus, strongly positive tests for HIV-1 antibody by EIA and the Western blot assay, negative tests for HIV-1 p24 antigen, a CD4+ lymphocyte count of 1381 per cubic millimeter, a CD8+ lymphocyte count of 3227 per cubic millimeter, and a negative Mantoux tuberculin skin test (using purified-protein derivative), with negative mumps and candida control skin tests. A blood sample obtained two weeks later was positive for HIV-1 by culture. The child's symptoms were attributed to acute (primary) HIV-1 infection.
The adenopathy and constitutional symptoms improved after two months but did not resolve completely. Oral thrush and herpes simplex stomatitis developed but responded promptly to therapy. Severe tinea capitis responded slowly. The CD4+ lymphocyte count four months after the first positive HIV-1 test was 851 per cubic millimeter; the CD8+ lymphocyte count was 2499 per cubic millimeter. CD4+ lymphocyte counts during the next several months ranged from 900 to 1100 cells per cubic millimeter.
Methods
Diagnostic HIV-1 PCR
Diagnostic PCR was performed essentially as described elsewhere11. There were two primer pairs: SK38 and SK39,12 and a pair corresponding to the HIV-1 long-terminal-repeat residues at positions 513 to 538 and 622 to 64413. Portions of the PCR products were assayed by hybridization with probes labeled with phosphorus-3211.
Direct Analysis of Proviral DNA Sequences
Blood samples for sequence analysis were obtained from Child 1 two months after Child 2's first positive HIV-1 test, from Child 2 two weeks and again six months after that same test, and from Child 2's mother. Samples were also obtained from perinatally infected control children with AIDS in our pediatric program.
Peripheral-blood mononuclear cells were purified from the blood samples by Ficoll-Hypaque density-gradient centrifugation and cryopreserved for later molecular analysis. DNA was prepared from the cryopreserved cells with a DNA extraction kit (Stratagene Cloning Systems, La Jolla, Calif.) according to the recommended procedure. To obtain PCR products, env and pol genes were amplified in two stages. The first amplification was carried out as described elsewhere,14 with primers with restriction sites appended to facilitate the subsequent cloning of amplified fragments. Approximately 100 ng of PCR product from the first amplification was then subjected to a second amplification for 30 cycles; in the second amplification, one primer was biotinylated at the 5' end15 to facilitate later purification for solid-state sequencing. Strict precautions were taken to prevent cross-contamination,16 and all samples were processed separately through the first stage of PCR amplification. In addition, separate blood samples were obtained from Children 1 and 2 and from Child 2's mother. These samples were sent directly to the Centers for Disease Control and Prevention (CDC) for analysis.
For env gene amplification of the samples from Children 1 and 2 and the local controls, the primers used were DV10 and DV11, and then DV7B and DV11 (Table 1). Sequences from the mother of Child 2 could not be amplified well with these primers, but they were amplified better after DV13 was substituted for DV10 and DV14B was substituted for DV7B. These primers yielded PCR products with approximately 550 base pairs (in the case of the children and the local controls) and 600 base pairs (in the case of the mother) that included the third hypervariable region (V3) of gp12017. To amplify the pol gene from the samples from Children 1 and 2 and the local controls, we used P5up and P7, and then P5B and P7. For the sample from the mother of Child 2 (which could not be amplified well with these primers) we used P5 and P7 first, and then P5B2 and P7 (Table 1). These primers yielded PCR products that included reverse transcriptase codons 22 through 279 (in the case of the children and the local controls) and codons 66 through 279 (in the case of the mother).
|
A consensus sequence of the env gene was obtained from the alignment of 27 env sequences constituting "Group B" (North American and European HIV-1 isolates) of an HIV-sequence data base13. The env gene sequences from Child 2 were compared pairwise with those from Child 2's mother, Child 1, and four local control children and with seven Group B sequences. The pol gene sequences from Child 2 were compared with those from Child 2's mother, Child 1, and two local control children. Pairwise differences were determined with the Gap and Pileup programs of the GCG Sequence Analysis software package (Genetics Computer Group, Madison, Wis.)19.
Direct sequencing of PCR-amplified HIV-1 proviral DNA yields a sequence representing a consensus of the many genotypes (the "quasispecies") present in a patient's viral population. Therefore, mixed sites are occasionally found in the sequence at which a substantial fraction of the viral population carries one nucleotide while the remainder of the population carries another. Mixed sites were scored as 50 percent concordant if one of the nucleotides in a mixed site in the sequence from one patient appeared in the corresponding position in the sequence from the other patient.
The sequences presented in this report have been submitted to Genbank (accession numbers L12751-5 and L19695-7).
Signature-Pattern Analysis
The env gene sequences obtained by direct sequencing were subjected to signature-pattern analysis to evaluate their evolutionary relatedness to other sequences20,21. In this method, a sequence of interest is first compared with a set of reference sequences to identify distinctive nucleotides. The sequence in the first sample obtained from Child 2 was scanned for nucleotides that occurred at homologous positions in fewer than 5 of the 27 Group B reference sequences from the HIV-sequence data base13. The distinctive nucleotides thus identified constituted the "signature pattern" of the sequence from Child 2 and were then compared with the nucleotides at the corresponding positions in the sequences from Child 1, Child 2's mother, and the four local control children.
Analysis of Cloned Proviral DNA Sequences
Clones were derived from the PCR products generated in the first amplification of the env and pol genes obtained from Children 1 and 2, as follows. Amplified DNA was digested with KpnI and PstI (pol) or EcoRI and HindIII (env). The products were then ligated into the similarly digested plasmid vector pUC18,22 and the ligation mixtures were used to transform Escherichia coli DH5alpha Max Efficiency Cells (GIBCO-BRL, Bethesda, Md.). Clones carrying HIV-1 sequences were identified as described elsewhere,14 with oligonucleotide probes corresponding to nucleotides 874 to 909 and 1100 to 1123 of the HIV-1 pol and env Group B consensus sequences, respectively13.
Cloned env gene products from Children 1 and 2 were sequenced as described above. The results were concordant with those of the direct sequencing and are not presented here. Cloned pol gene products from Children 1 and 2 were analyzed for their residues at codon 215 by colony hybridization with the oligonucleotide probes P9WT-P (5'GGGGGTTTACCACACCAGAC) and P9Y-P (5'GGGGGTTTTACACACCAGAC), to distinguish the wild-type reverse transcriptase codon 215 (ACC, or threonine) from that associated with in vitro zidovudine resistance (TAC, or tyrosine)10.
Epidemiologic Investigation
The children's medical records were reviewed, and interviews were conducted to determine the extent of possible exposure between the two children during the interval when transmission appeared to have occurred -- that is, the period beginning two months before Child 2's last negative HIV-1 test and ending with the child's first positive test.
Results
Sequence Analysis of HIV-1 Proviral DNA -- the env Gene
There was an overall difference of 1.3 percent between the env gene sequences in the blood sample from Child 1 and those in the first sample from Child 2 (Table 2). The two sequences were identical at 220 of the 226 nucleotide positions scored (Figure 1). They differed only at the mixed sites (Figure 2); the sequence in the sample from Child 1 contained six mixed sites, whereas the sequence in the first sample from Child 2 contained no mixed sites. At least one of the nucleotides at each of the mixed sites in the sample from Child 1 was present at the corresponding position in the first sample from Child 2. The second sample from Child 2 was identical to the first except that there was one mixed site (Figure 1).
|
|
|
To verify our results, separate blood samples were sent directly to the CDC for sequence analysis (see the Methods section). Direct sequencing of a 342-nucleotide stretch of PCR-amplified env DNA, including the V3 region, indicated a nucleotide divergence of 2.0 percent between the sequences from the two children, of 12.0 percent between Child 2 and the child's mother, and of 12.6 percent between Child 1 and Child 2's mother.
The signature pattern in the first sample from Child 2 consisted of eight nucleotides (marked with asterisks in Figure 1). All eight were present in the sequence from Child 1 at 100 percent frequency (i.e., none were mixed sites) (Table 3). In contrast, only one of the eight nucleotides occurred in the sequence from the mother of Child 2, and all eight were infrequent or absent in the sequences from the local controls (Table 3).
|
Direct sequencing of the reverse transcriptase gene from Child 1 showed the zidovudine-resistance codon TAC at position 215. Among the clones obtained from the primary PCR product, 32 yielded definitive results with either the mutant or wild-type probe; 28 (88 percent) were found to have the TAC codon by colony hybridization, whereas 4 were wild-type. Direct sequencing of the first sample from Child 2 showed that here, too, the predominant codon at position 215 was the zidovudine-resistance codon (TAC), and 11 of 16 clones (69 percent) derived from the primary PCR product had this mutation, whereas 5 were wild-type. In contrast, the wild-type sequence (ACC) predominated in the second sample from Child 2. Colony-hybridization data confirmed this result; the codon 215 mutation was found in only 2 clones, whereas 13 were wild-type. Direct sequencing of the samples from both children showed predominantly or exclusively wild-type nucleotides at the other four codons known to be associated with in vitro zidovudine resistance (codons 41, 67, 70, and 219)9,10. Direct sequencing of the sample from the mother of Child 2 showed only wild-type residues at codons 67, 70, 215, and 219 (the PCR product derived from this sample did not include codon 41).
The direct sequences of pol DNA from Child 2 differed from those of Child 1 by only 0.7 to 0.8 percent over a span of 603 nucleotides, from those of the mother of Child 2 by 3.4 to 3.8 percent over a span of 469 nucleotides, and from those of two control children by 3.0 to 4.0 percent over a span of 603 nucleotides (with the mutation at codon 215 excluded from consideration in all cases).
Epidemiologic Investigation
The interval beginning 2 months before Child 2's last negative HIV-1-antibody test and ending with the child's first positive test lasted 15 months. Throughout this period, the two children lived in the same household, and both were between two and five years old. Child 1 received zidovudine during the entire period.
During this period, no exposure of Child 2 to the blood or body fluids of Child 1 was witnessed. However, there were several opportunities for such exposure. Child 1 had frequent nosebleeds, associated with the use of a nasal steroid inhaler and resulting in an average loss of approximately 5 ml of blood; gum bleeding occurred almost daily with tooth brushing; and there was frequent otitis media with purulent otorrhea. Throughout most of the period, Child 2 had a rash consisting of 5 to 20 discrete pruritic, papulovesicular lesions 1 to 2 mm in diameter on the trunk and extremities that were frequently excoriated from scratching. Occasionally, the two children bit each other (although bleeding was never noted), slept in the same bed, and were thought to have used the same toothbrush.
Seven months before Child 2's first positive HIV-1 test, Child 1 received a 10-day course of antibiotics in the home through a peripheral intravenous catheter that was inserted at the hospital and replaced once at home by a nurse. Twice a day, antibiotics were administered and the catheter was flushed. Contaminated needles and syringes were stored in a puncture-resistant container on a shelf out of reach of the children. No needle-stick injuries were noted.
During the three months before Child 2's first positive HIV-1 test, Child 1 was struck by Child 2 and had a laceration that required sutures for closure; Child 1 had tympanostomy tubes inserted, resulting in an increased volume of purulent otorrhea, which sometimes drained onto Child 1's face. On one occasion, both children received carefully monitored two-hour intravenous infusions in the same clinic on the same day.
The two children were never hospitalized together, did not receive dental care from the same dentist, and never received immunizations on the same day. There was no history or physical evidence of sexual abuse of either child. Child 2 had no contact with the mother of Child 1.
Discussion
The clinical histories and epidemiologic circumstances of these children led us to suspect that HIV-1 may have been transmitted from Child 1 to Child 2; nucleotide-sequencing data strengthened this inference. The agreement between the V3 regions of the proviral sequences from the two children (with a difference of only 1.3 percent) was similar to or greater than that found in many previous studies of V3 sequence variation within infected subjects and between epidemiologically linked subjects20,23,24,25,26,27. Furthermore, the signature nucleotides of the two children were identical, confirming that they harbored very similar quasispecies21,28. The differences of 10 to 16 percent between the children's V3 sequences and those of the mother of Child 2, the local controls, and the HIV-1 data base are similar to those reported in epidemiologically unrelated subjects13,20,26. The V3 sequences from Child 1, who was infected perinatally, were more heterogeneous than those from Child 2, who was infected more recently. This is consistent with recent studies demonstrating increasing sequence divergence over time after sexual, parenteral, or perinatal transmission25,26,27.
Child 2 presented with an acute mononucleosis-like syndrome after laboratory evaluation, including negative diagnostic PCR, indicated that perinatal infection had not occurred. The negative PCR distinguished this child from the rare infants who revert from seropositive to seronegative but apparently harbor HIV-1 proviral DNA detectable by PCR testing29,30,31. A workup of Child 2's acute episode led to the presumptive diagnosis of a syndrome of primary HIV-1 infection. The frequency of this syndrome among HIV-1-infected adults may be as high as 50 to 70 percent32; it has been reported in a four-year-old child with hemophilia33.
Our findings with regard to the resistance mutation at codon 215 provide additional support for the close similarity between the two children's quasispecies and for the transmission from one to the other. Many earlier studies failed to detect zidovudine-resistance mutations in patients who had not received the drug34,35,36. The high frequency of the codon 215 mutation in the viral population in the first sample from Child 2 presumably resulted from the selective pressure of zidovudine therapy in Child 1.
This report documents with molecular epidemiologic evidence the apparent transmission of HIV with a zidovudine-resistance mutation. In one other patient, HIV-1 infection with a virus bearing a zidovudine-resistance mutation (also at codon 215) was thought to have been acquired sexually from a zidovudine-treated adult partner, but samples from the partner were not available for study37. In that patient, zidovudine treatment during the acute retroviral syndrome was followed by persistence of the codon 215 mutation with the selection of highly zidovudine-resistant virus. In our Child 2, during the five-month interval between the first and second blood samples, when the child was not receiving zidovudine, the codon 215 mutation largely disappeared from peripheral-blood mononuclear cells.
Our findings with regard to overall differences in pol sequences are also in accord with transmission to Child 2 from Child 1 rather than from Child 2's mother. However, we know of no previous molecular epidemiologic studies of pol in known cases of perinatal transmission, and there is only a brief description of a comparison of pol amino acids in a forensic study of sexual transmission38. Thus, overall differences in pol sequences cannot yet be considered as informative from an epidemiologic standpoint as the env data or the pol resistance-mutation results.
The efficacy of zidovudine in patients infected with viral variants that have in vitro zidovudine resistance is not well understood, although the emergence of such virus in patients receiving long-term therapy is a poor prognostic sign39. Whether primary infection with resistant virus carries a similarly poor prognosis is not known. The rapid overgrowth of resistant variants by susceptible virus, which we observed in Child 2, suggests that a selective disadvantage may keep resistant viruses in check when the selective pressure of zidovudine therapy is removed. Whether resistant variants will re-emerge at an accelerated rate when therapy is initiated later has yet to be determined.
The mode of transmission in this case is unknown, but transmission probably resulted from an unrecognized exposure to blood. Blood from Child 1's nose, gums, or laceration may have come in contact with Child 2's mucous membranes or excoriated skin lesions. One earlier case of suspected household transmission between young children2 was not well documented. In another case, transmission between young siblings with hemophilia was documented,3 but in that instance there was ample opportunity for a needle-stick injury. Considerable epidemiologic evidence indicates that transmission in the household in the absence of sexual or percutaneous exposure is rare1,4,5,6,7,8. Nonetheless, the blood and bloody body fluids of HIV-infected persons are infectious,40 no matter what the setting. Therefore, care should be exercised to prevent exposure to blood40 in all settings, including the home.
Supported in part by an American Foundation for AIDS Research/Pediatric AIDS Foundation Scholar Award (to Dr. Fitzgibbon) and Central New Jersey Pediatric AIDS Program Research Funds (to Drs. Gaur and Frenkel).
We are indebted to the New Jersey Department of Health for advice and cooperation, to Ms. Sheila Mazar and Ms. Rebecca Kuk for technical assistance, to Ms. Florence Szymanski for assistance in the preparation of the manuscript, and to Dr. Chi-Cheng Luo, Dr. Marcia Kalish, and Dr. Gerald Schochetman of the CDC for independently verifying the results of sequencing.
Source Information
From the Departments of Molecular Genetics and Microbiology (J.E.F., L.D.F., D.T.D.) and Pediatrics (S.G., L.D.F.), University of Medicine and Dentistry of New Jersey-Robert Wood Johnson Medical School, Piscataway, N.J.; the Division of Field Epidemiology, Epidemiology Program Office (F.L.), and the Division of HIV/AIDS, National Center for Infectious Diseases (B.R.E.), Centers for Disease Control and Prevention, Atlanta; and the New Jersey Department of Health, Trenton (F.L.).
Address reprint requests to Dr. Dubin at the Department of Molecular Genetics and Microbiology, UMDNJ-Robert Wood Johnson Medical School, 675 Hoes Ln., Piscataway, NJ 08854.
References
| |||||||||||||||||||||||||||||||||||||||
Related Letters:
Transmission of HIV-1 from One Child to Another
Alimenti A., Levy J., Fremont-Smith K., Dunn D., Newell M.-L., The European Collaborative Study , Friedland G. H., Gaur S., Frenkel L. D., Dubin D. T., Laraque F.
Extract |
Full Text
N Engl J Med 1994;
330:1313-1314, May 5, 1994.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
