|
|
| |||||||||||||||||||||||||||||||
Background The presence of a specific cellular receptor is thought to be necessary for susceptibility to viral infection. The erythrocyte P antigen is the cellular receptor for parvovirus B19. We hypothesized that the rare persons with the p phenotype, whose erythrocytes do not have this receptor, would be naturally resistant to B19 infection, which causes erythema infectiosum.
Methods Blood samples were collected from two populations in cross-sectional studies. We determined the P antigen phenotype of the red cells and tested plasma for anti-B19-specific antibodies. Bone marrow from donors of known P antigen phenotype was inoculated with parvovirus B19. Infectivity was measured by assays of erythroid progenitor cells, dot blot analysis, and in situ hybridization for B19 DNA, and an immunofluorescence assay for viral-capsid proteins.
Results Of the 17 subjects with the p red-cell phenotype, who did not have P antigen on their erythrocytes, none (0 of 11 and 0 of 6) had serologic evidence of previous parvovirus B19 infection. In contrast, the seropositivity rates in the two control groups were 71 percent (53 of 75, P<0.001) and 47 percent (32 of 68, P = 0.03). In vitro, bone marrow from donors with the p phenotype maintained normal erythropoiesis despite very high concentrations of virus, with no evidence of infection of erythroid progenitor cells by parvovirus B19.
Conclusions People who do not have P antigen, which is the cellular receptor for parvovirus B19, are naturally resistant to infection with this pathogen.
Replication of parvovirus B19 has been demonstrated only in human erythroid progenitor cells. In the laboratory the virus has been propagated only in bone marrow, peripheral blood, fetal liver, and a few hematopoietic cell lines with erythroid characteristics. The basis for the extraordinary tropism of the virus for erythroid cells has been clarified by the recent discovery of the cellular receptor of parvovirus B196. The virus binds to an antigen of the blood-group P system known as P antigen or globoside, and in tissue culture either excess soluble receptor or an antibody to globoside can protect erythroid progenitor cells from infection.
P antigen is present not only on erythrocytes and erythroblasts but also on megakaryocytes, endothelial cells, and placenta, fetal liver, and heart cells7,8. The tissue distribution of P antigen is consistent with the clinical syndromes caused by parvovirus B19 and with laboratory data. Although the virus has not been shown to replicate in megakaryocytes or myocardial cells, its entry into these cells has been inferred from viral inhibition of megakaryocytopoiesis in vitro and in situ hybridization of fetal heart tissue9,10. After its attachment to P antigen and entry into the cell, parvovirus B19 may replicate in a manner dependent on erythroid-specific transcription factors11.
The P blood group, discovered in 1927 by Landsteiner and Levine,12 contains two common antigens, P1 and P, and a third that is much rarer, Pk. p antigen was subsequently identified as globoside13. Red cells of persons with the blood-group P1 phenotype have both P and P1 antigens; persons with the P2 phenotype have P antigen alone (Table 1). The rare persons with P1k and p phenotypes have no P antigen on their red cells; persons with P1k have both P1 and Pk antigens13. The blood-group p phenotype was first identified in 1951 and originally called Tj(a-)14; persons with this phenotype have none of the three blood-group P antigens and are at risk for massive hemolysis if they are given transfusions of blood containing P antigen15. We have previously shown that parvovirus B19 will not hemagglutinate erythrocytes from persons with the blood group P1k or p phenotypes6. We therefore hypothesized that persons without P antigen would be naturally immune to B19 infection and that their bone marrow would be resistant to B19 infection in vitro.
|
Blood and Bone Marrow Samples
Blood and bone marrow samples were obtained under research protocols approved by the institutional review board of the National Heart, Lung, and Blood Institute.
Assays for Erythroid Progenitor Cells
Parvovirus B19 is highly inhibitory to the production of erythroid colonies from human late erythroid progenitor cells (colony-forming unit, erythroid) in methylcellulose culture, and this system was used as an assay for parvovirus B19 infection16. Bone marrow was obtained from two subjects with blood-group phenotype p, and their mononuclear cells were isolated and infected with parvovirus B19 in various concentrations. The cell mixtures were cultured in methylcellulose, and erythroid colonies were assayed by microscopical examination of the plates seven days later.
Virologic Studies
Evidence of past parvovirus B19 infection was based on the detection of virus-specific IgG antibodies with an antibody-capture test17.
Viral DNA and protein were assayed in infected marrow cultures. Mononuclear bone marrow cells were obtained from persons with either the p or the P blood-group phenotype; they were incubated with parvovirus B19 for two hours, washed to remove nonadherent virus, and maintained as a suspension culture18. Cells were harvested 2 hours after the incubation with parvovirus B19 or 48 hours later for the assay of B19 DNA production. Total DNA was extracted from cells (similar quantities were obtained from all extracts), and serial dilutions were quantitated by dot blot hybridization with a 32P-labeled pYT103 insert probe, a full-length clone of parvovirus B1919.
Cultured cells were also assayed for B19 DNA and RNA by in situ hybridization or for B19 proteins by an immunofluorescence assay. Cells were infected with parvovirus B19 as described above and harvested five days later. In situ hybridization was performed by a modification of previous methods20. Cells were fixed with paraformaldehyde, made permeable with Triton X-100 (0.01 percent), digested with proteinase K (100 µg per milliliter of solution), and refixed in paraformaldehyde. Hybridization was performed with a 367-nucleotide, digoxigenin-labeled probe to the coding region for the B19 nonstructural protein (nucleotides 1908 to 227419); the probe was generated by the polymerase chain reaction.
An immunofluorescence assay for B19 proteins was performed with human anti-B19 polyclonal serum as previously described, followed by fluorescein isothiocyanate-conjugated antiserum and counterstaining with Evans blue21.
Statistical Analysis
Epidemiologic data were analyzed with Fisher's exact test. Colony-culture data were analyzed with Student's t-test.
Results
B19 Seroprevalence in Populations with the P and p Phenotypes
The p and P1k phenotypes are very rare; the frequency of the p phenotype is only 1 in 200,00022. The Swartzentruber Amish have a higher prevalence of the p phenotype than the general population23. There is a large population of Amish in Ohio, where we performed two cross-sectional epidemiologic studies, one based on samples collected from blood-bank donors and a second, confirmatory clinic-based study.
Plasma samples from 11 blood donors with phenotype p had been collected in a Cleveland blood bank over a 15-year period, and in the first study these samples were assayed for evidence of previous parvovirus infection. Plasma samples from 75 persons who had P antigen on their red cells (either the P1 or the P2 phenotype) were used as controls: 53 of these people (71 percent) had detectable IgG antibodies specific for parvovirus B19, a rate similar to the seroprevalence elsewhere in the United States24. In contrast, none of the 11 blood donors with the p phenotype had antibody to parvovirus B19 (P<0.001).
A second, confirmatory study was undertaken to exclude possible bias due to the use of stored plasma or the recent introduction of the virus to this community. Blood samples from all Swartzentruber Amish patients undergoing phlebotomy for other clinical infections were tested for red-cell type and viral serologic status. Blood was collected from 74 patients. Among those with the P1 or the P2 phenotype, 32 of 68 (47 percent) had B19 IgG antibodies, but none of the 6 persons with the p phenotype had antibody to parvovirus B19 (P = 0.03). Among persons 15 or more years of age, 24 of 40 with the P1 or the P2 phenotype (60 percent) had antibody to B19, as compared with none of 3 with the p phenotype (P = 0.08).
These two epidemiologic studies showed that no patient with the blood-group p phenotype had evidence of previous infection with parvovirus B19, despite normal seroprevalence levels of B19 in the rest of the community.
Bone Marrow Infection of p and P Erythroid Cells
In a consistent, dose-dependent fashion, parvovirus B19 inhibits erythroid colony formation in cultures of normal bone marrow2. We tested the effects of the virus on bone marrow from two donors with the p phenotype, neither of whom was from the Cleveland Amish community (both were negative for antibodies to parvovirus B19). Data on the first donor and two control subjects with the P phenotype are shown in Figure 1. As expected, erythroid colony formation was completely inhibited by parvovirus B19 in concurrent cultures of normal bone marrow from the control subjects. When bone marrow cells from donors with the p phenotype were used, erythroid colony formation was unaffected even at very high concentrations of parvovirus B19, indicating that erythroid progenitor cells that did not express P antigen were not susceptible to the cytotoxic effect of parvovirus B19. The results were similar for the second donor with the p phenotype (data not shown). In this subject, infection with 1013 particles of parvovirus B19 had no effect on the mean (±SE) number of erythroid colonies formed per plate (bone marrow alone, 331 ±13; with parvovirus B19 added, 427 ±22). In contrast, infection of bone marrow from a donor with the P phenotype completely inhibited colony formation (colonies per plate with bone marrow alone, 536 ±3; with parvovirus B19 added, 0).
|
|
|
Red cells are known to contain more than 400 antigens, many being membrane proteins or oligosaccharide structures classified into families known as blood groups25. Despite their clinical importance in blood banking, the function of most of these structures is obscure. Some blood-group antigens, including the members of the P blood system, are attachment sites for pathogens, especially bacteria. Specifically, uropathogenic Escherichia coli have been shown to bind to the Pk antigen, and persons with the P1k phenotype are at increased risk of urinary tract infections and pyelonephritis26. The human malarial parasite Plasmodium vivax has been shown to bind to Duffy-positive red cells, and Duffy-negative persons are relatively resistant to infection with P. vivax27. Blood-group heterogeneity may have evolved to protect human populations against viral, bacterial, and protozoal infections.
The facts that persons with the p phenotype showed no evidence of prior infection with parvovirus B19 and that bone marrow from donors with this phenotype could not be infected in vitro with the virus confirm our recent identification of the receptor for parvovirus B19 as globoside or P antigen6. Ideally, we would also have liked to test people with the P1k phenotype for susceptibility to parvovirus B19 infection, because the genetic defect in these donors is more specific than that in persons with the p phenotype: P1k red cells do not have globoside (P), whereas p erythrocytes do not have either globoside or ceramide trihexoside (Pk) (Table 1). Unfortunately, the P1k phenotype is exceedingly rare, and we have been able to locate only a single P1k blood donor in the United States, who was unwilling to donate bone marrow. However, we have previously shown that parvovirus B19 does not bind to Pk (or to P1) cells6.
Strains of mice (i.e., SJL/J) have been described that do not have the receptor for mouse hepatitis virus and are highly resistant to infection28,29. Persons with the p phenotype are examples of a population that is entirely resistant to a human virus infection because the cellular receptor is absent. These results also suggest that preventing the binding of parvovirus B19 to its receptor or that of other animal parvoviruses to their respective receptors may serve as effective therapy.
We are indebted to Dr. S.G. Sandler and the American Red Cross for their assistance in finding suitable bone marrow donors; to Dr. B.J. Cohen, Virus Reference Division, Public Health Laboratory Service, London, for B19 serologic assays; and to Ms. J. Proctor, Blood Transfusion Medicine, National Institutes of Health, for red-cell typing.
Source Information
From the Hematology Branch, National Heart, Lung, and Blood Institute, Bethesda, Md. (K.E.B., J.R.H., G.G., S.M.A., N.S.Y.); Mount Eaton Clinic, Mount Eaton, Ohio (E.D.L.); and the American Red Cross, Cleveland (P.M.).
Address reprint requests to Dr. Brown at Bldg. 10, Rm. 7C218, National Institutes of Health, 9000 Rockville Pike, Bethesda, MD 20892.
References
| |||||||||||||||||||||||||||||||
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
