X-Linked Agammaglobulinemia Patients Are Not Infected with Epstein-Barr Virus: Implications for the Biology of the Virus

Abstract

Epstein-Barr virus (EBV) infects both B lymphocytes and squamous epithelial cells in vitro, but the cell type(s) required to establish primary and persistent infection in vivo has not been definitively elucidated. The aim of this study was to investigate a group of individuals who lack mature B lymphocytes due to the rare heritable disorder X-linked agammaglobulinemia in order to determine the role of the B cell in the infection process. The results show that none of these individuals harbored EBV in their blood or throat washings. Furthermore, no EBV-specific memory cytotoxic T lymphocytes were found, suggesting that they had not undergone infection in the past. In contrast, 50% of individuals were found to carry human herpesvirus 6, showing that they are infectible by another lymphotropic herpesvirus. These results add weight to the theory that B lymphocytes, and not oropharyngeal epithelial cells, may be required for primary infection with EBV.

Epstein-Barr virus (EBV) is a human herpesvirus carried by more than 90% of the world adult population as a lifelong, persistent infection (32). Primary infection usually occurs during childhood when it is generally asymptomatic; however, if primary infection is delayed until adolescence or early adulthood, infectious mononucleosis (IM) occurs in around 50% of individuals (16).

EBV transmission is almost exclusively via the salivary route, and virus can be recovered from throat washings of most seropositive individuals and IM patients (11). It is thought that, during primary infection, virus enters the body through initial infection of the oropharyngeal epithelium and thus infects B lymphocytes circulating through the lymphoid tissue at this site and subsequently disseminates and establishes persistence in the mature B-cell population (reviewed in reference 44). There is convincing evidence to suggest that B lymphocytes are the site of EBV persistence. For example, it has been shown in IM patients and individuals suffering from herpes zoster that EBV B-cell infection is maintained during acyclovir therapy, while virus shedding into saliva is almost completely abrogated (56, 57). Similarly, persistent infection was entirely lost by obliteration of B cells in an EBV-positive individual prior to bone marrow transplant from an EBV-negative donor. The recipient remained free from infection until more than 3 years posttransplant and subsequently acquired an EBV strain different from that of the original infection (14).

After primary infection, EBV can be detected in the peripheral blood of healthy seropositive individuals. Spontaneous outgrowth of virus carrying B cells from peripheral blood gives rise to B lymphoblastoid cell lines (LCLs) in culture (37) and PCR analysis of peripheral blood cell DNA can detect viral genomes. By these two methods, EBV-infected circulating B lymphocytes are detected at a frequency of approximately 1 to 60 per million (24, 31).

In healthy seropositive individuals, EBV-specific cytotoxic T lymphocytes (CTLs) to a wide range of EBV epitopes are readily detectable in the circulation (reviewed in reference 45). It is assumed that, in order to persist for the lifetime of the host, the virus must evade immune recognition by establishing a form of latent infection with minimal expression of viral proteins (reviewed in reference 54). However, for EBV to be successfully transmitted, it must effect egress from the host. Thus, a lytic infection with the production of cell-free virus particles must occur at or near an epithelial surface. This is thought to occur via a reversal of the route of viral entry, involving the reinfection of epithelial cells by virus carrying B lymphocytes circulating through the lymphoid tissue of the oropharynx. Virus replication is postulated to occur in the infected epithelial cells, with infectious virus particles released into the saliva, and, from this site of production, the virus spreads to other individuals (reviewed in reference 44). The mechanism of passage of virus from lymphocytes to epithelial cells is unclear because until recently, there was little evidence of lytic replication in B cells in vivo, although productive infection can be induced in LCLs in vitro (reviewed in reference 44). Support for this route of infection and transmission came from experiments showing that the virus can infect human cervical epithelial cells in vitro (51) and an in vivo demonstration that exfoliated oropharyngeal cells of IM patients, thought to be epithelial, contained replicating virus (25, 50).

Epithelial cell infection is also seen in vivo in the cells of the benign, EBV-associated lesion, oral hairy leukoplakia (OHL). Since this lesion mainly occurs in immunocompromised human immunodeficiency virus-infected individuals (reviewed in reference 15), where reactivation of persistent virus infections is common, it has been considered to represent an amplification of normal events. However, studies on OHL showing that infected epithelial cells are situated in the superficial layers of the epithelium and that no viral DNA or protein expression is detectable in the basal layer have suggested that this lesion is perhaps caused by repeated infection of mature epithelial cells by saliva-borne virus (36, 49; reviewed in reference 15).

Extensive research on naso- and oropharyngeal tissues from both IM and healthy donors has failed to identify any infected epithelial cells, although both latently (19, 34, 35) and lytically (33, 53) infected B lymphocytes are seen in the underlying lymphoid tissues, infiltrating the epithelial layers and also on the surface of the tonsillar epithelium (2). These findings prompted some workers to question the infection of epithelial cells in vivo (34, 53) and to suggest that the whole of the EBV life cycle, including entry to and egress from the host, can be accomplished by B-cell infection alone (19, 33, 54). This had been thought unlikely as there was little in vivo evidence of productive infection of B lymphocytes. However, it has recently been shown that B lymphocytes are capable of supporting lytic EBV replication in vivo both in peripheral blood (12, 41) and in those lymphocytes infiltrating the oropharynx (2, 33, 53). Thus, it no longer seems impossible that the B lymphocyte is the cell type exclusively exploited by EBV for primary and persistent infection of humans.

In light of the ongoing controversy regarding the role of oropharyngeal epithelial cells and B lymphocytes in EBV biology, our aim was to reassess the situation using a model where infection of B lymphocytes cannot occur. We have therefore looked for evidence of EBV infection in individuals with X-linked (Bruton’s) agammaglobulinemia (XLA). These individuals possess inherited mutations in the gene encoding Bruton tyrosine kinase (Btk). Defects in this cytoplasmic tyrosine kinase, involved in cell signalling, result in a block to B-cell development and consequently a lack of mature circulating B lymphocytes (reviewed in reference 38). The absence of B lymphocytes results in agammaglobulinemia. XLA patients thus have increased susceptibility to opportunistic bacterial infections, although their T-cell immunity and resistance to viral infections is not affected (reviewed in reference 38).

We have looked for evidence of EBV infection in these “human B-cell knockouts” to determine the fate of the virus in the absence of B lymphocytes, but where oropharyngeal epithelium is intact.

MATERIALS AND METHODS

XLA patients.

Whole-blood and throat washing samples were collected from individuals suffering from XLA, a rare primary immunodeficiency. A group of six male patients were studied, drawn from the south of England, who attended a clinic at the John Radcliffe Hospital, Oxford, United Kingdom, every 2 to 4 months. All had increased susceptibility to recurrent bacterial infections presenting in childhood between 2 and 7 years of age. Each received immunoglobulin replacement therapy from the time of their original diagnosis. Four patients first received intramuscular immunoglobulin which gave low serum immunoglobulin levels of 2 to 4 g/liter and were switched to intravenous therapy (serum immunoglobulin levels of 6 to 10 g/liter) at 11 to 39 years of age. All patients now receive intravenous infusions of pooled donor immunoglobulin two or three times a week. These infusions are generally self-administered at home. The ages and samples obtained from each patient are shown in Table 1.

TABLE 1.

XLA patient ages and summary of samples obtained from each patient over 18 months

XLA patient	Agea (yr)	Amt (ml) of whole blood	No. of throat washing samples
			Cell pellet	Supernatant
XLA1	13	25, 40	3	2
XLA2	34	35, 16	2	2
XLA3	23	25, 23, 10	3	3
XLA4	22	25, 40, 7.5	2	3
XLA5	42	35, 25, 7.5	3	2
XLA6	10	7.5, 8, 35, 20, 9	3	1

^aAge at time of collection of the first sample. 

B- and T-cell immunophenotyping.

Peripheral blood T-cell (CD3⁺, CD4⁺/CD8⁺) and B-cell (CD19,20⁺) numbers were enumerated by dual-color flow cytometry using a Becton-Dickinson (Oxford, United Kingdom) FACS. Monoclonal antibodies used were fluorescein isothiocyanate anti-CD3, phycoerythrin-labelled anti-CD4, anti-CD8, anti-CD19, and anti-CD20. All were obtained from Becton-Dickinson, and were used at standard dilutions according to the manufacturer’s instructions. EDTA-anticoagulated samples were analyzed and stained within 3 h of collection.

Indirect immunofluorescence staining.

Serological testing for immunoglobulin G (IgG) antibodies to EBV viral capsid antigen (VCA) was performed on four XLA patients and all healthy donors by routine methods (17). XLA patient sera were shown to contain infused antibodies to VCA at titers of no significant difference to those found in healthy seropositive donors (P > 0.2).

All XLA patients’ throat washing supernatants were also tested by the same method for the presence of anti-VCA IgG.

Maintenance of cell lines.

Control LCLs, B95-8 (American Type Culture Collection [ATCC], Rockville Md.) (29) and MLA-144 (ATCC) cell lines were maintained in standard culture medium containing RPMI 1640 supplemented with 2 mM l-glutamine, 100 IU of penicillin per ml, 100 μg of streptomycin per ml and 10% heat-inactivated fetal calf serum (FCS) (Gibco BRL, Middlesex, United Kingdom).

Isolation of PBMC for tissue culture and PCR.

Whole blood that had been treated with heparin was diluted 1:1 in RPMI 1640 containing 2% FCS (2% FCS/RPMI 1640) and layered onto an equal volume of Ficoll-Hypaque (Pharmacia Biotech, Uppsala, Sweden). Following centrifugation at 840 × g for 15 min, peripheral blood mononuclear cells (PBMC) were harvested from the plasma-Ficoll interface. Cells for tissue culture were washed twice in 2% FCS/RPMI 1640, counted, and either used immediately or suspended in 25% FCS/RPMI 1640 containing 10% dimethyl sulfoxide, frozen at −70°C and subsequently stored in liquid nitrogen until use. PBMC for PCR analysis were washed twice in phosphate-buffered saline and either used immediately for DNA extraction or quick-frozen and stored at −70°C for future use.

E-rosetting.

T-cell depletion of PBMC samples was achieved as previously described (23) by rosetting with 2-aminoethylisothiouronium bromide-treated sheep erythrocytes followed by centrifugation over a Ficoll-Hypaque gradient allowing separation of T cells (E+) from the residual mononuclear cell population. The B-cell-enriched fraction (E−) were harvested from the Ficoll-medium interface, washed in 2% FCS/RPMI 1640 and cultured immediately.

Generation of LCLs by infection of E− cells with EBV.

Concentrated preparations of the EBV strain B95-8 were prepared as described previously (23) and used to infect E− cells from all XLA patients and six healthy donors. Up to 2 × 10⁷ E− cells were incubated with virus preparation diluted 1:10 in 1 ml of standard culture medium for 1 h at 37°C and 5% CO₂. Cells were then suspended at 10⁶ per ml in standard culture medium and plated out at 2 ml per well in 24-well plates. Cultures were incubated at 37°C and 5% CO₂ and maintained by replacing half the volume of medium each week. Cells were monitored regularly for signs of immortalization until outgrowth occurred or cell viability was 0% (as determined by trypan blue staining) (1 to 6 weeks).

Spontaneous outgrowth assays.

Between 5 × 10⁵ and 2 × 10⁷ E− cells were plated at 10⁶ per ml in standard culture medium in 96-well U-bottom plates at 200 μl per well. These cells were then incubated at 37°C and 5% CO₂ and monitored regularly for signs of immortalization. Cultures were maintained by replacing half the volume of medium each week until either outgrowth occurred, or cell viability was 0% as determined by trypan blue staining (2 to 9 weeks).

Immunostaining.

LCLs obtained by either in vitro immortalization or spontaneous outgrowth were routinely screened for presence of viral nuclear antigens (EBNAs) by standard anticomplement immunofluorescence methods (42).

Preparation of throat washing specimens.

Throat washings were collected in 10-ml aliquots of isotonic (0.9%) saline. Samples were centrifuged at 600 × g for 10 min in order to separate cellular material from supernatant. Cellular material was quick-frozen and stored at −70°C until use or immediately used for DNA extraction as described blow. Supernatants were concentrated 100 times by ultracentrifugation at 25,000 × g for 2 h at 4°C using a Beckman L8-80 ultracentrifuge and a type 50 rotor (Beckman Instruments, Palo Alto, Calif.). Control tubes containing isotonic saline only were included in every centrifuge run to ensure that there was no contamination of samples by extraneous EBV DNA. The pellet was resuspended in 40 μl of sterile distilled water and incubated at 37°C for 1 h with 100 μg of proteinase K (Promega Corporation, Southampton, United Kingdom) per ml. The enzyme was then inactivated at 100°C for 10 min.

Extraction of DNA for PCR.

DNA was extracted from fresh or frozen PBMC, throat washing cell pellets, or control cell line samples by using the Easy-DNA kit according to the manufacturer’s instructions (Invitrogen, BV Leek, The Netherlands). DNA was dissolved in sterile distilled water, and the total DNA yield was established by spectrophotometry (GeneQuant II; Pharmacia Biotech). One microgram of DNA was taken to represent the yield from 1.5 × 10⁵ cells based on estimates that the human genome is 3 × 10⁹ bp long and that the average molecular mass of a nucleotide base is 330 Da.

Throat washing cell pellet DNA was also extracted using the Invitrogen Easy-DNA kit, but DNA was resuspended in a standard 50-μl volume of sterile distilled water and the total DNA yield was not established.

EBV W-repeat PCR.

The EBV W-repeat PCR detects the BamHI-W-repeat sequences of the EBV genome. The primers amplify a 298-bp fragment of the W repeats. 5′ primer; 5′-CTTTAGAGGCGAATGGGCGC-3′ (positions 14069 to 14088 inclusive). 3′ primer; 5′-AGGACCACTTTATACCAGGG-3′ (positions 14366 to 14347 inclusive). Primer and probe coordinates for EBV were based on the published prototype DNA sequence (3) (available from the GenBank nucleotide sequence database under accession number V01555).

PCR was performed in a final volume of 100 μl with either 1 μg of PBMC DNA, 10 μl of throat washing cell pellet DNA solution, or 10 μl of concentrated throat washing supernatant per reaction. LCL or B95-8 DNA (1 μg) was included as a positive control, and 1 μg of PBMC DNA from a seronegative donor was included as a negative control. Reagent-only reactions were also included to control for DNA contamination.

Samples were denatured at 95°C prior to PCR cycling. The amplification mixture contained the following: 1× PCR buffer (50 mM KCl, 10 mM Tris-HCl [pH 9.0] and 0.1% Triton X-100) (Promega) plus 1.5 mM MgCl₂ (Promega), 0.2 mM each of the four deoxynucleoside triphosphates (Pharmacia), 1 μM (each) primer (Oswel DNA Service, University of Southampton, Southampton, United Kingdom), 2.5 U (0.5 μl) of Taq polymerase (Promega) per reaction. Reactions were overlaid with mineral oil (Sigma Chemical, Poole, Dorset, United Kingdom). The reaction mixtures were incubated in an Omnigene thermal cycler (Hybaid, Middlesex, United Kingdom) under the following cycling conditions: 35 cycles of PCR, with 1 cycle consisting of 1 min at 94°C, 2 min at 45°C, and 2 min at 72°C.

To determine whether samples contained DNA suitable for PCR amplification and to test for false negatives, all samples were also screened for the human β-globin gene (48).

In order to determine the sensitivity of the W-repeat PCR, a plasmid, pBSW, containing a single copy of a W-repeat sequence, was diluted into human PBMC DNA from a seronegative donor. Having established accurate concentrations of a plasmid solution in sterile distilled water by spectrophotometry, the copy number per microliter was calculated based on the fact that pBSW is 6.03 × 10³ bp long and the estimate that the average mass of a nucleotide base is 330 Da. Serial dilutions were made of pBSW and between 0 and 10⁶ copies in a background of DNA from 5 × 10⁵ PBMC from a seronegative donor were used in W-repeat PCRs. The results of one such experiment are shown in Fig. 1A. The W-repeat PCR was consistently found to be capable of detecting one copy of the W fragment in a background of DNA from 5 × 10⁵ cells.

FIG. 1.

EBV DNA detection by PCR in throat washings and PBMC of healthy EBV seropositive donors. (A) W-repeat PCR sensitivity. PCR products obtained from a dilution of plasmid pBSW in a background of DNA from 5 × 10⁵ PBMC from seronegative donor NHD14 were analyzed by Southern blotting and hybridization. Lanes 1 and 2, PCR reagent controls containing water instead of template DNA; lanes 3 to 11, results obtained from PCR on plasmid dilutions ranging from 0 to 10⁶ starting copies per PCR (pBSW template copy number per PCR is indicated below each lane); lane 12, product from 0.1 μg of DNA obtained from the EBV-positive cell line B95-8. (B) PCR products obtained from concentrated throat washing supernatants of healthy donors. Lanes 1 and 2, PCR mix controls containing water instead of template DNA; lane 3, product obtained from 1 μg of PBMC DNA from seronegative donor NHD14; lanes 4 and 5, products from concentrated throat washing samples from seropositive donor NHD3; lanes 6 to 8, products from concentrated throat washing samples from seropositive donor NHD13; lanes 9 to 11, reactions using saline control included in the ultracentrifugation step as described in Materials and Methods; lane 12, PCR mix control containing water instead of template DNA; lane 13, product from 1 μg of EBV LCL DNA. (C) PCR screening of healthy seropositive donor PBMC DNA. Lanes 1 and 2, PCR mix controls containing water instead of template DNA; lane 3, reaction using 1 μg of PBMC DNA from seronegative donor NHD14; lanes 4 to 11, PCR products from eight replicate reactions, each done on 1 μg of DNA obtained from PBMC of seropositive donor NHD6; lane 12, PCR mix control containing water instead of template DNA. Lane 13, product from 1 μg of DNA obtained from an EBV LCL.

HHV-6 PCR.

DNA from PBMC was screened for human herpesvirus 6 (HHV-6) by PCR (13). In our hands, this had previously been shown capable of detecting HHV-6 in concentrated sera obtained from healthy donors and heart-lung transplant recipients (unpublished data). Briefly, a 233-bp product is generated using the primer set 5′-AAGCTTGCACAATGCCAAAAAACAG and 3′-CTAATCCCCAGAGCCGTATGAGCTC in a 100-μl reaction mixture using the same reaction conditions used in the EBV W-repeat PCR and the following cycling conditions: 40 cycles, with 1 cycle consisting of 1 min at 94°C, 1 min at 65°C, and 2 min at 72°C. Supernatant from the T-cell line J-jhan infected with HHV-6 was used as a positive control. DNA extracted from B95-8 was used as a negative control. Reagent-only reactions were also included to control for DNA contamination.

Gel electrophoresis and Southern blotting.

PCR products were separated by gel electrophoresis using 2.5% agarose (type II-A; Sigma) containing 1 μg of ethidium bromide (Sigma) per ml. The molecular weight marker φX174 digested with HinfI (Promega) end labelled with [γ-³²P]dATP (Amersham International, Little Chalfont, Buckinghamshire, United Kingdom) was always included. Following electrophoresis, gels were transferred and UV cross-linked to positively charged nylon membranes (Hybond-N+; Amersham) according to the manufacturer’s instructions for Southern blotting.

Blots were then prehybridized at 42°C for 2 h in a solution containing 50% deionized formamide, 5× Denhardt’s solution, 5× SSC (1× SSC is 0.15 M NaCl plus 0.015M sodium citrate), 0.5% sodium dodecyl sulfate (SDS), and 100 μg of denatured sheared salmon sperm DNA (CP Labs, Bishops Stortford, Hertfordshire, United Kingdom) per ml. DNA hybridization oligonucleotide probes used were as follows: for EBV W-repeat PCR products, 5′-TGACTTCACCAAAGGTCAGG-3′ (14227 to 14246 inclusive). For HHV-6 PCR products, 5′-AACTGTCTGACTGGCAAAAACTTTT-3′. (Oswel).

Probes were labelled with [γ-³²P]dATP (Amersham) using T4 polynucleotide kinase (Promega) by incubating 5 pmol of probe and 1.85 MBq of [γ-³²P]dATP for 30 min at 37°C with 10 U of enzyme in a buffer containing 70 mM Tris-HCl (pH 7.6), 10 mM MgCl₂, and 5 mM dithiothreitol (Promega). Labelled probes were purified by fractionation on a Sephadex column (NICK columns; Pharmacia) according to the manufacturer’s instructions. Labelled probes were then applied to the prehybridized membranes with the addition of a further 100 μg of denatured sheared salmon sperm DNA per ml and left to hybridize at 42°C for at least 12 h. Hybridized blots were washed in 2× SSC–0.1% SDS and 0.2× SSC–0.1% SDS and then exposed to Hyperfilm-MP (Amersham) for 7 days at −70°C.

HLA genotyping.

DNA was extracted from EDTA-anticoagulated peripheral blood by a rapid salting-out method (30). Genotyping of the HLA class I A and B loci was performed by using PCR and sequence-specific primers as previously described (6).

Bulk CTL reactivation and limiting dilution analysis (LDA).

On occasion, cells were stored frozen prior to these assays. Previous experience of the authors (S.R.B., R.K., and D.J.M.) and others (52) has shown that these cells respond normally after thawing.

Responder cells were either fresh PBMC (healthy donors), thawed PBMC, and/or T-cell-enriched E+ cells (XLA patients 1 and 2) (Table 2).

TABLE 2.

Summary of CTL experiments performed^ac

Patient or donorb (HLA type)	Responders:stimulators		Target cell type
	Bulk CTL reactivation	LDA
XLA1 (HLA-B8)	E+:E+ restimulated with allogenic HLA-matched LCL	E+:BL30/EBO-pLPP-Sig-FLRG (allogeneic)	Allogeneic HLA-matched PHA blasts
XLA2 (HLA-B44)	E+:E+ and PBMC:PBMC both restimulated with PBMC	PBMC:PBMC	PHA blasts
XLA3 (HLA-B44)	PBMC:PBMC restimulated with PBMC	ND	PHA blasts
XLA4 (HLA-B7)	ND	PBMC:PBMC	PHA blasts
XLA5 (HLA-B44)	PBMC:PBMC restimulated with PBMC	ND	PHA blasts
XLA6 (HLA-B44)	ND	E+:PBMC	PHA blasts
NHD 19 (HLA-B8)	PBMC:PBMC restimulated with allogenic HLA-matched LCL	PBMC:BL30/EBO-pLPP-Sig-FLRG (allogeneic)	Allogenic HLA matched PHA blasts.

^aResponder, stimulator, and target cell types for bulk CTL reactivations and LDA. 

^bXLA, XLA patient; NHD, normal healthy donor. 

^cAll cells autologous unless otherwise stated. E+, T-cell-enriched PBMC. ND, not done. 

Stimulator cells were either autologous fresh PBMC (healthy donors), thawed PBMC, and/or T-cell-enriched E+ cells (XLA patients 1 and 2) (Table 2). Stimulator cell pellets were resuspended in 100-μg/ml solutions of synthetic peptide diluted in RPMI 1640 and incubated at 37°C and 5% CO₂ for 45 min. Peptide-treated cells were then washed twice in RPMI 1640. Peptides were purchased from Chiron Mimotopes (Melbourne, Australia).

For bulk CTL reactivation, responder cells were cultured with autologous peptide-treated stimulator cells at a ratio of 3:1 (1.5 × 10⁶ to 0.5 × 10⁶) at 10⁶ per ml in standard culture medium, plated out at 2 ml per well in 24-well plates, and incubated at 37°C and 5% CO₂. After 7 days, responders were restimulated with a further 0.5 × 10⁶ autologous peptide-treated stimulator cells. In one case, 10⁵ gamma-irradiated (80 Gy) allogeneic HLA-matched LCL cells were used (Table 2). Cultures were expanded by addition of more standard culture medium when required. Cytotoxicity assay by chromium release was performed on day 10 or 12 of culture.

For LDA, responder cells were plated out in U-bottom 96-well plates at the following cell numbers per well in 24 replicates: 50 × 10³, 25 × 10³, 12.5 × 10³, and 6.25 × 10³. Autologous peptide-treated stimulator cells, prepared as described for bulk CTL reactivation, were gamma irradiated (20 Gy) and cultured with responders at 5 × 10⁴ per well. In one case, 10⁴ allogeneic gamma-irradiated (20 Gy) BL30/EBO-pLPP-Sig-FLRG cells were used as stimulators. The BL30/EBO-pLPP-Sig-FLRG line is a Burkitt’s lymphoma (BL) cell line transfected with an expression plasmid (EBO-pLPP-Sig-FLRG) encoding an endoplasmic reticulum translocation signal followed by the epitope FLRGRAYGL (EBNA-3A). This cell line has been shown previously to be a very efficient stimulator of memory T cells specific for FLRGRAYGL from healthy seropositives (20). The final culture volume was 100 μl in standard culture medium. On days 3 and 7, cultures were fed with 50 and 100 μl, respectively, of medium containing 10% FCS/RPMI 1640 supplemented with 2 mM l-glutamine, 100 IU of penicillin per ml, 100 μg of streptomycin per ml, 30% MLA-144 (ATCC) conditioned medium (filtered and heat inactivated), and 20 U of recombinant interleukin-2 per ml (47, 55) (referred to as T-cell medium). Where necessary, 100 μl of T-cell medium was replaced on day 10 or 11. Cytotoxicity assay by chromium release was performed between days 10 and 14.

Generation of PHA blasts.

Phytohemagglutinin (PHA) blasts were generated from fresh PBMC (healthy donors), thawed PBMC or E+ (XLA patients) by culturing cells in standard culture medium containing 20 μg of PHA (CSL, Melbourne, Australia) per ml at 5 × 10⁵ per ml. On day 3, cells were harvested, washed in RPMI 1640, and resuspended in T-cell medium at the same concentration. Cultures were expanded by doubling the volume of medium twice weekly.

Chromium release assay.

Cytotoxic activity of responder cells was assessed by standard chromium release assay. Target cells were autologous PHA blasts or allogeneic HLA-matched PHA blasts treated with peptide as described above (Table 2). Both peptide-treated and untreated targets were labelled with 3.7 MBq of ⁵¹Cr (sodium chromate; Amersham) at 37°C and 5% CO₂ for 90 min and washed twice in standard culture medium.

Bulk reactivation CTLs were harvested, their viability was assessed by trypan blue staining, and the CTLs were applied to targets at effector-to-target ratio of 20:1. Either two or three replicates were performed for each target cell type.

LDA effectors (50 μl) from each well were applied to 10⁴ cells of each target cell type.

Spontaneous release (targets plus standard culture medium only) and maximum release (targets plus medium containing 0.1% SDS) controls were included for each target cell type. Effectors and targets were incubated for 5 h at 37°C and 5% CO₂, and supernatants were then assessed for ⁵¹Cr levels by beta-scintillation counter (Topcount Microplate; Packard Instrument Co., Meriden, Conn.). The release of ⁵¹Cr was used to calculate the specific lysis of targets by effectors.

Statistical analyses.

The two-tailed Fisher exact χ² test and Student’s t test were used.

RESULTS

EBV cannot be detected in the throat washings of XLA patients.

In order to establish whether XLA patients were harboring EBV in the oropharynx, throat washing samples which had been separated into cell pellet and supernatant components were screened for EBV DNA using primers directed against the BamHI-W-repeat sequence. Cell pellet DNA was extracted prior to screening. Supernatants were concentrated by ultracentrifugation. Each ultracentrifugation run included a saline control in order to ensure that there was no potential contamination of samples by extraneous EBV DNA during this step; these samples always tested negative for EBV by PCR. Figure 1A shows the results obtained from an experiment carried out to assess the sensitivity of the PCR. The technique could consistently detect one copy of target DNA in a background of 5 × 10⁵ EBV-negative cells.

No EBV-positive throat washing samples were found for any XLA patient. This is in contrast with the 14 of 20 positive samples obtained from 10 healthy seropositive donors (Table 3). In total, none of the six XLA patients compared to 8 of 10 healthy seropositive donors were found to have EBV DNA in their throat washing samples (P < 0.01). Figure 1B shows an example of positive PCR results obtained with throat washing samples from healthy seropositive donors.

TABLE 3.

Results obtained from PCR screening of throat washing samples and PBMC DNA from XLA patients and healthy donors

Patient or donora	No. of PBMC screened (no. of reactions)b	PBMC screening resultc	Throat washing samples screened (no. positive/total no.)b
			Cell pellet	Supernatant
XLA1	1.0 × 107 (67)	−	0/3	0/2
XLA2	1.5 × 106 (10)	−	0/2	0/2
XLA3	4.0 × 106 (27) 4.5 × 105 (30)	−	0/3	0/3
XLA4	3.1 × 106 (21) 1.2 × 106 (8)	−	0/2	0/3
XLA5	5.4 × 106 (36) 2.7 × 106 (18)	−	0/3	0/2
XLA6	1.0 × 107 (67)	−	0/3	0/1
NHD1	6.3 × 106 (42)	+	ND	ND
NHD2	6.3 × 106 (42)	−	1/1	1/1
NHD3	ND	ND	0/1	1/1
NHD4	6.3 × 106 (42)	−	1/1	0/1
NHD5	6.3 × 106 (42)	−	ND	ND
NHD6	6.3 × 106 (42)	+	1/1	1/1
NHD7	6.3 × 106 (42)	+	0/1	0/1
NHD8	ND	ND	2/2	1/1
NHD9	ND	ND	0/1	1/1
NHD10	ND	ND	1/1	1/1
NHD11	6.3 × 106 (42)	+	ND	ND
NHD12	6.3 × 106 (42)	−	ND	0/1
NHD13	ND	ND	1/1	1/1
NHD14 (seronegative)	6.3 × 106 (42)	−	ND	ND

^aXLA, XLA patient; NHD, normal healthy donor. 

^bND, not done. 

^c−, negative result; +, positive result; ND, not done. 

XLA patients have normal T-cell counts, but no B cells infectible by EBV in the peripheral blood.

Flow cytometric analysis performed on PBMC obtained from each patient confirmed that they had no peripheral blood B lymphocytes, demonstrated by CD19 and CD20 staining (Table 4). Analysis for CD3, CD4, and CD8 however, showed that the XLA patients had T-lymphocyte populations within normal ranges (Table 4).

TABLE 4.

FACS analysis of PBMC from XLA patients

XLA patient	Cell count (% of total PBMC)
	CD3+	CD4+	CD8+	CD19+	CD20+
XLA1	88	61	26	0	0
XLA2	83	59	23	0	0
XLA3	90	60	29	0	0
XLA4	89	65	24	0	0
XLA5	72	40	30	0	0
XLA6	92	48	44	0	0

In addition, whereas EBV infection of T-cell-depleted PBMC gave rise to LCLs in 100% (six of six) healthy donors, no immortalized cell lines were obtained from any of the XLA patients, despite several attempts (P < 0.01) (Table 5). These results confirm that the XLA patients investigated in this study had no EBV-infectible cells in their peripheral blood.

TABLE 5.

Results obtained from immortalization and spontaneous outgrowth assays performed on XLA patients and healthy donors

Patient or donora	Resultb of assay
	Immortalization	Spontaneous outgrowth
XLA1	−	−
XLA2	−	−
XLA3	−	−
XLA4	−	−
XLA5	−	−
XLA6	−	−
NHD2	+	−
NHD5	+	+
NHD6	+	+
NHD8	+	−
NHD11	+	+
NHD21c	+	ND
NHD1, -3, -4, -7, -9, and -10	ND	−

^aXLA, XLA patient; NHD, normal healthy donor. 

^b−, negative result; +, positive result; ND, not done. 

^cSeronegative. 

XLA patients have no EBV in their peripheral blood.

The presence of EBV in the peripheral blood of the XLA patients and healthy donors was assessed by both culture of T-cell-depleted lymphocytes (E−) (spontaneous outgrowth assay) and PCR analysis of PBMC DNA.

(i) Spontaneous outgrowth assay.

Between 5 × 10⁵ and 2 × 10⁷ E− lymphocytes obtained from PBMC were cultured at 10⁶ per ml for between 2 and 9 weeks until cell viability was 0% or spontaneous outgrowth of an LCL occurred. No LCLs were generated from XLA patients by this method. This is in contrast to the results obtained with healthy seropositive donors, where 3 of 11 donors produced a spontaneous LCL (Table 5).

Since the spontaneous outgrowth assay depends on the presence of B cells in the culture to rescue virus particles released by infected cells (43), it is possible that virus was carried and released by non-B cells in the XLA culture and not detected in this assay. In addition, the method is not highly sensitive and gave positive results in only a small number of healthy EBV-infected donors (27%). PCR screening was therefore also employed.

(ii) PCR analysis.

BamHI-W PCR was performed on DNA extracted from approximately 1.5 × 10⁵ PBMC per reaction (1 μg). Replicate reactions were undertaken to utilize the maximum DNA available (from between 1.5 × 10⁶ and 10⁷ cells [Table 3]). By this method, all six XLA patients were negative for EBV in their PBMC DNA (Table 3). This is compared to detection of EBV in four of eight healthy seropositive donors (Table 3). Figure 1C shows an example of the positive PCR result found for seropositive donor NHD6. The seronegative donor tested was negative for EBV by PCR (Table 3). Comparison of the total numbers of PCRs gives no (of 282) positive results for XLA patients and 35 (of 336) positives (P < 0.01) (Table 3).

XLA patients have no memory CTLs specific for EBV.

Peripheral blood cell samples from each patient were assessed by bulk reactivation and/or LDA for the presence of CTLs specific for EBV. Generally, responder cells (PBMC or T-cell-enriched [E+] fractions) (Table 2) from each patient were cultured in the presence of stimulator cells presenting EBV synthetic peptides known to be presented by the HLA type of each patient and that had been shown to stimulate an EBV-specific CTL response from healthy seropositive donor PBMC in vitro (Table 6).

TABLE 6.

HLA types of XLA patients and healthy donors and HLA-restricted EBV epitopes and EBV antigens represented

Patient or donora	HLA type	EBV epitope	EBV antigen (reference)
XLA1	B8	FLRGRAYGL	EBNA-3A (8)
NHD19
		RAKFKQLL	BZLF-1 (4)
XLA2	B44	KEHVIQNAF	EBNA-3C (21)
XLA3
XLA5		EENLLDFVRF	EBNA-3C (7)
XLA6
NHD15,b NHD18		VEITPYKPTW (used experimentally as a peptide mix)	EBNA-3B (22)
XLA4	B7	RPPIFIRRL	EBNA-3A (18)
NHD16b
NHD17

^aXLA, XLA patient; NHD, normal healthy donor. 

^bSeronegative. 

Stimulator cells used were either E+ or PBMC treated with synthetic peptides or alternatively the BL30/EBO-pLPP-Sig-FLRG cell line presenting the epitope FLRGRAYGL (EBNA-3A) (20) (Table 2). Stimulated cells (effectors) were applied to peptide-treated autologous or allogeneic HLA-matched PHA blasts, and specific lysis of these target cells was assessed by chromium release cytotoxicity assay (Table 2).

Figure 2 shows that, for all XLA patients assessed by bulk CTL reactivation, the percentage of specific lysis of peptide-treated target cells was clearly not significantly different from background killing (shown by the lysis of untreated targets). This result indicates that XLA patients have no memory T cells to these EBV epitopes (P = 0.5). Bulk CTL reactivated from healthy seropositive donors produced specific lysis levels against peptide-treated targets which were significantly (P < 0.01) greater than background killing, indicating that a strong CTL response had been stimulated (Fig. 2).

FIG. 2.

Bulk CTL reactivation results. Specific lysis of PHA blast target cells by EBV-specific CTLs from XLA patients and healthy donors is shown. (A) Experiments using epitopes VEITPYKPTW (EBNA-3B) and KEHVIQNAF/EENLLDFVRF (EBNA-3C) for patients 2, 3, and 5 and NHD18 (HLA B44). (B) Experiments using epitopes RAKFKQLL (BZLF-1) and FLRGRAYGL (EBNA-3A) for XLA1 and NHD19 (HLA B8). Specific lysis of target cells was calculated as follows: (mean ⁵¹Cr release target cells − mean spontaneous release)/(mean maximum release − mean spontaneous release) × 100.

LDA is a highly sensitive method for the screening of memory CTLs. Figure 3 shows the results obtained from the LDA experiments. The proportion of wells containing EBV-specific CTLs for each original dilution of cells (50,000 to 6,250) was 0% in every case for XLA patients and healthy seronegative donors. However, for healthy seropositive donors, the proportion of wells containing EBV-specific CTLs for each original dilution of cells ranged from 54 to 8% for donor 17 and 100 to 62.5% for donors 18 and 19 (Fig. 3). Thus, XLA patients show no evidence of an EBV-specific memory CTL response in the peripheral circulation.

FIG. 3.

LDA results. The proportions of wells at each of the original cell concentrations shown to contain EBV-specific CTL by chromium release assay are shown. Specific lysis of target cells was calculated for each well: (mean ⁵¹Cr release of target cells − mean spontaneous release)/(mean maximum release − mean spontaneous release) × 100. Wells were considered to contain EBV-specific CTLs under the following criteria: (i) specific lysis was greater than 10%; (ii) specific lysis was at least 10% greater than that shown in the corresponding untreated target well; and (iii) specific lysis was always greater than 3 standard deviations above the mean spontaneous release from the peptide-treated target cells.

XLA patients have no mucosal immunity to EBV.

In order to establish the role of infused donor IgG in the passive immunization of XLA patients against EBV infection, throat washing samples from each patient were assessed for the presence of anti-EB VCA IgG. Since the infusions are known to contain only trace amounts of IgA, this was not screened for. Undiluted throat washing supernatants were tested by using an indirect immunofluorescence assay and were shown to contain no detectable anti-VCA IgG.

XLA patients are infectible by the lymphotropic HHV-6.

PCR analysis for HHV-6 was performed on DNA extracted from 10⁶ PBMC for each XLA patient and five healthy donors of unknown HHV-6 serological status in order to screen for the presence of another lymphotropic herpesvirus. Three of six patients and two of five healthy donors were PCR positive for HHV-6. These incidences concur with data published by other workers using PCR for detection of HHV-6 in normal donors (13). There is no significant difference between the results obtained from XLA patients and healthy donors (P = 1). Figure 4A and B show the PCR results obtained from both patients and donors.

FIG. 4.

Detection of HHV-6 DNA in PBMC of XLA patients and healthy donors. Products obtained from HHV-6 PCR performed on DNA extracted from PBMC were analyzed by Southern blotting and hybridization. (A) Two of five healthy donors are positive for HHV-6 by PCR. Lane 1, PCR mix control containing water instead of template DNA; lane 2, reaction using 1 μg of DNA from HHV-6 negative cell line B96-8; lanes 3 to 9, reactions each using 1 μg of PBMC DNA from NHD7; lanes 10 to 15, reactions each using 1 μg of PBMC DNA from NHD8; lane 16, reaction using 10 μl of supernatant from the T-cell line J-jhan infected with HHV-6. (B) Three of six XLA patients are positive for HHV-6 by PCR. Lanes 1 and 2, PCR mix controls containing water instead of template DNA; lane 3, reaction using 1 μg of DNA from HHV-6-negative cell line B95-8; lanes 4 to 10, reactions using 1 μg of PBMC DNA from XLA1; lanes 11 to 13, reactions using 1 μg of PBMC DNA from XLA2; lanes 14 to 20, reactions using 1 μg of PBMC DNA from XLA3; lane 21, reaction using 10 μl of supernatant from the T-cell line J-jhan infected with HHV-6.

DISCUSSION

In vitro EBV infection of squamous epithelial cells was reported in the early 1980s (51), and following this, EBV DNA was demonstrated in exfoliated oropharyngeal epithelial cells from patients with IM (50). These observations led to a universally accepted model of EBV biology in which productive infection of oropharyngeal epithelial cells played a key role in the primary infection and transmission of EBV and in the infection of B lymphocytes. It was also suggested that the lifelong persistence of EBV may result from infection of basal epithelium of the oropharynx with cellular maturation-linked viral gene expression culminating in full productive infection in the mature surface cells (1) in a fashion analogous to papillomavirus infection of the epidermis (28). However, experimental evidence to support this theory is scarce, and convincing data has accumulated to suggest that the virus persists in the host by establishing a latent infection in B lymphocytes. Recently, the role of epithelial cells in primary infection has also been challenged by several workers using in situ hybridization, immunohistology, and PCR techniques to detect EBV RNA, proteins, and DNA in oropharyngeal cells (2, 19, 33–35, 53). These experiments detect latent and lytic infection in intraepithelial B cells but no epithelial cell infection. Since it could be argued that these essentially negative data result from the insensitivity of the techniques used to detect EBV and/or the rarity of the infected cells in the epithelium, we have used an entirely different approach to investigate the cell type(s) required for EBV infection and transmission.

Six XLA patients with a rare genetic mutation of the cytoplasmic tyrosine kinase gene, which results in a complete absence of mature circulating B cells, were studied. Although immature B cells may be present in the bone marrow of XLA patients (38), by flow cytometry we found no CD19- or CD20-positive cells in the peripheral blood samples of our cohort, indicating that no mature circulating B cells were present. Similarly, we were unable to generate LCLs from their peripheral blood T-lymphocyte depleted cells by in vitro infection with EBV preparations known to produce LCLs from 100% of peripheral blood B-cell samples from healthy donors. These results indicate that the patients studied had no circulating B cells capable of supporting EBV infection.

Since XLA patients have no antibody-producing cells and are regularly infused with pooled immunoglobulin preparations from large numbers of donors, we could not use conventional serological testing to determine past exposure to EBV. However, using the tissue culture method of spontaneous outgrowth of LCL, we showed consistently negative results from all XLA patient samples which were in sharp contrast to those found with healthy seropositive donors where 27% of donors’ E− cells grew into an LCL. A sensitive PCR technique which regularly detects one copy of target EBV DNA in a background of DNA from half a million EBV-negative cells was also used to screen for virus in patient and healthy donor peripheral blood and throat wash samples. Where possible, DNA from ten million PBMC was examined from each patient and DNA from more than six million PBMC was examined for normal healthy donors. Whereas 50% of healthy seropositive donors gave a positive result, no EBV DNA was detected in the XLA patient PBMC or that of a seronegative donor. Similarly, throat washing supernatant and cell pellet samples which gave 80% positivity for healthy seropositive donors were uniformly negative for the XLA patients. Thus, our results show conclusively that none of the six XLA patients examined is persistently infected with EBV, whereas 15 of 16 (94%) of the normal seropositive donors had evidence of EBV infection in at least one of the assays used.

Between 5 and 10% of the adult population in the United Kingdom remain apparently uninfected by EBV for life with no immunological evidence of past or persistent infection (reviewed in reference 9). However, B cells from these individuals can be infected with EBV in vitro, and it is therefore assumed that they have escaped infection with the virus. It is possible that the six XLA patients examined here all fall into this category because their susceptibility to opportunistic bacterial infections has perhaps led to a sheltered lifestyle which has protected them from infection. In fact, these individuals all lead normal lives, attending school, pursuing careers, and forming relationships. Moreover, as EBV is largely an asymptomatic infection, except in the case of IM, there is no way of knowingly avoiding those individuals who are shedding infectious virus. In addition, serum samples from a small number of parents and siblings of the XLA patients were screened for antibodies to EBV. All parents and siblings tested were seropositive for EBV, indicating that some XLA patients could at least have been exposed to EBV from close family members.

XLA patients’ immunotherapy contains antibodies to EBV antigens which could reach epithelial surfaces via serous secretions and protect them from primary EBV infection. However, studies on healthy EBV-seropositive individuals have concluded that salivary antibody (IgG and IgA) responses are unlikely to confer immunity to EBV (58). Furthermore, the XLA patients’ intravenous immunoglobulin preparations contain only trace quantities of IgA, and we found no detectable IgG antibody to VCA in patients’ saliva. Similarly, as four of the patients received only intramuscular immunotherapy (resulting in low serum IgG levels) until the ages of 11, 12, 23, and 39 (XLA patients 4, 3, 2, and 5, respectively [Table 1]), there was a considerable window of opportunity within which EBV infection could have occurred, prior to the onset of potentially protective intravenous immunotherapy.

To investigate the potentially protective role of systemic IgG, we screened patients for HHV-6. This virus has many parallels with EBV in that it is a ubiquitous (5, 39) herpesvirus, exhibiting lymphotropism (26, 27) with salivary transmission (10, 40), although we recognize that HHV-6 infection typically occurs earlier in life than EBV (5, 9, 39). We found HHV-6 DNA at the same frequency (P = 1) in patient and healthy donor PBMC DNA. Thus, the results obtained from the HHV-6 screening show that XLA patients can be infected with a lymphotropic herpesvirus. In addition, based on the seroepidemiological profile of EBV taken from the general population (reviewed in reference 9), and accounting for the ages of these six XLA patients (Table 1), the likelihood of them all randomly being EBV negative is 3.44 × 10⁴:1.

The lack of persistent EBV infection in XLA patients suggested to us that either their lack of B cells renders them uninfectible with the virus or that they had undergone primary infection in the oropharyngeal epithelium but, because of the absence of B cells, no lifelong persistence had been established. To distinguish between these two possibilities, we used the most sensitive techniques available to examine their T-cell memory to EBV epitopes known to be presented by the particular HLA types of the patients. If the XLA patients had at any time experienced a primary infection in the oropharyngeal epithelium, CTLs should have been generated, although if healthy epithelial cells are infected in vivo, the viral gene expression is as yet unclear. We assume that the full range of both latent and lytic antigens would be expressed during productive infection, as representatives of both sets of genes have been shown to be coexpressed on OHL (49) including the EBNA-2 protein which is a powerful transactivator of the other latent viral genes (reviewed in reference 46). The experiments were limited by the known epitopes available for recognition by the HLA types of the patients, but the broadest range possible were tested for memory CTL responses. However, despite the patients having peripheral blood T-cell counts within the normal range and by screening against lytic and latent antigens when possible, we found no evidence of a memory CTL response to EBV from patient or healthy seronegative donor samples, whereas all seropositive donors tested gave strong CTL responses. That functional T cells were present in the samples was confirmed by the fact that PHA blasts were produced from all but one patient (where insufficient cells were available), indicating that normal lymphoproliferative responses to stimulation with a T-cell mitogen were intact. Our results suggest that the XLA patients are in fact uninfectible by EBV due to their lack of mature B lymphocytes and therefore that B cells and not epithelial cells are required for primary infection with EBV. However, the possibility that a prior infection of epithelial cells leading to a CTL response that has since waned and is now undetectable cannot be entirely ruled out by our findings.

Despite this, we believe that our findings support the data presented by other groups (2, 19, 33–35, 53) and suggest that the life cycle of EBV and its ability to infect virtually the entire population of the world are a result of infection of intra- and extraepithelial B cells in the oropharynx.

It can also be inferred that in the absence of a detectable latent EBV infection in these patients, the B cell must be the only site of viral persistence. However, if our assumption that EBV is unable to establish primary infection in these patients is correct, then it is presumably impossible for persistent infection to ensue. Therefore, an epithelial site of latency, however unlikely, cannot be entirely ruled out in healthy individuals by our study.