Adiba Isa, Department of Medicine, Solna, Infectious Disease Unit, Center for Molecular Medicine, CMM L8:03, Karolinska Institutet, Karolinska University Hospital, Solna, SE-171 76 Stockholm, Sweden. E-mail: email@example.com
The aim of this study was to characterize the proinflammatory and T helper (Th)1/Th2 cytokine responses during acute parvovirus B19 (B19) infection and determine whether an imbalance of the Th1/Th2 cytokine pattern is related to persistent B19 infection. Cytokines were quantified by multiplex beads immunoassay in serum from B19-infected patients and controls. The cytokine responses were correlated with B19 serology, quantitative B19 DNA levels and clinical symptoms. In addition to a proinflammatory response, elevated levels of the Th1 type of cytokines interleukin (IL)-2, IL-12 and IL-15 were evident at time of the initial peak of B19 viral load in a few patients during acute infection. This pattern was seen in the absence of an interferon (IFN)-γ response. During follow-up (20–130 weeks post-acute infection) some of these patients had a sustained Th1 cytokine response. The Th1 cytokine response correlated with the previously identified sustained CD8+ T cell response and viraemia. A cross-sectional study on patients with persistent B19 infection showed no apparent imbalance of their cytokine pattern, except for an elevated level of IFN-γ response. No general immunodeficiency was diagnosed as an explanation for the viral persistence in this later group. Neither the acutely infected nor the persistently infected patients demonstrated a Th2 cytokine response. In conclusion, the acutely infected patients demonstrated a sustained Th1 cytokine response whereas the persistently infected patients did not exhibit an apparent imbalance of their cytokine pattern except for an elevated IFN-γ response.
Parvovirus B19 (B19), a non-enveloped, single-stranded DNA virus, is one of the smallest viruses known to infect mammalian cells. The viral genome (5kbp) encodes for three major proteins, the non-structural protein 1 (NS1) and the viral capsid proteins (VP1 and VP2). B19 is associated with a variety of clinical manifestations. The infection can be asymptomatic or produce a mild febrile illness (erythema infectiosum) associated with transient anaemia, exanthema and arthralgia. The clinical manifestations can also be more severe, such as cytopenia in immunocompromised individuals and hydrops fetalis or intrauterine fetal death in pregnant women. B19 is classified as a lytic non-persistent virus that is transmitted normally through the respiratory route .
Until recently, it has thus been believed that B19 was cleared from the blood circulation shortly after production of neutralizing antibodies, a few weeks to months post-acute infection. However, by introducing a highly sensitive B19 DNA polymerase chain reaction (PCR) quantitative assay, we have shown previously that viral clearance is much slower than believed earlier and viral DNA can be detected in peripheral blood of infected individuals more than 2 years post-acute infection . Interestingly, a sustained, activated and mature antigen-specific CD8+ T cell response accompanies this viraemia and is also present in the circulation more than 2 years post-primary infection [3,4]. Thus, the classification of B19 as a rapidly clearing lytic infection in immunocompetent subjects should be reconsidered. Consequently, as infected individuals eventually clear the infection in blood but still maintain a low CD8+ T cell response [2,4], the virus may persist in other compartments and replicate only intermittently or at a low level in remotely infected individuals. A B19-specific T helper (Th) cell-mediated response has also been detected in remotely infected individuals [5,6]. It has been shown that B19 persists commonly in immunosuppressed individuals and in a few cases of immunocompetent individuals .
Cytokines mediate a number of important immunoregulatory functions and play a critical role in the modulation of the immune system. An aberrant proinflammatory cytokine profile or a later shift in the balance of the Th1/Th2 cytokine response may play a role in the control of the viral infection and lead to viral persistence . Primary B19 infection has been associated previously with a mixed Th1/Th2 profile, with elevated levels of interferon (IFN)-γ, tumour necrosis factor (TNF)-α and interleukin (IL)-6. An association between the symptoms related to B19 infection and elevation of several inflammatory cytokines was also shown [8,9]. The objective of the present study was to characterize the kinetics of the proinflammatory and Th1/Th2 cytokine pattern in consecutively followed acutely infected patients and to determine whether an imbalance of this response predominates in individuals with persistent B19 infection. A better understanding of cytokine responses in relation to viral load, cellular and humoral immune responses as well as severity of infection in the individual patient may allow new therapeutic strategies for severe B19 infection.
Materials and methods
Group I: acutely B19-infected patients
Eight previously healthy adults (eight females, mean age 43 years, range 38–54 years) with B19 infection presenting with at least three of the four symptoms (fever, arthralgia, fatigue and rash) were included. They had early resolution of symptoms, except for two patients with mild transient arthralgia lasting for several months. All patients were identified prospectively after their serum samples had been referred to the Clinical Virology Laboratory at the Karolinska University Hospital, and were followed for 20–130 weeks (median 80 weeks). All patients had detectable B19 DNA in serum by PCR and anti-B19 IgM and IgG in serum at the first time-point. Five of these patients have been presented in previous studies [2–4].
Group II: persistently B19-infected patients
Twenty-two (four males, 18 females, mean age 45 years, range 31–62 years) individuals were included, based on persistent B19 DNA in bone marrow (BM). This was defined as at least two repeatedly B19 DNA-positive BM samples collected with at least a 6-month interval. In total, two to six samples were collected per patient (mean time of follow-up 33 months, range 6–110 months). In the patients followed for only a few months, a recent asymptomatic infection could not be excluded as the virus may not yet have been cleared after primary infection. Clinical data on the persistently infected individuals were as follows: fatigue 19/22; chronic fatigue syndrome 14/22; arthralgia 20/22; myalgia 18/22 and paraesthesia 11/22. No patients had a history of repeated severe bacterial or other viral infections as a sign of a general immunodeficiency. All patients were B19 IgG-positive and two were also B19 IgM-positive for almost 2 years. Sixteen of these patients have been presented previously, in two separate studies [10,11].
Group III: healthy control subjects
Eighteen healthy B19 seropositive laboratory workers (12 male, six female, mean age 43 years, range 31–61 years) without any history of recent apparent symptoms were included in the study as controls. All were B19 IgG-seropositive, B19 IgM-seronegative and had no detectable B19 DNA in serum.
Ethical approval for the study was obtained from the local ethical committees at the University of Gothenburg and Karolinska Institutet, Karolinska University Hospital, Sweden.
Parvovirus B19 serology and PCR
B19 IgG and IgM were detected using enzyme immune assay (EIA; Biotrin International Ltd, Dublin, Ireland) according to the manufacturer's instructions. Presence of B19 DNA in serum and BM was confirmed using a nested PCR with a sensitivity of 103 geq/ml  and a quantitative PCR with a sensitivity of 102 geq/ml .
Analysis of the general immune status of the persistently infected individuals
To determine whether the persistently infected individuals (group II) had an underlying immunodeficiency, an extended evaluation was performed according to routines at the Clinical Immunology Laboratory, Sahlgrenska University Hospital, Gothenburg, Sweden.
Human leucocyte antigen (HLA) typing
HLA expression was analysed by the Tissue Typing Laboratory (Sahlgrenska University Hospital, Gothenburg, Sweden). HLA class I expression was determined by serology. Lymphocytes in peripheral blood were isolated and determinations of HLA were also performed using a complement-dependent cytotoxicity technique (CDC). HLA-DR genotyping was performed using a PCR–sequence specific primers (SSP) technique .
Immunophenotype of peripheral blood mononuclear cells
Fluorescence activated cell sorter (FACS) analysis was performed using peripheral blood mononuclear cells (PBMC) stained with directly fluorochrome-conjugated anti-human CD3, CD4, CD8, CD16, CD19, CD56 and HLA-DR (BD Biosciences, Mountain View, CA, USA). PBMC were incubated at 4°C for 15 min with antibodies. Cells were washed and later fixed with 1–2% formaldehyde. Cell acquisition was performed using FACSCalibur with CellQuest software (Becton Dickinson, Stockholm, Sweden). The absolute number of lymphocytes was determined using Sysmex K-4500 cell counter (TOA Medical Electronics Co, Nishi-Ku, Kobe, Japan). The results for each subpopulation were expressed as the percentage of lymphocytes and as the number of cells × 109/l.
Enzyme-linked immunospot (ELISPOT) assay
The ex vivo cellular immune responses were measured by IFN-γ ELISPOT. Briefly, 96-well nitrocellulose-bottomed plates (Multiscreen-HA plates; Milipore, Mulsheim, France) were coated with 15 µg/ml anti-IFN-γ monoclonal antibody overnight at 4°C (Mabtech, Stockholm, Sweden). After washing five times, duplicates of PBMC in different concentrations (5 × 104, 2 × 104 and 1 × 104 cells/well) supplemented in 200 µl medium were added. Thereafter, concanavalin A (ConA) was added at a final concentration of 25 µg/ml. The cells were incubated for 48 h at 37°C with 5% CO2. One hundred µl of 1 µg/ml biotin-conjugated anti-IFN-γ (Mabtech, Stockholm, Sweden) was added and incubated at room temperature for 2 h; 100 µl streptavidin-conjugated alkaline phosphatase (4 µg/ml) was added per well. Plates were washed five times with 0·05 M Tris buffer pH 9·5 and the spots were developed with 100 µl of bromochloroindolyl phosphate–nitroblue tetrazolium (BCIP-NBT) (Life Technologies, Heidelberg). One spot was considered to represent IFN-γ secreting cells, and counted using a dissection microscope. The results were reported as mean spot-forming cells (SFC) per million PBMC.
Quantification of cytokines
Twelve different cytokines, IL-1β, -2, -4, -5, -6, -8, -10, -12, -15, granulocyte–macrophage colony-stimulating factor (GM-CSF), IFN-γ and TNF-α, were quantified in serum from all individuals. A commercially available multiplex beads immunoassay, based on the Luminex platform (Biosource International, Inc, Camarillo, CA, USA) was used according to the manufacturer's procedure. All samples were run in duplicate. Briefly, beads with defined spectral property were conjugated to the analyte-specific capture antibodies. Beads, samples, standards and the controls were pipetted in a filter-bottomed 96-well plate and incubated for 2 h while shaking (550 rpm). After three washes the biotinylated detector antibodies were added to the beads and incubated for 1 h at the room temperature. Streptavidin conjugated with R-phycoerythrin (SA-PE) was added to the wells after several washings and incubated for 30 min. By monitoring the spectral properties of the beads and the amount of fluorescence associated with PE, the instrument measures the concentration of the analytes presented in the original specimens. The data (mean fluorescence intensity) were analysed using a Luminex reader (Luminex, Austin, TX, USA), and the mean concentration was calculated as pg/ml serum.
Statistical analyses were performed using GraphPad Instat and EpiInfo statistical analysis.
Dominance of a Th1 cytokine response in acute B19 infection
Eight individuals (group I) were studied regarding B19 serology, B19 DNA and cytokine profile during and following acute B19 infection (20–130 weeks' follow-up). The first serum sample was collected within 1–5 weeks after onset of acute symptoms. B19 IgM was detected in all patients from study start and up to 17 weeks post-primary infection, whereas B19 IgG was present in all serum samples throughout the study. B19 DNA was detected for the entire period for all but one patient, who cleared the infection after 108 weeks.
The presence and concentration of cytokines were tested longitudinally in the serum samples representing all individuals. Several proinflammatory cytokines (IL-1β, GM-CSF, IL-6, TNF-α and chemokine IL-8) were elevated about 2–65-fold within the first weeks of acute infection in two of eight individuals (Fig. 1). Thereafter the concentrations decreased to low or undetectable levels. The cytokine levels in the other six patients were < 5 pg/ml during the entire follow-up except for the chemokine IL-8, which rose transiently in three patients at one or two time-points each without obvious relations to any other markers. All proinflammatory cytokines were very low in the healthy seropositive individuals, except for two subjects who showed high levels of IL-8 (data not shown).
The concentrations of the Th1 cytokines IL-2, IL-12 and IL-15 were elevated in the first weeks following infection in some of the patients (Fig. 2). These three Th1 cytokines peaked in patient P2 in a sample collected at 3 weeks following onset of acute symptoms (Fig. 2a). IL-2 and IL-12 levels decreased at week 20, whereas IL-15 levels remained at about 600 pg/ml. Patient 3 was not sampled until 5 weeks post-onset of symptoms but had higher levels of IL-2, IL-12 and IL-15 compared with the sample collected at week 10 (Fig. 2b). Interestingly, patient P1 suddenly peaked in viral load (from 104 to 1·5 × 105 DNA copies/ml) after about 1 year of stable viral titres. This coincided with a peak of IL-2, IL-12 and IL-15, followed by a loss of viraemia some months later (Fig. 2c).
The Th1 cytokine responses in the other five patients were high and stable. IL-2, IL-12 and IL-15 were elevated in patients P4 and P7, IL-12 and IL-15 were elevated in patients P6 and P8 and only IL-12 was elevated in patient P5 (data not shown).
However, in contrast to the other Th1 cytokines, the levels of IFN-γ (except P2 at weeks 1–3) were not elevated. The Th2 cytokines IL-10, IL-4 and IL-5 were < 5 pg/ml in all patients (except for a raised value of IL-10 in patient P4 at week 5) (data not shown).
Persistently B19-infected individuals do not have a general immunodeficiency
Extensive investigations were performed to evaluate the HLA pattern and immunological condition of the persistently B19-infected patients (group II). HLA class I and II typing, immunophenotyping of PBMC and also the capacity of producing IFN-γ after ConA stimulation were performed. The frequency of HLA distribution was compared with 40928 individuals in the Tobias Registry (the Swedish register of BM donors). No significant differences in frequencies of HLA class I types were found, with one exception (Fisher's exact test). HLA-A9 was more common in persistently infected individuals (eight of 22), which was statistically significant (36% and 18%, P-value < 0·047, Fisher's exact test). The frequency of HLA-DR did not show any significant difference compared to a study on healthy Swedish individuals . The frequencies of lymphocytes and natural killer (NK) cells did not reveal any abnormalities compared to a reference group . Six study patients had subnormal levels of CD19+ B cells (< 0·1 × 109), but gel electrophoresis did not reveal any deficiency in immunoglobulin distribution. The capacity of IFN-γ production after ConA stimulation displayed no difference between persistent B19-infected individuals and controls (blood donors, Mann–Whitney). All individuals mounted high IFN-γ responses to ConA, more than 400 spots/106 PBMC.
These assays were performed at an accredited clinical immunology laboratory and the results were interpreted according to standard reference values. In summary, we could not detect any general immunodeficiency, which could explain the viral persistence of B19 in this group.
Cytokine responses: some persistently B19-infected individuals have an elevated IFN-γ response
All persistently B19-infected individuals (group II) had detectable B19 DNA in BM for at least 6–110 months as tested by nested PCR. B19 DNA in serum could be detected in only seven of the 22 individuals. Cytokine levels in serum showed significantly higher concentrations of IFN-γ (P-value 0·004, unpaired t-test) compared with healthy seropositive controls (group III) (Fig. 3). However, only 10 of 22 patients had levels above 100 pg/ml of IFN-γ. The mean concentration of the other proinflammatory or Th1/Th2 associated cytokines did not differ significantly when comparing persistently infected individuals with healthy seropositive controls (data not shown).
Acute B19 infection is associated with extensive replication of B19 virus in erythroid precursor cells in the bone marrow. We reported recently that this results in persistent viraemia despite development of neutralizing antibodies and a B19-specific CD8+ cell response . The lack of viral resolution is contradictory and we have therefore dissected the fine-tuned cytokine immune response in consecutive samples from eight acutely infected patients and followed them for a minimum of 20 weeks. An initial peak of proinflammatory cytokines (IL-1β, TNF-α, IL-6 and the chemokine IL-8) was found at onset of acute B19 infection in two of the eight patients. This peak may have been missed in the other patients as the samples were collected within 1–5 weeks after onset of symptoms. However, an induction of the Th1 type of cytokines IL-2, IL-12 and IL-15 was already seen in the earliest available samples and this expression was sustained in many patients during the follow-up period. In contrast, cytokines associated with a Th2 type of immune response (IL-4, IL-5 and IL-10) as well as an IFN-γ response remained low during the observation time. Despite the lack of these Th2 cytokines the patients developed normal B19-specific IgG levels. These antibodies may have been induced by a low-grade IL-6 response as well as other cytokines not studied here (i.e. TGF-β).
With regard to the adaptive immune response, IL-12 was the first detectable Th1 type of cytokine followed by IL-2 and IL-15 elevations. The induction of these cytokines was linked in time to some of the patients who have previously been assessed for B19-specific CD8+ T cell responses which increased in magnitude, matured and remained activated over time . Together with IL-7, IL-15 appears to be important in up-regulation of Bcl-2 in antigen-specific activated T cells and thus of critical importance for the survival of the effector CD8+ T cells during the contraction phase . However, IL-15 does not seem to be mandatory for the expansion of the effector T cell during a viral infection . Together with sustained IL-2 and IL-12 expression, the IL-15 expression presented here is thus in line with our previous results with the detection of an activated and mature B19-specific CD8+ effector T cell (CD38+ CD57+ perforin+) population, long after resolution of B19-related symptoms, but during the time of a continuous low-level B19 viraemia. This type of vigorous cytokine and CD8+ T cell response has not been seen previously for lytic viruses in humans but, rather, resembles antigen-persistent virus infections such as Epstein–Barr virus (EBV).
However, despite the presence of a Th1 cytokine response there was no detectable serum concentration of IFN-γ. This was indeed a surprising finding, because IL-12 and IL-2 production is linked normally to significant IFN-γ synthesis. We also assessed samples from patients with chronic inflammatory conditions and found significantly increased IFN-γ levels in these patients (data not shown). We therefore do not believe that lack of detectable IFN-γ in our acutely infected patients was due to technical reasons. Our previous study showed that short-term in vitro culture with B19 antigens induced significant IFN-γ production as measured by ELISPOT during more than 2 years following acute B19 infection . Together, these data indicate that the persistent B19 viraemia in our acutely infected patients may be associated with an aberrant cytokine profile in vivo with a selective IFN-γ deficiency despite Th1 cytokine induction.
It has indeed been possible to link the cytokine responses to previous studies on the CD8+ T cell responses in detail in some of the patients. For example, one of the patients (P1) had a second viraemic period (> 1·5 log increase of viral load) about 1 year after the acute B19 infection. Interestingly, the peak of viraemia coincided with a peak of the CD8+ T cell response  and peak IL-2, IL-12 and IL-15 levels. A few weeks before this peak, a loss of one of the three epitope-specific CD8+ T cell responses were noted and the frequency of B19-specific CD8+ T cells were decreased to much lower levels around this time, as defined by tetramer staining . It is not possible to determine whether the increase of the viral load was due to a new infection or reactivation of the primary infection, the loss of one of three epitope specific CD8+ T cells, or a result of a viral mutation to escape the immune response. Surprisingly, this patient was the only individual in this study who cleared the viraemia during follow-up (at 108 weeks), perhaps as a result of the boosted immune response.
One previous study has correlated B19 infection with a mixed Th1/Th2 profile with elevated levels of IFN-γ, TNF-α and IL-6 at acute B19 infection . Another study demonstrated expression of mRNA for IFN-γ and IL-2 appriximately 2 weeks post-symptom presentation in a child with acute B19 infection . These differences to our study could possibly be explained by the timing of the first sample taken in relation to onset of primary infection as well as to severity of acute infection, and to the fact that we followed our patients with consecutive samples for a long time-period. Our patients were previously healthy and resolved their B19-associated clinical symptoms within 2 months, thus representing the vast majority of acutely infected patients by not developing chronic fatigue or long-lasting severity arthralgia. The pathogenesis of B19 infection with transient symptoms but persistence of viraemia for more than 2 years is unusual when compared to most systemic viral infections. In most settings, chronic viraemia is associated with clinical symptoms or findings. For instance, chronic hepatitis B infection (deficiency in cytotoxic T cell response) and chronic hepatitis C infection (IFN-α deficiency) results in liver damage, HIV-1 infection (lack of polyfunctional CD8+ T cells) in CD4+ T cell decline, EBV (lack of perforin and induction of viral IL-10) in lymphoproliferative conditions and severe cytomegalovirus (CMV) [down-regulated major histocompatibility complex (MHC) class I and cytotoxic T cells] in a broad spectrum of symptoms. In all these examples abnormal immune responses result in impaired T cell-mediated cytotoxic response. All these viruses are associated with selective disturbances in their cytokine response pattern [7,17–19]. Furthermore, complete induction of Th1 immune responses is associated normally with elimination of the pathogen via cytotoxic activity. Our finding of the lack of IFN-γin vivo during a prolonged period from acute B19-infection is therefore surprising, and abnormal in the context of other generalized viral infections.
We were not able to follow the primary B19-infected patients for more than 2 years and could thus not determine at which time-point Th1 cytokines, T cell responses and viraemia decreased to the undetectable levels observed in seropositive healthy individuals . Nor did we follow a large enough group of acutely infected patients to determine whether an imbalance of the cytokine pattern was correlated with a persistent status of the infection, as seen in a small percentage of infected patients [20,21]. One may speculate that activated B19-specific CD8+ T cells do not migrate to the bone marrow that harbour the virus-infected cells or that the cells may not be functioning fully in vivo. Our and other studies [3,5,6] have demonstrated B19-specific cytotoxic and helper T cell activity after in vitro stimulation. In order to evaluate the relevance of these assays it is of vital interest to study these phenomena in vivo at the site of virus replication, in this case the bone marrow, as the concentration of cytokines at different sites may vary.
We next analysed the cytokine pattern in a group of persistently B19-infected individuals who remained B19 DNA-positive in bone marrow samples for at least 6 months, several of them for years, following acute infection. Initially we performed an extensive immunological examination to determine whether these individuals suffered from a general immunodeficiency that could explain the B19 persistence. Some patients had subnormal levels of NK cells and CD19+ B cells in blood (compared to established reference values), but without any immunoglobulin distribution deficiencies, and the evaluation did not reveal any general underlying immunodeficiency disorders in the patients. However, HLA A9 was more common among persistently B19-infected patients than in controls. A relation between HLA-DR*4 and symptomatic acute B19 infection has been described previously . As seen for many other infections, individuals can be sensitive to microbial agents due to selective properties in their immunological repertoire. Our finding, of overrepresentation of HLA A9 in persistent B19 infection, has not been described previously and is difficult to interpret. However, by extensive evaluations of genetic (unpublished data), humoral  and cellular  immune responses we have found only one type of defect that discriminates persistently infected individuals from healthy seropositive subjects: a shift in the in vitro T cell immune response from the B19 NS1 to the B19 viral capsid proteins . In the present study we aimed to determine whether a difference in the cytokine response was associated further with the persistent status of the infection. An imbalance of Th1/Th2 cytokines has been suggested to be associated with the chronicity of hepatitis C virus (HCV) . However, only higher levels of IFN-γ compared with acutely and remotely B19 infected patients were found in our persistently infected individuals. This probably indicated an unspecific response and cannot possibly explain the persistent status of the B19 infection or the aberrant T cell response described above.
It seems clear that the previous designation of B19 as a lytic non-persistent virus needs modification as the infection leads neither to rapid viral clearance nor contraction of the initial T cell burst [23,24]. Rather, the virus seems to have characteristics of both persistent and lytic non-persistent viruses, now also demonstrating that the Th1 cytokine levels were sustained at high levels. This further confirms our hypothesis of describing B19 infection as a new type of host–virus relationship not seen in any other ‘acute’ human viral infection.
This study was supported financially by the Alice Swenson Foundation for Scientific Research, the Fokus Borås Foundation, Research and Development Council in Södra Älvsborg, Research and Development at Södra Älvsborg Hospital Borås, the Tobias Foundation, the Swedish Children's Cancer Foundation, the Swedish Cancer Foundation, the Swedish Research Council, the Wellcome Trust, the Medical Research Council, UK and the Commission of the European Communities (QLK2-CT-2001–00877).