SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

Congenital cytomegalovirus (CMV) infection is the most common congenital infection causing childhood morbidity. The pathogenetic mechanisms behind long-term sequelae are unclear, but long-standing viremia as a consequence of the inability to convert the virus to a latent state has been suggested to be involved. Whereas primary CMV infection in adults is typically rapidly controlled by the immune system, children have been shown to excrete virus for years. Here, we compare T cell responses in children with congenital CMV infection, children with postnatal CMV infection and adults with symptomatic primary CMV infection. The study groups included 24 children with congenital CMV infection, 19 children with postnatal CMV infection and eight adults with primary CMV infection. Among the infants with congenital CMV infection, 13 were symptomatic. T cell responses were determined by analysis of interferon gamma production after stimulation with CMV antigen. Our results show that whereas adults display high CMV-specific CD4 T cell responses in the initial phase of the infection, children younger than 2 years have low or undetectable responses that appear to increase with time. There were no differences between groups with regard to CD8 T cell function. In conclusion, inadequate CD 4 T cell function seems to be involved in the failure to get immune control of the CMV infection in children younger than 2 years of age with congenital as well as postnatal CMV infection.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

Cytomegalovirus (CMV) is the most common congenital infection, with a prevalence of 0.1–2.2% [1-6] all live births in developed countries. Congenital infection with CMV has been linked to the development of long-term sequelae; in fact, this disease is responsible for more disabilities in children in the United States than are other better-known conditions such as foetal alcohol syndrome and spina bifida [7]. The majority of children with intrauterine-acquired CMV infection are asymptomatic during the neonatal period. In the remaining 10–15%, petechiae, thrombocytopenia, microcephaly, jaundice and hepatosplenomegaly are the most commonly observed symptoms [8]. Children with symptomatic disease are at high risk for long-term sequelae, such as hearing deficit, visual impairment and neurological disabilities [9, 10]. In a recent European study, 38% of the symptomatic children had some kind of deficit (e.g., sensorineural hearing loss, cerebral palsy, developmental delay) at 2 years of age [11]. To a lesser extent (10–20%) [2, 12, 13], sequelae also develop in asymptomatic infants. The mechanisms beyond progressive congenital CMV disease are poorly understood [14].

Primary CMV infections frequently result in intermittent shedding of virus in the urine for months after seroconversion [15]. Although viruria occurs only sporadically in immunocompetent adults [16], children of preschool age may shed virus for extended periods and therefore represent a major societal reservoir of CMV [17, 18].

The most protracted viral shedding is seen in children with congenital CMV infection, who may excrete virus in the urine for up to 10 years [19-21]. This prolonged shedding is a sign of an inability to convert the virus to a latent state. It has been speculated that persistent viral replication in these children also contributes to the development of sequelae [22], a theory that is supported by reports of improved hearing or maintenance of normal hearing after ganciclovir treatment [23, 24]. It has further been suggested that prolonged antiviral treatment may improve audiological and neurological outcome by suppression of viral load during a sensitive period of development [25, 26]. The significance of CD8 (T-killer) and CD4 (T-helper) cells in the process of converting a primary CMV infection to latency and preventing viral replication has been demonstrated in both rodents [27, 28] and humans [29, 30]. The recent development of new analytical techniques has made the study of virus-specific cells in humans easier and more accurate. Here, intracellular cytokine staining after CMV antigen pulsing was used to investigate the CMV-specific T cell responses in children. The CD4 T cell responses were also compared with adults with primary CMV infection.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

Study population

The study was approved by the local ethics committee. Subjects (for paediatric patients, the parents or legal guardians) received oral and written information before giving informed consent.

Infants and children

Congenital CMV infection

Congenital CMV (cCMV) infection was defined by the presence of CMV-DNA in the plasma [31], urine [31], leucocytes [32] or dried blood spots [33] during the first 3 weeks of life.

In an epidemiological screening study for congenital CMV infection (October 2003–June 2004; August 2004–October 2004) in the Stockholm area, 12 infants with congenital CMV infection were identified [34]. Parents to one infant declined participation; thus, 11 were included in the immunological study. Further 13 infants/children with symptomatic CMV disease were included. Clinical characteristics of the congenitally infected children are summarized in Table 1 [35].

Table 1. Children with congenital cytomegalovirus (CMV) infection
SubjectGenderSample age (months)CMV-related symptoms
  1. SNHL, sensorineural hearing loss; AS, asymptomatic; M, male; F, female.

cCMV 2F8.5; 30Bilateral SNHL, head circumference deviating > −2 SD, mental retardation, autism
cCMV 5M2Unilateral SNHL
cCMV 6M1.5; 16; 22; 35AS
cCMV 8M6; 21AS
cCMV 9M13; 27AS
cCMV 10F1.5; 13; 34AS
cCMV 12F3.5; 12AS
cCMV 13M6; 14; 17AS
cCMV 16F19AS
cCMV 17F13AS
cCMV 18F101Bilateral SNHL
cCMV 19F52Bilateral SNHL
cCMV 20M55Bilateral SNHL
cCMV 21M106Bilateral SNHL
cCMV 22F40Bilateral SNHL
cCMV 23M19.5AS
cCMV 26F24AS
cCMV 40M1.5Haemolytic anaemia
cCMV 59F4Thrombocytopenia, elevated transaminases, unilateral SNHL
cCMV 62F0.8Respiratory distress syndrome, anaemia, coagulase-negative staphylococcus infection
cCMV 71M0.5; 1.5Seizures
cCMV 73M8; 9Bilateral SNHL
cCMV 75M18Feeding problems, hypotonia, bilateral SNHL
cCMV 77M9AS
Table 2. Numbers of children, blood samples and median lymphocyte counts in different age groups
 Age interval, monthsCongenital cytomegalovirus (CMV)Postnatal CMVCMV-negative controls
  1. a

    Median values in blood from healthy children: 0–6 months: 4100–6000; 6–24 months: 5500–6000; 2–5 years: 3300 [35].

Children (n) 241921
Symptomatic (n) 1300
Total number of blood samples0–69710
6–241969
>24972

Lymphocytes per μL blood

Median (range)a

0–66850 (1200–8500)6400 (1950–7000)5100 (2900–8500)
6–246150 (1850–11,700)7900 (4800–10,900)4800 (2599–10,300)
>243900 (2000–6500)2700 (1000–4900)2000 and 2500
Postnatal CMV infection

For the purpose of this study, postnatal CMV (pCMV) infection was defined as detectable αCMV-IgG (Enzygnost AntiCMV/IgG; Dade Behring, Deerfield, IL, USA) in plasma. As the youngest children (<12 months) may have had residual placenta-transferred maternal αCMV-IgG, these children were included if CMV-DNA was detected in the urine.

Two children were diagnosed in clinical practise. A boy was identified because his twin brother had congenital CMV infection detected by newborn CMV screening. A girl was assessed for CMV infection in the diagnostic evaluation of hearing deficit. She had positive virus isolation from urine, but the CMV-PCR analysis of the DBS sample was negative.

Sixteen children were identified in a group of paediatric patients visiting Uppsala University Hospital. Thirty-seven children, admitted to the hospital for elective surgery, gastroscopy, cystoscopy or follow-up due to premature birth, were recruited for the study. They were analysed without knowledge of their CMV status, and blood plasma or urine was frozen for retrospective analysis. The 21 children without CMV-αCMV-IgG or positive CMV-PCR were used as negative controls.

Data from infants and children were divided into three groups based on age when the sample was taken: Below 6, 6–24 and over 24 months (Table 2).

Adults with primary infection

Eight CMV-infected adults with a mean age of 30 years (range: 16–44) were included in the study. These adults, recruited from the Department of Infectious Diseases at Uppsala University Hospital, were all immunocompetent and had symptomatic (fever, headache and/or myalgia and elevated liver transferases) primary CMV infection, as verified by IgM and low IgG levels. Five of these patients have been described previously [36].

Data were analysed based on time point after symptoms first appeared: 0–3 and 3–6 months.

Blood sampling

Blood samples were collected at the Departments of Neuropediatrics at Karolinska University Hospital, Huddinge, Stockholm and the Departments of Infectious Disease and Neonatal Care at the University Hospital, Uppsala. Samples were obtained from children only in conjunction with sampling for clinical purposes. Therefore, repeated sampling at later time points was only feasible in 11/43 infants. The amount of blood was in each case determined by the physician in charge and therefore varied with the age and size of the infants. Samples were kept at room temperature during transport, and analyses were begun within 6 h of venipuncture.

Cell counting

Lymphocytes were counted using a Medonic CA620 cell counter (Clinical Diagnostic Solutions, Plantation, FL, USA). Absolute numbers of CD8 and CD4 T cells were calculated from the percentages that were obtained for each cell population from the intracellular staining and flow cytometry analyses, as described below. Throughout this paper, CD3+ CD8− cells are used to approximate CD4 cells.

CMV-specific T cell activation and intracellular cytokine analysis

Functional assays of CMV-specific CD4 (for adults and children) and CD8 (children only) T cells were performed as previously described [36, 37], using protocols initially developed by Kern et al. [38] and Nomura et al. [39]. In brief, sodium-heparinized whole blood was incubated with CMV antigen, peptide pools [15 amino acids (aa) overlapping by 11 aa] spanning the entire pp65 and IE1 proteins, respectively (JPT Peptide Technologies GmbH, Berlin, Germany), or CMV lysate (Advanced Biotechnologies, Columbia, MD, USA) and αCD28/CD49d costimulatory antibodies (BD Biosciences, San Jose, CA, USA). After 2 h, Brefeldin A (Sigma-Aldrich, St. Louis, MO, USA) was added, and the tubes were placed in a programmable water bath and incubated for 4 h then kept at 10 °C until the next day. After erythrocyte lysis and cell permeabilization, antibody staining was performed using αCD3-APC, αCD8-PerCP, αCD69-PE and αIFNγ-FITC (all from BD Biosciences).

Flow cytometric analysis

Flow cytometric analyses were performed using a four-colour FACSCalibur (BD Biosciences, San Jose, CA, USA) instrument with CellQuest Pro Software (BD Biosciences, San Jose, CA, USA). The analyses were carried out immediately after cell staining, and 20,000 to 30,000 CD3+ lymphocytes were collected. The result from a control tube without antigen was subtracted from the corresponding experimental results to correct for background cytokine release. Results below the detection limit of 0.05% CD4 or CD8 cells were assigned a value of 0.025%. The percentage values were re-calculated as positive cells per μl blood using the cell counting results.

Statistics

All statistical analyses were performed using GraphPad Prism version 4 for Macintosh (GraphPad Software, San Diego, CA, USA).

If samples were drawn twice from the same patient within the same age or time interval, the average of these two values was used. For comparisons of the frequencies IFNγ-positive cells between groups, the nonparametric Kruskal–Wallis test with 95% confidence interval was used.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

CMV-specific CD4 T cell function

Intracellular IFNγ after CMV lysate stimulation was measured in 34 CMV-infected children (22 cCMV, 12 pCMV) and eight adults with symptomatic primary CMV infection. Additional blood samples could be obtained in 11 of the children.

Median values of IFNγ-producing CD4 T cells at 0–6, 6–24 and >24 months of age were 0, 1.3 and 4.1 cells/μl in infants with congenital CMV infection and 0, 1.8 and 1.3 cells/μl in infants with postnatal CMV (Fig. 1).

image

Figure 1. Cytomegalovirus (CMV)-specific CD4 T-cell responses in different study groups. Asterisks denote significant differences from levels in adults 0–3 months after symptoms (*P < 0.05, **P < 0.01, ***P < 0.001).

Download figure to PowerPoint

One of the eight adults with primary CMV infection could only be analysed once. In the remaining seven, the levels of IFNγ-producing CD4 T cells were generally high initially and then decreased with time (Fig. 2). The median level within the first 3 months after appearance of symptoms was 9.7 IFNγ+ cells/μl.

image

Figure 2. High initial and decreasing levels of CD4 T-cell responses in eight adults with primary cytomegalovirus (CMV) infection. Intracellular IFN-γ is measured after in vitro stimulation with CMV lysate (see 'Materials and methods').

Download figure to PowerPoint

We found no difference between children with congenital and postnatal CMV infection in any of the three age groups (P > 0.05), and there was no IFNγ production in CD4 T cells in any of the CMV-negative children.

CMV-specific CD8 T cell function

Overlapping peptide pools covering the entire pp65 and IE1 proteins, respectively, were used to activate CMV-specific CD8 T cells in children with congenital and postnatal CMV infection. Results are displayed in Fig. 3. In children under 2 years of age, intracellular IFNγ was detectable in response to pp65 peptides in 22 of 24 children with congenital CMV infection. The median values for pp65 peptides in children 0–6, 6–24 and >24 months of age with congenital CMV infection were 1.9, 4.8 and 1.9 cells/μl (cCMV) and 4.6, 7.8 and 3.2 cells/μl (pCMV). For the IE1 peptides, the corresponding median values were 2.8, 8.6 and 1.5 cells/μl (cCMV) and 12.5, 7.4 and 9.4 cells/μl (pCMV). No differences could be found between children with congenital and postnatal CMV infection or between different age groups (P > 0.05).

image

Figure 3. CD8 T-cell responses to pools of peptides spanning the entire pp65 and IE1 proteins in children with congenital (CCMV) and postnatal (PCMV) CMV infection.

Download figure to PowerPoint

Results in the control group with negative αCMV-IgG or CMV-DNAuria were below detection limits in 13 of 17 children (pp65) and 14 of 14 children (IE1) and never exceeded 0.3% of the CD8 T cells.

Analysis of samples from three pairs of twins

Two boys (cCMV73 and cCMV 77; homozygotic twins) were born after a normal pregnancy of 37 gestational weeks. One of them (cCMV73) had bilateral deafness discovered by routine hearing screening. Both were diagnosed with congenital CMV infection after DBS analysis. At 9 months of age, no CD4 T cell activity could be detected. The twin with hearing deficiency had higher IE1 responses than his brother did (2.9% or 73 cells/μl versus 0.73% or 22 cells/μl).

The second pair of twins (pCMV7 and cCMV8; dizygotic boys) was born after a normal pregnancy in week 36 + 4. Retrospective serum analysis showed that the mother was CMV seropositive in early pregnancy, suggesting a reactivated CMV infection during pregnancy. DBS screening revealed that one of the boys was congenitally infected with CMV. At 2 months, CMV virus isolation in the urine was positive of the other twin, suggesting a postnatally acquired CMV infection. Immunological analyses were performed at 6 and 21 months, showing higher levels of IFNγ-producing CD8 T cells at 21 months than at 6 months. Similary, IFNγ-producing CD4 T cells increased from undetectable (in the twin with cCMV) or just above detection limits (in the twin with pCMV) at 6 months, to approximately 20 cells/μl in both boys at 21 months (Fig. 4). Both boys were healthy and to date – at age five – have shown no signs of sequelae.

image

Figure 4. Cytomegalovirus (CMV)-specific CD4 and CD8 T cell responses in all children with at least two sampings with more than one month inbetween. IFNγ-producing CD8 T-cells when stimulated with peptide pools from CMV proteins IE-1 (white bars) and pp65 (grey bars) and CD4 T-cells in response to CMV lysate (•). CCMV 7 and pCMV 8 are dizygotic twins.

Download figure to PowerPoint

The third pair of twins, a girl (pCMV60) and a boy (pCMV61), was enrolled at 4 months of age based on detectable viruria. Their CMV-specific CD8 T cell function was very similar, with 4.2 and 5 IFNγ-producing cells/μl (0.6 and 0.26%), respectively. The boy lacked detectable CD4 function, but the girl responded with IFNγ production (0.11%, 2.5 cells/μl) after stimulation with CMV lysate.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

The immune responses to CMV have been studied in adults, but mainly in immunocompromised patients and in healthy blood donors with latent infection. There are only scarce data on infants/children and hardly any data from immunocompetent adults with primary CMV infection. Given the childhood morbidity caused by congenital CMV infection, studies for better understanding of pathogenic mechanisms of the immune defence are needed. The lack of studies in infants is likely a result of the practical and ethical difficulties. Firstly, congenital CMV infection is asymptomatic in the majority of newborns, and the infection often goes unrecognised. Newborn screening for congenital CMV infection is the only way to identify the majority of infected infants [6]. Further, limited blood amounts, difficulties in blood sampling and the absolute need for accurate information to the parents are among the factors that make this type of studies a challenge. In the studies performed, CMV-specific immune responses in infants with postnatal CMV infection have been investigated [17, 40], as well as the CD8 T cell responses in children with congenital CMV infection [41, 42]. To our knowledge, this is the first study to investigate both CD4 and CD8 T cell responses in a population of asymptomatic congenitally infected infants identified by screening as well as children with symptomatic congenital CMV infection.

Our data show that in infants up to 2 years of age with congenital or postnatal CMV infection, the CMV-specific CD4 responses are inferior compared with adults during their first 3 months of symptomatic, primary CMV infection. In adults with primary CMV infection, CD4 T cell function peaked at 2–4 weeks after the first symptoms and then declined. Similar data have been reported for immunosuppressed patients who develop primary CMV infection after transplantation [43]. These results are analogous to the pattern observed for CD8 T cells. Our group has previously published data showing that CMV-specific CD8 T cell responses peak within a few weeks after the appearance of symptoms. The levels of HLA A*0101 and A*0201 tetramer-binding cells increased vigorously during the first weeks and then rapidly diminished [35]. Although the magnitude of the up-regulation of CD4 T cell function seems to be modest when compared with the situation in CD8 T cells, it is possible that this pattern is crucial for gaining rapid control over the virus. From data shown in Figs. 2 and 4, it appears that infants fail to do this, but rather seems to slowly increase their CMV-directed CD4 T cell function with time. In adult recipients of renal transplants with primary CMV infection, there is a delay in the CD4 T cell response in the patients with symptomatic infection [44]. It could be speculated that this is also true for infants with congenital CMV, but we found no differences between symptomatic and asymptomatic children – possibly because the groups were too small.

A feasible explanation for the decreased CD4 T cell function in infants is tolerance induction as a result of the intrauterine virus infection. It has been shown that foetuses are capable of generating functional suppressive regulatory T cells directed to non-inherited maternal antigens, and that these cells prevail long after birth [45]. If the same is true for viral antigens, presence of suppressing CMV-specific Treg cells may explain the lack of response in CMV-specific cells. Yet, another explanation is immune escape mechanisms. In the process of achieving successful co-existence with its host, CMV has developed multiple strategies to escape the host's immune system. Among reported mechanisms with direct or indirect effect on CD4 T cells are down-regulation of MHC II expression on dendritic cells [46, 47], decreasing numbers of Langerhans-type dendritic cells [48] and inhibition of differentiation of monocytes into dendritic cells [49]. It is possible that only a mature immune system is capable of overriding these mechanisms, whereas an immature developing immune system is not. The fact that not only infants with congenital infection but also those with CMV infection acquired after birth displayed a low level of CD4 activity (as previously shown by Tu et al. [17]) favours this hypothesis.

One theory has been that passively transferred maternal antibodies inhibit immune responses to antigen stimulation in infancy by antibody interference. It is likely that most of the infants with postnatal CMV infection in our study had maternal CMV-IgG antibodies because the most common way of acquiring CMV infection in infancy is transmission from the mother during birth or through breast feeding [5]. In recent studies on measles and mumps, however, the concept of antibody interference has been questioned. Young infants were shown to elicit an impaired humoral response to viral vaccine, but this happened irrespective of present maternal antibodies. A cellular response was evoked, but the specific CD4 response was lower in infants than in adults [50].

Interestingly, there are obviously children who manage to mobilize efficient CD4 T cell defences against CMV. One asymptomatic congenitally infected infant in our study, cCMV6, already displayed high CD4 activity at 1.5 months of age. This IFNγ production slowly diminished over the course of the next 2 years (Fig. 4). Not surprisingly, the levels and activity of his CMV-specific CD8 T cells were also consistently high.

High levels of CMV-specific CD8 T cells were seen in many of the children. CMV-specific T cell responses have been reported in foetuses as early as 22 weeks of gestation [51]. We found that the CMV-specific CD8 T cell function in children with ongoing CMV infection was similar to that of latently infected adults [36, 38]. This finding is consistent with previously published reports [40, 42, 51].

The relative magnitude of the CD8 responses directed towards IE1 and towards pp65 has been showed to differ between individuals with different HLA types [36, 52]. When the pp65 and IE1 responses were compared in 37 children, the IE1-directed response was higher (0.1% or more) in 16 of the children, the pp65-response was dominant in 12, and the responses were equal in 9. Gibson et al. have reported that IE1-specific responses predominate by 1 year of age in infants with either pCMV or cCMV [42]. In the present study, four of the five congenitally infected children younger than 6 months had undetectable responses towards the IE1 peptides. The impact of age and HLA type is illustrated by the three pairs of twins in our study. The homozygotic brothers (cCMV73, cCMV77) had almost identical CD8 T cell responses at 9 months of age, and the dizygotic brother and sister with pCMV also had similar data. Not surprisingly, the least similar responses were seen when one twin had been infected intrauterinely (cCMV8) and the other twin after birth (pCMV7).

Although congenitally infected infants may suffer permanent sequelae, postnatally infected infants usually do not. It is also much less common that children with congenital maternal recurrent infection develop neurological disabilities compared with children with maternal primary infection. Studies of the immunological response in congenitally infected infants compared with postnatally infected infants may provide information for a better understanding of the pathogenesis of long-term sequelae.

In this study, no difference in specific T cell responses was noted between young infants with congenital CMV infection and postnatally acquired infection. A reason for the higher risk of neurological damage in children with congenital infection might be that the central nervous system of a foetus is more vulnerable to viral infection than that of a full-term newborn. Also, in primary maternal infection, there is a time-window before the mother has developed IgG antibodies which can pass the placenta and protect the foetus. This might lead to a higher viral load and a more longstanding viral replication. Neonatal viremia has been suggested to be a predictive factor of neurological outcome [26].

Without CD4 help, CMV-specific CD8 T cells are apparently unable to halt viral replication [40]. If the virus has time to spread efficiently, producing high levels of viral antigen over a long period of time, this persistence may, in turn, cause clonal T cell exhaustion. Because of current technical limitations, it is not clear whether CMV-specific CD4 T cells in children are indeed formed but are not functional, or whether they fail to develop at all. This ambiguity could potentially be resolved using MHC II tetramers, but the availability of these reagents, in contrast to MHC I tetramers, is limited.

It is unclear how the size and activity of CMV-specific T cell clones are affected in the long run; however, the results presented here indicate that the levels slowly increase to match those observed in adults with latent CMV infection.

In summary, we have found that CD4-function in infants with congenital and postnatal CMV infection is impaired compared with immunocompetent adults. CD8 T cell function, on the other hand, is comparable to the response previously described in adults. These data contribute to a better understanding of the immunological mechanisms involved in congenital and postnatal CMV infection, which will be a prerequisite for new treatment strategies of young children to improve long-term outcome as well as for vaccine development.

Acknowledgment

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

The authors wish to thank Kajsa Gustavsson, Cecilia Ewald and Ylva Bäckman at the children's departments of Uppsala University and Karolinska University Hospitals for including children and providing clinical specimens and Susanne Lindblom for outstanding technical and logistic assistance. We acknowledge Mats Bengtsson and Niclas Ljung for scientific and methodological discussions and Deborah McClellan for editorial assistance. Finally, we express our warmest gratitude to the participating families for making this study possible. The study was supported by grants from the Swedish Research Council, Samariten Foundation, and Olinder-Nielsen Foundation.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References