Cytomegalovirus Infection in Gambian Infants Leads to Profound CD8 T-Cell Differentiation

JOURNAL OF VIROLOGY, June 2007, p. 5766–5776 Vol. 81, No. 110022-538X/07/$08.00�0 doi:10.1128/JVI.00052-07Copyright © 2007, American Society for Microbiology. All Rights Reserved.

Cytomegalovirus Infection in Gambian Infants Leads to Profound CD8T-Cell Differentiation�

David J. C. Miles,1* Marianne van der Sande,2 David Jeffries,1 Steve Kaye,3 Jamila Ismaili,4Olubukola Ojuola,5 Mariama Sanneh,1 Ebrima S. Touray,1 Pauline Waight,6

Sarah Rowland-Jones,1 Hilton Whittle,1 and Arnaud Marchant7

MRC Laboratories Gambia, P.O. Box 273, Banjul, The Gambia1; National Institute for Public Health and the Environment (RIVM),Bilthoven, The Netherlands2; Jefferiss Trust Laboratory, Imperial College London, London, United Kingdom3;

I.B.L. Institut Pasteur, Lille, France4; Department of Pediatrics, Bronx Lebanon Hospital Center,1650 Selwyn Avenue, Bronx, New York 104575; Immunisation Department, Health Protection Agency

Centre for Infections, London, United Kingdom6; and Institute for Medical Immunology,Universite Libre de Bruxelles, Charleroi, Belgium7

Received 9 January 2007/Accepted 9 March 2007

Cytomegalovirus (CMV) infection is endemic in Gambian infants, with 62% infected by 3 months and 85%by 12 months of age. We studied the CD8 T-cell responses of infants to CMV following primary infection.CMV-specific CD8 T cells, identified with tetramers, showed a fully differentiated phenotype (CD28� CD62L�

CD95� perforin� granzyme A� Bcl-2low). Strikingly, the overall CD8 T-cell population developed a similarphenotype following CMV infection, which persisted for at least 12 months. In contrast, primary infection wasaccompanied by up-regulation of markers of activation (CD45R0 and HLA-D) on both CMV-specific cells andthe overall CD8 T-cell population and division (Ki-67) of specific cells, but neither pattern persisted. At 12months of age, the CD8 T-cell population of CMV-infected infants was more differentiated than that ofuninfected infants. Although the subpopulation of CMV-specific cells remained constant, the CMV peptide-specific gamma interferon response was lower in younger infants and increased with age. As the CD8 T-cellphenotype induced by CMV is indicative of immune dysfunction in the elderly, the existence of a similarphenotype in large numbers of Gambian infants raises the question of whether CMV induces a similarlydeleterious effect.

Human cytomegalovirus (CMV) is endemic throughout theworld, with infection typically occurring earlier in life in low-income countries than in the industrialized nations where it hasbeen most intensively studied (29). Congenital infection is of-ten associated with pathology, while early postnatal infection,which is usually transmitted by breast-feeding, is normallyasymptomatic (17).

The CD8 T-cell response has been associated with protec-tion from CMV disease following some organ transplants (5, 6,44). However, a vigorous CD8 T-cell response alone fails tocontrol primary CMV infection transmitted during kidneytransplantation.

The phenotype of CMV-specific CD8 T cells in chronicinfection is well documented and shows characteristics distinctfrom cells specific for other viruses. The vast majority loseexpression of the B7 ligand CD28, and a large number loseexpression of the costimulatory molecule CD27 (1) and expressthe RA isoform of CD45 during chronic infection (7, 50).Many CMV-specific CD8 T cells express Ki-67 in the earlystages of infection (10), indicating that they are in the processof mitotic division (12), and the number of dividing cells wanesover time.

Chronic infection with CMV typically leads to large expan-

sions of CMV-specific T cells. An average of around 4% ofCD8 T cells can produce gamma interferon (IFN-�) in re-sponse to CMV (39). This probably underestimates the size ofthe CMV-specific CD8 population, as clones of a similar mag-nitude that are specific for single peptides have been detectedby tetramer staining in children and adults (8, 24). The pro-found effect that CMV has on the overall CD8 T-cell popula-tion is evident from the substantially higher proportion ofdifferentiated CD28� cells present in CMV-infected comparedto non-CMV-infected adults (26, 28, 48).

Considerably fewer studies have been carried out on chil-dren than adults, although the accumulation of differentiatedCD28� CD8 T cells has been shown in children as young as 18months (8) and in umbilical cord blood samples from congen-itally infected infants (27). In the latter study, the CD8 T cellsshowed cytolytic activity and were able to produce IFN-�,implying that their mature phenotype was matched by somedegree of functionality. However, studies on other viral infec-tions have shown that infant CD8 T cells are less able toproduce IFN-� (19, 25, 35) in response to antigenic stimulationthan those of adults, implying that infant CD8 T cells are notas functionally mature as their phenotype suggests.

More than half of infants in The Gambia are infected withCMV by 6 months of age (2). Consequently, CMV is one of thefirst challenges that their immune systems encounter, and in-vestigation of the response of infant T cells to CMV may offerinsights into how infants respond to natural infections.

We recruited subjects from a birth cohort established in

* Corresponding author. Present address: MLW Clinical ResearchProgramme, P.O. Box 30096, Chichiri, Blantyre 3, Malawi. Phone: 2651 87 6444. Fax: 265 1 87 5774. E-mail: [email protected].

� Published ahead of print on 21 March 2007.

5766

at University of Liverpool Library on M

ay 13, 2007 jvi.asm

.orgD

ownloaded from

http://jvi.asm.org

Sukuta, The Gambia, in order to determine the onset of in-fection in infants and to evaluate the subsequent CD8 T-cellresponse.

MATERIALS AND METHODS

Subjects. Subjects were recruited at birth from the maternity ward of SukutaHealth Centre as part of an ongoing birth cohort. Informed consent was obtainedfrom their mothers and documented by signature or thumb print. Recruitmentwas restricted to healthy singleton infants, defined by a birth weight of at least1.95 kg and no congenital abnormalities. The catchment area of Sukuta HealthCentre is periurban and characterized by low income and crowded living condi-tions. The human immunodeficiency virus statuses of the study subjects wereunknown, but adult prevalence in the region was 1 to 4% at the time of the study(National AIDS Control Program, unpublished data) and so was not thought tobe a significant confounder of this study.

Mononuclear cells, plasma, and buffy coats were separated from umbilicalcord blood and stored. Limited molecular human leukocyte antigen (HLA)typing was carried out on the buffy coats (4) in order to identify subjects withHLA-A2, -B7, -B8, and -B35, for which known epitope peptides were available.

In total, 258 infants were followed up for 12 months to establish the age atCMV infection. A subset of 116 infants was selected for studies on the devel-opment of the CD8 T-cell response to CMV, which was biased toward HLAtypes for which peptides were available.

The prevalence of CMV in the adult population was established by serologicaldetection of CMV-specific immunoglobulin G (IgG) in a subsample of 194mothers, using ELICYTOK-G enzyme-linked immunosorbent assay (Diasorin).All samples were seropositive.

The study was approved by the Medical Research Council-Gambian Govern-ment Ethics Committee.

Diagnosis of CMV infection and blood-sampling schedule. To determine theonset of CMV infection, urine samples were collected at birth, fortnightly forthe first 3 months of life, and monthly until the age of 12 months and tested forthe presence of CMV DNA by PCR. If it was detected, a second sample wastaken immediately to confirm infection, and urine sampling was discontinued.Excretion in the urine is usually continuous for at least 6 months followinginfection (D. J. C. Miles and S. Kaye, unpublished data). Congenital infectionwas defined by CMV detection in a urine sample collected in the first 2 weeks oflife.

Following CMV diagnosis, a blood sample was taken as soon as possible andfollow-up samples were scheduled for 4 and 12 months after diagnosis. Sampleswere collected in heparin and transported to the laboratory within 4 h. If aninfant reached 12 months of age without CMV DNA being detected, a bloodsample was collected at that time.

Whole blood was used for flow cytometry, and peripheral blood mononuclearcells (PBMCs) were separated on lymphoprep (Axis-Shield PoC AS) columns foruse in enzyme-linked immunospot (ELISpot) assays.

Flow cytometry. All flow cytometry reagents were from Becton Dickinson. TheCD8 population was identified using peridinin chlorophyll-conjugated anti-CD8antibody. Preliminary experiments demonstrated that it was possible to identifya pure population of CD3� CD8high T cells and that cells binding tetramersexpressed CD8 at the same high level. Further phenotyping for surface markerswas carried out using allophycocyanate-conjugated antibodies to CD27,CD45R0, and CD62L and fluorescein isothiocyanate-conjugated antibodies toCD28, CD45RA, CD95, and the class II HLA-D molecules DR, DP, and DQ.Intracellular markers were identified using fluorescein isothiocyanate-conjugatedantibodies to Bcl-2, Ki-67, perforin, and granzyme A. All antibodies were titratedprior to use to ensure that they were used at their optimal concentrations. Wherepossible, CMV-specific cells were identified with phycoerythrin-conjugated B7-TPRVGGGAM and A2-NLVPMVATV tetramers, synthesized as previouslydescribed (13, 27). New batches of tetramer were initially used in parallel witholder batches to ensure consistent staining. Red blood cells were lysed usingFACSlyse solution, and lymphocytes were permeabilized using FACSperm II.

Statistical analysis of flow cytometry data. As the study involved the collectionof over 1,800 flow cytometry samples over a prolonged period of time, and thestudy design involved both cross-sectional and longitudinal data, objectivelydefining subpopulations of interest was a major challenge. In order to eliminatethe inconsistencies and possible bias inherent in a human operator repetitivelyanalyzing large amounts of data, automatic algorithms were used to defineclusters for various combinations of markers using MATLAB 6.1 (The Math-works Inc). For each marker set, the same algorithm was applied to each subjectto maintain objectivity and consistency of the flow cytometry analysis. As no

objective criteria exist for the definition of the subpopulations described in thepresent study, we made the assumption that the algorithms were valid if theyrecognized the same subpopulations as an experienced flow cytometry operator.

Lymphocyte populations were identified using a combination of contouringand morphological operators on the pixels of the front versus side scatter plot.After gating this population, similar algorithms were used to identify CD8 T-cellpopulations and tetramer� cells. This part of the process was tested on a smallnumber of samples stained for CD3 and CD8, and it was established that whilethere was a substantial overlap between the expression of CD8 on CD3� andCD3� cells, �90% of the CD8� cells identified were CD3� T cells, and thispopulation included an average of 72% of all CD8� CD3� cells (data notshown).

Cells stained with CD27 and CD28, CD27 and perforin, and CD27 and gran-zyme A were divided into three subpopulations, and cells stained with CD95 andCD62L were divided into two subpopulations using a fuzzy c-means clusteringalgorithm. The extremities of the population region were defined by the furthestoutliers.

Cells expressing Ki-67 and cells expressing either high or low levels of Bcl-2were identified using morphological operators to locate a region of highest eventdensity. Where this was not possible, a weighted average of the events was usedto separate the two regions.

For cells stained with HLA-D and CD45R0, an area of highest event densitywas identified where possible using algorithms similar to those used for Ki-67 andBcl2, and the populations were then classified using a cross-hair placed withhorizontal and vertical positions defined by the extreme borders of the densestregion.

For cells stained with CD45RA and CD45R0, a breakpoint regression throughthe smoothed trend of CD45RA expression was used to identify at which pointonly CD45RA was expressed, thus defining a CD45RAbright CD45R0� popula-tion. To further explore the phenotype of the cells as defined by the expressionof CD45 isoforms, a line of regression was plotted through the population of cellsexpressing both CD45RA and CD45R0, and the midpoint was calculated. A“CD45 ratio” was calculated by dividing the distance from the midpoint of theregression line to the border with the CD45RAbright CD45R0� subpopulation bythe distance from the midpoint to the event with the lowest expression ofCD45RA as defined by fluorescence intensity. A relatively high value for theCD45 ratio indicated a high level of expression of CD45R0 and a low level ofexpression of CD45RA.

A conservative approach was used for clustering definition, and for eachalgorithm limits were set to define a successful and meaningful cluster/regiondefinition. Of the 1,886 samples scanned, 130 (6.9%) were discarded becausethey could not be characterized after an iterative process in which scans weremanually checked to identify obvious algorithm failure.

For each set of markers, appropriate region statistics were defined, and de-pending on their distribution, they were summarized using means and medians.The within-subject repeated measurements were accounted for using panel-averaged models (fitted using the generalized estimating equation function inStata 8). Where necessary, the region statistics were appropriately transformedto improve the normality and constant variance assumptions. Due to the smallnumber of time points, the most appropriate correlation structure was the ex-changeable model, which acknowledges the correlation but assumes it is thesame between all time points. The significance of differences over time wasquantified by a Wald test, and if significant at the 5% level, appropriate contrastswere calculated and P values were adjusted using a step-down Bonferronimethod.

Diagnosis of EBV infection. A subset of plasma samples from 12-month-oldinfants was tested for Epstein-Barr virus (EBV) seropositivity. To allow for thepersistence of maternal IgG, an infant was regarded as infected if he/she ex-pressed IgM to the viral capsid antigen (VCA) at 12 months or a higher level ofIgG to the VCA antigen than was present in umbilical cord plasma collected. Thetests were carried out using the ETI-VCA-G and ETI-EBV-M enzyme-linkedimmunosorbent assays (Diasorin).

ELISpot. A kit for the detection of IFN-� (Mabtech) was used in ELISpotassays. The peptides used were the A2-restricted MLNIPSINV, NLVPMVATV,VLGPISGHV, YILEETSVM, and IAGNSAYEYV; the B7-restricted RPHERNGFTV, TPRVTGGGAM, and CRVLCCYVL; the B35-restricted IPSINVHHY and VFPTKDVAL; and the B8-restricted DANDIYRIF (11, 22, 33, 38,42, 47, 49).

A detectable response was defined as a mean treatment spot count that wasmore than 10 spots 105 cells�1 greater than twice the mean negative control.These criteria have previously been found to provide high specificity while al-lowing reasonable sensitivity in the detection of peptide-specific responses (18).Responses were measured as specific spot-forming units, calculated as the dif-

VOL. 81, 2007 INFANT CD8 T-CELL RESPONSE TO CMV 5767


ay 13, 2007 jvi.asm

.orgD

ownloaded from

http://jvi.asm.org

ference between the mean treatment spot count and the mean negative controlspot count. Data were analyzed using Stata 8 and MINITAB 14.

RESULTS

Most infants are CMV infected by 12 months of age. Con-genital infection, defined as excretion of CMV in the urinewithin the first 2 weeks of life, was detected in 4.1% of theinfants. Prevalence increased most rapidly between 6 and 12weeks of age, by which time 62% were infected. Thereafter,prevalence continued to rise steadily until sampling concludedat 12 months, by which time 85% were infected (unpublisheddata).

Sampling for CD8 T-cell studies. Longitudinal studies onthe CD8 T-cell response to CMV infection were carried out on98 infants. From these subjects, 64 samples were taken imme-diately after diagnosis, 92 4 months later, and 79 12 monthslater, so that 235 of the 294 scheduled samples were collectedsuccessfully. A further 18 samples were collected from infantswho had reached 12 months of age without being infected withCMV, so the total cohort numbered 116 infants.

Frequencies of peptide-specific CD8 T cells were very vari-able. Of 67 subjects who were identified as expressing HLA-A2, at least one sample containing CD8 T cells that bound theA2-NLVP tetramer was collected from 20 (30%). Of the 37subjects who expressed HLA-B7, at least one sample thatbound the B7-TPRV tetramer was collected from 16 (43%).The mean frequency of CD8 T cells that bound either tetramerwas 3.4%, but there was considerable variability, with frequen-cies ranging from 0.05% to as high as 32%.

The age of CMV infection had no influence on the frequencyof tetramer� cells, and there were no significant differences infrequency between the samples collected immediately after diag-nosis and those collected 4 or 12 months later (data not shown).

The phenotypes of tetramer� CD8 T cells followed the sametrends as the overall population. The statistical interactionbetween the phenotypes of tetramer� cells and the overallpopulation was assessed to establish whether the time elapsedsince diagnosis had different effects on the phenotypes of tet-ramer� cells and the overall population. The only significantinteraction found was in the case of the proportion of cellsexpressing Ki-67 (P � 0.015), which is discussed below. For allother parameters measured, changes in the phenotype of thetetramer� cells were reflected by changes in the phenotype ofthe overall CD8 T-cell population, although the actual propor-tions of cells in different subpopulations at any given samplingtime were often very different.

Age at time of diagnosis had no effect on the magnitudes ofeither the tetramer� subpopulations or any of the subpopula-tions defined by phenotype (data not shown).

Activation was highest in the early stages of infection intetramer� cells and the overall CD8 T-cell population. Cellswere stained for CD45RA and CD45R0, and a breakpointregression was used to divide them into a CD45RAbright

CD45R0� subpopulation and a subpopulation expressing ap-proximately reciprocal proportions of both isoforms. Nearly alltetramer� cells expressed CD45R0 (P � 0.0001), which istypically expressed by functional memory or effector cells (37,43), considerably more than the overall CD8 T-cell population(Table 1 and Fig. 1A and B). However, the CD45 ratio among

the subpopulation that expressed both isoforms was not differ-ent between tetramer� cells and the overall CD8 T-cell pop-ulation (Table 1 and Fig. 1C), indicating that while a higherproportion of the tetramer� subpopulation expressed CD45R0than the overall CD8 T-cell population, the level of CD45R0expression did not differ between these populations.

The proportion of cells expressing both isoforms was at itshighest immediately postdiagnosis and was substantially lowerin the samples taken 4 and 12 months later (P � 0.012) (Table1 and Fig. 1B). The CD45 ratio showed the same pattern (P �0.008), indicating that not only the proportion of cells thatexpressed CD45R0 but also the level of CD45R0 expressionwas higher immediately after diagnosis than in the followingmonths.

Cells were also stained with CD45R0 and HLA-D, and thecells that expressed either or both were identified. As withCD45R0, the proportion of cells that expressed HLA-D wasgreater among tetramer� cells than among the whole CD8T-cell population (P � 0.0001), and expression was higherimmediately after diagnosis than 4 or 12 months later (P �0.004) (Table 1 and Fig. 1D and E). The proportion of cellsexpressing both CD45R0 and HLA-D molecules followed thesame pattern (Table 1 and Fig. 1F).

The proportion of CD8 T cells in the mitotic cycle washighest immediately after infection. The age at the time ofdiagnosis did not affect the proportion of cells expressing Ki-67(i.e., in the mitotic cycle) (12) at any of the sampling times(data not shown).

Mitosis was the only variable for which the development ofthe tetramer� population followed a different pattern than theoverall CD8 T-cell population. Immediately after diagnosis,the tetramer� cells exhibited significantly more mitosis than ineither of the two subsequent samples (P � 0.0006) (Table 1and Fig. 2A and B), and more tetramer� cells were undergoingmitosis than in the overall CD8 T-cell population (P � 0.0001).However, the proportion of tetramer� cells in mitosis halvedfrom 31% immediately after diagnosis to 15% 4 months laterand ceased to be different from the overall CD8 T-cell popu-lation.

Among the overall CD8 T-cell population, the highest pro-portion of CD8 T cells in mitosis never exceeded 12%, andthere were no significant differences between the proportion ofcells expressing Ki-67 at different sampling times (Table 1 andFig. 2B).

Expression of CD62L, CD95, and Bcl-2 remained constantin the year following infection with CMV. The CD8 T-cellpopulation was divided into an undifferentiated CD62L� sub-population (20, 43), expressing various levels of CD95 (Fas),and a differentiated CD62L� subpopulation, in which nearlyall cells expressed CD95 (Fig. 3A and B). Nearly all of thetetramer� cells were in the latter subpopulation (P � 0.0001).Differentiation status could also be defined by high and lowlevels of Bcl-2 expression, and nearly all tetramer� cells werein the more differentiated Bcl-2low subpopulation (P � 0.0001)(23) (Table 1 and Fig. 3C and D).

In contrast to CD45 isoforms, HLA-D, and Ki-67, the rela-tive sizes of the subpopulations defined by CD62L and Bcl-2remained constant for 12 months following diagnosis in boththe tetramer� cells and the overall CD8 T-cell population.

5768 MILES ET AL. J. VIROL.


ay 13, 2007 jvi.asm

.orgD

ownloaded from

http://jvi.asm.org

The proportion of differentiated CD8 T cells is constant inthe year following CMV infection, but CD27 is lost progres-sively. Cells stained for the costimulatory molecules CD27 andCD28 were divided into naıve (CD27� CD28�), intermediate(CD27� CD28�), and fully differentiated (CD27� CD28�)subpopulations (1, 16, 43). Few tetramer� cells expressed anyCD28, so they nearly all fell into the last two subpopulations(Table 1 and Fig. 4A). Among the CD8 T-cell population as awhole, cells in the intermediate and fully differentiated popu-lations nearly all expressed perforin (Table 1 and Fig. 4B) andgranzyme A (data not shown).

Age at diagnosis did not affect the relative distribution ofcells between the three populations or their subsequentchanges in phenotype. Similarly, the proportion of cells ex-pressing the naıve phenotype remained constant in the 12months following diagnosis. However, there was a progressive,albeit nonsignificant, reduction of the intermediate populationand a more pronounced increase in the fully differentiatedpopulation among both the tetramer� cells and the overallCD8 T-cell population (P � 0.0003), indicating loss of CD27expression from the differentiated CD8 T-cell population (Ta-ble 1 and Fig. 4A).

The changes that were apparent when the populations were

defined by CD27 and CD28 expression remained when thestain for CD28 was replaced with stains for perforin (Table 1and Fig. 4B) or granzyme A (data not shown).

Production of IFN-� by peptide-specific cells is impaired inyounger infants. The IFN-� production of the PBMCs of sam-ples that responded to tetramers was assessed by ELISpotassay, using the peptide incorporated into the tetramer as thestimulant. The proportion of subjects in which a response wasdetected was higher in the samples collected 12 months afterdiagnosis (16/19) than in those collected immediately (6/11) or4 months (2/11) after diagnosis, though the only significantdifference was between the 4- and 12-month postdiagnosisresponses (Fisher’s exact probability test; P � 0.0012). Asfrequencies of CMV-specific cells remained constant with timesince diagnosis, the increased number of IFN-�-producingPBMCs indicated an increase in the proportion of functionalCD8 T cells. Among the samples in which responses weredetected, there were no effects of either age at time of sam-pling or time since diagnosis on the magnitude of the response(data not shown).

To establish whether the same trend was apparent in a largerdata set, the ELISpot response to peptides with the appropri-ate HLA restriction was evaluated using PBMCs from 18 of the

TABLE 1. Relative sizes of subpopulations of interest within the CD8 T-cell population

Subpopulation CD8 T-cell population

Size

Infected (time postdiagnosis)aUninfectedat 12 moImmediately 4 mo 12 mo

CD45RAbright CD45R0� Tetramer� 10.3 � 8.1Ab 14.5 � 7.6Bb 18.8 � 10.5Bb

Overall 33.0 � 11.1A 37.7 � 10.9B 40.9 � 10.8B 41.3 � 19.3HLA-D� Tetramer� 67.5 � 24.4Ab 55.1 � 26.1Bb 56.6 � 23.9Bb

Overall 28.3 � 16.6A 23.6 � 11.8B 24.0 � 13.6B 22.4 � 20.1CD45RO� HLA-D� Tetramer� 42.0 � 26.0Ab 30.8 � 20.1Bb 34.3 � 20.1Bb

Overall 14.3 � 11.6A 10.7 � 7.6B 9.1 � 6.9B 9.9 � 11.8Ki-67� Tetramer� 31.3 � 22.3Ab 14.5 � 14.5B 10.4 � 9.9B

Overall 12.5 � 9.7 8.4 � 5.2A 8.5 � 5.5A 13.4 � 14.7CD62L� Tetramer� 16.6 � 13.0Ab 10.3 � 8.9Ab 11.9 � 16.3Ab

Overall 64.8 � 12.4A 62.6 � 14.9A 64.8 � 14.7A 73.2 � 13.4c

Bcl-2low Tetramer� 97.1 � 3.2Ab 89.5 � 15.5Ab 84.3 � 23.6Ab

Overall 56.0 � 19.1A 59.3 � 19.7A 55.4 � 22.9A 40.2 � 18.7CD27� CD28� Tetramer� 26.7 � 26.9Ab 15.6 � 18.6Ab 14.4 � 13.0Ab

Overall 59.3 � 19.4A 60.7 � 17.9A 62.8 � 19.8A 86.9 � 20.4c

CD27� CD28� Tetramer� 61.4 � 23.9Ab 60.8 � 18.4ABb 58.3 � 16.4Bb

Overall 28.4 � 14.4A 22.9 � 10.7AB 20.0 � 9.9A 10.6 � 16.7c

CD27� CD28� Tetramer� 10.8 � 5.8Ab 22.2 � 14.6Bb 26.6 � 14.5Bb

Overall 12.3 � 8.1A 16.5 � 10.2B 17.3 � 12.4B 2.5 � 3.8c

CD27� perforin� Tetramer� 31.7 � 28.0Ab 33.4 � 28.0Ab 29.9 � 27.1Ab

Overall 63.5 � 20.0A 60.8 � 18.5A 65.1 � 19.7A 83.8 � 21.3c

CD27� perforin� Tetramer� 53.7 � 25.6Ab 44.1 � 20.9Ab 46.4 � 20.4Ab

Overall 23.8 � 13.8A 21.5 � 10.0A 20.0 � 13.8A 12.2 � 16.1c

CD27� perforin� Tetramer� 14.1 � 8.6Ab 21.8 � 16.0Bb 22.8 � 14.9Bb

Overall 12.6 � 8.4A 17.7 � 10.2B 14.9 � 10.2B 4.0 � 5.3c

CD27� granzyme A� Tetramer� 6.7 � 9.8Ab 6.4 � 12.8Ab 2.7 � 2.7Ab

Overall 51.8 � 20.0A 51.5 � 18.4A 52.7 � 20.6A 69.3 � 22.8c

CD27� granzyme A� Tetramer� 72.7 � 20.7Ab 66.7 � 16.8Bb 67.0 � 15.1Bb

Overall 33.6 � 15.5A 27.8 � 11.8B 26.6 � 11.8B 24.6 � 18.9c

CD27� granzyme A� Tetramer� 17.6 � 10.3Ab 26.1 � 16.5Bb 29.1 � 15.9Bb

Overall 14.7 � 8.7A 20.8 � 10.8B 20.7 � 13.5B 6.1 � 5.1c

a Represented as mean percentages � standard deviations. Different uppercase letters are placed next to subpopulations that are significantly different betweensample times.

b Significant differences in the relative size of the relevant subpopulation between the tetramer� and the overall CD8 T-cell populations.c Significant differences between the relative sizes of the relevant population between infants uninfected with CMV at 12 months and infected infants of the same

age.



ay 13, 2007 jvi.asm

.orgD

ownloaded from

http://jvi.asm.org

50 infants assessed immediately postdiagnosis who had a de-tectable response to at least one peptide, while 17 of 34 (50%)responded at 4 months postdiagnosis and 29 of 65 (44%) re-sponded at 12 months postdiagnosis, indicating no increase inthe number of detectable responses over time since diagnosis.

However, when the magnitudes of the response in thosecases where the response was detectable at all were considered,the age at diagnosis correlated with the IFN-� response at 12months postdiagnosis (Spearman’s r � 0.46; n � 29; P �0.012), although not in the samples collected immediately and4 months postdiagnosis (Fig. 5A, B, and C).

As the age at which the 12-month-postdiagnosis sample wastaken depended on the age at diagnosis, it was necessary toestablish whether the correlation was a direct consequence ofthe age of infection or simply due to IFN-� production increas-ing as the infants aged. To eliminate the effect of age at sam-pling on the magnitude of the IFN-� response, a subset ofinfants who were all the same age of 40 to 60 weeks wasidentified, but it included infants that had been infected either4 or 12 months previously. The infants who were sampled 4months postdiagnosis had been infected at 17 to 39 weeks of

age, and those sampled 12 months postdiagnosis had beeninfected at 4 to 10 weeks of age. There was no difference in themagnitude of response by Mann-Whitney U test (Fig. 5D),indicating that age at diagnosis was not a determinant of theresponse in itself.

In order to establish whether there was an effect of age atsampling, the magnitudes of response at 4 and 12 monthspostdiagnosis were compared longitudinally among the subsetdescribed above. Nearly all showed a higher magnitude ofresponse at 12 months postdiagnosis than at 4 months postdi-agnosis (Wilcoxon statistic � 59; n � 11; P � 0.023) (Fig. 6E).As the increase in the magnitude of the response increasedwith age at sampling but was not related to the age at diagno-sis, the results indicate that the function of the CD8 T cells waslow in early infancy and increased with age.

There was no effect of time since diagnosis on the magni-tudes.

The range of peptides recognized does not alter with age.The number of subjects with the appropriate HLA type thatresponded to any given peptide at any given time point variedfrom 0 to 59%, with an average of 18.5%. The PBMC samples

Per

cen

tag

e

CD

45R

Ab

righ

t CD

45R

O-

0

1.5

3.0

4.5

6.0

CD

45ra

tio

CD

45R

O

CD45RA

Per

cen

tag

e

HL

A-D

+

Per

cen

tag

e

CD

45R

O+H

LA

-DR

+

CD

45R

O

HLA-D

Imm

edia

tepost

dia

gnosi

s4

month

spost

dia

gnosi

s12

month

spost

dia

gnosi

s

A B Cp=0.012

p=0.0003

p=0.008

p=0.0009

p=0.004

p=0.0003

p=0.0003

p=0.0003

CD8+Tetramer- CD8+Tetramer+

D E

Imm

edia

tepost

dia

gnosi

s4

month

spost

dia

gnosi

s12

month

spost

dia

gnosi

s

total tet+ total tet+ total tet+Population:

0

20

40

60

100

80

0

20

40

60

100

80

0

20

40

60

100

80

F

n = 62 21 74 24 75 24 62 21 74 24 75 24

n = 61 20 76 26 75 24 61 20 76 26 75 24total tet+ total tet+ total tet+

26%74% 11%89%

10%12% 25%23%

7% 17%

Imm

edia

tepost

dia

gnosi

s4

month

spost

dia

gnosi

s12

month

spost

dia

gnosi

s

Imm

edia

tepost

dia

gnosi

s4

month

spost

dia

gnosi

s12

month

spost

dia

gnosi

s

total tet+ total tet+ total tet+Population: total tet+ total tet+ total tet+

FIG. 1. CMV-specific CD8 T cells are more activated than the overall CD8 T-cell population. (A) Expression of CD45RA and CD45R0 bytetramer� and tetramer� CD8 T cells, showing the line of best fit and the cutoff used to define the CD45RAbright CD45R0� population. (B) Meanpercentage of CD45RAbright CD45R0� CD8 T cells in tetramer� and total populations for each sampling time. (C) Mean CD45 ratios of tetramer�

and total CD8 T cells at each sample time. (D) Expression of CD45R0 and HLA-DR on tetramer� and tetramer� CD8 T cells, showing the cutoffsused to define the CD45R0� and HLA-DR� populations. (E) Mean percentages of HLA-D� and (F) CD45R0� HLA-DR� CD8 T cells amongtetramer� and total populations at each sample time. The percentages in the scatter plots refer to cells within that quadrant. The boxes indicatemedians and interquartile ranges.



ay 13, 2007 jvi.asm

.orgD

ownloaded from

http://jvi.asm.org

collected immediately after diagnosis responded on average to16% of the peptides with the appropriate restriction, but sam-ples collected 4 months later responded to only 12%. Abroader response was apparent 12 months later, as PBMCsresponded to an average of 28% of peptides tested (data not

shown). These differences were not significant by the Kruskal-Wallis test.

Infection with CMV drives differentiation of the CD8 T-cellsubset over the first year of life. It was not possible to ascertainwhether the changes in the CD8 T-cell subpopulations were a

FIG. 2. CMV-specific CD8 T cells divide more than the overall CD8 T-cell population during the early stages of infection. (A) Expression ofKi-67 by tetramer� and tetramer� CD8 T cells in a representative sample collected immediately postdiagnosis. (B) The expression of Ki-67 forall subjects where tetramer staining was carried out, with differences in expression of Ki-67 between tetramer� cells indicated. There were nodifferences between the overall populations. The boxes indicate medians and interquartile ranges.

FIG. 3. CMV-specific CD8 T cells have a more apoptosis-prone phenotype than the overall CD8 T-cell population. (A) Expression of CD62Land CD95 on tetramer� and tetramer� CD8 T cells, showing the definition of the CD62L� and CD95� CD62L� populations. (B) Meanpercentages of CD62L� CD8 T cells among tetramer� and tetramer� populations at each sample time. (C) Expression of Bcl-2 on tetramer� andtetramer� CD8 T cells, showing the definition of the Bcl-2high and Bcl-2low populations. The percentages refer to proportions of Bcl-2low cells.(D) Mean percentages of Bcl-2low CD8 T cells among tetramer� and tetramer� populations at each sample time. Differences between phenotypesof the total CD8 T-cell population at different sampling times are indicated. The boxes indicate medians and interquartile ranges. The percentagesin the scatter plots refer to adjacent regions.



ay 13, 2007 jvi.asm

.orgD

ownloaded from

http://jvi.asm.org

consequence of CMV infection or of the ontogeny of the CD8T-cell subset by considering only changes within the CD8 T-cell populations of infected infants. Consequently, the overallCD8 T-cell phenotypes of 51- to 60-week-old uninfected in-fants were compared with those of infants of a similar age whohad been diagnosed at least 43 weeks previously.

Over 90% of the CD8 T cells of 10 of 16 (63%) uninfectedinfants from whom data on CD27 and CD28 were collected fellinto the naıve (CD27� CD28�) population, which was a higherproportion of naıve cells than was observed in any CMV-infected infant (P � 0.003) (Table 1 and Fig. 6A). Similarresults were found when the stain for CD28 was replaced witha stain for perforin or granzyme A (Fig. 6B and C).

The more differentiated phenotype of the CD8 T cells ofinfants infected with CMV was also indicated by the greaterexpression of CD95 and loss of expression of CD62L (P �0.024) (Table 1 and Fig. 6D). No differences in the relativeexpression of CD45 isoforms, Bcl-2, HLA-D, or Ki-67 weredetected.

The possibility remained that the differences between in-fants infected or uninfected with CMV were not a direct resultof CMV infection but were due to exposure to a wider range of

immune challenges than those that remained uninfected. IfCMV status is indeed a correlate of exposure, the same isprobably true of EBV. The EBV status of the subset of infantsthat were included in the analyses of the effect of CMV on theCD8 T-cell phenotype was determined, and 29 out of 46 (63%)were found to be infected at the time of sampling. All of thephenotypes that were different between CMV-infected anduninfected infants were tested to establish whether there weresimilar differences according to EBV status, but none werefound (data not shown).

DISCUSSION

The interactions between the CD8 T-cell population ofadults and CMV are well documented (1, 10, 13, 26, 39).Considerably less information is available on the response ofthe CD8 T-cell population of infants. We studied the develop-ment of the CD8 T-cell response of infants in a prospectivecohort. A major technical challenge that we encountered wasthe analysis of the very large volume of flow cytometry dataproduced. There were a number of technical sources of uncer-tainty over nearly 3 years of data collection, as well as the

FIG. 4. CMV-specific CD8 T cells are more differentiated and show higher cytotoxic potential than the remainder of the CD8 T-cell population.(A) Scatter plots of representative samples of tetramer� and tetramer� CD8 T cells stained for CD27 and CD28 with the indicated populationsand the percentage of cells within each population, box plots of all samples indicating medians and interquartile ranges, and differences betweenthe distributions of all populations between sample times indicated. (B) The same data for cells stained for CD27 and perforin. (C) The same datafor cells stained for CD27 and granzyme A. The populations are identified, and the percentages indicate the proportion of cells in each of them.



ay 13, 2007 jvi.asm

.orgD

ownloaded from

http://jvi.asm.org

subjectivity inherent in the usual method of fitting regionsaccording to the opinion of the operator. To address theseissues, we used a set of algorithms that defined subpopulationsby applying repeatable criteria. When developing the algo-rithms, we regarded them as valid if they were in agreementwith the opinions of experienced T-cell biologists, who exam-ined every plot produced. The algorithms had the advantage ofbeing able to rapidly and consistently apply the same criteriafor the definition of a given subpopulation to the entire dataset, which was particularly important given the longitudinalnature of the data collection.

Gambian infants are typically infected with CMV within thefirst year of life, so CMV is one of the first immunologicalinsults that they encounter. Studies on neonates infected con-genitally have shown a mature CD8 T-cell response to CMV(27), and the suggestion that the infant CD8 T-cell responsedevelops in a similar way to that of adults is supported by theobservation that there was no effect of age on the frequency of

tetramer� cells. Further, there was no effect of age at diagnosison any phenotypic marker.

Cells specific for CMV showed higher levels of HLA-D andCD45R0 than the overall CD8 T-cell population, indicating a

FIG. 5. The magnitude of IFN-� responses was greater in olderinfants. There was no correlation between age at the time of diagnosisand the magnitude of IFN-� response in those cases where a peptideinduced a peptide-specific response (A) immediately after diagnosis or(B) 4 months later, but (C) there was a positive correlation by 12months following diagnosis. (D) Samples collected from subjects be-tween 40 and 60 weeks of age who had been infected 4 months pre-viously had an IFN-� response similar to that of infants who had beeninfected 12 months previously. The bars indicate means. (E) Twelve-month samples were collected from infants diagnosed with CMV at 4to 10 weeks when they reached 40 to 60 weeks of age, and 4-monthsamples were collected from infants diagnosed at 17 to 39 weeks whenthey reached the same age. When the 4- and 12-month samples fromsubjects who fell into either of these two categories were compared, agreater IFN-� response was elicited from the 12-month sample innearly all cases. The absence of a bar indicates that the sample wascollected but no response was detected.

FIG. 6. The CD8 T-cell population of CMV-infected infants is moreactivated and differentiated than that of uninfected infants. (A) Expres-sion of CD27 and CD28 in representative samples and all infants sampled,drawn from the groups of CMV-infected and non-CMV-infected infantsat 12 months of age. (B) Expression of CD27 and perforin in represen-tative samples and all infants sampled, drawn from the groups of CMV-infected and non-CMV-infected infants at 12 months of age. (C) Expres-sion of CD27 and granzyme A in representative samples and all infantssampled, drawn from the groups of CMV-infected and non-CMV-in-fected infants at 12 months of age. (D) Expression of CD62L and CD95in representative samples and all infants sampled, drawn from the groupsof CMV-infected and non-CMV-infected infants at 12 months of age. Thepercentages in the scatter plots refer to cells within the adjacent region.The boxes indicate medians and interquartile ranges.



ay 13, 2007 jvi.asm

.orgD

ownloaded from

http://jvi.asm.org

relatively high state of activation. Activation markers declinedafter the first few months of infection, implying a transitionfrom the acute to the chronic phase of infection. It is remark-able that the pattern of change with time after diagnosis thatwas seen in the phenotypes of the tetramer� cells was reflectedin the overall CD8 T-cell population.

A limitation of the tetramers is that they could identify onlythe subpopulation of CMV-specific cells that responded to thepeptide they incorporated, so it was impossible to determinewhat proportion of the CMV-specific population was repre-sented in each individual. As CMV is a large virus composed ofover 200 proteins, that proportion probably varied widely be-tween individuals. Both the proportion of CD8 T cells thatrespond to chronic CMV infection and the range of availableepitopes are typically considerable (39, 45, 46, 49). In thepresent study, the proportion of CMV-specific CD8 T cells wasprobably similarly high, as the median proportion of tetramer�

cells was 3.4% and several individuals sustained tetramer�

populations of over 10% of the total CD8 T-cell populationthroughout the year following infection.

The same trends were seen in the overall populations as inthe tetramer� populations in all characteristics that were stud-ied except for division. It is possible that there was a highenough proportion of CMV-specific cells within the overallpopulation to dominate the trends shown by that population,although it would be impossible to be certain without a methodfor identifying all CMV-specific CD8 T cells. However, it isalso possible that CMV infection generated a cytokine envi-ronment that nonspecifically affected the activity and differen-tiation of T cells. Such bystander activation has been demon-strated in a number of contexts (21, 30, 40).

Tetramer� CD8 T cells took on a differentiated phenotype, asdefined by loss of CD27, CD28, and CD62L (1, 34, 37, 43), and incontrast to the transient expression of markers of activation, suchas HLA-D, the proportion of differentiated cells remained stablefor the year following infection. This observation supports theimplication of the consistent CMV-specific subpopulations that apopulation of CMV-specific CD8 T cells is produced very soonafter infection and maintained thereafter.

The observation that the relative size of the differentiatedCD8 T-cell population was much smaller in uninfected 12-month-old infants than in infected infants suggests that theaccumulation of differentiated CD8 T cells is largely driven byCMV itself. Studies on adults have shown that while olderpeople tend to have more CD28� CD8 T cells than youngerpeople, CMV-infected individuals tend to have a much largerpopulation of CD28� CD8 T cells than their uninfected peers(26, 28, 48). Our findings show that such a difference beginswith primary CMV infection, even when this occurs very earlyin life.

Differentiated CD8 T cells, defined by the loss of CD28,nearly all expressed perforin and granzyme A, indicating thatthe differentiated cells have the potential to function as cyto-toxic effector cells (27). Further, the proportion of cells ex-pressing CD95 or low levels of Bcl-2 remained constant for theyear following diagnosis. Although both of these markers areassociated with activation-induced cell death (3), these sub-populations remained constant while the proportion of divid-ing cells did not, so it is unlikely that apoptosis took place at a

rate as constant as the expression of CD95 and Bcl-2 wouldsuggest.

In the light of the stability of the differentiated CD28�

population, the reduction in the proportion of cells expressingCD45R0 indicated that a substantial proportion of the differ-entiated cells must be CD28� CD45RAbright CD45R0� by 12months after infection. It has been shown that a high propor-tion of CMV-specific CD8 T cells express CD45RA duringchronic infection (7), and it seems that this population ap-peared gradually over the course of the year following infec-tion among both the tetramer binding subpopulations and theoverall CD8 T-cell population.

In spite of the emergence of a population of differentiatedcells immediately after infection, the IFN-� responses wereimpaired in early infancy, even when the presence of a constantpopulation of cells specific for a given peptide was detected bytetramer staining. The difference was particularly evident inthe sample collected 12 months following infection, and thelack of difference between children infected for different peri-ods of time but sampled at the same age indicated that thedifference was due to the age of the infants when sampledrather than the age at infection in itself.

Although an IFN-� response to CMV peptides has beendetected in the umbilical cord blood of congenitally infectedinfants (27), the ability of the T cells of infants to produceIFN-� in response to antigenic stimuli is impaired by the poorinterleukin-12 (14) and CD40 (9) expression of infant dendriticcells, and indeed, a reduced IFN-� response to human immu-nodeficiency virus has been shown in infants below 1 year ofage (25, 35). In spite of the phenotypic maturity of the CD8�

T-cell responses of some infants, it is evident that the cells werefunctionally impaired.

The expansion of the overall CD28� CD8 T-cell subpopu-lation observed in infants under 12 months of age raises someconcerns. A subpopulation of a similar size is characteristic ofAmericans of approximately 70 years of age (15) or 40- to60-year-old Swedes infected with CMV (28). Further, the pres-ence of expanded CD28� CD45RA� CD8 T cells is indicativeof failure to respond to influenza vaccination in elderly people(36). Our findings indicate that CMV infection and the pres-ence of a large population of CD28� CD45RAbright CD8 Tcells are so commonplace as to be normal in Gambian infants.This is probably the case throughout sub-Saharan Africa, asour observations are consistent with the finding that largeproportions of the CD8 T-cell population of young Ethiopianchildren have lost CD27 expression (41). As CMV infection inour study commonly occurred before the age of 12 weeks andimmediately induced a CD28� CD8 T-cell population, mostinfants exhibited indicators for impaired immune function (31)while receiving childhood vaccinations and before their firstexposure to endemic diseases of infancy, such as malaria,pneumonia, and rotavirus. Whether CMV infection interfereswith vaccine responses is not known, although it has beensuggested that there is a finite amount of immunological“space” and that the large clonal expansions of T cells thatCMV induces may occupy so much of it that there is little roomfor the expansion of clones to newly encountered antigens (32).It is also possible that the presence of a large proportion ofdifferentiated CD8 T cells and vaccine failure are unrelated



ay 13, 2007 jvi.asm

.orgD

ownloaded from

http://jvi.asm.org

consequences of events at a more fundamental level of theimmune system.

Although a causal link between CMV infection and im-paired vaccine responses has not been formally demonstratedand studies of infants in the United States (8) and Japan (24)have not associated postnatal CMV infection with poor health,the ubiquity of CMV infection in the African context makes itimportant to establish whether the effect of CMV on CD8T-cell phenotypes is matched by the effect that studies on theelderly imply.

In summary, we found that the CD8 T-cell populations ofCMV-infected infants exhibited high but transient levels ofactivation and that large subpopulations of differentiated cellsemerged rapidly after infection and remained stable for at leasta year following infection. Consequently, CMV-infected in-fants showed considerably more differentiated CD8 T cellsthan uninfected infants of the same age. The phenotypes of theCMV-specific CD8 T cells of infants appeared essentially sim-ilar to those reported in adults, and neither their phenotypenor their subsequent development was affected by the age ofinfection. In spite of the mature phenotype of CMV-specificCD8 T cells, considerably fewer of the cells of younger infantswere able to produce IFN-� in response to CMV-derived pep-tides, indicating that phenotypic maturity was not synonymouswith functional maturity.

ACKNOWLEDGMENTS

We thank the staff of the Sukuta birth cohort, Omar Badjie, FatouBah, Saihou Bobb, Janko Camara, Sulayman Colley, Isatou Drammeh,Mam Maram Drammeh, Baboucarr Jobe, Momodou Jobe, AlbertMagnusen, John S. Mendy, Ngui Ndow, Faalou Njie, Bala Musa Sam-bou, Sarjo Sanneh, and Mamadi Sidibeh. Akram Zaman, Paul Snell,and Sarah Crozier helped with data handling. We are extremely grate-ful for the support of Sally Savage of Sukuta Health Centre and AbiKhan of Western Division Health Team. Louis-Marie Yindom andAssan Jaye carried out HLA typing.

This work was supported by GSK Biologics and MRC.

REFERENCES

1. Appay, V., P. R. Dunbar, M. Callan, P. Klenerman, G. M. Gillespie, L.Papagno, G. S. Ogg, A. King, F. Lechner, C. A. Spina, S. Little, D. V. Havlir,D. D. Richman, N. Gruener, G. Pape, A. Waters, P. Easterbrook, M. Salio,V. Cerundolo, A. J. McMichael, and S. L. Rowland-Jones. 2002. MemoryCD8� T cells vary in differentiation phenotype in different persistent virusinfections. Nat. Med. 8:379–385.

2. Bello, C., and H. Whittle. 1991. Cytomegalovirus infection in Gambian moth-ers and their babies. J. Clin. Pathol. 44:366–369.

3. Budd, R. 2001. Activation-induced cell death. Curr. Opin. Immunol. 13:356–362.

4. Bunce, M., C. O’Neill, M. Barnardo, P. Krausa, M. Browning, P. Morris,and K. Welsh. 1995. Phototyping: comprehensive DNA typing for HLA-A,B, C, DRB1, DRB3, DRB4, DRB5 and DQB1 by PCR with 144 primermixes utilizing sequence-specific primers (PCR-SSP). Tissue Antigens 46:355–367.

5. Bunde, T., A. Kirchner, B. Hoffmeister, D. Habedank, R. Hetzer, G.Cherepnev, S. Proesch, P. Reinke, H.-D. Volk, H. Lehmkuhl, and F. Kern.2005. Protection from cytomegalovirus after transplantation is correlatedwith immediate early 1-specific CD8 T cells. J. Exp. Med. 201:1031–1036.

6. Chalandon, Y., S. Degermann, J. Villard, L. Arlettaz, L. Kaiser, S. Vischer,S. Walter, M. Heemskerk, R. van Lier, C. Helg, B. Chapuis, and E. Roosnek.2006. Pretransplantation CMV-specific T cells protect recipients of T-cell-depleted grafts against CMV-related complications. Blood 107:389–396.

7. Champagne, P., G. Ogg, A. King, C. Knabenhans, K. Ellefsen, M. Nobile, V.Appay, G. Rizzardi, S. Fleury, M. Lipp, R. Forster, S. Rowland-Jones, R.-P.Sekaly, A. McMichael, and G. Pantaleo. 2001. Skewed maturation of mem-ory HIV-specific CD8 T lymphocytes. Nature 410:106–111.

8. Chen, S., W.-W. Tu, M. Sharp, E. Tongson, X.-S. He, H. Greenberg, T.Holmes, Z. Wang, G. Kemble, A.-M. Manganello, S. Adler, C. Dekker, D.Lewis, and A. Arvin. 2004. Antiviral CD8 T cells in the control of primaryhuman cytomegalovirus infection in early childhood. J. Infect. Dis. 189:1619–1627.

9. De Wit, D., V. Olislagers, S. Goriely, F. Vermeulen, H. Wagner, M. Goldman,and F. Willems. 2004. Blood plasmacytoid dendritic cell responses to CpGoligodeoxynucleotides are impaired in human newborns. Blood 103:1030–1032.

10. Gamadia, L., E. Remmerswaal, J. Weel, F. Bemelman, R. van Lier, and I.Ten Berge. 2003. Primary immune responses to human CMV: a critical rolefor IFN-�-producing CD4� T cells in protection against CMV disease.Blood 101:2686–2692.

11. Gavin, M. A., M. J. Gilbert, S. R. Riddell, P. D. Greenberg, and M. J. Bevan.1993. Alkali hydrolysis of recombinant proteins allows for the rapid identi-fication of class I MHC-restricted CTL epitopes. J. Immunol. 151:3971–3980.

12. Gerdes, J., H. Lemke, H. Baisch, H. H. Wacker, U. Schwab, and H. Stein.1984. Cell cycle analysis of a cell proliferation-associated human nuclearantigen defined by the monoclonal antibody Ki-67. J. Immunol. 133:1710–1715.

13. Gillespie, G., M. Wills, V. Appay, C. O’Callaghan, M. Murphy, N. Smith, P.Sissons, S. Rowland-Jones, J. Bell, and P. Moss. 2000. Functional hetero-geneity and high frequencies of cytomegalovirus-specific CD8� T lympho-cytes in healthy seropositive donors. J. Virol. 74:8140–8150.

14. Goriely, S., B. Vincart, P. Stordeur, J. Vekemans, F. Willems, M. Goldman,and D. de Wit. 2001. Deficient IL-12(p35) gene expression by dendritic cellsderived from neonatal monocytes. J. Immunol. 166:2141–2146.

15. Gorozny, J., J. Fulbright, C. Crowson, G. Poland, W. O’Fallon, and C.Weyland. 2001. Value of immunological markers in predicting responsive-ness to influenza vaccination in elderly individuals. J. Virol. 75:12182–12187.

16. Hamann, D., M. Roos, and R. van Lier. 1999. Faces and phases of humanCD8� T-cell development. Immunol. Today 20:177–180.

17. Hamprecht, K., J. Maschmann, M. Vochem, K. Dietz, C. Speer, and G. Jahn.2001. Epidemiology of transmission of cytomegalovirus from mother to pre-term infant by breastfeeding. Lancet 357:513–518.

18. Hill, P., R. Brookes, A. Fox, K. Fielding, D. Jeffries, D. Jackson-Sillah, M.Lugos, P. Owiafe, S. Donkor, A. Hammond, J. Otu, T. Corrah, R. Adegbola,and K. McAdam. 2004. Large-scale evaluation of enzyme-linked immuno-spot assay and skin test for diagnosis of Mycobacterium tuberculosis infectionagainst a gradient of exposure in The Gambia. Clin. Infect. Dis. 38:966–973.

19. Jaimes, M., O. Rojas, A. Gonzalez, I. Cajiao, A. Charpilienne, P. Pothier, E.Kohli, H. Greenberg, M. Franco, and J. Angel. 2002. Frequencies of virus-specific CD4� and CD8� T lymphocytes secreting gamma interferon afteracute natural rotavirus infection in children and adults. J. Virol. 76:4741–4749.

20. Kaech, S., E. Wherry, and R. Ahmed. 2002. Effector and memory T-celldifferentiation: implications for vaccine development. Nat. Rev. Immunol.2:251–262.

21. Kanai, T., E. Thomas, Y. Yasutomi, and N. Letvin. 1996. IL-15 stimulates theexpansion of AIDS virus-specific CTL. J. Immunol. 157:3681–3687.

22. Kern, F., I. P. Surel, N. Faulhaber, C. Frommel, J. Schneider-Mergener, C.Schonemann, P. Reinke, and H.-D. Volk. 1999. Target structures of theCD8�-T-cell response to human cytomegalovirus: the 72-kilodalton majorimmediate-early protein revisited. J. Virol. 73:8179–8184.

23. Khan, N., N. Shariff, M. Cobbold, R. Bruton, J. Ainsworth, A. Sinclair, L.Nayak, and P. Moss. 2002. Cytomegalovirus seropositivity drives the CD8 Tcell repertoire toward greater clonality in healthy elderly individuals. J Im-munol 169:1984–1992.

24. Komatsu, H., A. Inui, T. Sogo, T. Fujisawa, H. Nagasaka, S. Nonoyama, S.Sierro, J. Northfield, M. Lucas, A. Vargas, and P. Klenerman. 2006. Largescale analysis of pediatric antiviral CD8� T cell populations reveals sus-tained, functional and mature responses. Immun. Ageing 3:11.

25. Lohman, B., J. Slyker, B. Richardson, C. Farquhar, J. Mabuka, C. Crudder,T. Dong, E. Obimbo, D. Mbori-Ngacha, J. Overbaugh, S. Rowland-Jones,and G. John-Stewart. 2005. Longitudinal assessment of human immunode-ficiency virus type 1 (HIV-1)-specific gamma interferon responses during thefirst year of life in HIV-1-infected infants. J. Virol. 79:8121–8130.

26. Looney, R., A. Falsey, D. Campbell, A. Torres, J. Kolassa, C. Brower, R.McCann, M. Menegus, K. McCormick, M. Frampton, W. Hall, and G.Abraham. 1999. Role of cytomegalovirus in the T cell changes seen in elderlyindividuals. Clin. Immunol. 90:213–219.

27. Marchant, A., V. Appay, M. van der Sande, N. Dulphy, C. Liesnard, M. Kidd,S. Kaye, O. Ojuola, G. Gillespie, A. Vargas Cuero, V. Cerundolo, M. Callan,K. McAdam, S. Rowland-Jones, C. Donner, A. McMichael, and H. Whittle.2003. Mature CD8� T lymphocyte response to viral infection during fetallife. J. Clin. Investig. 111:1747–1755.

28. Olsson, J., A. Wikby, B. Johansson, S. Lofgren, B.-O. Nilsson, and F.Ferguson. 2000. Age-related change in peripheral blood T-lymphocyte sub-populations and cytomegalovirus infection in the very old: the Swedish lon-gitudinal OCTO immune study. Mech. Ageing Dev. 121:187–201.

29. Pass, R. F. 1985. Epidemiology and transmission of cytomegalovirus. J. In-fect. Dis. 152:243–248.

30. Patki, A., M. Quinones-Mateu, D. Dorazio, B. Yen-Lieberman, W. Boom, E.Thomas, and M. Lederman. 1996. Activation of antigen-induced lymphocyteproliferation by interleukin-15 without the mitogenic effect of interleukin-2that may induce human immunodeficiency virus-1 expression. J. Clin. Inves-tig. 98:616–621.



ay 13, 2007 jvi.asm

.orgD

ownloaded from

http://jvi.asm.org

31. Pawelec, G., A. Akbar, C. Caruso, B. Grubeck-Loebenstein, R. Solana, andA. Wikby. 2005. Human immunosenescence: is it infectious? Immunol. Rev.25:257–268.

32. Pawelec, G., and C. Gouttefangeas. 2006. T-cell dysregulation caused bychronic antigenic stress: the role of CMV in immunosenescence? Aging Clin.Exp. Res. 18:171–173.

33. Retiere, C., V. Prod’homme, B.-M. Imbert-Marcille, M. Bonneville, H. Vie,and M.-M. Hallet. 2000. Generation of cytomegalovirus-specific human T-lymphocyte clones by using autologous B-lymphoblastoid cells with stableexpression of pp65 or IE1 proteins: a tool to study the fine specificity of theantiviral response. J. Virol. 74:3948–3952.

34. Sallusto, F., D. Lenig, R. Forster, M. Lipp, and A. Lanzavecchia. 1999. Twosubsets of memory T lymphocytes with distinct homing potentials and effec-tor functions. Nature 401:708–712.

35. Sandberg, J., N. Fast, K. Jordan, S. Furlan, J. Barbour, G. Fennelly, J.Dobroszycki, H. Spiegel, A. Wiznia, M. Rosenberg, and D. Nixon. 2003.HIV-specific CD8� T cell function in children with vertically acquiredHIV-1 infection is critically influenced by age and the state of the CD4� Tcell compartment. J. Immunol. 170:4403–4410.

36. Saurwein-Teissl, M., T. Lung, F. Marx, C. Gschosser, E. Asch, I. Blasko, W.Parson, G. Bock, D. Schonitzer, E. Trannoy, and B. Grubeck-Loebenstein.2002. Lack of antibody production following immunization in old age: asso-ciation with CD8� CD28� T cell clonal expansions and an imbalance in theproduction of Th1 and Th2 cytokines. J. Immunol. 168:5893–5899.

37. Seder, R., and R. Ahmed. 2003. Similarities and differences in CD4� andCD8� effector and memory T cell generation. Nat. Immunol. 4:835–842.

38. Solache, A., C. L. Morgan, A. I. Dodi, C. Morte, I. Scott, C. Baboonian, B.Zal, J. Goldman, J. E. Grundy, and J. A. Madrigal. 1999. Identification ofthree HLA-A*0201-restricted cytotoxic T cell epitopes in the cytomegalovi-rus protein pp65 that are conserved between eight strains of the virus.J. Immunol. 163:5512–5518.

39. Sylwester, A., B. Mitchell, J. Edgar, C. Taormina, C. Pelte, F. Ruchti, P.Sleath, K. Grabstein, Hosken, F. Kern, J. Nelson, and L. Picker. 2005.Broadly targeted human cytomegalovirus-specific CD4� and CD8� T cellsdominate the memory compartments of exposed subjects. J. Exp. Med.202:673–685.

40. Tough, D. F., P. Borrow, and J. Sprent. 1996. Induction of bystander T cellproliferation by viruses and type I interferon in vivo. Science 272:1947–1950.

41. Tsegaye, A., D. Wolday, S. Otto, B. Petros, T. Assefa, T. Alebachew, E. Hailu,F. Adugna, W. Measho, W. Dorigo, A. L. Fontanet, D. van Baarle, and F.

Miedema. 2003. Immunophenotyping of blood lymphocytes at birth, duringchildhood, and during adulthood in HIV-1-uninfected Ethiopians. Clin. Im-munol. 109:338–346.

42. Utz, U., S. Koenig, J. E. Coligan, and W. E. Biddison. 1992. Presentation ofthree different viral peptides, HTLV-1 Tax, HCMV gB, and influenza virusM1, is determined by common structural features of the HLA-A2.1 mole-cule. J. Immunol. 149:214–221.

43. van Lier, R., I. ten Berge, and L. Gamadia. 2003. Human CD8� T-celldifferentiation in response to viruses. Nat. Rev. Immunol. 3:1–8.

44. Walter, E., P. Greenberg, M. Gilbert, R. Finch, K. Watanabe, E. Thomas,and S. Riddell. 1995. Reconstitution of cellular immunity against cytomeg-alovirus in recipients of allogenic bone marrow by transfer of T-cell clonesfrom the donor. N. Engl. J. Med. 333:1038–1044.

45. Weekes, M., A. Carmichael, M. Wills, K. Mynard, and J. Sissons. 1999.Human CD28� CD8� T cells contain greatly expanded functional virus-specific memory CTL clones. J. Immunol. 162:7569–7577.

46. Weekes, M., M. Wills, K. Mynard, R. Hicks, J. Sissons, and A. Carmichael.1999. Large clonal expansions of human virus-specific memory cytotoxic Tlymphocytes within the CD57� CD28�CD8� T-cell population. Immunol-ogy 98:443–449.

47. Weekes, M. P., M. R. Wills, K. Mynard, A. J. Carmichael, and P. J. G.Sissons. 1999. The memory cytotoxic T-lymphocyte (CTL) response to hu-man cytomegalovirus infection contains individual peptide-specific CTLclones that have undergone extensive expansion in vivo. J. Virol. 73:2099–2108.

48. Wikby, A., B. Johansson, J. Olsson, S. Lofgren, B. O. Nilsson, and F.Ferguson. 2002. Expansions of peripheral blood CD8 T-lymphocyte sub-populations and an association with cytomegalovirus seropositivity in theelderly: the Swedish NONA immune study. Exp. Gerontol. 37:445–453.

49. Wills, M. R., A. J. Carmichael, K. Mynard, X. Jin, M. P. Weekes, B. Plachter,and P. J. G. Sissons. 1996. The human cytotoxic T-lymphocyte (CTL) re-sponse to cytomegalovirus is dominated by structural protein pp65: fre-quency, specificity, and T-cell receptor usage of pp65-specific CTL. J. Virol.70:7569–7579.

50. Wills, M. R., G. Okecha, M. P. Weekes, M. K. Gandhi, P. J. Sissons, and A. J.Carmichael. 2002. Identification of naive or antigen-experienced humanCD8� T cells by expression of costimulation and chemokine receptors:analysis of the human cytomegalovirus-specific CD8� T cell response. J. Im-munol. 168:5455–5464.



ay 13, 2007 jvi.asm

.orgD

ownloaded from

http://jvi.asm.org

Documents

Cytomegalovirus Infection in Gambian Infants Leads to Profound CD8 T-Cell Differentiation