Latent CMV Infection Is Associated With Lower Influenza Virus-Specific Memory T-Cell Frequencies, but Not With an Impaired T-Cell Response to Acute Influenza Virus Infection

Latent infection with cytomegalovirus (CMV) is assumed to contribute to the age-associated decline of the immune system. CMV induces large changes in the T-cell pool and may thereby affect other immune responses. CMV is expected to impact especially older adults, who are already at higher risk of severe disease and hospitalization upon infections such as influenza virus (IAV) infection. Here, we investigated the impact of CMV infection on IAV-specific CD8+ T-cell frequencies in healthy individuals (n=96) and the response to IAV infection in older adults (n=72). IAV-specific memory T-cell frequencies were lower in healthy CMV+ older individuals compared to healthy CMV- older individuals. Upon acute IAV infection, CMV serostatus or CMV-specific antibody levels were not negatively associated with IAV-specific T-cell frequencies, function, phenotype or T-cell receptor repertoire diversity. This suggests that specific T-cell responses upon acute IAV infection are not negatively affected by CMV. In addition, we found neither an association between CMV infection and inflammatory cytokine levels in serum during acute IAV infection nor between cytokine levels and the height of the IAV-specific T-cell response upon infection. Finally, CMV infection was not associated with increased severity of influenza-related symptoms. In fact, CMV infection was even associated with increased IAV-specific T-cell responses early upon acute IAV infection. In conclusion, although associated with lower frequencies of memory IAV-specific T cells in healthy individuals, CMV infection does not seem to hamper the induction of a proper T-cell response during acute IAV infection in older adults.


INTRODUCTION
The worldwide population is ageing rapidly. With age, deleterious changes in the immune system arise, referred to as 'immunosenescence', which impair responses against infectious diseases and vaccinations (1,2). Age-related changes of the immune system mainly occur in the T-cell pool, including an increase in the number of CD45RA + memory T cells (3,4) and decreased diversity of the T-cell receptor repertoire (5,6). About 2 decades ago, latent infection with cytomegalovirus (CMV) was implicated as a possible driving force of these age-related changes (7)(8)(9). CMV seropositivity was identified as part of the so-called 'immune risk profile', predictive of early mortality in older adults (7,10). Moreover, CMV seropositivity is the largest non-heritable factor influencing differences among humans in the immune profile (11). CMV is generally thought to establish this large effect by its frequent attempts to reactivate during life-long carriage, thereby gradually affecting the immune system (12). Since the 2000s, it is often hypothesized that CMV might impair the human immune response to a heterologous challenge (6,12,13), as was shown in mice (14)(15)(16).
With age, the risk for serious complications and hospitalization after influenza virus (IAV) infection increases (17). Vaccination is an important tool to prevent infection, however also the efficacy of influenza vaccination decreases with age (2,18). The role of CMV infection on influenza vaccination efficacy has been well studied. These studies have yielded conflicting results as a negative effect of CMV (19)(20)(21), a positive effect (1) as well as no effect of CMV on the vaccine response have been reported (22,23). A systematic review by our group including a meta-analysis showed no clear evidence for a negative effect of CMV infection on the antibody response to influenza vaccination (24).
Both CMV and ageing primarily affect the T-cell compartment (3,4,(25)(26)(27). During IAV infection, an effect of CMV would therefore mainly be expected on the T-cell response. T cells play an important role in clearance of the IAV (28). These T-cell responses are predominantly specific for the internal viral proteins, such as matrix protein-1 (29), which are conserved among influenza strains that have undergone antigenic drift. Indeed, pre-existing and early IAV-specific T-cell responses are associated with lower disease severity of influenza (30,31), while delayed T-cell responses to IAV are thought to induce prolonged inflammation and delayed viral clearance and recovery (31,32). Upon activation, IAV-specific CD8 + T cells show increased expression of activation markers (33,34), produce proinflammatory cytokines and kill virus-infected cells by releasing perforin and granzyme B (29,35).
Whether CMV infection attenuates the immune response to IAV infection in older adults remains unclear. There are two major hypotheses explaining how CMV infection may affect the immune response to a heterologous virus. First, it has been proposed that large clonal expansions of terminally differentiated CMV-specific CD8 + T cells, which are a hallmark of CMV infection and which can take up to 30%-90% of the CD8 + Tcell memory pool (36)(37)(38), may fill the 'immunological space' (10,13,39). Thereby CMV infection may hamper the induction of other immune responses (10,13,39). IAV-specific T cells may thus be outcompeted by CMV-specific T cells in their competition for proliferation and survival factors (13). Secondly, it has been suggested that CMV is linked to 'inflammaging', the lingering low-graded level of inflammation occurring with ageing (40). The production of pro-inflammatory mediators has been shown to enable CMV reactivation (41,42).
Upregulation of TNF-a, IL-6 and CRP have been observed in CMV-infected individuals, as well as increased production of IL-10 (11,43,44). Especially the increase of IL-10 combined with a decrease of IFNg has been associated with reduced cytolytic capacity of CD8 + T cells responsible for clearing IAV, which also fits with the observed lower levels of granzyme B (45)(46)(47). Importantly, even though CMV has been suggested to diminish the T-cell response to IAV, there is no clinical evidence of a direct link between CMV infection and the T-cell response against IAV infection in humans.
We had the unique opportunity to study the effect of CMV infection on IAV in humans in the in a relatively large cohort of natural influenza infected older adults. We first investigated the effect of CMV infection on the presence of IAV-specific memory T cells in healthy young and healthy older individuals. Next we assessed the effect of latent CMV infection on the IAV-specific Tcell response in older adults undergoing an acute IAV infection. Our data show that CMVinfection is associated with reduced frequencies of IAV-specific T cells in healthy older adults, whereas in healthy younger adults no association with CMV infection is observed. Nevertheless, CMV infection does not hamper the T-cell response to acute IAV infection in older adults.

Healthy Young and Older Adults
Samples of healthy individuals covering a broad age range were combined from two cohorts. Samples of young adults (n=34), between 18 and 52 years of age, from unvaccinated controls or pre-vaccination participants were selected based on age and sex from a study carried out in 2009-2011 (the Pandemic influenza vaccination trial, Netherlands Trial Register NL1952) (48

IAV A Infected Older Adults During IAV Infection
Laboratory-confirmed Influenza virus A infected older adults were selected from the same study as the healthy older adults. In this prospective observational study participants were monitored for influenza-like-illness (ILI) in the influenza season of 2014-2015 (NL4666, Kaaijk et al., submitted). Study design of the Influenza-like-illness-3 study was comparable to previous studies as described in Van Beek et al. (49). In short, participants were instructed about influenza-like-illness (ILI) symptoms according to the Dutch Pel criteria, defined by fever (≥37.8°C) with at least 1 other symptom of headache, muscle pain, sore throat, coughing, runny nose, or chest pain (50) and to report ILI as soon as possible after onset. Nasopharyngeal and oropharyngeal samples were obtained within 72 hours of reporting ILI by standard procedures (49). IAV infection was laboratory confirmed, and subtyped by PCR and sequencing in n=72 individuals by methods described previously (49). The 72 IAV confirmed patients were included in the current study. The H3N2 strain was detected in the majority of patients (n=64, of which n=20 clade 3C.3b, n=37 clade 3C.2a, n=7 not determined), and the H1N1 strain in the remaining individuals (n=8). Blood samples were collected within the first 72 hours of fever onset, and followed up after 2 weeks and 8 weeks.

PBMC and Serum Isolation
Peripheral blood mononuclear cells were obtained by Lymphoprep (Progen) density gradient centrifugation from heparinized blood, according to the manufacturer's instructions. PBMCs were frozen in 90% fetal calf serum and 10% dimethyl sulfoxide at -135°C until further use. Serum was isolated out of tubes with clot-activation factor and stored at -80°C until further use.

Cytomegalovirus (CMV)-Specific Antibodies
Anti-CMV IgG antibody concentrations were measured either using a commercial ELISA (IBL international GMBH) according to manufacturer's instructions or by an in-house-developed multiplex immunoassay (51), depending on the cohort. For healthy young adults, CMV-specific antibody levels were measured using a commercial ELISA. Recommended cutoffs of the commercial ELISA kit were followed. Participants with a CMV antibody level of ≥12 U/ml or higher were considered CMV + , those with a level of ≤8 U/ml were considered CMV -, and those with a level between 8 and 12 U/ml were considered inconclusive and hence excluded for further analysis. For older healthy adults and IAV-infected individuals, CMV-specific antibody levels were measured in serum by our in-housedeveloped multiplex immunoassay. Cutoff were based on previous calculations: Individuals with a CMV-specific antibody level of ≤4 arbitrary units/ml were considered to be CMVand individuals with an antibody level > 7.5 RU/ml were considered CMV + , and those with a level between 4 and 7.5 arbitrary units/ml were considered inconclusive and hence excluded from further analysis (52). To reduce inter-assay variation, all samples from the same individual were measured on the same plate.

Antigen-Specific T Cells by Flow Cytometry
Healthy Individuals Acquisition was performed on a LSRFortessaX20 and data analysis was performed using FlowJo (Treestar). tSNE-analyses were performed using Cytobank (www.cytobank.org) (53) with for every donor 10.000 CD8 + T cells. Donors with less than 10.000 CD8 + T cells were excluded from t-SNE analysis (n=3).

IAV-Infected Individuals
Of all HLA-A2 positive individuals undergoing IAV infection (n=36), both IAV-specific T cells and CMV-specific T cells were assessed within the first 72 hours 2 and 8 weeks after fever onset. PBMCs were stained using the HLA-A2 dextramer for the GILG epitope of the M1 protein of IAV (A*0201/GILGFVFTL-APC, Immudex) (± 8 mln PBMCs) (detected in n=17) and the HLA-

IAV-Specific and CMV -Specific IFNg T-Cell Response by ELISpot
Virus-specific T-cell responses were quantified using the IFNg enzyme-linked immunospot (ELISPOT) assay at all three time points for those of which enough cells were available (respectively n=66, n=61, n=58 out of n=72

Cytokine and Chemokine Levels in Serum
Cytokines and chemokine levels in serum were assessed for all IAV-infected individuals within 72 hours, and 2 and 8 weeks after fever onset. Levels were measured by bead-based multiplex LEGENDplex ™ (BioLegend) according to the manufacturer's instructions. The pro-inflammatory cytokines IL-6, IFNy, IL-10 and CRP were analyzed for this study. Stimulations were performed in duplicate. Analysis was performed on a Canto II flowcytometer. Data were analyzed via Legendplex V8.0 software (Biolegend). All data were transformed into averages of the logarithms of two measurements, and each data point was corrected by subtraction of the intra-assay averages to correct for batch effects.

Preparing TCRbeta cDNA Libraries for High Throughput Sequencing
T-cell receptor analysis was performed as described previously (54), with minor modifications. Briefly, mRNA was isolated with the RNA microkit (Qiagen) according to manufacturer's protocol. Isolated mRNA was used for cDNA synthesis with 5'RACE template switch technology to introduce a universal primer binding site, and unique molecular identifiers (UMI's) were added at the 5' end of the cDNA molecules using the SMARTScribe Reverse Transcriptase (TaKaRa). cDNA synthesis was followed by an AMPure XP bead-based cleanup (Beckman Coulter). Purified cDNA molecules were amplified in two subsequent PCR steps using the Q5 ® High-Fidelity DNA Polymerase (New England BioLabs), with an AMPure XP bead-based cleanup in between. PCR products were size selected on gel and purified using the Nucleospin PCR cleanup kit (Machery-Nagel). The PCR products were sequenced via Illumina MiSeq paired end using 2x250 bp sequencing.

Analysis of TCRbeta Clonotype Analysis
Raw sequencing data was processed using the 12nt UMIs to correct for amplification biases and error-correction of reads. RTCR (55) was used to identify both the UMI sequence and clonotype information from the reads. Because of the relatively small number of cells per sample, additional filtering steps were followed to minimize cross-sample contamination and biases introduced by errors in the UMI sequence. Sequences were only accepted if their UMI was observed in at least 40 sequencing reads. Sequences with identical UMIs in multiple samples were removed if they did not occur in at least 1000 sequencing reads and also if their absolute frequency was lower than 10% of the maximum frequency in the other sample. UMIs were clustered within each sample within a Hamming Distance of 3.

Severity of Symptoms Assessment
Symptom assessment was performed by self-reporting by the participants. As soon as a fever occurred, participants reported the already experienced symptoms up until then and subsequently reported the last day of each symptom until full recovery up until a maximum of timepoint three (8 weeks after start of fever). This way, the presence and duration (start date and end date) of the following symptoms were collected: fever (≥37.8°C), cough, sore throat, runny nose, headache, pain while breathing and muscle pain. Symptoms of an IAV-infected individual up until ten days before onset of fever were considered to be IAV infection related and included in the analysis in this paper. The duration of each symptom in days was normalized by transformation of these values to Z-values, e.g. the number of standard deviations by which the value of the score of an individual is above or below the mean value. Finally, to assess the overall severity of symptoms during IAV infection, the average of the Z-values of the seven symptoms was calculated.

Statistical Analysis
Differences between groups (for example CMVcompared to CMV + ) were assessed using Mann-Whitney U test, and comparisons within the same individuals (for example to compare time points in response upon influenza virus infection) with the Wilcoxon singed-rank test. Differences between groups in categorical variables were tested by chisquare test and corrected for multiple testing if applicable and indicated in figure legends. Correlations were tested with Spearman's rank correlation coefficient. For all analyses p values < 0.05 were considered statistically significant. Data were analyzed using GraphPad Prism 8.3 and SPSS statistics 22 for Windows (SPSS Inc., Chicago, IL, USA).

Characteristics of Study Population
Healthy individuals were on average 59.2 years old (range 21-82 year) (n=99). They were categorized into young (21-52 years old) (n=34) and old (>60 years old) (n=65) individuals, of whom respectively 55.9% and 58.5% were CMV-infected ( Table 1). No significant differences in age or sex were observed between CMVand CMV + individuals ( years, range 60-88 years) with confirmed IAV infection were included in this study (n=72). Also for the IAV-infected older adults, no significant differences in age or sex were observed between CMV + and CMVindividuals ( Table 1). The majority of individuals were infected with the H3N2 strain (n=64/72), while some were infected with the H1N1 strain of influenza (n=8/72). All individuals in the H1N1 infected group turned out to be CMV -.

CMV Induces an Increase in Senescence-Associated Markers in the T-Cell Pool in Older Adults
We assessed the effect of latent CMV infection on the CD8 + Tcell pool of all healthy individuals, by performing a cluster analysis (tSNE) based on memory T-cell markers CD27, CCR7, CD95, CD45RO, and CXCR3, and senescence-associated T-cell markers CD57 and KLRG-1, known to be altered in CMV infection (56). Cluster analysis conformed the large differences between the CMV -(n=40) and CMV + (n=56) group ( Figure  1A), and six different clusters were identified. Clusters 1-3 containing non-senescent CD27 high CCR7 high CD57 low cells were predominantly present in CMVindividuals ( Figure 1B and Supplementary Figure 1A). In contrast, clusters 4-6 containing the more differentiated cells expressing KLRG-1 high CD57 medium were more pronounced in CMV + individuals ( Figure 1B and Supplementary Figure 1A). We next performed more detailed analysis on the expression of senescenceassociated markers within the two age groups: young adults (20-52 years) and older adults (older than 60 years). CMV infection was associated with significantly increased expression of CD57 and KLRG-1 among older, but not in young, adults ( Figures 1C, D). Likewise, expression of PD-1 was only significantly reduced in CMV + individuals (as compared to CMV -) in the older group, and not in the younger group ( Figure 1E). In line with this, CMV infection was associated with significantly increased frequencies of T EMRA cells specifically in older individuals ( Figure 1F). Together, this indicates that CMV infection establishes large changes in the CD8 + T-cell pool by inducing terminally differentiated and senescent T cells in older adults.

Frequency of IAV-Specific T-Cells Is Decreased in CMV + Individuals, but Only in Older Adults
To investigate the hypothesis that CMV infection may negatively influence the immune response to other pathogens by outcompeting other antigen-specific T cells, the frequency of IAV-specific T cells in healthy individuals was determined using HLA-A2 dextramers containing the matrix protein-1 GILG-epitope. The frequency of IAV-specific T cells was significantly lower in CMV + compared to CMVindividuals (P=0.0005) ( Figure 1G). Importantly, this lower percentage of IAV-specific T cells was solely explained by a lower frequency in the older group, where some donors even had no detectable HLA-A2 GILG-specific T cells. Among the young adults no differences in the frequencies of IAV-specific T cells between CMVand CMV + individuals were observed ( Figure 1H). Thus, CMV infection results in lower frequencies of IAV-specific memory T cells, but only in older adults.

Characterization of the T-Cell Response During IAV Infection in Older Adults
To investigate the effect of CMV infection on the T-cell response during IAV infection, we characterized the T-cell response to IAV infection during infection in older adults (N=72), i.e. within 72 hours after start of fever, and 2 and 8 weeks later. The frequency of IAV-specific CD8 + T cells was determined dextramer staining for the HLA-A2 GILG-epitope. IAVspecific CD8 + T-cell frequencies were increased upon IAV infection at the 2 week time point compared to <72 hours after fever onset (median 0.03% to 0.15% of total CD8 + T cells respectively), after which the response contracted (8 weeks post infection, median of 0.08/total CD8 + T cells) ( Figure 2A). The increase in IAV-specific CD8 + T-cell frequencies was mainly explained by the expansion of IAVspecific effector memory T cells (T EM ) ( Figure 2B), while the frequencies in the other subsets (i.e. naïve (T N ), central memory (T CM ) and T EMRA ) did not increase (data not shown). IFNg responses after in vitro stimulation of PBMCs with peptide pools covering the influenza matrix-protein-1 revealed similar results ( Figure 2C).
This was true for both H3N2 influenza-infected and H1N1 influenza-infected individuals, and no difference in increase in the IFNg response between the two strains was observed (data not shown). When the IAV-specific IFNg response was calculated as percentage of CD3 + T cells instead of PBMCs, the IAV-specific IFNg response also tended to be higher in response to IAV infection (2 weeks) compared to steady state (8 weeks), albeit not significant (p=0.059) ( Supplementary Figure 2A left panel). Differences between groups were compared by Mann Whitney U test. **P < 0.01, ***P < 0.001, ****P < 0.0001. ns, non significant.

T-Cell Response to IAV Infection in Older Adults Is Not Impaired by CMV Infection
To investigate whether the T-cell response during IAV infection (N=72) is influenced by CMV infection, we analyzed CMV + and CMVindividuals separately. No significant association of CMVseropositivity on IAV-specific CD8 + T-cell percentages or on the percentage of T EM cells among the IAV-specific T cells was observed ( Figures 2D, E). Likewise, IAV-specific IFNg responses were not negatively associated with CMV seropositivity ( Figure 2F) or with the level of CMV-specific antibodies in CMV + individuals at any of the time points (Supplementary Figure 2B). Surprisingly, in the acute phase (<72 hours after fever onset), CMV + individuals even showed a significantly higher IAV-specific IFNg T-cell response than CMV -individuals (P=0.013) ( Figure 2F). Also when corrected for the percentage of T cells among all lymphocytes, the same trend was observed ( Supplementary Figure 2A right panel). This higher IAV-specific IFNg T-cell response in CMV + individuals could not be explained by the moment of sampling within the 72 hours after fever onset (data not shown). A possible explanation for the higher IAV-specific IFNg Tcell response in CMV + individuals might be related to differences in T-cell migration. We therefore also investigated the total frequency of T cells in the blood and their expression of activation and migration markers upon IAV infection in CMVand CMV + individuals. The frequency of total CD3 + T cells in the blood of influenza-infected individuals was significantly lower early after infection compared to 2 or 8

cells upon influenza infection for CMVand CMV + individuals. (E) The percentage effector memory (EM) cells of IAV-specific T cells upon influenza virus infection for CMV + and CMVindividuals. (F) IAV-specific IFNg T-cell response upon influenza infection for CMVand CMV + individuals. (G) Frequencies of CD3 + T cells in the blood of IAV-infected individuals at < 72 hours, 2 weeks and 8 weeks after infection (left panel) and in CMVand CMV + individuals (right panel). (H) Percentage of CXCR3 + within total CD8 + T cells upon infection (left panel) and in CMVand CMV + individuals (right panel). Wilcoxon test was used to compare T-cell responses of individuals in time and
Mann-Whitney U test was used to compare CMVand CMV + individuals. Correlation between CMV-specific immune responses and IFNg response to influenza virus was tested by Spearman correlation. T-cell IFNg responses are presented per 1 x 10 6 PBMCs. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. ns, non significant. weeks later ( Figure 2G left panel). Comparing CMV + individuals with CMVindividuals, the CD3 + T-cell frequency was also consistently lower in CMVindividuals ( Figure 2G right panel). Next, the migration of CD3 + T cells outside the blood, early after influenza infection, was further assessed by comparing the relative decrease within CMVand CMV + individuals. The drop in CD3+ T-cell frequency early after IAV infection was significantly larger in CMVthan in CMV + individuals (Supplementary Figure 3A). Thus, the higher IFNg IAV-specific responses in blood in CMV + individuals early after influenza virus infection ( Figure 2F) may at least partly be explained by higher frequencies of total T cells in CMV + as compared to CMVindividuals. The higher frequencies of total T cells may be a result of less migration of T cells. Investigating migratory capacity of T cells, we indeed found that early after IAV infection CXCR3 expression by CD8 + T cells, which is associated with migratory capacity of these cells to the lungs, was lower in CMV + individuals compared to CMVindividuals ( Figure 2H). These data suggest that T cells in CMV + individuals may be less prone to be recruited to the lungs early after IAV infection and thereby accumulate in the blood leading to higher frequencies.
However, on IAV-specific CD8 + T cells, only a small trend of increased CXCR3 expression on (p=0.11) was observed in CMVcompared to CMV + individuals (Supplementary Figure 3B) early after IAV infection. In addition, the activation status of IAV-specific T cells by the frequencies of HLA -DR + CD38 + or CD127 -KLRG -1 + revealed no difference between CMVand CMV + individuals. IAV-specific T cells from CMVand CMV + individuals both showed increased expression of these activation markers early after IAV infection (<72 hours after start fever) as compared to 2 and 8 weeks later, irrespective of CMV serostatus. This pattern was not seen for CMV-specific T cells, which served as a control. Figure 3C). Overall, these data suggest that migration of T cells to the lungs might play a role in the enhanced IAV-specific IFNg response observed in CMV + individuals early after IAV infection ( Figure  2F) although evidence remains circumstantial.

Large CMV -Specific T-Cell Responses Are Not Associated With Impaired IAV-Specific T-Cell Responses
As CMV might only affect the immune system through competition for "limited immunological space" in individuals with large CMV-specific T-cell responses, we next investigated the association between CMV infection and the T-cell response to IAV by taking into account the magnitude of the IFNg CMVspecific T-cell responses within CMV + individuals. First, we assessed the IFNg T-cell response to IAV infection. No negative correlation between expanded CMV-specific T-cell responses and T-cell response to IAV infection was observed at any of the three time points ( Figure 3A). Surprisingly, we even observed a significant positive correlation between the height of the CMV-specific IFNg T-cell response and the height of the IAV-specific IFNg T-cell response at 2 and 8 weeks after fever onset ( Figure 3A) (R: 0.52, p=0.016 and R:0.45, p=0.014 respectively. In addition, we further assessed the quality of the IAV-specific T-cell response by investigating TCR diversity of the IAV-specific T-cell response, by sorting the HLA-A2-GILGspecific T cells and sequencing the TCRb chain. The clonotype distribution was analyzed over time after IAV infection in two CMVand two CMV + individuals. In general it was observed that 89% of the different TCR sequences expressed the Vb-19 segment, of which 74% consisted of the highly conserved -RSmotif (See Supplementary Tables 1A, B for the V and J segments). Interestingly, all donors showed a shift in dominance after IAV infection, however large heterogeneity was observed between the donors with some repertoires becoming more diverse and other less diverse over time. No indication for CMV-related differences were present, as no differences between the CMVand CMV + individuals were observed. Based on these data no differences were observed in the TCR diversity of the IAV-specific T-cell repertoire between CMVand CMV + individuals ( Figure 3B). These lines of evidence do not support the competition between CMVspecific memory T cells and a proper IAV-specific T cells during IAV infection.

Cytokine Levels in Serum of Influenza-Infected Individuals Are Not Affected by CMV and Not Associated With IAV-Specific T-Cell Responses
As it is suggested that CMV is linked to "inflammaging", we questioned whether pro-inflammatory mediators may be enhanced in CMV + individuals during IAV-infection. We investigated the levels of pro-inflammatory cytokines in serum of CMV + and CMVindividuals, and their potential association with the IAV-specific T-cell response. At the early phase of IAV infection, the inflammation-associated factors IL-6 and CRP were elevated in serum compared to 2 and 8 weeks later ( Figure 4A). No significant differences in IL-6 and CRP levels were observed between CMV + and CMVindividuals, also at the peak of the T-cell response (2 weeks later) and at steady state (week 8 after fever onset) ( Figure 4A). We also measured the IFNg:IL-10 ratio, as it was suggested that a shift in this ratio leads to a decline in IAV-specific T-cell responses with age and is associated with decreased protection against IAV (46,57). No difference was observed ( Figure 4B) in the IFNg:IL-10 ratio between CMV + and CMVindividuals at any of the time points, even though both IFNg and IL-10 levels were elevated at the earliest timepoint (Supplementary Figure 4). Although increased levels of the cytokines IL-6, CRP, IFNg and IL-10 were observed in the acute phase of IAV infection, no significant association between these cytokine levels and the height of the IAV-specific T-cell response was observed <72 hours after start of fever ( Figure 4C) or 2 and 8 weeks later (data not shown). Together, this suggests that CMV infection in older adults does not affect cytokine levels in serum upon IAV infection.

Severity of Symptoms of IAV Virus Infection Is Not Increased by CMV
Since early T-cell responses during IAV infection play an important role in limiting disease severity, we also investigated whether the height of the IAV-specific T-cell response was associated with severity of symptoms of IAV infection. In IAV-infected older adults, the number and duration of the following symptoms was assessed: fever (≥37.8°C), cough, sore throat, runny nose, headache, pain while breathing and muscle pain. The severity of symptoms of IAV infection was positively associated with the height of the IAV-specific T-cell response at timepoint 3 (8 weeks after start of fever) ( Figure 5A), but not within 72 hours or 2 weeks after start of fever. This association was mainly based on the number of symptoms, and not on their duration (Supplementary Figure 5). Thus, the IAV-specific Tcell response 8 weeks after IAV infection, and not the early T-cell response, was linked to the severity of symptoms of the IAV infection in this cohort. We next investigated whether CMV infection was associated with more severe symptoms during IAV infection. No significant differences between CMVand CMV + individuals were observed in the number of symptoms during IAV infection ( Figure 5B), total duration of symptoms ( Figure 5C), or severity of symptoms as assessed by a combination of duration and number of symptoms ( Figure 5D). When investigating the IAV infection symptoms individually, CMV + individuals tend to suffer less from muscle pain ( Figure 5E) and although frequency of coughing was not different between CMV + and CMV -, CMV + individuals tended to suffer less long from coughing as a symptom compared to CMVindividuals (data not shown). When the frequency of individuals coughing was plotted over time, we indeed found a indication for a faster decline in the frequency of coughing individuals amongst CMV + individuals compared to CMVindividuals ( Figure 5F). Of note, both the increased frequency of muscle pain and the longer persistence of coughing among CMVindividuals could not be ascribed to a difference in IAV strain infection between CMVand CMV + individuals (data not shown). The severity of symptoms was also not associated with CMVspecific antibody levels (data not shown). Together, this suggests that CMV infection at least does not worsen the number and duration of IAV infection symptoms.

DISCUSSION
In this study we investigated the impact of CMV infection on the immune response to IAV. We found that CMV infection is associated with a more differentiated and senescent phenotype of CD8 + T cells. In healthy younger individuals, no difference in IAV-specific T-cell frequencies were observed, but, CMV + older individuals had lower frequencies of IAV-specific memory CD8 + T cells compared to CMVolder individuals. Nevertheless, the induction of an IAV-specific T-cell response during active IAV infection in older adults was not impaired. Also, severity of IAVassociated symptoms was not negatively affected by CMV infection. We did not find any evidence for a negative effect of CMV infection on the IAV immune response by 'limited immunological space' or by increased levels of proinflammatory mediators. In contrast, there seemed to be a positive association between CMV infection and the IAVspecific T-cell response early after IAV infection (<72 hours after fever onset).
To the best of our knowledge, we are the first to investigate the potential effect of CMV infection on the T-cell response to a heterologous infection in humans. A negative effect of CMV infection on the functioning of the immune system is often intuitively explained by competition between T cells for 'limited immunological space'. Indeed, CMV infection has a profound impact on the composition of the overall CD8 + T-cell pool in healthy individuals, by increasing the number of highly differentiated memory cells, especially in older adults (25). We show that CMV infection leads to a decrease in the frequency of memory IAV-specific T cells in healthy older adults. Relatively low frequencies of IAV-specific T cells in older adults compared to younger adults have been reported before (58,59) and are considered to be a key determinant of a diminished T-cell response in IAV infection (60). Note, that the observed frequencies and characteristics of the IAV-specific response are based on one HLA-A2 epitope (GILGFVFTL). Although this epitope is immunodominant and conserved between most influenza strains, it remains to be determined whether it reflects the T-cell total IAV-specific T-cell response. Unfortunately, absolute T-cell numbers could not be investigated in this study, leaving the possibility that the decreased frequency of IAV-specific T cells merely reflected a relative increase in CMV-specific T cells in older adults. A decline in the frequency -and not in the number-of IAVspecific T cells may explain why decreased IAV-specific T-cell frequencies in healthy older adults do not seem to result in reduced T-cell responses upon acute influenza virus infection. This was also observed in the MCMV mouse model, were reduced frequencies were observed, but the absolute counts of CD8 + T cell against an heterologous virus infection were maintained in MCMV infected mice (61). Alternatively, the number of IAV-specific T cells may have been lower in CMV + older adults. A similar effect has previously been reported for EBV-specific T cells (37). During IAV infection, we observed that IAV-specific T cells in CMV + and CMVindividuals responded equally well and no substantial changes in IAV-specific T-cell repertoire were observed in CMVand CMV + individuals. Previously, to influenza vaccination, other studies reported impaired IAV-specific T-cell responses by CMV infection (20,62,63). In contrast, other groups showed a positive effect of CMV on the vaccine-induced IAV-specific T-cell response in humans (64). In addition, in mice, MCMV infection was shown to enhance the diversity of the T-cell repertoire against Listeria monocytogenes expressing OVA (65), whereas we found no differences in the repertoire diversity between CMVand CMV + individuals. This could probably be explained by the small sample size of our repertoire analyses, or by the difference in epitope, as the HLA-A2-GILG specific T-cell repertoire is suggested to be more restricted compared to other antigenspecific repertoires (66). We also found that the magnitude of the CMV-specific T-cell response was positively associated with the magnitude of the IAV-specific T-cell response 2 weeks and 8 weeks after fever onset of IAV infection. Together, this suggests that CMV + individuals have a sufficient amount of IAV-specific T cells able to respond to IAV infection. Pre-existing T cells and timing of the IAV-specific CD8 + Tcell response are thought to play an important role in the reduction of severity of IAV-related symptoms (30,31). An early T-cell response has been proposed to accelerate viral clearance, whereas a delayed and prolonged T-cell response may lead to high and prolonged levels of inflammation and increased severity of disease (31,32). Unfortunately, pre-existing T-cell responses could not be investigated in this study. Our data suggest that CMV + individuals have an increased early or accelerated IAV-specific T-cell response compared to CMVindividuals. We speculate that this may lead to earlier viral clearance, and thereby might explain the faster recovery of coughing and decreased frequency of muscle pain. When we assessed severity of symptoms of symptoms, we found a significant positive correlation between the magnitude of the IAV-specific T-cell response and severity of symptoms 8 weeks after fever onset, and not early after fever onset or 2 weeks later. This suggests that a prolonged T-cell response, still present 8 weeks after fever onset, is increasing symptom severity of IAV infection.
As CMV infection was suggested to induce a more inflammatory environment (67), we also studied the potential association between CMV and inflammatory markers and how this might be related to IAV-specific T-cell responses. Although the levels of pro-inflammatory cytokines and chemokines measured here were significantly increased in the acute phase of IAV infection (<72 hours after fever onset), they were similar between CMV + and CMVindividuals. Since pro-inflammatory responses were similar between CMV + and CMVindividuals, it may (partly) explain the lack of difference in the severity of clinical symptoms between these groups. Furthermore, these levels did not seem to be associated with the IAV-specific Tcell response at any time point, although, we cannot rule out the possibility that other immune modulatory factors may play a role.
Most studies claiming a negative effect of CMV on an immune response to a heterologous challenge have been performed in mice. Several mouse studies have shown that only lifelong infection with MCMV leads to decreased immunity against heterologous infections (14)(15)(16). The magnitude of the effect of CMV infection might thus be linked to the duration of CMV infection and the experienced amount of viral reactivation in an infected host. We found no evidence for decreased immunity against a heterologous infection in humans infected with CMV. Also, high CMVspecific antibody levels were not associated with the height of the IAV-specific T-cell response. One of the reasons for the observed differences between mice and men, could be the order of infections by CMV and IAV. In mouse studies, mice are typically first infected with MCMV long before they are challenged with a heterologous acute infection. Many humans, in contrast, may have undergone their first IAV infection before they were infected with CMV, which may lead to the presence of IAV-specific memory T cells before the CMV-specific immune response is established. A potential harmful effect of CMV might be less pronounced in a host who already has a proper immune response against IAV. Furthermore, as mouse models of CMV are almost exclusively done in specific pathogen free mice, and humans are exposed to dozens of infections and triggers during life, it might be that the effect of CMV is magnified in mice. Even if CMV can modulate other immune responses, there is no substantial evidence that CMV impacts the function of the immune system by hampering immune responses against heterologous infections in humans.
The increased IFNg T-cell response to IAV infection that we observed in CMV + older individuals remains partially unexplained. Previously, enhanced influenza vaccine responses in humans and mice were explained by an increase in IFNg in serum (1). However, we did not observe a difference in IFNg levels in serum between CMV + and CMVindividuals in our study. However, an important difference between our study and the study by Furman et al. is the age of the individuals, as the latter observed this increase in IFNy levels mainly in younger adults. As the positive effect of CMV was observed in IAV IFNg ELIspot assays, we cannot exclude the possibility that CMV infection may only affect CD4 + T cells, which may respond in the assay as well. Another explanation for the difference in the early IAV-specific T-cell response between CMVand CMV + individuals could be related to the migratory capacity of the responding T cells, as we observed a trend towards enhanced CXCR3 expression of IAV-specific CD8 + T cells in CMVindividuals. This may lead to early migration towards tissues such as the lungs. The effect of CMV on IAV-specific T-cell responses at the site of infection instead of the blood would be of great interest, and requires further research. Also, other factors like apoptosis or proliferation of the IAV-specific T cells may play a role in the difference in IAV-specific T-cell response between CMVand CMV + individuals.
In conclusion, identification of the driving forces that induce age-related changes in the immune system is important to protect the growing population of older adults against infectious diseases. Especially IAV leads to more increased disease burden in older adults, e.g. severe symptoms and higher risk of complications, hospitalization and mortality. Our study shows that despite the lower frequency of IAV-specific memory T-cell responses in older adults, CMV infection does not seem to impair the T-cell response against acute IAV infection. This work supports the view that CMV acts, if anything, as an immune modulatory mediator, rather than having a true negative impact on the immune system.

DATA AVAILABILITY STATEMENT
Datasets are available on request: The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by ethical committee METC Noord Holland. The patients/participants provided their written informed consent to participate in this study.

AUTHOR CONTRIBUTIONS
The original idea for this study was from DV. SV performed the majority of experiments, gathered data and analysis. JL performed a large part of the experiments, gathered data and analysis. RJ, MH, MV, NN, and RV performed and evaluated experiments. JV designed the original study of IAV infected individuals. JD with help from JB and DV supervised the project. SV and JL prepared figures and wrote the manuscript with contributions and review from JD, DV, and JB. All authors contributed to the article and approved the submitted version.

FUNDING
Funded by Dutch Ministry of Health Welfare and Sport.