HCMV carriage in the elderly diminishes anti-viral functionality of the adaptive immune response resulting in virus replication at peripheral sites

Human cytomegalovirus (HCMV) infection and periodic reactivation is, generally, well controlled by adaptative immune responses in the healthy. In older people, overt HCMV disease is rarely seen despite the association of HCMV with increased risk of mortality; evidence from studies of unwell aged populations suggest that HCMV seropositivity is an important co-morbidity factor. HCMV genomes have been detected in urine from older donors, suggesting that the immune response prevents systemic disease but possibly immunomodulation due to lifelong viral carriage may alter its efficacy at peripheral tissue sites. Previously we have demonstrated that there were no age-related expansions of T cell responses to HCMV or increase in latent viral carriage with age and these T cells produced anti-viral cytokines and viremia was very rarely detected. To investigate the efficacy of anti-HCMV responses with increasing age, we used an in vitro Viral Dissemination Assay (VDA) using autologous dermal fibroblasts to determine the anti-viral effector capacity of total PBMC, as well as important subsets (T cells, NK cells). In parallel we assessed components of the humoral response (antibody neutralization) and combined this with qPCR detection of HCMV in blood, saliva and urine in a cohort of young and old donors. Consistent with previous studies, we again show HCMV specific cIL-10, IFNγ and TNFα T cell responses to peptides did not show an age-related defect. However, assessment of direct anti-viral cellular and antibody-mediated adaptive immune responses using the VDA shows that older donors are significantly less able to control viral dissemination in an in vitro assay compared to young donors. Corroborating this observation, we detected viral genomes in saliva samples only from older donors, these donors had a defect in cellular control of viral spread in our in vitro assay. Phenotyping of fibroblasts used in this study shows expression of a number of checkpoint inhibitor ligands which may contribute to the defects observed. The potential to therapeutically intervene in checkpoint inhibitor pathways to prevent HCMV reactivation in the unwell aged is an exciting avenue to explore.


Supplementary Figure S6 -Breadth and donor frequency of HCMV specific T cell responses
The breadth of the three cytokines response to the 5 different HCMV protein mixes Latency associated proteins (LAT: UL138, US28, LUNA, vIL-10), pp65 and UL144, IE1 and IE2 (IEs), pp71 and US3 and gB proteins in the Aquaria donor cohort are illustrated (A), the breadth was calculated by summing the number of HCMV proteins with an above threshold cytokine (100 sfu/ million CD3+ T cells) response for each donor the mean of the young and old responses is shown as a dashed line. All donors irrespective of age had a CD8+ and CD4+ T cell IFNγ response to a minimum of 3 of the HCMV protein mixes, with the majority of both old and young donors responding to all 5 HCMV protein mixes. The frequency of donors responding to the 5 different HCMV protein mixes for all three cytokines are illustrated for CD8+ T cells (B) and CD4+ T cells (C). The data is presented as scatter dot plots showing the geomean and geometric standard deviation for each age group on both graphs. All donors made a CD8+ and CD4+ T cell IFNγ response to pp65 and UL144 and gB stimulation, the highest frequency TNFα response was to pp71 and US3 stimulation in both T cell subsets, and stimulation by the latency associated proteins and also pp71 and US3 produced an IL-10 responses in CD8+ and CD4+ T cells.

Supplementary Figure S7 -Viral Dissemination assay and Area Under the Curve (AUC) calculations.
The protocol for the viral dissemination assay are shown (A), 96 half-well plates were seeded with dermal fibroblasts and then infected with the dual fluorescence tagged strain of Merlin (UL36-mCherry and UL32-GFP) virus at a low multiplicity of infection. Twenty-four hours post infection autologous immune cell subsets are added at a range of effector to target (E:T) ratios and then cocultured with the fibroblasts for 11 days. Fibroblasts were harvested and the mCherry and GFP expression was measured by flow cytometry with typical results at day 11 observed by flow cytometry and microscopy shown, illustrating immediate early gene expression of UL36 (mCherry+GFP-fibroblasts) indicative of viral spread and entry into cells and late gene expression of UL32 (mCherry+GFP+ fibroblasts) indicative that viral DNA replication has occurred; a full time course of infection with this virus is illustrated in figure 2 of Houldcroft et al. 2020 (3) showing the kinetics of expression of mCherry and then GFP over a 10 day time course. Viral spread in each well was determined as a percentage of control infected wells without effector cells and control uninfected well to determine background fluorescence. The normalized viral spread data was plotted for all the E:T ratios and representative curves for early phase infection present in PBMC co-culture for a young CMV seropositive (turquoise circles), old CMV seropositive (purple squares) and CMV seronegative (green triangles) are shown (B). To allow comparison of viral control between individual donors, area under the curve calculations were performed and the calculated area is shown in part C (shaded area and value on each graph). Increasing AUC values reflects decreasing control of viral spread as shown by the seronegative (green shaded area) donor AUC value compared to the seropositive donor AUC values (turquoise and purple shaded areas).

Supplementary Figure S8 -Anti-viral activity of PBMC, NK cells, CD8+ T cells and CD4+ T cells from Young and Old Seropositive and Seronegative donors
Ex vivo donor cells over a range of effector:target (E:T) ratios were co-cultured with autologous dermal fibroblasts infected with the dual fluorescence tagged Merlin strain of HCMV. After 11 days the cultures were harvested and analysed for mCherry and GFP expression by flow cytometry. Shown are the mean +/-standard error of the mean of the normalized viral dissemination for the three cohorts analysed, young positive n=6 (turquoise points and shading), old positive n=9 (purple points and shading) and seronegative donors n=9 (green points and shading) for both Early (mCherry+) gene expression (LH graphs) and late (GFP+ and mCherry+) gene expression (RH graphs). The results for PBMC (Early (A) and Late (B)), NK cells (Early (C) and Late (D)), CD8+ T cells (Early (E) and Late (F)) and CD4+ T cells (Early (G) and Late (H)) are shown. This clearly illustrates that PBMC, CD8+ and CD4+ T cells from the young donor cohort are more effective at controlling viral spread at Early gene expression time points compared to the old donor and seronegative donor cohorts.

Supplementary Figure S9 -Correlation of VDA anti-viral activity with HCMV specific T cell responses
The anti-viral activity measure using the VDA (AUC values) of CD4+ (A & B) and CD8+ T cells (C & D) were correlated with the total HCMV specific IFNγ response (summed sfu/1x10 6 CD3+ T cells response from the 5 protein pool stimulations) for young (turquoise circles) and old (purple squares) seropositive donors for both Early and Late gene expression. The correlation of the HCMV specific T cell response with anti-viral activity of the individual T cell subsets was analysed using Spearman rank correlation (Spearman rs and p value for young and old groups are indicated with the line of best fit (solid) and 95% CI (dotted lines) also shown on the graphs (Young turquoise lines and old purple lines)). Overall, this analysis shows that there is no correlation between the magnitude of the HCMV specific IFNγ secreting T cell response and the anti-viral activity in the viral dissemination assay.

Supplementary Figure S11 -Neutralisation activity of all donor cohorts and Avidity of serum HCMV gB and pentamer specific IgG interactions
Neutralization assays with heat inactivated sera were performed on fibroblasts (A) and epithelial cells (B), serum was diluted and pre-incubated with the fluorescence tagged virus prior to adding to the cells. The mean +/-standard error of the mean of the normalized viral infection for the three cohorts analysed, young positive (turquoise points and shading), old positive (purple points and shading) and seronegative donors (green points and shading) for late (GFP+ and mCherry+) gene expression for both cell types are shown. The avidity of the interaction of serum IgG specific to HCMV gB and pentameric complex (gH/gL/pUL128-130-131 (5)) proteins in young and old HCMV seropositive donors was measured by the addition of a urea treatment to displace weakly bound antibody. 96 well ELISA plates were coated overnight at 4°C with either gB protein (Abcam, Cambridge, UK) or Pentamer protein (Native Antigen Company, Oxford, UK) and then blocked (2% FCS in PBS) for 1 hour at room temperature. Pre-diluted serum samples from the AQUARIA donor cohort were added in duplicate for each condition alongside relevant sample controls and incubated for 1 hour at room temperature. Urea treated wells then had the serum samples removed and the 6M urea (Thermo Fisher Scientific) in PBS wash added to the wells for 8 minutes, followed by 3 subsequent wash steps using the urea wash buffer. The untreated wells were then washed with normal wash buffer (0.1% tween in PBS), All wells were incubated with 1:10 000 dilution of anti-human HRP conjugated antibody (goat-anti-human IgG, Dianova via Stratech, Ely, UK) and incubated for 1 hour at room temperature and then washed. After washing, tetramethylbenzidine peroxidase substrate was added to each well for 30 minutes, diluted 1:1 in peroxidase substrate solution B (KPL, USA). The reaction was stopped by adding 100 µl of 1M phosphoric acid to each well. The optical density at 450 nm (OD450) was determined using an Emaxmicroplate reader (Eurofins MWG Operon). The percentage loss of antibody was determined by the following equation: % = (mean OD of urea-treated sera ÷ mean OD urea-untreated sera) × 100 The results of this calculation are shown as a box and whisker minimum to maximum plot (with median and upper and lower quartiles represented by the box) (C) for each protein with the young HCMV positive donors in turquoise and the old HCMV positive donors in pink shading and points.

Supplementary Figure S12 -Identification of Saliva HCMV DNA+ donors in absolute count and Fibroblast inhibitory ligand molecules
Absolute count data for the Monocytes, B cells, T cells and NK cells according to the young and old seropositive and seronegative groups are shown with donors AQU007 (orange points) and AQU022 (red points) highlighted (geomean and geometric standard deviation shown) (A). Dermal fibroblast inhibitory ligand expression by young and old groups as box and whisker minmax plots are shown with donors AQU007 (orange points) and AQU022 (red points) highlighted (B).