Signatures of immune dysfunction in HIV and HCV infection share features with chronic inflammation in aging and persist after viral reduction or elimination

Significance Chronic inflammation contributes to morbidity and mortality in aging, but whether similar mechanisms underlie dysfunction in infection-associated chronic inflammation is unclear. Using a multicohort systems immunology approach, we identified signatures of immune dysfunction that are shared in aging and chronic viral infections, namely HIV and hepatitis C virus. We show that these shared dysfunctions persist despite viral clearance, and we describe the changes in functional coordination that occur during viral eradication. Finally, we highlight a partial restoration in interferon-α sensitivity across all major immune cell lineages as viral load drops. Our findings suggest a broad and persistent functional remodeling and deterioration of the human immune system despite removal of a chronic pathogenic burden that shares features of chronic inflammation in aging.


Human subjects and sample collection
The Stanford-Ellison longitudinal cohort on aging was started in 2007 and continues under the direction of MMD and Dr. Scott Boyd. The cohort is part of the Stanford 1,000 Immunomes Project lead by MMD and DF. The data used here is from 2009 and was collected in that same year by the Stanford Vaccine Center directed by CLD, and analyzed by the Human Immune Monitoring Center at Stanford, directed by HTM. In 2009, the cohort consisted of 89 healthy individuals (N = 60 aged 61-90, and 29 aged 33 as controls). Subjects provided informed consent for participation in the study, and PBMCs and serum were drawn at time of recruitment. Exclusion criteria at time of enrollment were an active systemic or serious concurrent illness, a history of immunodeficiency, any known or suspected impairment of immunologic function, including clinically significant liver disease, diabetes mellitus treated with insulin, moderate to severe renal disease, blood pressure >150/95 mmHg at screening, chronic hepatitis B or C, recent or current use of immunosuppressive medication. None of the volunteers were recipients or donors of blood or blood products within the past 6 months and 6 weeks respectively, nor showed any signs of febrile illness on the day of blood draw.
The HCV cohort was recruited by AA and EAP from the Division of Gastroenterology and Hepatology in the Department of Medicine at Stanford Health Care in Palo Alto, CA, USA. PBMCs and serum samples were collected from 14 HCV-infected patients prior to initiating standard of care DAA therapy (Sofosbuvir and Ribavirin, Sofosbuvir and Ribavirin and IFN-, or ledispavir/sofosbuvir) during treatment, and after treatment. Ten patients underwent at least one previous treatment with interferon, the other four were treatment naïve. Thirteen patients experienced SVR after twelve weeks of therapy, with one relapsing following the initial course. For uninfected controls, PBMCs were collected from 11 and serum from 10 healthy individuals from the Stanford Blood Center, Palo Alto, CA, USA.
The HIV cohort was recruited by PMG at the Division of Infectious Diseases in the Department of Medicine at Stanford Health Care in Palo Alto, CA, USA. This cohort comprises 24 virologically suppressed, HIV-infected individuals ages 26-78, and 45 uninfected healthy controls. All HIVinfected individuals had undetectable viral loads while on cART at the time of PBMC and serum specimen collection.
All serum specimens were aliquoted and stored at -80 °C and all PBMC specimens were aliquoted and stored at -180 °C. All human subjects provided written informed consent, and all study protocols were reviewed and approved by the Stanford University Institutional Review Board.

CMV serology
CMV serology for each study participant was determined using a commercially available ELISA kit (CMV IgG, Gold Standard Diagnostics, Davis, CA, USA) as per manufacturer's instructions. Briefly, sera stored at -80 °C were thawed to room temperature (20-25 °C) and diluted 1:51 in kit diluent. Diluted samples were added to wells coated with CMV antigen from strain AD169 and incubated at room temperature for 30 minutes. Wells were washed and drained, followed by the addition of goat anti-human IgG antibodies labeled with calf alkaline phosphatase, and incubated at room temperature for 30 minutes. Wells were washed again and drained, followed by the addition of p-nitrophenyl phosphate substrate, and incubation at room temperature for 30 minutes. After the addition of 0.5M trisodium phosphate stop solution, the absorbance of each well at 405 nm was read and results were analyzed using the manufacturer's instructions.

Enrichment analysis
Cluster enrichment analyses were performed by computing p-values for under-or over-enrichment based on the cumulative distribution function of the hypergeometric distribution using the hypergeometric p-value calculator available at https://systems.crump.ucla.edu/hypergeometric/. Mathematically, the hypergeometric probability is expressed as: where is the number of successes within a sample of size drawn from a population of size containing total successes.
When samples are drawn with replacement, as in the case of node enrichment for network hubs and/or bottlenecks, the probability function follows a binomial distribution and is expressed as: where is the number of successes in trials, and is the probability of a success. In these cases p-values were calculated with https://stattrek.com/online-calculator/binomial.aspx.

Hierarchical clustering analysis
Raw immunological measurements were intra-cohort min-max features scaled in R, and hierarchical clustering and heatmaps were generated in R with the gplots library. The optimal number of clusters were defined by gap statistic (33) method. Briefly, this method clusters the input data, varying the number of clusters from k = 1, , K and computes the total within-cluster dispersion Wk for each k. Next, B reference data sets are generated from a random uniform distribution, and each is clustered as above and within-cluster dispersions * , b = 1, , B, k = 1, , K are computed. The estimated gap statistic is then computed as follows: and the standard deviation of the statistics sk is also computed. Finally, the number of clusters is chosen as the smallest value of k such that the gap statistic is within one standard deviation of the gap at k+1: ≥ + 1 − +1 . Gap statistics were computed on scaled data and plotted in R with the factoextra and NbClust libraries with 500 Monte Carlo bootstrap samples. Clusters were then determined by partitioning the clustering dendrograms at the highest level that yielded the optimal number of clusters k. Approximately unbiased p-values for clusters were computed by multiscale bootstrap resampling (10,000 bootstrap samples) using the pvclust R package, and clusters meeting the significance threshold of = 0.05 ere highlighted.

Additional statistical analyses
Two-class paired significance analysis of microarray (SAM) were performed with MultiExperiment Viewer available at http://mev.tm4.org/. FDR was determined by q-values generated by running the analysis with all 16,384 possible permutations. All additional statistical tests were performed with Graphpad Prism software v7.0d and v8.2.1 for Mac OS X (La Jolla, CA, USA).

Data availability
The data and R code used to generate the figures in this manuscript are available at https://github.com/CesarLopezAngel/AgingHCVHIV.