Introduction
Viral respiratory infections remain a predominant cause of morbidity and mortality in aged adults. The elderly have heightened susceptibility to infection, an increased risk of developing severe viral-induced pulmonary disease and have slower recovery rates [
1]. Several physiological parameters are thought to contribute to the poor outcomes of infectious disease in the elderly population, including the aging immune as well as respiratory system. Almost all components of the immune system have been shown to undergo age-associated restructuring that greatly impacts immune function [
2‐
5]. The decline in immune function with age also results in reduced vaccine efficacy, further enhancing susceptibility to infection in the elderly [
6‐
8].
Age-associated alterations in the mucosal immune system are thought to occur at distinct times and in a distinct manner relative to systemic immunity [
9]. Data suggests that immunosenescence may occur earlier in the mucosa than the systemic immune system with a dramatic shift, with age, in the proportion of distinct T cell subsets and a decrease in total B lymphocytes [
3,
10]. Advanced age has also been associated with a reduction in antigen-specific IgA, an important protective antibody predominantly localized to the mucosa [
9]. In addition to the immunological remodeling as a function of age, there are also major alterations in respiratory physiology. The aging lung has been shown to undergo structural changes which include a loss in static recoil forces, a stiffening of the chest cavity and diminished alveolar surface area, ultimately resulting in reduced vital capacity [
11‐
13]. In addition, respiratory muscle strength consistently declines with age making it more difficult for an elderly person to breath even when not suffering from a respiratory infection.
The limitations of the aged immune and respiratory systems likely contributed to the increased mortality observed in elderly patients (>60 years old) with severe acute respiratory syndrome coronavirus (SARS-CoV). The SARS-CoV epidemic in 2002–2003 resulted in over 8000 human infections with an estimated 10% mortality rate [
14]. Advanced age and comorbidities were significantly associated with increased risk of SARS-CoV related death, due to acute respiratory distress syndrome [
15‐
18].
It is well appreciated that pulmonary damage in SARS-CoV infection is caused by direct viral effects as well as immunopathological factors [
15], however the pathogenic mechanisms in the vulnerable aged populations remain poorly defined. Several aged animal models of SARS-CoV infection have been established to evaluate the response and elucidate mechanisms for increased SARS-CoV pathogenicity in the aged host. Recombinant infectious clones and mouse passaged isolates of SARS-CoV show increased severity of disease and lethality in aged as compared to young mice [
19‐
21]. Interestingly, the aged and young hosts show similar levels of SARS-CoV replication in most experimental infections [
22]. Thus the increased acute lung injury in SARS-CoV-infected aged animals is thought to be related to the over exuberant immune responses and not heightened viral-mediated damage. However, many aspects of the elderly immune response and how it may differ from the young adult are still unclear. Furthermore, our study represents one of only two SARS-CoV infection studies in aged nonhuman primates as almost all aged SARS-CoV experiments to date have been conducted in mouse models with mouse-adapted viral strains. Although murine models are often advantageous and informative, nonhuman primates may be better suited for studying the aging immune and respiratory systems. Unlike mice, nonhuman primates show a high level of genetic homology to humans, are not inbred, have longer life spans and their lungs are more structurally similar to humans than other laboratory animals [
23]. Importantly, studies have shown that nonhuman primates undergo immune senescence similar to what has been described for humans [
24,
25]. The aim of this study was to determine how the peripheral and mucosal immune responses to SARS-CoV infection compare in the aged and juvenile nonhuman primate host and to determine how this may impact viral replication levels. We report that SARS-CoV virus titers were significantly higher in the nasal cavity of aged monkeys at day 1 post infection but, by day 3, the difference in titers between age groups was negligible. Although SARS-CoV virus levels were similar in aged and juvenile monkeys at later time points post infection there were significant age-dependent differences in systemic and mucosal immune responses to SARS-CoV.
Discussion
The elderly (those over 65), are considered the fastest-growing demographic in the United States and are expected to make up 19% of the population by 2030 [
31]. It is well recognized that elderly individuals incur enhanced severity of respiratory infections and according to the Center for Disease Control and Prevention, an estimated 9 out of 10 flu-related deaths in the United States occur in people 65 and older. When SARS-CoV emerged in the human population in 2002, the elderly were also disproportionately affected, with individuals over 65 making up 50% of the total fatal cases [
32,
33]. We have a poor understanding of the relationship between aging and the host response to respiratory virus infection. Most of our understanding of the biological changes that occur with aging in humans has been limited to studies of peripheral blood, which may not reflect the immune dynamics of the respiratory tract. Our study aimed at gaining insight into the kinetics and magnitude of the systemic and mucosal immune response to SARS-CoV infection in aged nonhuman primates, an important translational model for immune-aging research.
The significantly higher levels of viral replication in the nasal secretions of aged monkeys as compared to the juvenile animals at 1 d.p.i. was unexpected as most aged SARS-CoV studies have detected no age-dependent differences in infection levels. Higher SARS-CoV titers may have been missed in the previous aged nonhuman primate experiment as 2 d.p.i. was their earliest sampling time point, [
22] and by day 3 in our study, aged and juvenile infection levels were similar. Although we cannot be certain that we were measuring replicating virus and not virus delivered from the inoculum, the dramatic difference in the amount recovered suggests that the aged nasal epithelium may support higher levels of SARS-CoV, or that the nasal cavity drainage mechanisms are impaired in the aged host. Despite the early age-associated differences in viral shedding at mucosal sites, by 5 d.p.i., levels of SARS-CoV in respiratory tract tissues were only slightly higher in the aged monkeys. Similar to previous reports, virus was not recovered from any animal at 10 d.p.i., suggesting that the kinetics of viral clearance may have been similar in both age groups. However, additional sampling between days 5 and 10 post infection would be necessary to confirm this observation.
The aged monkeys in our study exhibited significantly decreased total white blood cell counts which is consistent with the inverse correlation of WBCs with age that has been observed at steady state in several rhesus and human studies [
25,
34]. Of the leukocyte populations in the blood, lymphocyte numbers were most dramatically affected by age, with flow cytometric results showing significantly reduced CD8 T cells and B cells in aged compared to juvenile monkeys. In contrast, peripheral cytokine responses showed only minimal age-related differences to SARS-CoV infection which may be related to the high variability, particularly in the aged group, which is consistent with previous reports in elderly rhesus monkeys [
5,
25]. In regards to the mucosal inflammatory reactions, proinflammatory cytokines IL-1beta, IL-6, IL-15, IL-12 and IL-18 were all significantly lower in the lungs of SARS-CoV-infected aged animals at either day 5 or 10 post infection. In particular, the reduced levels of IL-1beta and IL-18 are in line with recent reports of impaired NLRP3 inflammasome function in elderly mice during influenza infection [
35]. The low cytokine response in aged monkeys may also be a reflection of the reduced total inflammatory cell numbers found in the lungs of aged as compared to juvenile animals as cells and cytokines were evaluated in adjacent lung regions. Our data is not consistent with the exacerbated acute inflammatory responses shown to promote disease pathogenesis in aged SARS-CoV-infected mouse models [
19,
21]. As we did not sample earlier than day 5 post infection in the lung, the increased inflammation in our aged monkeys may have been missed. Our sampling schedule may have also precluded collections at optimal inflammatory response time points, as a biphasic pattern occurring in waves, at day 2 and 7 post infection with mouse-adapted SARS-CoV strains has been described in aged rodents [
20,
36].
In addition to the differences between our findings and those from aged mice, most of our results are also dissimilar to the previously reported aged nonhuman primate SARS-CoV study by Smits et al., [
22] as we did not observe stronger innate immune responses or more severe pathology in our aged monkeys. In comparing our results, there are several experimental design aspects to take into consideration, beyond differences in dose and route of SARS-CoV infection. First, cynomolgous macaques were used in the previous study whereas our SARS-CoV infection model utilizes African green monkeys. In addition, timing of lung sample collection differed between the two studies, days 2 and 4 were assessed in the cynomolgous macaques while days 5, and 10 post infection were evaluated in our study. Furthermore, a standardized lung collection scheme was used in our immunology and pathology analysis, to give a more comprehensive perspective whereas their study assessed gene changes in specific regions of high SARS-CoV replication. Despite the major differences in study design, there were several results that were comparable, including the similarity of SARS-CoV viral titers in the aged and juvenile lung at later infection time points (days 4 and 5). In addition, several chemokines, including CCL5, CXCL12 and CCL20 were significantly increased in the lungs of SARS-CoV-infected aged monkeys here, which is consistent with the finding of elevated chemokine signaling genes in the aged lungs of cynomolgous macaques during SARS-CoV infection by microarray [
22].
Immune senescence has been shown to impact innate as well as adaptive branches of the immune system, both contributing to the diminished immunity observed in the elderly [
24,
37]. In examination of the kinetics of lung and lymph node cell expansion and/or trafficking following SARS-CoV infection, we found significantly reduced lung macrophages, DCs, CD8 T cells and B cells in the aged animals relative to their juvenile counterparts. Not only were there fewer lung macrophages and DCs in the aged animals, but the frequency of costimulatory CD86+ cells were also significantly reduced. In contrast, the lymph node reaction showed no age-specific differences. Interestingly, although lung leukocyte numbers were reduced in aged monkeys there were disproportionately high levels of the chemokines capable of recruiting these cells into the lung, CCL5, CXCL11, CXCL12 and CCL20. This prompted our examination of the expression of the corresponding ligand-binding receptors on peripheral blood leukocytes to determine if there were age-related differences in the migratory capacity of these cells into the lung. For CCL5 receptors, we found significantly reduced frequencies of CCR3 but not CCR1+ monocytes in aged animals and similar frequencies of CCR5+ CD8 T cells in both age groups. The frequency of aged DCs expressing the CCL20 receptor, CCR6 showed the most dramatic age-specific reduction which is reasonable, given that CCR6 is predominantly expressed by immature DCs and may reflect a reduction of these cells in our aged animals. The observed differences in lung chemokine and ligand receptor expression may reflect age-specific modulation of leukocyte trafficking into the lung. Although we did not confirm these observations with mechanistic in vitro experiments, previous studies have shown that DCs acquired from older individuals display significantly impaired ability to migrate in response to chemokines [
38].
We also examined age-specific differences in both T and B cell responses to SARS-CoV infection systemically and at mucosal sites. Not surprisingly, the proportion of naïve CD8 T cells in the lung and lymph node was significantly reduced in aged compared to juvenile animals in SARS-CoV infection. This corresponded to significantly lower levels of CD8 T cell proliferation at these sites in aged animals as well. Interestingly, the frequency of cytotoxic enzyme positive (granzyme B) T cells was higher at 1 d.p.i. in the periphery of aged compared to juvenile animals. However, aged peripheral granzyme + T cell frequencies remained unchanged by SARS-CoV infection whereas juvenile cytotoxic T cells showed infection-induced fluctuation. In contrast to peripheral blood, the lung and lymph node of aged animals had lower frequencies of granzyme + T cells compared to juvenile animals at 5 and 10 d.p.i.. The humoral response was also greatly reduced in aged SARS-CoV-infected monkeys with significantly reduced serum neutralizing antibodies and mucosal anti-SARS-CoV IgA in the aged host. This limited antibody response may be related to the dramatically reduced lung and peripheral B lymphocyte populations observed in the aged monkeys which is consistent with reports in elderly humans [
39]. Despite the deficiencies observed in the aged adaptive responses, viral titers were only significantly higher in aged monkeys at day 1 post infection. This suggests that innate or other compensatory immune responses are sufficient for viral control in aged monkeys and serves to further support the minimal involvement of cytotoxic T cells and neutralizing antibodies in SARS-CoV clearance as demonstrated in mouse depletion models [
36].
Taken together, our data indicate that systemic and mucosal immunity to SARS-CoV infection differs in the aged as compared to the juvenile host. This knowledge will be important to consider in the design of effective intervention strategies for SARS-CoV and potentially other respiratory infections. Our experimental results support future mechanistic studies to identify adjunctive strategies capable of overcoming the immune deficits of the aged airway mucosa, which may ultimately translate into novel approaches to enhance efficacy of vaccines against respiratory pathogens for the elderly population.
Conclusions
In this study we found that viral titers in aged, as compared to juvenile monkeys were significantly higher early after infection, but levels became comparable in both age groups at later infection time points. We observed significant age and infection-dependent differences in both the systemic and mucosal immune compartments with more dramatic changes in cytokine levels and leukocyte frequencies in lung as compared to peripheral blood and tracheobronchial lymph nodes. In regards to the exacerbated SARS-CoV responses previously reported in aged nonhuman primates, we did not observe enhanced host immunity or pathology in our aged monkeys. Instead, we detected less inflammatory cytokines and total lung leukocytes, as well as reduced adaptive responses with increased age. Although our results are discrepant, both studies indicate that pulmonary immunity is intrinsically different in the aged as compared to the juvenile host, warranting further exploration given the important implications this has for vaccine and therapeutic design for the elderly.
Methods
Ethics statement
All procedures were conducted under protocols approved by the Institutional Animal Care and Use Committee (IACUC) at Lovelace Respiratory Research Institute (LRRI), all facilities were accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC), and guidelines for nonhuman primates described in the Guide for the Care and Use of Laboratory Animals, National Research Council, were strictly adhered to.
Nonhuman primates
The animals used in this study were adult African green monkeys obtained from the Barbados Primate Research Centre, Barbados, West Indies through Global Research Supply, LLC (Reno, NV), a Class B USDA licensed animal dealer. A permit to import the animals into New Mexico was acquired through the Fish and Wildlife Service by LRRI. The aged African green monkeys were determined to be ~10-20 yrs old while the remaining subjects were young adults (~2-6 years old). Additionally, two female African green monkeys of 17 yrs and 18 yrs were acquired from the Wake Forest University Primate Center (Medical Center Boulevard, Winston-Salem). As there is an estimated 3.2-fold difference for relating age between humans and rhesus macaques [
26,
27], the aged African green monkeys in our study are expected to represent approximately, 50 year old humans while the juvenile African green monkeys correspond to 6–12 year olds. The rhesus-to-human age conversion was used given that the estimated age conversion is not as well established for African green monkeys and there is 95% genetic similarity [
40] between the two Old World monkey species belonging to the Super Family Cercopithecoidea. Randomization and group assignments for the aged monkey studies were performed with Provantis Integrated Preclinical Software (Instem Life Science Systems Ltd.). All monkeys were quarantined for 6 weeks prior to the study, during which they were tested for tuberculosis. Animals were given a subcutaneous microchip for identification and temperature measurements (IPTT-300 implantable temperature transponder and a WRS-6007 Handheld Wireless Reader System (Bio Medic Data Systems, Inc, Seaford, Delaware)). During the infectious portion of the study, animals were individually housed indoors in stainless steel cages with wire mesh bottoms. Temperature and humidity ranges were controlled along with 12 h light and dark cycles. Animals were also given environmental enrichment including toys twice per day. Tap water from the institutional watering system was available ad libitum and animals were fed twice a day Harlan Teklad Certified 20% Monkey Diet (4 to 6 biscuits/kg body weight) with bananas or apples for enrichment 3 times per week.
SARS-CoV infection of African green monkeys
The SARS-CoV strain HKU-39849 was provided by Dr. Leo Poon (Department of Microbiology, University of Hong Kong, Hong Kong, China) and viral stocks were generated in Vero E6 cells. Experimental infection with 10
7 plaque-forming units (PFU) of SARS-CoV instilled intranasally was conducted as previously reported [
28]. Monkeys were examined by trained Laboratory Animal Technicians twice per day (at least 6 h apart) on each day of the study for parameters including appetite, appearance, activity, stool, posture, neurological signs, respiration, ocular discharge and nasal discharge. Onset of any abnormal clinical sign was documented and a score sheet completed each day beginning from the date the animal fell ill. Euthanasia to alleviate suffering was conducted based on the clinical observation score sheet and consultation with the Staff Veterinarian. Animals were sacrificed at 5 and 10 d.p.i. (n = 6 for juvenile and n = 5 for aged animals at each time point) by intravenous overdose of Euthasol® (Virbac AH, Inc., Fort Worth, TX). Virus-free cell culture medium was used to inoculate mock-infected controls (n = 6 for juveniles and n = 2 for aged). Some of the immunology, virology and pathology data for the juvenile animals were previously reported [
28] but are included in this manuscript for comparison with aged monkeys.
Plaque and neutralization assays
Virus titers were measured in plaque assays by applying serial dilutions of homogenized tissue supernatants onto Vero E6 cell monolayers as previously described [
28]. To measure SARS-CoV neutralizing antibodies in serum, samples were serially diluted and incubated with 2,000 PFU/ml SARS-CoV overnight before inoculation onto Vero E6 cells. Titers are expressed as the reciprocal of the highest dilution at which the cytopathic effect was completely inhibited.
Flow cytometry
Standardized collected lung tissue from the proximal portion of the right caudal lobe was processed into single-cell suspensions for flow cytometry. Of note, the samples for flow cytometric analysis and cytokine protein evaluation were acquired from adjacent tissue regions. Single cell suspensions were prepared as detailed in [
28]. Briefly, mechanical and enzymatic digestion with Liberase (Roche, Pleasanton, CA) and DNAse (Sigma, St. Louis, MO) solution was performed prior to overlaying samples on a Percoll (Sigma) gradient for 20 min at 500 × g with no brake. Tracheobronchial lymph nodes were collected in a standardized manner in which the same region and size of tissue was collected from each animal. Single-cell suspensions were made by mechanical disruption. Peripheral blood mononuclear cells (PBMC) were isolated from heparinized blood with a Ficol-Hypaque (Sigma) gradient as previously reported [
41]. All flow cytometric analysis was conducted on previously frozen PBMC, lung or lymph node leukocytes. The following antibodies were used in 4- or 6- color staining panels: CD3 (clone SDP34-2); CD8 (clone SK1); CD11c (clone SHCL-3); CD14 (clone M5E2); CD20 (clone L27); CD68 (clone KP1, Santa Cruz Biotechnologies, Dallas, TX); CD86 (clone 2331Fun1); Granzyme B (clone GB11, Invitrogen, Carlsbad CA); HLA-DR (clone L243); Ki67 (clone B56); CCR1 (clone 53504); CCR3 (clone 61828); CCR5 (clone 3A9); CCR6 (clone 11A9) and CCR7 (clone 150503, R&D Systems, Minneapolis, MN) all conjugated to fluorochromes FITC, PE, PerCP, PerCpCy5.5, APC, PECy7, Alexa647 or APCCy7 (BD Biosciences, San Jose, CA unless specified). For intracellular antigens, BD FACS lyse and permeabilization solutions were used according to the manufacturer’s instructions. CD4 was not assessed given the downregulation of this marker on African green monkey T lymphocytes [
42‐
44]. Antibody-stained samples were fixed for 16 hr in 1% paraformaldehyde and 3% FBS in phosphate buffered saline. Sample data were acquired on a FACS-Calibur or FACS-Canto flow cytometer instrument (BD Biosciences) and data files analyzed utilizing FlowJo software 7.6.1 (TreeStar, Medford, OR). For all flow cytometric analysis, within one experiment, the gates applied to each sample were identical (See gating strategies in Additional file
1: Figure S1, Additional file
2: Figure S2 and Additional file
4: Figure S3). When appropriate and possible, similar gates were used across the two different experiments (juvenile and aged samples). However, based on changes to flow cytometer settings and fluorescence minus one controls, some gates were adjusted between the two experiments. Markers were chosen for characterization of specific leukocyte subsets in PBMC, lymph node, and lung based on previously published studies in humans or nonhuman primates [
45‐
53].
ELISAS and cytokine bead-based assays
Lung tissue was acquired in a standardized collection scheme in which a section of the proximal right caudal lobe was divided into two portions, one for homogenization to assess cytokines and antibodies and virus while lung leukocytes were isolated from the other portion for flow cytometry (see previous section). For homogenization, a volume of RPMI media (Gibco, Life Techologies, Grand Island, NY) equal to 10% of the lung tissue weight was added. SARS-CoV IgG and IgA antibodies were measured by ELISA detailed in [
28]. Briefly, ELISA plates were coated with purified recombinant SARS-CoV S protein in carbonate-coating buffer (S protein NR-686 obtained through NIH Biodefense and Emerging Infections Research Resources Repository) and nonspecific binding blocked with PowerBlock (Biogenex, San Ramon CA). Starting at a 1:25 dilution, serially diluted lung tissue homogenate supernatants (clarified by centrifugation) were incubated on S-protein coated plates overnight at 37°C. An anti-monkey IgG or IgA HRP-conjugated antibody (KPL Inc., Gaithersburg, MD) was applied followed by substrate development with ABTS Microwell peroxidase substrate system (KPL) and absorbance reading at 405 nm using a Thermo Electron Corporation plate reader (Thermo Electron Corporation, Houston, TX). Data was acquired with Ascent software (Ascent Software, London, UK) and the ELISA antibody titer recorded for each sample was the reciprocal of the highest dilution in which the optical density (O.D.) reading for S-protein bound wells was at least two-fold higher than that of the nonfat milk control. The O.D. of the highest titer chosen also had to fall within the linear range of the serial dilutions.
Cytokine and chemokine protein was measured in lung tissue homogenates and serum using both human and nonhuman primate multiplex bead-based array kits (Millipore, Billerica, MA). The assays were performed according to the manufacturer’s instructions with an overnight incubation of the samples in antibody-immobilized beads. Data was collected on the Bio-Plex System (BioRad, Hercules, CA) and a weighted 5-parameter logistic curve-fitting method used to calculate the concentration of individual analytes. All measurements were performed in duplicate and data reported as pg/ml. CD40L was undetectable in serum of either age group and IL-5, TNF-alpha, IL-17, IL-2, CCL24 and CCL26 were below the level of detection in lung tissue homogenates.
Histopathology
Sampling of tissues for histopathology was performed as previously reported [
28] in a standardized manner with random assessment. All organs were fixed in 4% paraformaldehyde solution prior to embedding in paraffin. 5 μm thick tissue sections were stained with hematoxylin and eosin for examination microscopically by a board certified veterinary pathologist. Histologic lesions were graded for severity (0 = normal, 1 = minimal, 2 = mild, 3 = moderate, 4 = marked) and distribution, (focal, multifocal, diffuse).
Statistical analysis
Infection and age differences were evaluated using a 2-way ANOVA with Bonferroni post-tests or two-tailed student T-tests, where appropriate. All data for statistical evaluation was log transformed and analyzed with GraphPad Prism 5.0 software (GraphPad, La Jolla, CA). A p value of 0.05 or less was considered statistically significant.
Competing interests
The authors declare no financial or non-financial competing interests in the work presented in this study.
Authors’ contributions
CCC data collection, analysis and interpretation of data, figure preparation, manuscript writing; ND data collection, analysis, manuscript editing; NF conception and design, analysis and interpretation of data; JBK data collection, analysis, manuscript editing; KO conception and design, study director, manuscript editing; JT data collection, data analysis, manuscript editing; JVW data collection, analysis, manuscript editing; FH board-certified pathologist, histopathology analysis, manuscript editing, KSH conception and design, interpretation of data, final approval of the manuscript. All authors read and approve the final manuscript.