Title
Obesogenic environment increases virulence of serially passaged influenza
Abstract
Introduction
Seasonal and pandemic outbreaks of influenza are a leading cause of morbidity and mortality worldwide with up to 15% of the world's population infected each year[1, 2]. Influenza A virus (IAV) infects the respiratory epithelium and causes a sore throat and cough, fever, myalgia, and fatigue and may lead to severe pneumonia and death[3]. A carefully coordinated immune response is required to clear the virus and protect the host from immunopathology[4, 5]. Host differences in the immune response to viral infection can exacerbate severe disease. Sex[6, 7], immunocompromised status[8], and chronic conditions[9, 10] can impact both innate and adaptive immune responses leading to differences in the viral immune response. Host status has also been implicated as a driving force in the emergence of new viral variants[11-13].
During the 2009 H1N1 pandemic, obesity was shown to be a risk factor for developing severe disease association with IAV infection worldwide[14-17]. Increased hospitalization, ICU duration, secondary infections, and mortality was observed in populations of BMI greater than 30[14-18]. The obese host provides a unique microenvironment for disease pathogenesis characterized by a state of chronic, low-grade inflammation[19, 20]. Increased adiposity results in increased production of pro-inflammatory cytokines and adipokine levels as well as reduced antiviral cytokines, eliciting suppressed innate and adaptive immune responses[10, 21, 22].
Altered immune surveillance and first-order cytokine and chemokine production is characteristics of the obese state. This impacts the cellular innate response, increasing the circulation of pro-inflammatory M1 macrophages, reduced mature macrophages in the lung, and impaired neutrophil responses[10, 20, 23]. The defective innate response impacts downstream adaptive responses as both B-cell and T-cell responses are compromised in the obese state[21, 24]. These baseline changes in the immune response may predispose obese patients to infection[20]. Obese patients infected with hepatitis[25, 26], HIV[27], respiratory viral infections[28], and other microbial pathogens[29-31] show increased morbidity and mortality, poorer immune memory, impaired vaccine response, and reduced antiviral effectiveness[32-34].
Previous studies in mouse models of obesity have demonstrated that the obese state impairs the immune repose to IAV infection. In a diet-induced model of obesity (DIO), mice had a 7 times greater mortality with IAV that was associated with increased lung infiltration, pathology, and delayed wound repair as compared to lean controls[23, 35-38]. Secondary immune responses were also dysregulated[23, 39, 40]. Studies in genetically obese mice lacking the leptin signaling molecule (OB) also suffered increased disease severity, reduced immune responses, and impaired wound healing[41].
IAV, like other RNA viruses, replicates by means of an RNA-dependent RNA polymerase (RdRp) which lacks proofreading functionality[42, 43]. This affords a high error rate during viral replication and leads to the emergence of many minor variants in the IAV within-host population that may affect viral fitness[44], which is a tenant of the viral quasispecies theory. RNA viruses, like IAV, exist as a quasispecies and exhibit large populations sizes, quick replication time, and low RdRp fidelity, which generates many low-frequency viral variants that may differ only slightly in sequence[45]. Based on the far-reaching effects obesity has on the host immune response and the knowledge that host status can impact the emergence of viral variants, we hypothesized that passaging influenza virus through an obesogenic lung microenvironment would evolve differently as compared to passage through a lean environment. The unique immune pressures in the obese host can impact the viral population itself, leading to the generation of novel minor variants leading to a more virulent population.
Serial passaging of A/California/04/09 pH1N1 in OB and DIO mice increased subsequent disease severity and mortality in reinfection models as well as led to novel mutations associated with increased virulence. To further these results, a clonal, plaque purified A/California/04/09 was used to infect OB and WT mice to elucidate the evolutionary forces unique to the two hosts. Understanding how the emergence of new infectious diseases and old diseases with new pathogenic properties could be influenced by the obesogenic state is paramount for the increasingly obese world.
Methods
Ethics statement
Viruses and animals
Generation of viruses: Serial passaging in diet-induced obese (DIO), lean (LN), genetically obese (OB) or wild-type (WT) mice used egg grown A/California/04/2009 (CA/09, pandemic H1N1 [pH1N1]) propagated in Maine-Darby canine kidney (MDCK) cells as the starting parental virus (P0). For the second passaging experiment, a reverse-genetics derived clonal parental virus (cP0) was used to elucidate differences in immune pressures in obese and lean mice. Something about transfection in 293T cells and generation of a clonal P0 virus. Titers were determined through tissue culture infectious dose 50% (TCID50) in MDCK cells. Infectious virus was calculated using the Reed-Muench endpoint[46].
Genetically obese mice: Eight-week old lean male C57BL/6 and B6.V-Lepob/ob genetically obese (OB) mice were obtained from Jackson Laboratory (Bar Harbor, ME). All animals were provided with food and water ad libitum.
Generation of diet-induced obese mice: Eight-week old male, C57Bl/6 mice were fed ad libitum either a high-fat (60% kcal%, D12492) or low-fat (10% kcal%, D12450B) diet (Research Diets) for 10 weeks to generate diet-induced obese (DIO) or lean (LN) mice respectively. Water was provided ad libitum.
Passaging experiments: Lean (wild-type or diet control) and obese (DIO or OB) mice (n=3-5/group) were anesthetized under isoflurane and infected with starting P0 or cP0 viruses at 103 TCID50 in 25 uL of phosphate-buffered saline (PBS). Mice were observed for clinical signs, morbidity, and weight loss. Three days post infection (dpi), mice were sacrificed under CO2 euthanasia and the lungs were collected, scored for (0=normal, 1=focal hemorrhage, 2=extensive hemorrhage) homogenized and virus titers were determined. Pooled virus was then used to infect a new set of lean and obese mice respectively for a total of 5 or 10 passages (Figure 1).
MLD50 and reinfection experiments: Quantification of the virulence of WT, LN, OB, and DIO mice passaged viruses was accomplished by intranasally infecting n=5/virus/dilution 6-week-old C57Bl/6 with P5 viruses (P0 – 102, 103, 104, 105 TCID50 units; LNP5 – 102, 103 TCID50 units; DIOP5 – 102, 103, 104, 105 TCID50 units; WTP5 – 102, 103, 104 TCID50 units; OBP5 – 102, 103, 104 TCID50 units) in 25 uL of PBS to discern the mean lethal dose 50% (MLD50). Mice were euthanized at >30% weight loss and clinical scores of 4. To understand the disease progression of the passage 5 viruses, recipient wild-type C57Bl/6 mice were infected intranasally with 103 TCID50 passage 1 and passage 5 viruses. Mice were monitored for morbidity and mortality over 10 or 14 days. On 3, 10 and 14 dpi, samples were collected, lung appearance scored, either frozen at -80°C for future assays or perfused with 4% paraformaldehyde for histology. Weights, clinical signs (0=normal, 1=ruffled fur, 2=hunched back, slow circular movement, 3=trembling, 4=paralysis and moribund), and breathing scores (0=normal, 1=labored, 2=shallow, 3=rapid) were recorded over the course of the study. Haematoxylin and eosin (H&E) staining and scoring of lung sections was performed by the St. Jude Children's Research Hospital Veterinary Pathology Core Facility.
In vitro experiments
Viral replication: Infections were performed as described previously. Briefly, MDCK were cultured in Eagle's minimum essential medium (MEM, MediaTech, Manassas, VA) supplemented with 2 mM glutamine and 10% fetal bovine sera (FBS, Gemini BioProducts, West Sacramento, CA) and grown at 37°C under 5% CO2. MDCK cells were infected with either a high multiplicity of infection (MOI) of 1.0 or low MOI of 0.01 for 1 hour at 37°C. Cells were washed three times to remove unbound virus and infected cells were cultured in appropriate media containing 0.075% BSA and 1 μg/ml TPCK-treated trypsin. Aliquots of culture supernatants were collected at 4, 8, 12, 16, and 24 hpi and immediately stored at -80°C for determination of virus titers as described previously.
Plaque diameter analyses: MDCK cells were plated in 6-well plates and grown to 100% confluence. Cells were washed twice with PBS and incubated with serial 10-fold dilutions of P0, WTP5, LNP5, OBP5, and DIOP5 viruses in DMEM at 37°C and 5% CO2 for one hour. Media was aspirated and cells washed twice with PBS and covered in MEM overlay containing 0.3% BSA, 1ug/mL TPCK-trypsin, and 2% agarose. Plates were incubated for three days at 37°C and 5% CO2. Three days post inoculation, the overlay was removed and cells fixed in 10% formaldehyde for 30 min at room temperature. Cells were stained with 1% crystal violet in 5% ethanol. Plaques were counted manually from duplicate wells and plaque diameter measured using ImageJ software.
Sequencing and PCR
RNA Extraction: Viral RNA from lung homogenates was extracted using QIAmp Viral RNA Mini Kit (Qiagen 52904; Hilden, Germany). RNA concentration was measured by Nanodrop.
MS RTPCR: Multi-segment polymerase chain reaction (MS RT-PCR) was then performed using SuperScript™ III One-Step RT-PCR System with Platinum™ Taq High Fidelity DNA Polymerases (ThermoFisher 12574-035; Waltham, MA) and Opti1 set of primers (Opti-F1 5' GTTACGCGCCAGCAAAAGCAGG, Opti-F2 5' GTTACGCGCCAGCGAAAGCAGG, Opti-R1 5' GTTACGCGCCAGTAGAAACAAGG). 5uL of RNA was added and placed into a thermocycler paused at 55°C. The following cycling parameters were then followed: 1 cycle of 55°C/2min; 1 cycle of 42°C/60min; 94°C/2min; 5 cycles of 94°C/30s, 44°C/30s, 68°C/3.5min; 26 cycles of 94°C/30s, 57°C/30s, 68°C/3.5min; 1 cycle of 68°C/10min; and then held at 4°C. 5uL of the reaction was analyze by 0.8% agarose gel electrophoresis to verify all genomic segments are present, with PB1 and PB2 migrating together at 2.3Kb and minimal non-specific amplification below 800bp.
Amplicon Purification: The MS RT-PCR reaction was purified using Agencourt AMPure XP Kit (Beckman Coulter, A63880). Briefly, 40uL of the DNA amplicons were transferred to a 96-well microplate (Biorad #HSP9601; Hercules, CA) and mixed with Agencourt AMPure XP beads for magnetic separation of the amplicons. The bound beads were washed twice with 100% ethanol and the purified amplicons were resuspended in 1x Tris buffer plus EDTA (TE). The concentration of the purified DNA measured using Nanodrop prior to storage at -20°C.
Next-generation sequencing and bioinformatics: DNA amplicons were deep sequenced using Illumina MiSeq technology. Paired-end reads of 300bp/multiplexing other fancy words like flowcell amplicons from influenza samples and 15 million reads total.
M gene RT-qPCR: M gene qRT-PCR was performed on samples using the 4x Taqman Fast virus mix (ThermoFisher, 4444432) and influenza A virus primers and probes at 20uM. The primrs used were as follows: (F) 5'-GACCRATCCTGTCACCTCTGAC-3', (R) 5'-AGGGCATTYTGGACAAAKCGTCTA-3'. The probe used is as follows: (R) 5'-TGCAGTCCTCGCTCACTGGGCACG-3'). Real time thermocycler parameters are as follows: 1 cycle of 50°C/5min, 40 cycles of 95°C/20s, 1 cycle of 95°C/3s, and 1 cycle of 60°C/30s. Cycle threshold (Ct) values were converted to copy numbers based on output of Influenza A gBlock standard curve. The gBlock sequence used is as follows: 3'-TCGCGCAGAGACTGGAAAGTGTCTTTGCAGGAAAGAACACAGATCTTGAGGCTCTCATGGAATGGCTAAAGACAAGACCAATCTTGTCACCTCTGACTAAGGGAATTTTAGGATTTGTGTTCACGCTCACCGTGCCCAGTGAGCGAGGACTGCAGCGTAGACGCTTTGTCCAAAATGCCCTAAATGGGAATGGGGACCCGAACAACATGGATAGAGCAGTTAAACTATACAAGAAGCTCAAAAGAGAAATAACGTTCCAT-5'
M gene end point PCR: End point PCR was performed to analyze defective particle composition using the One Step Qiagen RT-PCR kit (Qiagen, 210212). M gene primers (F) 5'-ATATCGTCTCGTATTAGTAGAAACAA and (R) 5'-TATTCGTCTCAGGGAGCAAAAGCAGG were used. Thermocycler parameters were as follows: 1 cycle of 50°C/50min; 1 cycle of 94°C/15min; 30 cycles of 94°C/30s, 52°C/45s, 72°C/3min; 1 cycle of 72°C/10min; and a 4°C hold. Samples were analyzed by 1% agarose gel electrophoresis in tris base, acetic acid and EDTA (TAE) with ethidium bromide and imaged under UV light.
Results
Disease severity increases in OB and DIO passaged viruses and is independent from increases in viral load.
Adult (8-14 week old) obese (OB or DIO) and lean (WT or LN) mice were infected intranasally with egg grown A/California/04/09. At 3dpi, mice were sacrificed and lungs homogenized to recover passage 1 viruses. After titer on MDCKs, pooled virus from obese or lean animals was used to infect the subsequent set and the process repeated up to passage 5 (Figure 1). There was no significant difference in titer from recovered obese or lean passaged viruses at each passage (Figure 2 A, D) nor were there any observable increases in titer across all passaged 1 to 5 (Figure 2 A, D). However, obese animals showed more severe disease as clinical scores increased for OB (Figure 2 B) and DIO (Figure 2 E) mice infected with obese passaged viruses over the lean animals. Lung appearance scores were also higher for the obese groups (Figure 2 C, F). There were no differences in viral load as measured by TCID50 nor genome copies present in the OB versus WT (Figure 2 H) or the DIO versus LN (Figure 2 I) passaged virus. However, a decrease in M gene copy number was observed for the OBP1 virus but this result was not statistically significant.
Passage of IAV through an obese host increases virulence and subsequent morbidity and mortality upon infection of recipient wild-type mice.
Wild-type mice were infected with 103 TCID50/mL of parental P0, obese-passaged, or lean-passaged P5 viruses and monitored for survival and morbidity through clinical scores and weight loss over 10 or 14 days. Mice infected with obese-passaged viruses (OBP5 or DIOP5) showed significantly increased mortality (Figure 3 A, D). In mice infected with OBP5, none survived past 8dpi (Figure 3 A). Significant weight loss and high clinical scores accompanied the increase in mortality (Figure 3 B-C) as compared to P0 and WTP5. WTP5 infected mice did show increases in morbidity and mortality but the phenotype was not as extreme as seen with the OBP5 virus. The 50% mean lethal dose (MLD50) in wild-type C57Bl/6 mice was, > 105 for the parental virus (P0) virus, > 103 for the LNP5 virus, and 102.5 for the DIOP5 virus. Need data for OB and WT MLD50 (supplemental table 1).
DIOP5 virus infected mice showed a similar, but not as extreme, phenotype as OBP5 passaged virus. Mortality (Figure 3 D) and morbidity (Figure 3 E, F) were increased over mice infected with P0 and LNP5 virus. In analysis of M gene copy numbers present, OBP5 viruses had modest increase in copy numbers as compared to P0 and WTP5 (Figure 3 H). No differences in copy numbers were evident for DIOP5 viruses in reinfection experiments (Figure 3 I).
Wild-type mice infected with obese-passaged IAV show increased lung pathology.
Increased lung damage and inflammation was observed in mice infected with obese passaged viruses (Figure 5). Mice infected with P0 and P5 viruses were sacrificed, lungs perfused with paraformaldehyde, and whole lung sections prepared. Staining with H&E revealed increased inflammation and significantly greater pathology in mice infected with obese-passaged viruses (Figure 4). In wild-type mice infected with OBP5, there is increased need histology report (Figure 4 A). Quantification of pathology shows increased scores and more severe disease for mice infected with OBP5 (Figure 4 B). These findings were also observed in lean mice infected with DIOP5 or LNP5 viruses (Figure 4 C, D).
Obese passaged viruses show increased replication and plaque diameter in MDCK cells.
Next-generation sequencing reveals mutations associated with virulence in obese-passaged viruses.
Clonal virus experiment results.
Discussion
The two major experimental designs, using a mixed population starting virus (P0) (Figure 1 A) and a clonal virus (cP0) (Figure 1 B), helps to strengthen conclusions garnered from this work. Natural contact with IAV results in a mixed quasispecies of related viruses infecting the respiratory epithelium. Evolutionary forces act on the population as a whole and can lead to the emergence of several minor variants that can impact virulence. However, to better understand the distinct pressures in the obese and lean hosts, a clonal virus was used. With theoretically only a single viral species used to infect the initial passage 1 obese and wild-type hosts, the changes observed would demonstrate how the two host environments would mold the resulting viral population.
Obesity has often been described as inducing premature aging of the immune system. In fact, serial passaging of a coxsackievirus through an aged host results in increases in virulence that mirror the increased virulence seen in obese-passaged viruses in this study.