SORVEGLIANZA DELLE INFEZIONI RESPIRATORIE ACUTE (ARI): APPROCCI INNOVATIVI PER L'IDENTIFICAZIONE E LA CARATTERIZZAZIONE DEGLI AGENTI VIRALI COINVOLTI

Surveillance of acute respiratory infections (ARI): innovative approaches for the identification and characterization of viral agents Introduction. Acute respiratory infections (ARI) are ubiquitous, air-borne transmitted and highly contagious infections, characterized by typical epidemic pattern. Though ARI causative agents may be bacterial or viral, viruses are by far the most common causes of ARI. In recent years, the global epidemiological scenario has been enlivened by the identification and emergence of many airborne pathogens (such as influenza viruses A/H5N1, A/H7N7, A/H7N9, coronaviruses SARS and MERS), which have been co-circulating along with the already known viral strains (i.e. seasonal influenza viruses, parainfluenza viruses, respiratory syncytial viruses, etc.), thus highlighting the public health concern about emerging infectious disease. Objectives. The project aimed at applying new molecular biology techniques, "virus discovery" methodology and bioinformatics analyses to investigate the etiology of ARI. The specific objectives were: 1) Identification of viral pathogens responsible of severe acute respiratory infections (SARI) and acute respiratory distress syndrome (ARDS) during the pandemic and post-pandemic (2009-2011). On purpose, the proportion of SARI/ARDS cases and deaths due to A(H1N1)pdm09 infection and the impact of other respiratory pathogens were evaluated during the pandemic and post-pandemic period in Lombardy. Additionally, unknown viruses were investigated in those cases for which diagnosis remained negative by using VIDISCA-454 methodology, a “virus discovery” technique. This analysis was performed at the Laboratory of experimental virology, Academic Medical Center (AMC), University of Amsterdam, where I completed an internship under the supervision of Prof. Lia van der Hoek. 2) Evaluating the genetic variability and molecular evolution of respiratory syncytial virus (RSV). In order to reconstruct the origin and phylodynamic history of RSV, the genetic diversity and evolutionary dynamics of RSV A and RSV B identified in respiratory specimens collected from children aged ≤ 3 years hospitalized in Lombardy for ARI over six epidemic seasons (2006 to 2012) were analyzed. Materials and methods 1) From October 2009 to December 2011, 206 respiratory samples were collected from patients (61.2% males, median age: 44.3 years) hospitalized for SARI/ARDS. Nucleic acids were purified by NucliSENS® easyMAG® (bioMérieux, France), and analyzed by real-time RT-PCR assay to identify influenza virus. The clinical specimens that resulted negative to influenza virus detection were then screened by real-time RT-PCR/PCR for a panel of respiratory pathogens (Respiratory MWS r-gene™ Real-time PCR, bioMérieux, France) to detect: RSV A and B; human metapneumovirus (hMPV) A and B; human rhinovirus (hRV) and enterovirus (hEV); adenovirus (AdV); human bocavirus (hBoV) 1-4; human coronavirus (hCoV) 229E, NL63, OC43, HKU1; human parainfluenza virus (hPIV) 1-4; Chlamydophila pneumoniae; Mycoplasma pneumoniae. Cases resulted negative to all diagnostic assays were further investigated by VIDISCA-454 (virus discovery cDNA-AFLP) technique. This is a virus discovery method based on recognition of restriction enzyme cleavage sites, ligation of adaptors and subsequent amplification by PCR combined with high-throughput sequencing 454 FLX/Titanium system of Roche. 2) RSV A (n=23) and RSV B (n=12) sequences obtained from oro-pharyngeal swabs of RSV-infected children aged ≤3 years hospitalized for ARI from 2006 to 2012 were considered for molecular characterization. Sequences were obtained by multiplex-PCR to amplify a fragment of RSV G gene and several bioinformatic programs were used for the phylogenetic and phylodynamic analysis. Phylogenetic trees of RSV A and RSV B sequences were constructed by MEGA 5 program using the Neighbor-Joining method to identify RSV genotypes circulating and a bootstrap re-sampling analysis was performed to test tree robustness. To clarify RSV variability, amino acid mutations analysis was performed, and potential N-glycosylation and O-glycosylation sites were predicted by NetNGlyc 1.0 and NetOGlyc 3.1 programs, respectively. To evaluate site-specific selection pressure, different Maximum Likelihood approaches were applied (SLAC, FEL/IFEL, and MEME) by DATAMONKEY. To assess the evolutionary dynamics of RSV A and B, dated trees and evolutionary rates were estimated by BEAST with a Bayesian Markov Chain Monte Carlo (MCMC) approach. This analysis allowed identifying the time of the most recent common ancestor (tMRCA). Results and discussion 1) During the pandemic and post-pandemic different pathogens have co-circulated and were associated with clinical cases in the study, but influenza virus A(H1N1)pdm09 showed the greatest impact. Influenza A(H1N1)pdm09 virus was detected in 58.3% (120/206) of SARI/ARDS cases (61.7% males; 13.3% aged ≤5 years, 67.5% aged 6-64 years). A(H1N1)pdm09 was identified in 77.8% of fatal ARDS cases. The impact of respiratory pathogens other than A(H1N1)pdm09 was 19.4% (40/206) (65% males; 30% aged ≤5 years, 47.5% aged 6-64 years). The influence of other respiratory viruses was significantly lower (19.4% vs. 58.3%, p<0.0000001): hRV/hEV were the most commonly detected viruses, but also A(H3N2) influenza virus has played a significant role. Forty-six (46/206: 22.3%) SARI/ARDS cases (including two fatalities) resulted negative to all diagnostic assays and were further investigated by VIDISCA-454 that revealed no sequence reads that could belong to a novel virus or viral variant; however it enabled the identification of one case of undiagnosed measles. VIDISCA-454 methodology proved to be a sensitive and specific method, successfully applied to the monitoring of viral respiratory infections. Anyway, nearly 22% of SARI/ARDS cases did not obtain a definite diagnosis. In clinical practice, great efforts should be devoted to improve diagnosis of severe respiratory infections and to reduce such “diagnostic gap”. The advantage from relying upon more accurate diagnosis could benefit the patient - in term of receiving the more appropriate antiviral drugs -, and could provide more detailed information on viruses circulating in the community, thus making public health authorities aware so as to adjust their policies accordingly. 2) From phylogenetic analysis resulted that 3 RSV A sequences clustered in genotype GA2 and all the other isolates clustered with NA1 genotype. Compared to the reference strain, 31 amino acid substitutions were identified. Phylogenetic analysis of RSV B sequences showed that study sequences were included in BA genotype tended to cluster within a clade including also reference sequence BA4 and 8 amino acid substitutions were identified among all of them. Similar mean evolutionary rates for RSV A and RSV B were estimated, 2.1x10-3 subs/site/year (95% highest density probability (HPD): 1.7-2.5x10-3) and 3.03x10-3 subs/site/year (95%HPD: 2.1-3.9x10-3), respectively. The tMRCA for the RSV A tree root was 71 years (95%HPD: 60-85), suggesting an origin of the currently circulating strains back to 1940s. The study strains within clade NA1 shared a single significant internal node with an estimated mean tMRCA of 7 years ago (2005). The three GA2 strains had a significant tMRCA dating back to 16 years ago (95%HPD = 14-18). The dated tree obtained with RSV B strains showed a temporal-structure similar to RSV A. In particular, the tree root had an estimated mean tMRCA of 55 years ago (1957). All the studying strains clustered within a significant subclade, having a tMRCA estimate of 9 years ago (2003), including also a single BA4 strain. The RSV A Bayesian skyline plots (BSP) showed a first pick of the number of effective infections in the second half of 1980s, followed by a decrease of transmission events ending in about 2005, when a sharp growth restored the original viral population size. The RSV B BSP showed a similar trend, with a decrease in the effective number of infections occurring between the mid-1980s and 1990s, followed by a rapid growth in early 2000s. The RSV A site-specific selection analysis identified 10 codons under positive selection and a total of 39 sites were found to be under negative selection. The RSV B codon-specific selection analysis identified only one positively selected site and 27 negatively selected. Different patterns of O- and N-glycosilation sites were found by analyzing RSV A and RSV B studied sequences. Phylogenetic and phylodynamic analysis permitted to better understand the natural history of the virus, even if RSV remains difficult to control due to its antigenic diversity. G protein variability may play a significant role in RSV pathogenesis by allowing immune evasion. It is important monitoring changes in sequences coding this protein to permit the identification of future epidemic strains and to implement both vaccine and therapy strategies. This study thus contributed to have a better knowledge on the molecular epidemiology of RSV in Lombardy and provided data for a comparative analysis with other strains circulating in other regions of the world. Conclusions. This project showed that respiratory virology needs a constant update of diagnostic procedures and surveillance scheme, essential support in the study and monitoring of viral infections of both classic and, above all, emerging or newly identified viruses. The application of cutting-edge technologies can be a successful tool for public health to face such emerging threats.