The inflammatory response to bacteria requires the interaction of pattern recognition receptors with bacterial surface constitu- ents, and humans deficient in components of inflammatory sig- nalling pathways such as IRAK4 are prone to invasive pneumococcal disease. Pneumolysin is a well-recognised viru- lence factor for Streptococcus pneumoniae that has multiple effects on the host immune response that are primarily thought to be pro-inflammatory; including causing IL1b release due to pore formation, and epithelial cell layer breakdown. We hypoth- esised that pneumolysin deficient TIGR4 (a serotype 4 strain) would induce less inflammatory cytokines than wildtype from human monocyte derived macrophages. While both pore form- ing and non-cytolytic purified pneumolysin induced dose dependent inflammatory cytokine release, the pneumolysin defi- cient bacteria induced greater TNF and IL6 than wildtype, by qPCR and ELISA measurement of protein. This was reduced by inhibition of phagocytosis with cytochalasin D. Given the pore forming effects of pneumolysin we assessed whether differential cell death contributed to the differences in inflammatory response. While wildtype bacteria caused more cell death at 24 h, inhibition of caspases had no effect on the cytokine response suggesting that apoptosis pathways don’t directly influence the early inflammatory response. Transcriptome analysis confirmed increased pro-inflammatory and interferon gene signalling with the mutant strain, with reduction of the inflammatory and inter- feron signature with inhibition of phagocytosis. Wildtype bacte- ria induced less NFkB translocation, but more IRF3 translocation than Dply. An in vivo intranasal mouse infection showed wildtype was more virulent, with more bacteria recov- ered from bronchoalveolar lavage fluid at 4 h. However, this was associated with reduced TNF compared to Dply. Neutralis- ing TNF intranasally abrogated the difference in bacteria recov- ered between wildtype and Dply. Thus, the early inflammation dampening effects of pneumolysin released within the phagoly- sosome may be an important contribution to its virulence by allowing bacterial replication at mucosal surfaces. This may be due to IRF3 mediated inhibition of inflammatory cytokine tran- scription. Better understanding of the biology of pneumolysin may aid in adjuvant treatment of S. pneumoniae. New developments in cough S87 ARE ORAL STEROIDS EFFECTIVE IN TREATING THE SYMPTOMS OF ACUTE LOWER RESPIRATORY TRACT INFECTION IN NON-ASTHMATIC ADULTS? THE ORAL STEROIDS FOR ACUTE COUGH (OSAC) PLACEBO- CONTROLLED RANDOMISED TRIAL 1 AD Hay, 1 HE Downing, 1 ST Brookes, 1 G Young, 2 A Harnden, 1 SP Hollinghurst, 3 D Kendrick, 4 P Little, 1 MT May, 4 MV Moore, 3 E Orton, 2 K Wang, 5 MJ Thompson. 1 University of Bristol, Bristol, UK; 2 University of Oxford, Oxford, UK; 3 University of Nottingham, Nottingham, UK; 4 University of Southampton, Southampton, UK; 5 University of Washington, Seattle, USA 10.1136/thoraxjnl-2015-207770.93 Background The majority of UK adults experience at least one lower respiratory tract infection (LRTI, or acute bronchitis) a year. Despite an absence of evidence in this patient group, some GPs prescribe inhaled or oral corticosteroids. OSAC sought to demonstrate ‘proof of concept’ symptomatic effectiveness of a moderate dose of oral corticosteroid for adults without asthma or COPD with acute LRTI. Methods OSAC was a double blind, placebo controlled RCT set in GP practices in England, powered to investigate if oral predni- solone reduces the duration of moderately bad or worse cough and/or the severity of its associated symptoms, when compared to placebo, by at least 20%. Adults (18 years) with acute (£28 days) cough, for whom same-day antibiotics were not clinically indicated, and without asthma or COPD, received 40 mg oral prednisolone or matched placebo for 5 days. Symptom diaries, completed for up to 28 days, measured two primary outcomes: the duration of moderately bad or worse cough; and the average severity of all symptoms on days 2 to 4 on a scale of 0–6. We sought to demonstrate a minimum clinically important reduction of 20% in each outcome. Results 398 participants were randomised to either prednisolone or placebo tablets (198 and 200 respectively) from 54 UK pri- mary care sites. Attrition was lower than expected, giving over 85% power for the two primary outcomes. Data were analysed on an intention-to-treat basis. The median duration of moder- ately bad or worse cough was 5 days in both groups (IQRs 2–8 and 3–8 for prednisolone and placebo respectively). Adjusting for trial centre and baseline characteristics, this gave a hazard ratio of 1.11 (95% CI 0.89 to 1.39, p = 0.35). Symptom severity was lower in the prednisolone group (mean 1.99 vs 2.16), adjusted difference -0.090 (-0.212 to 0.003, p = 0.152). Conclusions We found no evidence that a moderately high dose of oral corticosteroid reduced either duration of moderately bad (or worse) cough, or symptom severity at days 2 to 4 in adults without asthma or COPD with LRTI not requiring immediate antibiotic treatment. Lower dose oral or high dose inhaled corti- costeroids are also unlikely to be beneficial. S88 THE VIRAL MIMIC POLYINOSINIC: POLYCYTIDYLIC ACID (POLY I:C) INDUCES TRPA1 CHANNEL HYPER- RESPONSIVENESS IN AN ADULT HUMAN STEM CELL- DERIVED SENSORY NEURONAL MODEL 1 R Clarke, 1 K Monaghan, 2 I About, 1 I El Karim, 1 JG McGeown, 1 SL Cosby, 1 TM Curtis, 1 FT Lundy, 1 L McGarvey. 1 School of Medicine, Dentistry and Biomedical Sciences, Queen’s University, Belfast, Northern Ireland, Belfast, N Ireland; 2 Aix Marseille Université, Marseille, France 10.1136/thoraxjnl-2015-207770.94 Background Changes in airway neuronal activity are likely to underpin the heightened irritant responses such as excessive cough and wheeze which accompany respiratory virus induced exacerbations of airways disease. The mechanisms responsible are unknown but we hypothesised that neurons express patho- gen recognition receptors such as toll-like receptors (TLR) through which viruses may alter neural function. Investigating this is hampered by the lack of suitable human tissues with both nerve endings and cell bodies present. We have refined an adult human neural crest stem cell-derived sensory neuronal model to overcome this obstacle. Methods Human dental pulp stem cells (hDPSCs) were differen- tiated towards a neuronal phenotype, termed peripheral neuro- nal equivalents (PNEs). Using molecular and immunofluorescent techniques, together with whole cell patch clamp electrophysiol- ogy, we investigated the expression and function of TLRs and the transient receptor potential (TRP) channels TRPV1 and TRPA1 on PNEs. We then assessed the effects of exposure to a viral mimic, the synthetic TLR3 agonist (polyI:C), on cytokine Spoken sessions A50 Thorax 2015;70(Suppl 3):A1–A254