Article Text

PDF

S23 Multidimensional phenotypes of asthma
  1. TSC Hinks1,
  2. XY Zhou1,
  3. P Lum2,
  4. KJ Staples1,
  5. B Dimitrov3,
  6. C Smith3,
  7. J Ward3,
  8. PH Howarth1,
  9. AF Walls1,
  10. SD Gadola1,
  11. R Djukanovic3
  1. 1Academic Unit of Clinical and Experimental Sciences, University of Southampton Faculty of Medicine, Sir HenryWellcome Laboratories, University Hospital Southampton, Southampton, UK
  2. 2Ayasdi Inc., 636 Ramona Street, Palo Alto, USA
  3. 3Southampton NIHR Respiratory Biomedical Research Unit, University Hospital Southampton, Southampton, UK

Abstract

Introduction and Objectives Asthma is a complex disease involving many cell types including CD4+ and CD8+ T cells, eosinophils, basophils and mast cells, and their soluble mediators. Poor understanding of disease heterogeneity and mechanisms underlying distinct clinico-pathological endotypes limits progress.

Abstract S23 Figure 1.

Bayesian Network Analysis showing statistically significant relationships between clinical and immunological parameters measured in the study. The strength of the interactions (as Euclidean distance) is shown by line weight. Figure 1. Bayesian Network Analysis showing statistically significant relationships between clinical and immunological parameters measured in the study. The strength of the interactions (as Euclidean distance) is shown by line weight.

We hypothesised that identification of distinct multidimensional endotypes using combined clinical and pathological parameters would enable a better understanding of asthma mechanisms and phenotypes. We undertook to analyse a range of cell types and mediators across a wide clinical spectrum of asthma and health with application of a novel analytical approach: topological data analysis (TDA).

Methods 76 adult subjects underwent extensive phenotyping, using clinical assessment, lung function, methacholine challenge, sputum induction, phlebotomy, and bronchoscopy. Samples were analysed for measurement of CD4+, CD8+ and invariant T cell subsets, of T cell cytokines, of mast cell and basophil mediators, bradykinin and vitamin D, using 9-colour flow-cytometry, RT-PCR, multiplex ELISA and mass spectrometry. We used Bayesian network analysis (BNA) to define the inter-connectivity between the key biological parameters investigated and TDA to identify novel multi-dimensional endotypes of asthma.

Results We generated a Bayesian belief network using 37 key parameters on a subset of 62 subjects with the most complete data (figure), revealing asthma severity as a multidimensional feature, influenced by a wide variety of parameters, including Th2 inflammation, mast cells and basophils.

A TDA network generated using only pathological features revealed distinct pathological endotypes including i) mast cell / basophil mediated ii) severe, obese, high bradykinin iii) minimal inflammation, without allergic rhinitis iv) Th2 mediated, milder disease. Significant differences were observed between clusters in carboxypeptidase, bradykinin, basogranulin, Th2 cells, and IL-4,-5,-12p70.

A TDA network generated using only clinical features revealed a clear dichotomy between steroid responsive and steroid refractory disease. Differences were observed between clinical groups in tryptase, chymase, eosinophils and Th2 and mucosal associated invariant T (MAIT) cells.

Conclusions This study highlights the diversity of aberrant immune responses associated with different clinical and pathological phenotypes of asthma. Attention should focus on further defining these distinct subtypes using such unbiased techniques and characterising under-researched pathways such as CD8+ T cells, basophils and bradykinin.

Statistics from Altmetric.com

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.