Immuno-genomic profiling of paediatric inflammatory bowel disease

Abstract

INTRODUCTION
Inflammatory bowel disease is a complex polygenic disease with a multi-factorial aetiology. Several immune signalling pathways regulated by multiple genes have been implicated in the pathogenesis of the disease. The current categorisation of IBD is based on the clinical phenotype which is blind to the immunological phenotype. The central aim of this project was to assess the functional integrity of key immune signalling pathways in paediatric patients with IBD through immunological assays and examine the relationship between altered immune function and patient-specific genetic mutations. The secondary aim of this project was to explore the potential for patient stratification based on their immuno-genomic profiles in order to facilitate the application of targeted therapeutics in IBD.

HYPOTHESES
1. Of the multiple immune pathways implicated and confirmed by genetic studies, individual patients present with compromised function of one or more specific immune signalling pathways.
2. Aberrant immune pathway function identified by immunoassays can be correlated with individual mutation profiles across genes within these pathways.
3. Integrated immuno-genomic profiling can stratify patients into novel subtypes that will permit better mechanistic and therapeutic evaluation, and facilitate targeting of therapies to molecularly stratified patients.

OVERVIEW OF METHODS

In this thesis, a multiplex-based immunological assay was developed and optimised to assess the molecular integrity of biological pathways implicated in IBD. Based on preliminary experiments assessing the suitability of different cell activation systems, extensively published literature and practicalities, cryopreserved (frozen) PBMCs were used for conducting immuno-assays. A standard operating procedure for conducting assays was developed following optimisation experiments in control samples. The optimised assay was conducted in a treatment naïve paediatric IBD cohort of twenty-two patients alongside ten adult and ten paediatric controls. The cytokine data generated were subjected to hierarchical clustering, an unsupervised machine-learning application to assess patient stratification based on immunological profiles. Clusters generated were then interrogated against exome sequencing data to identify genomic signals which may be contributing to the clustering pattern.
In addition to the immunological assays investigating innate immune signalling in IBD, the research work presented in this thesis includes the assessment of periostin protein as a potential biomarker of disease activity, immuno-histochemistry of periostin expression in the GI mucosa and the interrogation of exome sequencing data in the context of clinical outcomes and plasma levels of periostin.
The thesis also includes the applicability of genomic analysis to assess thiopurine-induced drug toxicity in paediatric patients with IBD.

MAIN RESULTS
A systematic review of published literature across a panel of genes implicated in IBD, conducted at the start of my PhD project showed a paucity of immuno-genomic studies, thereby signposting a clear direction for my research integrating immunological work-up with exome sequencing data.
Following extensive dose-finding and optimisation experiments conducted in healthy volunteers, a standard operating procedure was established using cryopreserved PBMCs. In the optimised assay, PBMCs were selectively stimulated for NOD2, TLR1-2 and TLR4-mediated pathways using MDP, Pam3CSK4 and LPS respectively. Effector responses of IL-10, IL-1b, IL-6 and TNF-a were assessed through multiplex assays to evaluate corresponding NF-kB, MAPK and NLRP3-inflammasome pathways.
A significant difference in immune induction was observed between the paediatric and adult healthy controls for two out of the twelve cytokine responses (TLR4-induced IL-10 and TNF-a). Therefore, only age-matched paediatric controls were used in the final analysis to interpret biological variations in paediatric patients. Combined innate immune responses in patients across twelve effector responses were significantly reduced compared to paediatric controls (p=0.003) and were driven primarily by ‘hypofunctional’ TLR responses. Application of hierarchical clustering to the immunological data generated eight distinct patient clusters based on their induced immune profile. Further interrogation of the clustering pattern using a whole gene pathogenicity score ‘GenePy’, identified a statistical excess of mutations in the TNF-a signalling pathway in one immune profile-based cluster, with a significant contribution from the USP21 gene.
In the periostin study, an inverse correlation was observed between plasma levels of periostin and disease activity for Crohn’s disease, with significantly higher levels observed during remission rather than active disease, suggesting a more prominent reparative role rather than inflammatory for this protein in IBD. The utility of plasma periostin as a biomarker of inflammation in IBD could not established as there were no significant differences in the plasma levels of this protein observed between patients and paediatric controls.
In the study investigating thiopurine drug-induced toxicity, a highly pathogenic novel variant was identified in the TPMT gene, demonstrating the strength of next generation sequencing as a powerful tool in identifying pathogenic variants, not detected through standard genotyping used for predicting thiopurine drug toxicity. A significant association was observed between the MOCOS gene and TPMT biochemical enzyme activity, highlighting the importance of genes other than TPMT in thiopurine-induced toxicity. Although NGS has the ability to detect rare or novel variants in the TPMT and other implicated genes, there is no clear advantage of genomic testing over the biochemical test in predicting toxicity to thiopurine drugs.

CONCLUSIONS

This research supports an immunological profiling led genomic analysis approach to a complex polygenic disorder. It provides a rationale for patient stratification based on their immuno-genomic profiles, signposting new mechanistic insights with a futuristic goal of personalised therapy. Currently, there is an unmet need for novel groupings in IBD based on the immunological profiles of individual patients. However, with the emergence of robust high throughput technologies coupled with the integration of multi-omic data, patient-stratification based on the molecular patterns of the disease looks increasingly possible. Judicious integration of multi-omic data will pave the way for personalised treatment approaches in IBD, leading to improved health outcomes and reducing the morbidity associated with this debilitating condition.

More information

Identifiers

ORCID for Sarah Ennis: orcid.org/0000-0003-2648-0869

Catalogue record

Export record