Abstract
Inflammatory Bowel Disease (IBD) is undergoing a transformation with an expanding repertoire of drugs targeting various aspects of the immune response. In recent years, we have witnessed the emergence of three novel classes of drugs that target leukocyte trafficking to the gut (vedolizumab), antibodies that neutralize key cytokines (ustekinumab) and inhibitors of cytokine signalling pathways (tofacitinib). IL-23 is considered a key cytokine in IBD, and drugs targeting its specific p19 subunit are in advanced development. Insights into how IL-23 might regulate mucosal immune cells are now needed.There is also a pressing need to develop precision medicine approaches since all agents assessed so far are hampered by primary and secondary loss of response. It is desirable to develop personalised strategies to inform which patients should be treated with which drugs. Stratification by clinical parameters alone lacks sensitivity, and alternative modalities are now needed to deliver precision medicine in IBD.
The rapid decrease in cost of whole tissue transcriptomics combined with increased expertise in bioinformatics has permitted the use of complex tools to analyse RNA like never before. In this thesis I have investigated the use of transcriptomics to probe the biological impact of IL-23 signalling in UC. Particular focus was given on the transcriptional effects of IL-23 in ex vivo cellular models using colonic tissue from. diseased colons of UC patients.
We created two models namely whole colonic biopsy and lamina propria mononuclear cells (LPMCs) using tissue from active ulcerative colitis and exposed biological replicates to IL-23 or standard culture media and then performing RT-qPCR and RNAseq. We then identified statistically differentially expressed genes from the IL-23 exposed compared to non-exposed samples. Using Ingenuity Pathway Analysis permitted us to identify relevant up and downregulated pathways and upstream regulators to enhance the meaning of our discovery of differentially expressed genes.
We aimed to:
1. identify the IL-23 responsive genes, pathways, and functions in UC
2. establish whether these IL-23 responsive genes were upregulated in active ulcerative colitis in reposited datasets
3. determine whether IL-23 responsive genes were able to impact response to UC therapies
IL-23 stimulation of LPMCs from active UC orchestrated a strong pro-inflammatory transcriptional programme with significant upregulation of IL-22 and IFNγ by RT-qPCR. As expected IL-23 induced the Th17 pathway but also induced novel pathways such as HMGB1, cancer and angiogenesis and promotes cell trafficking and chemotaxis, previously unheralded functions. Network analysis revealed IL-17A, IL-17F, IL-22 and IFNG appear to function as central nodes involved in the coordination of the inflammatory process.
By utilising publicly available reposited data sets from active colonic tissue from patients with UC, we were able to interrogate these samples for enrichment of our gene sets and with allied clinical data we could judge the effectiveness of our gene signature to predict response to therapies. The IL-23 LPMC enrichment scores were statistically significant higher in active disease.
We had access to the UNIFI dataset courtesy of Janssen. The IL-23 LPMC gene signature showed statistically significantly higher enrichment scores in a number of endpoints including endoscopic healing and clinical remission at week 8. Furthermore, stratification by enrichment scores increased the likelihood of achieving endpoints and was further enhanced by including CRP in the analysis. For example, unstratified patients had 12.7% chance of meeting the endpoint of clinical and endoscopic remission but when IL23 LPMC enrichment score <0 and CRP <5 22.8% of patients achieve this endpoint. We also showed that the IL23 LPMC enrichment score was statistically different in patients who received placebo and who responded, which highlights an important group who may not require therapy at all.
Surprisingly, enrichment scores were not higher in anti-TNF non-responders in available reposited datasets. Utilising data from the PURSUIT trial (152 patients of golimumab in UC) data courtesy of MSD – the IL-23 LPMC signature had statistically higher enrichment scores in patients who did not achieve week 6 endoscopic remission but did not show a difference in other endpoints. We showed that by stratifying enrichment scores we can improve the chance of predicting response such as 18% if enrichment score <0 achieved clinical remission and endoscopic healing vs 12% if unstratified. Furthermore, when combined with CRP <5 the proportion can be
increased further to 20%.
The analysis of colonic explants has raised more questions than answers them. We found that IL-22 was significantly upregulated when investigated with RT-qPCR but relevant pro-inflammatory pathways such as the Th17 pathway or T cell activation were not found. Instead, we found activation of IBD relevant oxidative phosphorylation ad PPARγ pathways and downregulation of aryl hydrocarbon receptor pathways. Together these pathways do not form a coordinated pro-inflammatory effect of IL-23 as expected but create a series of opposing effects.
The colonic explant signature showed higher enrichment scores in active UC though showed reduced enrichment in anti-TNF non-responders contrary to the hypothesis.
We demonstrate the generation a gene signature by exposing biological replicates of active colonic tissue from patients with UC to IL-23 and identifying the differentially expressed genes. We demonstrate that the colonic explant model was not successful in inducing a coordinated pro-inflammatory response but the lamina propria experiment induced appropriate and expected Th17 cytokines as well as novel findings suggestive of co-activation with IL-1 cytokines. Furthermore, we demonstrate the IL-23 LPMC signature can differentiate active from inactive UC and healthy controls and is statistically significantly more expressed in patients achieving clinical endpoints when receiving anti-IL12p40 antibodies in UNIFI trial. We show through stratification of the enrichment scores and addition of biomarkers we can further improve the rates of achieving endpoints. We have shown the potential of a transcriptomic based gene signature biomarker in UC though further work will be required to improve the sensitivity and to ratify our results before it can be utilised in a clinical setting.
| Date of Award | 1 Jun 2024 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | Nick Powell (Supervisor) & Natalie Prescott (Supervisor) |