P041 Characterization of cytokine and drug concentrations in serum, mucosa and faeces during induction treatment of moderate-to-severe ulcerative colitis with anti-TNF monoclonal antibodies

Van Oostrom, J.(1);Hanzel, J.(1,2);Verstockt, B.(3,4);Singh, S.(5);Smith, J.(5);Stylli, J.(5);Vermeire, S.(3);D'Haens, G.(1);

(1)Amsterdam UMC- Meibergdreef 9- 1105 AZ Amsterdam- The Netherlands, Gastroenterology, Amsterdam, The Netherlands;(2)UMC Ljubljana- Japljeva ul. 2- 1000 Ljubljana- Slovenia, Gastroenterology, Lubljana, Slovenia;(3)University Hospitals Leuven- KU Leuven- Leuven- Belgium, Department of Gastroenterology and Hepatology, Leuven, Belgium;(4)KU Leuven- Leuven- Belgium, Department of Chronic Diseases and Metabolism-, Leuven, Belgium;(5)Progenity Inc, Progenity laboratories, San Diego, United States;


It remains unclear why up to 30% of patients with moderate-to-severe ulcerative colitis (UC) do not respond to anti-TNF treatment (anti-TNF). In this study, we explored multi compartment cytokine and anti-TNF concentrations in peripheral blood (PB), faeces (F) and colonic mucosa (CM) before, during and after anti-TNF induction treatment in UC.


91 UC patients starting infliximab or adalimumab were included at two tertiary centres (table 1). Patients underwent clinical and endoscopic assessment with CM, PB, and F sampling at baseline. Follow-up CM samples were taken at first follow-up endoscopy during the first year. Follow-up PB and F samples were taken at follow-up endoscopy and optionally on day 3, 7, 14, and 28.

Response (R) was defined as Mayo endoscopic score (MES) 0-1 (assessed by a blinded independent reader) and subsequent anti-TNF therapy for > 1 year. All other patients were classified as non-responders (NR).

TNFα, IL6, IL10 and drug levels were measured in all compartments and OSM and IL12p40 in PB and CM (F anti-TNF with ELISA and cytokines and PB and CM anti-TNF with a bead-based immunoassay). CM samples were normalised for gram of protein (gop).

To differentiate pharmacokinetic (PK) from pharmacodynamic (PD) failures, we calculated CM anti-TNF/ TNFα. We classified values above the median as sufficient PK (Good PK) and values below as insufficient PK (Bad PK).


We analysed CM, PB and F samples from 71, 91 and 32 patients respectively. Baseline cytokine concentrations were similar in all compartments.

At follow-up (table 2), PB IL6 was lower in R (p=0.045). All F cytokines were significantly higher in R (TNFα: p<0.001; IL6: p<0.001; IL10: p<0.001). In CM, IL6 and OSM were lower and IL12p40 higher in R (IL6: p=0.004; OSM: p=0.01; IL12p40: p=0.021).

After stratification of CM samples for PK readouts (table 3, figure 1), IL6 was highest in Good PK NR vs Good PK R (6905 vs 2980 ng/gop, p = 0.002) and OSM was lowest in R vs NR groups (2466 vs 36111 ng/gop in Good PK R vs Good PK NR, p = 0.002). Trends were observed towards highest IL10 in Bad PK R and Highest IL12p40 in Good PK R, although not significant.


Baseline cytokine concentrations in PB, CM and F did not predict response to anti-TNF induction treatment in UC. At follow-up, response (R) was characterised by lower PB IL6; higher F cytokines; lower CM IL6 and OSM and higher IL12p40.

After stratification for PK, pharmacodynamic failure (Good PK NR) was characterised by highest CM IL6, indicating it is a driver to inflammation alternative to anti-TNF. R was characterised by lowest OSM regardless of PK, indicating OSM mainly reflects disease activity. Further research focuses on cytokine-PK interplay contributing to anti-TNF (non-) response.