ILC1 distribution predicts response to ustekinumab

Innate lymphoid cells (ILC) are recently identified immune cells with a high cytokine-producing capacity at mucosal barriers. In patients with active CD, a shift is observed from homeostatic ILC3 towards pro-inflammatory ILC1 in the intestines. Ustekinumab – which targets the p40 subunit shared by IL-12 and IL-23 – was recently approved by the FDA and EMA for treatment of moderate-to-severe CD.

IL-12 and IL-23 play distinct roles in the plasticity of ILC1 and ILC3. Creyns et al. evaluated how biological therapy impacts ILC populations in peripheral blood by studying the effect of ustekinumab on ILC populations in blood [13].

A total of 46 CD patients (68% female; median age 42 years) was included. They were refractory to anti-tumour necrosis factor (TNF) therapy and vedolizumab with a median SES-CD of 16.5. Treatment initiated with ustekinumab (6 mg/kg intravenous (IV) at induction, followed by subcutaneous (SC) ustekinumab 90 mg q8w thereafter).

It emerged that patients with (n=6) and without (n=40) endoscopic response at week 24 had a similar inflammatory burden. This was reflected by similar median FCP (1,800 vs. 1,225 μg/g, P=0.44) and CRP (23.9 vs. 10.3 mg/L, P=0.10) levels at baseline. However, baseline endoscopic activity was much higher in patients responding to ustekinumab, compared with non-responders (median SES-CD 22 vs. 14, P=0.02). Before the start of therapy, the baseline contribution of ILC1 in the total ILC pool was significantly lower in responders compared with non-responders (7.39% vs. 16.90%, P=0.017). In contrast, at baseline ILC2 contribution was elevated in responders compared with non-responders (70.1% vs. 40.8%, P=0.02). There was no significant difference in (-) ILC3 (P=0.12).

After four weeks of treatment, a significant increase in natural cytotoxicity receptor (-) ILC3 frequency was observed compared with baseline, independent of response (P<0.001). This trend persisted at eight weeks (P=0.02) but could no longer be observed after 24 weeks (P=0.19). Thus, increased levels of ILC1 in peripheral blood at baseline may be a predictive biomarker for non-response to ustekinumab. Non-response may be explained by an increased reservoir of pro-inflammatory ILC1 in the circulation, which can migrate toward the gut to annihilate treatment effects.

Validation is needed in a larger and independent cohort with inclusion of biopsies. In general, these findings may guide individualised selection of biological agents in CD, and provide mechanisms of primary (non-)response to ustekinumab.