The experimental protocol was approved by the Ethics Committee of La Paz University Hospital (Madrid, Spain) (local code PI-4732) and written informed consent was obtained from the patients and healthy donors for participation in this study.

From the induction cohort, a longitudinal cohort of CD patients with inflammatory activity was prospectively selected, who initiated treatment with Ustekinumab for remission induction, as chosen by their physician. Patients received their first dose of ustekinumab intravenously, adjusted by weight (6mg/kg: < 55 kg, 260 mg; 55–85 kg, 390 mg; > 85 kg, 520 mg) and subsequently followed maintenance treatment with subcutaneous ustekinumab 90 mg every 8 to 12 weeks at the discretion of their physician.

Clinical activity was assessed using the Harvey-Bradshaw Index (HBI). Clinical activity was defined as a decrease of ≥ 3 points from baseline in the HBI score, and clinical remission as a HBI score <5. Quality of life was also assessed using the Inflammatory Bowel Disease Questionaire-9 (IBDQ-9).

Biological activity was considered as increased faecal calprotectin (≥ 250µg/g; ≥ 100 in the case of interventional CD). Biological remission was defined as faecal calprotectin level < 250µg/g (< 100µg/g in patients with ileocecal resection). In patients with a baseline faecal calprotectin level < 250µg/g, biological remission was defined as a C-reactive protein value < 5g/dL.

Endoscopic activity was assessed using the Simple Endoscopic Score for Crohn´s Disease (SES-CD). Endoscopic response was defined as a ≥ 50% reduction in the SES-CD score from baseline, and endoscopic remission as a score ≤ 2. In patients with ileocelcal resection, postoperative recurrence was assessed using the Rutgeerts index. Postoperative recurrence was defined as a Rutgeerts score ≥ i2, endoscopic response as a decreased of ≥ 1 point, and endoscopic remission as a score ≤ i1.

At week 24, patients were classified into four response groups: Group 1, patients achieving clinical, biochemical and endoscopic remission; Group 2, patients with combined clinical and biochemical remission but persisting endoscopic activity; Group 3, patients achieving only one domain of remission (either clinical or biochemical) without endoscopic remission; and Group 4, patients without clinical, biochemical or endoscopic remission.

Serum samples were obtained from whole blood by centrifugation.

Peripheral blood mononuclear cells (PBMCs) were isolated using the Ficoll-Hypaque density gradient centrifugation (Amersham-Pharmacia-Biotech, Buckinghamshire, UK) of whole-blood samples obtained from the patients or healthy volunteers.

Colonoscopy sample processing for isolation of epithelial cells and immune cells from the lamina propria: biopsy samples were placed in Hank’s balanced salt solution (HBSS) without Ca²⁺ and Mg²⁺ (HBSS; Gibco, Grand Island, NY, USA) supplemented with 10% FCS (Gibco) and antibiotic/antimycotic (Gibco) for 30 min at 4 °C and gently stirred. The biopsy samples were then placed in HBSS without Ca²⁺ and Mg²⁺ 0.5mM EDTA (Gibco). The epithelial and immune cells were separated from the rest of the tissue by vigorous shaking at 37 °C in HBSS without Ca²⁺ and Mg²⁺0.5mM EDTA (Gibco). The biopsy tissues in HBSS were then cut with collagenase type IV (Gibco) and incubated for 30 min at 37 °C with shaking. After centrifuge the supernatant enriched with epithelial cells is discarded, leaving only the pellet of lamina propria cells.

The T-cell subpopulation samples were incubated with antibodies against LIVE/DEAD™ Fixable Yellow (Thermo Fisher Scientific, Waltham, MA, USA), CD3 FITC or BV510, CD4 BUV395 or APC, CD8 PerCP-Cy5.5, CD45RA BV650 or FITC, CD45RO BV510 or APC-Vio700, CXCR3 BV421, CCR6 BV605, CD25 BV786, CD127 PE and CD31 PE-Cy7 (BD Biosciences, San Jose, CA, USA) for 20 minutes. The cells were washed and resuspended in 0.2 mL 1x PBS for acquisition in a BD FACSCelesta™ Cell Analyser. The data were analysed with FlowJo™ v10.8.

Events were then gated on a lymphocyte gate based on forward-scatter (FSC-A) versus side-scatter area (SSC-A). Initial gating involved the use of a forward-scatter width (FSC-W) versus height (FSC-H) plot and side-scatter width (SSC-W) versus height (SSC-H) plot to remove doublets. Events were further gated on CD3, CD4 and CD8 for T cells. Following their identification, the T cells were analysed and classified into regulatory T cells, Th1, Th2 and Th17 cells, based on their expression of CD45RO, CD45RA, CD25, CD127, CXCR3 and CCR6. Flow cytometry gating strategy is described in Supplementary Figure 1.

Cytokine production was assessed in serum samples in accordance with the manufacturer’s instructions and the concentration of IL-6, IL-17A and IL-12/IL-23 was determined by enzyme-linked immunoassay (ELISA) (Biolegend, San Diego, CA, USA).

Mean values ± SD were calculated from the measurements of all patients in each group, which in turn were obtained from the mean of three independent experiments.

The Kruskal-Wallis test was used to analyse Th cells subpopulations, comparing the differences between healthy donors and Group 1 or 4 at 24 weeks of treatment.

For the comparison between weeks, Friedman’s test for dependent samples was used for studies of the IL-17-A cytokine, comparing the differences between week 0 and week 24 of treatment for each group. A mixed-effects model was used to analyse differences between groups (healthy controls and patient groups) over time (0, 8 and 24 weeks), with repeated measures taken from the same subjects. The Geisser-Greenhouse correction was applied and post hoc group comparisons at each time point were performed using Tukey’s multiple comparisons test for the studies of IL-12/IL-23 cytokines.

Differences between the samples were considered statistically significant if p < 0.05. The statistical analysis was performed with GraphPad Prism version 10.3.1.

23 patients selected for this study suffered from CD and had been treated with ustekinumab at the dosage indicated in Methods (see the Methods section) to study T cell subpopulations in PBMCs and the lamina propria at the start of treatment (0w), at 8 weeks (8w) and 24 weeks (24w). Cytokines levels (IL-6, IL-17A and IL-12/IL-23) were also measured at the same time. Table 1A shows the clinical data of the selected patients, Supplementary Table 1 shows the analytical data and Four groups of patients were established (Table 1B) based on different criteria or response to treatment: Group 1 were patients with the best response to treatment, showing clinical, biological and endoscopic remission (Tables 1B, Supplementary Table S2); Groups 2 and 3 had an intermediate response to treatment (Tables 1B, Supplementary Table S2), showing both clinical and biological, but not endoscopic, remission (Group 2), or only one remission, clinical or biological, (Group 3); Group 4 had the poorest response to treatment, with no kind of remission at 24 weeks of treatment (Tables 1B, Supplementary Table S2). The baseline characteristics are summarised in Supplementary Table 3. Supplementary Table 1 shows the analytical data of all the patients included in the study. It is worth mentioning that all the patients in Group 4 had been previously treated with adalimumab and were considered non-responders.

In order to obtain homogeneous data, CD patients with lymphopenia were excluded from the study (Table 2). We decided to study CD4+ T cell subpopulations in PBMC over a 24-week treatment period (0w, 8w and 24w). For the Th subpopulations, we analysed Th1, Th2, Th17 and Th22 cells (Figures 1A, B). As in previous studies (24), Th1 levels rose in the patients in Group 1 at 8 weeks of treatment, but levels dropped by 24 weeks. In general, all the groups had more Th1 cells than the controls, except Group 1 at 24 weeks of treatment, whose percentage of Th1 cells dropped, in a pattern similar to the controls, with no significant differences between the controls and Group 1 (Figure 1B). Unlike Group 1, Th1 levels increased in Group 4 over the weeks of treatment, showing opposite patterns in good and poor responders. These differences in Th1 levels were also found when compared with the controls (Figure 1B).

With Th2 cells, levels were generally did not rise, unlike Th1 and Th17 cells in CD (25). We observed higher Th2 levels in the controls than in all the patient groups and, as expected, at 24 weeks of treatment, Group 1 (good responders) had recovered Th2 levels similar to the controls (Figure 1B), with significant differences compared to the controls (Figure 1B). This was not the case with Group 4, who showed significant differences in Th2 levels at 24 weeks of treatment compared to the controls (Figure 1B).

The study of Th17 cells showed that Group 1 had lower levels of these cells at 24 weeks of treatment, with similar percentages to the controls, there being no significant differences (Figure 1B). However, over the course of treatment, levels of Th17 did not fall in Group 4 patients, highlighting significant differences with the controls and indicating a poor response to treatment (Figure 1B).

Th22 cells have also been described as enhancing mucus barrier protection by promoting mucin production (Figure 1B) (26). Th22 levels increased in Group 1 at 24 weeks of treatment, but this result could not be confirmed as it would require testing in the lamina propria, where Th22 levels are too low to be conclusive (Figure 2B).

Finally, a number of studies have found that Treg cells are not altered by ustekinumab treatment (24). Observed levels did not change during the treatment in any of the Groups (Figure 1C). Different subpopulations of Tregs were analysed (Supplementary Figure 2), but their low numbers cells precluded any conclusions being drawn from these subpopulations.

From the T cell subpopulations studies in blood, we can conclude that in Group 1 (good responders), the percentages of Th subpopulations were modified during treatment, and at 24 weeks of treatment they had a similar pattern of percentages to healthy donors in terms of Th1, Th2 and Th17, while Group 4, the poor responders, showed no change to their percentages of Th subpopulations during treatment.

The next step was to analyse the T cell subpopulations in the affected tissue and measure progression over the course of treatment. To do this, we studied CD4+ T cell subpopulations in the lamina propria obtained from endoscopic biopsies in inflamed and non-inflamed tissue over a 24-week treatment (0w and 24w) (Figure 2). With regard to Th subpopulations, we analysed Th1, Th2, Th17 and Th22 cells (Figures 2A, B). This study does not permit conclusions to be drawn regarding treatment, as it is a description of Th subpopulations in the lamina propria of CD patients. Group 1 were patients with endoscopic remission and Groups 2, 3 and 4 were patients with no endoscopic remission. Regarding the colon samples, all patient samples were collected from non-inflamed tissue. We observed similar patterns of Th1 and Th2 levels between the groups and healthy donors. Despite there being non-inflamed tissues, the patients with no endoscopic remission (Groups 2, 3 and 4) harboured Th17 cells. These were not observed in Group 1 or healthy donors (Figure 1B). All that can be said about the ileum samples is that inflamed tissue showed Th17 levels that were not detected in non-inflamed tissues (Figure 2B). Finally, we detected higher levels of Treg cells in the inflamed ileum of patients with endoscopic remission (Group 1) (Figure 2C). Similar higher levels were detected in the inflamed ileum of Groups 2 and 3 (patients without endoscopic remission), so we cannot draw any conclusions from this observation. Different Treg subpopulations were analysed (Supplementary Figure 3), but the insufficient number of cells means we cannot draw any conclusions from these subpopulations.

The last step was to analyse the levels of different cytokines in the patient serum to monitor their progression during treatment with ustekinumab. We measured IL-6, IL-17 and IL-12/IL-23 in patient serum over a 24-week treatment period (0w, 8w and 24w) (Figure 3). With regard to IL-6 levels, Group 1 had the lowest levels, being the patients with the best response to treatment (Figure 3A). For IL-17 levels, Group 1 showed a steady, statistically significant drop in levels of this cytokine during treatment, while the groups with a poorer response showed no significant reduction in IL-17 levels (Figure 3B). IL-12/IL-23 was the cytokine blocked by ustekinumab, as shown in Figures 3C, where levels of IL-12/IL-23 rose after the treatment in all groups. Group 1, with the best response to treatment, maintained the highest levels of IL-12/IL-23 in serum during treatment, thus proving it is the group in which the treatment was most efficient, with the highest statistically significant differences compared to healthy donors at 8 weeks of treatment, generating the highest block of IL-12/IL-23 receptors. By contrast, Group 4 had the lowest levels of IL-12/IL-23 in serum, showing that blocking is not as efficient in these patients. There was a correlation between IL-12/IL-23 levels in serum and the response to treatment: the higher the levels of IL-12/IL-23 after treatment, the better the response to ustekinumab treatment.