2.1 Isolation of PBMC

A total of 20–60 ml of peripheral blood in trisodium citrate solution (S-Monovette; Sarstedt) were collected from the arm vein of UC and CD patients as described previously.¹⁰ The blood was diluted with Hank's balanced salt solution (Sigma Aldrich) in a 1:2 ratio and lloaded onto LeucoSep tubes (Greiner Bio One). Tubes were centrifuged at 400g for 30 min without acceleration and break and PBMC were extracted from the interphase and diluted with phosphate buffered saline (PBS) to a final volume of 40 ml. Cells were counted and centrifuged at 1400g for 5 min. The cell pellet was resuspended in PBS at a concentration of 4 × 10⁶ cells in 100 µl.

2.2 Flow cytometric analysis

All antibodies (Table S1) were purchased and used according to the manufacturer's instructions (Biolegend). Flow cytometry was performed using a Thermofisher Attune NxT (Thermo Fisher Scientific) and analysed with FlowJo 10.1-Software (FlowJo LLC).

2.3 Seahorse analysis

PBMCs were isolated from 10 ml of blood and seeded into the Agilent Seahorse XFp Cell Culture Miniplates (Agilent Technologies Inc.) to a density of 250000 cells/well. The assay medium consisted of Roswell Park Memorial Institute (RPMI) Medium (Thermo Fisher Scientific) RMPI, 100 nM pyruvate, 100 nM glutamine, and 100 g/L glucose (Agilent Technologies Inc.). Postseeding, the cells were kept at 37°C for 1 h, before the assay was performed. Patient samples were measured in triplicates or quintuplicates in the ground state without further activation. To measure the impact of copanlisib, cells were pre-incubated with anti-CD3 (61 ng/ml; BioLegend) for 1 h at 37°C in the presence or absence of copanlisib (100 nM/ml). Energy state was measured in the ground state for 1 h at 10 time points. For measurements, a Seahorse XFp Extracellular Flux Analyzer (Agilent Technologies Inc.) was used.

2.4 Cell culture

PBMCs from healthy donors were isolated from 10 ml blood. The cell pellet was resuspended in RPMI Medium (Thermo Fisher Scientific) at a concentration of 1 × 10⁶ cells/ml as described previously.²⁴ Cells were incubated with additional 500 μl RPMI with 10% fetal calf serum (FCS) and 1% penicillin-streptomycin (Sigma-Aldrich). Wells incubated in the absence or presence of anti-CD3 (61 ng/ml; BioLegend) served as negative and positive controls, respectively. Copanlisib was added at a concentration of 100 nM.

2.5 Study protocol

Mice were obtained from Charles River Laboratories and kept under specific pathogen-free conditions in individually ventilated cages in a facility controlled according to the Federation of Laboratory Animal Science Association guidelines as described previously.²⁵ Six- to eight-week old NOD.cg-Prkdc^SCID Il2rg^tm1Wjl/Szj mice (abbreviated as NOD-scid IL-2Rγ^null, NSG) were engrafted with 100 µl PBMC cell solution (4 × 10⁶) into the tail vein on Day 1 as previously described²⁵ and presensitized by rectal application of 150 µl 10% ethanol on Day 7 using a 1 mm catheter (Henry Schein). The catheter was lubricated with lidocaine 2% gel (AstraZeneca).²⁵ Rectal application was performed under general anesthesia using 4% isoflurane. Following application mice were kept at an angle of 30° to avoid ethanol dripping. On Day 14 mice were additionally challenged with 50% ethanol following the protocol of Day 8. Copanlisib was applied intraperitoneally (i.p) at a concentration of 6 mg/kg in 0.5% methylcellulose gel in PBS (Firma Cat# M0512; Merck KGaA) on Days 7, 8, 14, 15, and 16. Mice were sacrificed on Day 18.

2.6 Clinical activity score

The assessment of severity of colitis was performed daily as previously described²⁶: Loss of body weight: 0% (0), 0%–5% (1), 5%–10% (2), 10%–15% (3), 15%–20% (4). Stool consistency: formed pellet (0), loose stool or unformed pellet (2), liquid stools (4). Behaviour: normal (0), reduced activity (1), apathy (4) and ruffled fur (1). Body posture: Intermediately hunched posture (1), permanently hunched posture (2). The scores were added daily into a total score with a maximum of 15 points per day. Animals who suffered from weight loss more than 20%, rectal bleeding, rectal prolapse, self-isolation or a severity score more than 7 were euthanized immediately and not taken into count. All scores were added for statistical analysis.

2.7 Macroscopic colon score

The colon was removed, and the colon was scored.²⁶ Pellet: formed (0), soft (1), liquid (2); length of colon: more than 10 cm (0), 8–10 cm (1), less than 8 cm (2); Dilation: no (0), minor (1), severe (2); Hyperemia: no (0), yes (2); Necrosis: no (0), yes (2).

2.8 Histopathology

Sections from distal parts of the colon were fixed in 4% formaldehyde for 24 h, stored in 70% ethanol and embedded in paraffin as described previously.²⁵ A total of 3 µm sections were cut and stained with haematoxylin and eosin (HE), periodic acid-Schiff (PAS) and Masson-Goldner trichrome (MGT, all from Morphisto GmbH).

Epithelial erosions were scored as described prviously²⁶: no lesions (1), focal lesions (2), multifocal lesions (3). Inflammation was scored as follows: infiltration of few inflammatory cells into the lamina propria (1), major infiltration of inflammatory cells into the lamina propria (2), confluent infiltration of inflammatory cells into the lamina propria (3), infiltration of inflammatory cells including tunica muscularis (4). Fibrosis was scored as follows: focal fibrosis (1), multifocal fibrosis and crypt atrophy (2) and general fibrosis and crypt atrophy (3). The presence of edema was scored as follows: focal (1), multifocal (2), general (3). Hyperemia was scored with one additional point. The scores for each criterion were added into a total score. Images were taken with an AxioVert 40 CFL camera (Zeiss) using the Zeiss ZE n2 lite software. Representative longitudinal sections in original magnification are shown. To enhance contrast of the pictures, Adobe Photoshop CC a tonal correction was used.

2.9 Isolation of human leucocytes

Spleens were minced and cells filtrated through a 70 µl cell strainer followed by centrifugation at 1400g for 5 min and suspended in FACS buffer (1X PBS, 2 mM ethylenediaminetetraacetic acid, 2% FCS) to isolate human leukocytes as previously described.²⁶ Cell suspensions were additionally filtrated using a 35 µm cell strainer and then labelled for flow cytometry analysis.

2.10 Statistical analysis

Statistical analysis was performed with R: A language and environment for statistical computing (R Foundation for Statistical Computing, Vienna, Austria; URL https://www.R-project.org/). Variables in the supplementary were represented with mean, SD, N, difference and 95% confidence interval (CI). A two-sided Student's t test and a CI of 95 was used to compare binary groups and for more than two groups, analysis of variance (ANOVA) followed by the post hoc Tukey's honestly significant difference (TukeyHSD) test was conducted. Variables subjected to ANOVA were tested for normal distribution. All variables fulfilled this requirement. Cummingplots were generated using the dabestr package and are used in this study for data presentation and comparison. Cumming plots are a new generation of data analysis with bootstrap-coupled estimation plots that move beyond p values.²⁷ These plots are used to analyze large samples and multiple groups. The estimation graphics mainly offer five key advantages: (1) The plot of the full sampling-error curve of the effect size presents the distribution's graded nature, (2) the difference axis increases transparency to the comparison being made and (3) in comparison to p values, which conflate the magnitude and precision in a single number, the relative size of the CI provides a specific measure of its precision.²⁷ (4) The sampling-error curve is derived with bootstrapping, which makes the method robust and versatile and (5) the difference diagram encourages the quantitative reasoning about the system under study by focusing on an effect size.²⁷ Orthogonal partial least square discrimination analysis (oPLS-DA) was performed using the ropls package.

3.1 Ex vivo analysis of IBD and non-IBD donors revealed differences in T cell activation and metabolism

To analyze the activation status of T cells, PBMCs from IBD patients (N = 97) patients and non-IBD donors (N = 35) were isolated and subjected to flow cytometric analysis as described in Section 2. Frequencies of CD4⁺ T cells expressing the early activation marker CD69⁺ or the late activation marker CD134⁺ (Ox40) were determined (For basic patient demographics see Table 1, for gating strategy see Figure S1). Data are presented as Cumming plots.²⁸ As shown in Figure 1A, frequencies of both cell types were highly variable in all groups reflecting the dynamics of inflammatory processes.²⁹ As indicated by nonoverlapping CI of the unpaired mean differences frequencies both cell types were significantly different in the IBD group as compared to non-IBD group (CD4⁺ CD69: p = 1E-04; CD4⁺ CD134: p = 7E-03; Welch two-sample t test). To evaluate whether the activation status was reflected in cell metabolism, PBMCs of IBD patients (n = 10) and non-IBD donors (n = 6) were subjected to OCR and ECAR analysis as described in Section 2 (For OCR and ECAR mean values and donor characteristic see Table S2). In this assay, oxidative phosphorylation is determined by measuring the OCR and glycolysis by measuring the ECAR caused by lactate as a byproduct of glycolysis. As shown in Figure 1B, The energy map of two different IBD patients in the ground state suggested increased glycolytic activity as indicated by slightly higher ECAR values as compared to the two different non-IBD donors. No correlation of the OCR or the ECAR was observed with age or body mass index. As the variability of ECAR and OCR was high in both groups, the correlation of the ECAR and OCR rates was analyzed by Pearson's product-moment correlation analysis. In contrast to the non-IBD group, OCR and ECAR were significantly correlated in the IBD group suggesting that in this group, oxygen consumption was coupled to glycolysis (non-IBD: cor = 0.37, p = .12, 95% CI = −0.11–0.71 IBD: cor = 0.79, p = 2e-10, 95% CI = 0.65–0.88).

3.2 Copanlisib suppresses glycolysis in activated T cells

As copanlisib exerts its activity on the AKT pathway, we would expect that the inhibitory activity of copanlisib is reflected in cell metabolism. To prove this hypothesis, the energy consumption of PBMCs was analyzed by Seahorse technology as described in Section 2. PBMCs were seeded into the wells at a concentration of 250000 cells/well. Experiments were performed with two different donors (N = 2) in single measurements or duplicates. Anti-CD3 monoclonal antibodies (61 ng/ml) were used to activate T cells and OCR and ECAR values were measured for 1 h at 10 consecutive time points. Three groups were compared: An untreated control group (control, n = 20), a group pre-incubated for 2 h with anti-CD3 (anti-CD3, n = 20), and a third group preincubated with anti-CD3 and treated with 100 nM copanlisib (anti-CD3 + copanlisib, n = 40). The energy map (Figure 2A) indicates increased glycolysis upon activation of T cells and suppressed glycolysis by copanlisib. This snapshot was confirmed by consecutive analysis presented as Cumming plots (Figure 2B) which showed significantly increased ECAR rates upon activation (control: 39.70 ± 23.92, anti-CD3: 58.24 ± 29.06) and significantly decreased ECAR rates in the presence of copanlisib (43.16 ± 20.23, ANOVA, p < .000). In contrast, OCR rates were only slightly affected, corroborating the assumption that copanlisib specifically impairs glycolysis.

3.3 Copanlisib suppresses activation CD4⁺ T cells activation

To analyse the impact of the copanlisib treatment on the activation of T cells, PBMCs were isolated donors and 1 × 10⁶ cells were incubated in the absence or presence of anti-CD3 antibodies (61 ng/ml) and copanlisib (100 nM) for 72 h as described in Section 2. Cells were subjected to flow cytometric analysis as described in Section 2 for frequencies of the T cell activation markers CD25, CD69, CD134, and CD103 (for gating strategy see Figure S1). Measurements were performed in triplicates or in quadruplicates with PBMC from seven donors (N = 7). Three groups were compared: untreated cells (control: number of samples n = 23), activated T cells (anti-CD3, number of samples n = 23) and activated T cells treated with copanlisib (copanlisib, number of samples n = 23). As shown in Figure 3A exposure to anti CD3 antibodies resulted in significantly increased frequencies of CD4⁺ T cells expressing CD25, CD103, CD134, and CD69 (ANOVA followed by TukeyHSD: CD4⁺ CD25⁺: 42.15 ± 21.46, mean ± SD, p = .000; CD4⁺ CD69⁺: 48.97 ± 25.24, p = .000; CD4⁺ CD103⁺: 8.21 ± 6.32, p = .002, CD4⁺ CD134⁺: 51.87 ± 26.5, p = .003) indicating a general activation of CD4⁺ T cells. Copanlisib significantly impaired the activation of CD25⁺ (26.06 ± 21.82, p = .013) and CD103⁺ (3.29 ± 3.70, p = .002) T cells whereas frequencies of CD69⁺ (50. 05 ± 20.66) and CD134 (44.02 ± 23.68) were unaffected.

As activation of T cells by anti-CD3 antibodies is known to result in secretion of ILs,³² the effect of copanlisib on cytokine levels was examined in supernatants of the cell cultures by Luminex analysis (control: number of donors N = 2, number of samples n = 6; anti-CD3: number of donors N = 3, number of samples n = 9; copanlisib; number of donors N = 3, number of samples n = 10). As shown in Figure 3B, activation of T cells with anti-CD3 antibodies resulted in significantly increased expression of IFNγ (control: anti-CD3: 15139.0 ± 10263.12 pmol/ml, mean ± SD, p < .000), IL-4 (49.20 ± 16.67 pmol/ml, p < .000), IL-17 (26733.0 ± 8387.89 pmol/ml, p < .000) and IL-10 (16002.0 ± 3773.3 pmol/ml, p < .000), corroborating the general activation by anti-CD3 observed in the previous experiment. As expected, copanlisib significantly suppressed the secretion of all examined cytokines of IFNγ (6332.0 ± 5707.61 pmol/ml, p = .018), IL-4 (14.35 ± 3.04 pmol/ml, p < .000), IL-17 (2529.9 ± 2392.19 pmol/ml, p < .000) and IL-10 (2769.0 ± 1892.36 pmol/ml, p < .000).

3.4 Copanlisib treatment ameliorates pathological phenotype in vivo

To further validate copanlisib as a potential therapeutic for IBD, NSG-UC mice were treated with copanlisib in two different experiments. Mice were reconstituted with PBMCs from two different UC patients in relapse who were currently treated with vedolizumab and tofacitinib or therapeutically naïve (for patient characteristics and groups of mice, see Table S3). NSG-UC mice were challenged according to a standard protocol as described in Material and Methods. Seven days post reconstitution the mice were divided into three groups, an unchallenged control group (control, n = 8), a group challenged with ethanol and treated intraperitoneally with the carrier methylcellulose (EtOH, n = 14) and an ethanol-challenged group treated with 6 mg/kg copanlisib in methylcellulose (copanlisib, n = 14). Except for the unchallenged control mice, all mice were rectally challenged with 10% ethanol on Day 7, followed by challenge with 50% ethanol on Day 14. Copanlisib in 100 µl methylcellulose and 100 µl methylcellulose alone were injected intraperitoneally on Days 7, 8, 14, 15, and 16 into the respective groups.

As previously observed,²³ untreated NSG-UC mice challenged with EtOH experienced mild weight loss and in some cases, diarrhea, as opposed to unchallenged control mice which appeared normal. Clinical symptoms were classified according to a clinical score as described in Section 2. As shown in Figure 4A, the clinical symptoms increased significantly in ethanol-challenged mice as compared to nonchallenged mice (control: 0.88 ± 0.84; mean ± SD; ethanol: 2.5 ± 0.94; ANOVA followed by TukeyHSD, p = .006) and the score slightly improved in response to treatment with copanlisib, albeit not significantly (1.5 ± 1.23, ethanol vs. copanlisib, p = n.s.). Macroscopic inspection of the colon corroborated the clinical score. In contrast to colon of control mice, colons of ethanol-challenged mice exhibited unformed pellets, dilatation and colon shortening. Copanlisib-treated mice slightly benefitted from treatment (Figure 4B). Appearance of the colons were classified according to a colonic score as described in Section 2. As shown in Figure 4A, the colon score of copanlisib-treated mice decreased, albeit not significantly (control: 0.5 ± 0.75; ethanol: 2.79 ± 1.37, ANOVA followd by TukeyHSD EtOH vs. control, p = .001; copanlisib: 2.14 ± 0.77, copanlisib vs. EtOH, p = .069). To further examine the effect of copanlisib, the histopathological manifestations were examined by three stainings: HE stain to colon architecture, MGT for connective tissue (green), and PAS stain to visualize goblet cells.^33-35 As shown in Figure 4C, inflammation in the colon of ethanol-challenged mice were characterized by edema, influx of a mixed infiltrate into the mucosa and submucosa, epithelial erosions, loss of goblet cells, altered crypt architecture and fibrosis. Treatment with copanlisib reversed the appearance and with the exception of ongoing fibrosis, resulted in an almost normal phenotype. The histological manifestations were classified according to a histological score as described in Section 2. As shown in Figure 4A, the score significantly increased upon challenge with EtOH (control: 3.25 ± 1.83; ethanol: 8.5 ± 3.81, p = .001) and significantly decreased in response to treatment with copanlisib (copanlisib: 4.57 ± 2.82, EtOH vs. copanlisb, p = .006.).

To examine the effect of copanlisib on activation of T cells, human leukocytes were isolated from spleen and subjected to flow cytometric analysis. As shown in Figure 5, frequencies of CD4⁺ CD25⁺, CD4⁺ CD69⁺, CD4⁺ CD103, and CD4⁺ CD134⁺ T cells of unchallenged mice were in a similar range as unchallenged PBMC in vitro (Figure 3). Unlike in the in vitro experiment, however, challenge with ethanol solely affected frequencies of CD4⁺ CD103⁺ (control: 0.53 ± 1.02; ethanol: 3.46 ± 2.19, ANOVA followed by TukeyHSD, p = n.s.) and CD4⁺ CD134⁺ (control: 16.92 ± 6.66, mean ± SD; ethanol: 25.23 ± 6.73, p = .04 whereas frequencies of CD4⁺ CD69⁺ and CD4⁺ CD25⁺ remained unaffected. Treatment with copanlisib significantly decreased CD4⁺ CD69⁺ T cells (copanlisib: 8.86 ± 4.17, ethanol vs. copanlisib, p < .000), whereas frequencies of CD4⁺ CD25 cells even increased albeit not significantly. This observation is in contrast to the results obtained in vitro where frequencies of CD4⁺ CD25⁺ T cells diminished upon exposure to copanlisib and CD4⁺ CD69⁺ cells remained unaffected.

3.5 oPLS-DA analysis to evaluate the in vivo efficacy of copanlisib

To evaluate the overall efficacy of copanlisib treatment, an orthogonal partial least square analysis (oPLS-DA) was performed using the clinical-, colon- and histological scores and frequencies of CD4⁺ CD69⁺ and CD4⁺ CD134⁺ T cells as variables. In this analysis, each data point represents one mouse, combining values of the selected variables.

As depicted in Figure 6

oPLS-DA revealed a clear distinction between the control versus the ethanol challenged group and ethanol challenged group versus copanlisib treated group. The values of the fraction of the variation of the X variables explained by the model (R2X), of the fraction of the variation of the Y variables explained by the model (R2Y) and of the fraction of the variation of the Y variables predicted by the model (Q2Y) were high in both comparisons, indicating a good fit of the model on the given data. For significance diagnostics, the pR2Y and pQ2 values were compared with values after random permutation of the y response. The obtained values of p = .05 corroborated the validity of the oPLS-model and indicate that both separations from the ethanol-challenged group were significant.

The separation of the control group from the ethanol challenged group was significant as indicated by the R2X-value of 0.839, the R2Y-value of 0.714, the Q2R-value of 0.53 and the low root mean square error of estimation [RMSEE] value of 0.292). The lower R2X, R2Y, and Q2Y values (0.855, 0.548, and 0.285, respectively) and the higher RMSEE value (0.371) suggested, however, a less pronounced distinction between the copanlisib and the ethanol-challenged group.