UvA-DARE (Digital Academic Repository)

Macrophages as a therapeutic target in inflammatory bowel disease Lessons learned from anti-TNF therapy Houttuijn Bloemendaal, F.M.

Publication date 2020 Document Version Other version License Other Link to publication

Citation for published version (APA): Houttuijn Bloemendaal, F. M. (2020). Macrophages as a therapeutic target in inflammatory bowel disease: Lessons learned from anti-TNF therapy.

General rights It is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons).

Disclaimer/Complaints regulations If you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Ask the Library: https://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. You will be contacted as soon as possible.

UvA-DARE is a service provided by the library of the University of Amsterdam (https://dare.uva.nl)

Download date:30 Sep 2021

CHAPTER 3

TNF-anti-TNF immune complexes inhibit IL-12/IL-23 secretion by inflammatory macrophages via an Fc-dependent mechanism

Felicia M. Bloemendaal, Pim J. Koelink, Karin A. van Schie, Theo Rispens, Charlotte P. Peters, Christianne J. Buskens, Jarmila D. van der Bilt, Willem. A. Bemelman, Hannelie Korf, João G. Sabino, Cyriel Y. Ponsioen, Anje A. te Velde, Geert R.A.M. D’Haens, Severine Vermeire, Gijs R. van den Brink, Manon E. Wildenberg

Journal of Crohn’s and Colitis, 2018 Aug 29;12(9):1122-1130 Chapter 3

ABSTRACT

Background and Aims We have recently shown that the mode of action of IgG1 anti-TNF antibodies in inflammatory bowel disease (IBD) requires Fcγ-receptor (FcγR) engagement on macrophages. Here we examine the effect of Fcγ-receptor signaling by anti-TNF on macrophage IL-12/IL-23 secretion.

Methods Cytokine production by human inflammatory macrophages was assessed at the level of RNA and or protein. TNF-anti-TNF immune complex formation was determined by size-exclusion chromatography and signaling visualized by immunofluorescence. IL-12/IL- 23p40 was measured in CD14+ lamina propria cells from IBD patients.

Results Infliximab and adalimumab potently suppressed IL-12/IL-23 production by inflammatory macrophages, but Fab’ fragment certolizumab did not. IL-12/IL-23 suppression depended on Syk activity and was mediated at the level of IL-12/IL-23p40 mRNA. Etanercept, a soluble TNF receptor fused to an Fc-region, did not inhibit IL-12/L-23 secretion, suggesting that the presence of an Fc-region was not sufficient. Infliximab and adalimumab formed immune complexes with soluble TNF while etanercept did not, suggesting that FcγR mediated suppression of IL-12/IL-23 required the formation of immune complexes. Indeed, non-specific IgG1 immune complexes but not uncomplexed IgG1 similarly suppressed IL-12/IL-23 secretion. Finally, infliximab significantly decreased IL-12/IL-23p40 production in myeloid cells isolated from the lamina propria of IBD patients.

Conclusions TNF-anti-TNF antibody immune complexes potently inhibit IL-12/IL-23 expression by inflammatory macrophages. Our data suggest that anti-TNFs and antibodies against IL-12/IL-23 may therefore have partially overlapping modes of action in patients with IBD.

56 Anti-TNF immune complexes inhibit IL-12/IL-23 secretion

INTRODUCTION

In contrast to rheumatoid arthritis, the mode of action of anti-TNFs in inflammatory bowel disease (IBD) is not exclusively related to therapeutic blockade of TNF.1 We have previously demonstrated that the therapeutic efficacy of anti-TNF was completely dependent on Fc- Fcγ-receptor (FcγR) interaction in a preclinical model of colitis. Additionally, we have shown that IgG1 anti-TNFs polarize macrophages to a regulatory and wound healing phenotype in an Fc dependent manner in humans and mice.2-5

Intriguingly, it has previously been shown that activation of FcγR by immune complexes can potently inhibit IL-12 secretion by macrophages.6-8 IL-12 is a dimeric cytokine which shares its p40 subunit with IL-23. Whether immune complexes also inhibit IL-23 is not known as the 3 IL-23 specific p19 subunit had not yet been identified at the time these experiments were performed.

IL-12 and IL-23 are mainly produced by macrophages and dendritic cells and play a key role in the pathogenesis of IBD. In the inflamed intestine of patients with Crohn’s disease, monocytes polarize towards an inflammatory macrophage phenotype and produce pro-inflammatory cytokines including TNF and IL-23.9 The blockade of IL-12p40 with ustekinumab, which neutralizes both IL-12 and IL-23, is currently approved as a treatment for patients with Crohn’s disease and more recently, the selective IL-23p19 inhibitors risankizumab and brazikumab showed promising results in clinical trials in patients with Crohn’s disease.10-13 Studies in ulcerative colitis are ongoing. The advent of this novel class of antibodies poses the question if patients with Crohn’s disease responding to anti-TNFs and IL-12/IL-23 blockade represent distinct or overlapping groups. TNF, IL-12 and IL-23 promote inflammation via distinct pathways, suggesting that therapies blocking these cytokines may establish mucosal healing via different routes and therefore work in different patient subgroups. However, given the fact that we have previously found that FcγR signaling is involved in the therapeutic effect of anti-TNFs and the ability of FcγR signaling to suppress IL-12 secretion, we aimed to examine how different anti-TNF blocking strategies affected macrophage IL-12/IL-23 secretion.

MATERIALS AND METHODS

Anti-TNF compounds Therapeutic anti-TNF compounds used in this study are adalimumab (Humira, AbbVie), infliximab (Remicade, MSD), golimumab (Simponi, MSD), certolizumab pegol (Cimzia, UCB Pharma), and etanercept (Enbrel, Pfizer). Infliximab Fab fragments were generated using the Pierce Fab micro Preparation Kit (Thermo Fisher, Landsmeer, The Netherlands).

57 Chapter 3

In vitro macrophage differentiation Peripheral blood mononuclear cells (PBMC) from healthy volunteers were isolated by Ficoll Paque density-gradient centrifugation. After washing, monocytes were isolated by Percoll density-gradient centrifugation.14 Inflammatory (M1) macrophages were obtained by culturing monocytes with IFNγ (50 ng/mL, PeproTech, London, UK) for 6 days in RPMI supplemented with 10% heat-inactivated FCS. Macrophages were then washed, reseeded and stimulated for 48 hours with LPS 100 ng/ml (Sigma, Aldrich, Zwijndrecht) and anti-TNF or IgG1 (Genetex, Irvine, CA). Complexed IgG1 was obtained by heating IgG1 at 63°C for 1 hour. Culture supernatants were harvested and cytokine levels were measured by ELISA for IL-23, IL-12p40 and IL-10 (Duoset, R&D) or by Cytometric Bead Array (Human Inflammatory Cytokine Kit, BD Biosciences). For pSYK staining, M1 macrophages were incubated with LPS and anti- TNF for 30 minutes, fixed using Fixation/Permeabilization Solution Kit (BD Biosciences) and stained for anti-pSYK-PE or isotype control staining (both Cell Signaling). Syk was inhibited with 1 μM R406 (Invivogen) added 1 hour prior to the addition of LPS and anti-TNF, cytokine production in this case was measured after 24 hours.

Heat-killed bacteria Heat-killed E. coli 0111:B4 was purchased from invivogen (catalog no. tlrl-hkeb2). E. faecalis (catalog no. 29212; ATCC) was cultured in brain-heart infusion (BHI) medium. Bacteria were harvested and washed twice with PBS. Then, bacterial suspension was heated at 80°C for 30 minutes, washed, resuspended in PBS, and stored at –80°C. Complete killing was confirmed by a 72-hour incubation at 37°C.

Preparation of LPMCs Intestinal mucosa was obtained from surgically resected specimens from patients with IBD, diagnosed on the basis of endoscopic and histological findings according to established criteria. All experiments were performed with mucosa from macroscopically inflamed specimens. All experiments were approved by the Medical Ethical Committee of the Academic Medical Center, Amsterdam. Written informed consent was obtained from all patients. Lamina propria mononuclear cells (LPMC) were isolated from intestinal specimens as follows; Briefly, dissected mucosa was incubated in calcium and magnesium-free HBSS containing 5 mM EDTA (Sigma-Aldrich) for 20 minutes at 37°C to remove epithelial cells. Tissues were then cut very fine and incubated in RPMI medium containing 10% FCS, 1 mg/ ml collagenase D (Roche) 1 mg/ml soybean Trypsin inhibitor (Sigma) and 50 µg/ml Dnase I (Roche) for 60 minutes at 37°C. The fraction was pelleted and resuspended in a 30% Percoll solution, then layered on 60% Percoll before centrifugation at 500 g for 10 minutes at room temperature. Viable LPMC were recovered from the 30%–60% layer interface. LPMC were then stimulated over night with heat-killed E. coli and E. faecalis both 1 × 108 cells/

58 Anti-TNF immune complexes inhibit IL-12/IL-23 secretion

ml in the presence of IgG1 or anti-TNF 10 µg/ml. Golgistop (BD Biosciences) was added after 1 hour. The next day, LPMC were stained for LIVE/DEAD® Fixable Green Dead Cell Stain Kit (Invitrogen), CD45-AF700 (Sony), CD14-PE-Cy7 (eBioscience), HLA-DR-PE, CD11b- BV421 (both Biolegend)) and then fixed using Fixation/Permeabilization Solution Kit (BD Biosciences). After fixation, cells were stained for IL-12p40-APC (Biolegend) and analyzed by FACS Fortessa (BD Biosciences) then analyzed using FlowJo software (Treestar Inc).

High-performance size-exclusion chromatography (HP-SEC) Recombinant human TNF was obtained from Active Bioscience. Therapeutic anti-TNF compounds (60 µg/ml, diluted in PBS) were incubated for at least one hour with 2 µg/ml recombinant human TNF. Samples were analyzed by applying 200 μl to a Superdex 200 HR 10/300 column (GE Healthcare, Uppsala Sweden), which was connected to an ÄKTAexplorer 3 HPLC system (GE Healthcare). Elution profiles were monitored by measuring absorbance at 215 nm. The Superdex column was calibrated using the Gel Filtration Markers Kit (29-700 kDa range, Sigma Aldrich). qPCR mRNA isolation was performed using the Bioline ISOLATE II RNA Mini kit (BIO-52073, Bioline) according to manufacturers’ instructions. First-strand cDNA was synthesized using Oligo-dT (Invitrogen), random hexamer primers (Promega, Madison, IL), RiboLock RNase and RevertAid reverse transcriptase (both Thermo Scientific). Quantitative RT-PCR was performed using sensifast SYBR No-ROX Kit (GC-biotech Bio-98020) on a BioRad iCycler. Relative gene expression was calculated using the 2−delta Ct method. GAPDH was used as reference gene.

Cohort studies of patients with Crohn’s disease treated with anti-TNF Intestinal biopsy specimens were collected in patients suffering from Crohn’s disease before anti-TNF treatment was initiated and during treatment at the Academic Medical Center in Amsterdam and at the University Hospital in Leuven. Endoscopy was performed in all patients prior to the onset of anti-TNF therapy and during follow-up. Seven responders and eight non-responders were included, where response versus non-response was determined by the clinician’s assessment of the endoscopic evaluation of the mucosa. Response was determined prior to analysis of the tissue samples (Patient characteristics in Supplementary Table 1). When comparing mRNA levels on consecutive timepoints within one patient, we compared biopsies within one anatomical location (i.e. colon versus colon and ileum versus ileum). The study protocols were approved by the Medical Ethical Committee of the Academic Medical Center, Amsterdam, the Netherlands and of the University Hospital, Leuven, the Netherlands, all participants provided written informed consent.

59 Chapter 3

Statistical Analysis For statistical analysis, unpaired T-test or ANOVA was used followed by Bonferroni post-test. Normalized data were analyzed with Wilcoxon signed rank test or Kuskal-Wallis followed by Dunn’s post-hoc analysis. Results were considered significant when P value < 0.05.

RESULTS

Infliximab inhibits IL-12p40 and IL-23 production by inflammatory macrophages Given the pro-inflammatory state of macrophages in the intestine of Crohn’s disease patients,9 we investigated the effects of infliximab on a priori polarized inflammatory macrophagesin vitro. Inflammatory macrophages were generated by culturing human monocytes with IFNγ. For clarity, we will hereafter refer to IFNγ-induced macrophages as M1 macrophages. M1 macrophages were washed and reseeded with LPS in the presence of infliximab or IgG1 isotype control. To explore the effect of infliximab on the pro-inflammatory cytokine profile of M1 macrophages we measured IL-12p40, IL-6, IL-8 and IL-1β production. While IL-6, IL-8 and IL-1β were comparable between infliximab and IgG1 treated M1 macrophages, IL-12p40 was specifically inhibited by infliximab. (Figure 1A) Further analyses revealed that the decrease in IL-12p40 was dose dependent. (Figure 1B). As IL-23 is composed of subunits IL-12p40 and IL-23-p19, we next measured IL-23 secretion and we found that IL-23 was also dose dependently decreased by infliximab. (Figure 1C)

Analysis by qPCR showed that infliximab inhibited the expression ofIL-12B mRNA, the gene encoding subunit IL-12p40. The expression of IL-12A and IL-23A, encoding subunits IL-12p35 and IL-23p19 respectively, was not affected by infliximab nor was the induction ofTNF itself. (Figure 1D) Overall, these data suggest that the suppression of the IL-23 protein is mediated by inhibition of IL-12B transcription and not IL-23p19.

Inhibition of IL12/IL-23 is Fc-dependent To investigate if the inhibition of IL-12p40 and IL-23 was Fc-dependent, we compared infliximab to certolizumab, a TNF neutralizing Fab’ fragment. Both anti-TNF compounds blocked TNF to undetectable levels. (Supplementary Figure 1) Hereby, we could distinguish between Fc-dependent mechanisms and effects that should be attributed to TNF neutralization. Again, infliximab treatment decreased IL-12p40 and IL-23. However, the production of these cytokines remained high in certolizumab treated M1 macrophages, indicating that the inhibition of IL-12p40 and IL-23 is mediated via the Fc-region of infliximab rather than neutralization of TNF. (Figure 2A). Similarly, infliximab Fab fragments did not inhibit IL-12p40 production, further supporting a role for the Fc-region (Supplementary Figure 2).

60 Anti-TNF immune complexes inhibit IL-12/IL-23 secretion

3

Figure 1. Infliximab specifically inhibits IL-12p40 and IL-23 production by inflam- matory macrophages. Human monocytes were cultured with IFNγ 50 ng/ml for 6 days to generate M1 macrophages. M1 macrophages were then reseeded with infliximab or IgG1 (both 10 µg/ml) in the presence of LPS 100 ng/ml (A) Inflammatory cytokine production by M1 macrophages cultured for 48 hours in 6 different donors. Each dot represent individual donor. Wilcoxon signed rank test. (B/C) IL-12p40 and IL-23 production by M1 macrophages measured by ELISA after 48 hrs. ANOVA followed by Dunnet’s Multiple comparisons test (D) mRNA expression of indicated genes normalized to GAPDH expression measured by RT-PCR. Mann Whitney U-test. Data are representative of at least 3 independent experiments with different donors and bars show means, error bars indicate standard error of the mean, * p<0.05, ** p<0.01, *** p<0.001

Interestingly, others have shown that FcγR binding on macrophages by IgG immune complexes not only inhibits IL-12 production, but also increases IL-10 production.6-8 We found that infliximab enhanced the production of IL-10 compared to certolizumab, confirming an Fc-dependent increase in IL-10 production. (Figure 2B) However, TNF neutralization with certolizumab reduced IL-10 production compared to cultures treated with LPS alone. Thus,

61 Chapter 3

IL-10 production is regulated by both TNF blockade and Fc-engagement in opposite ways and the outcome may differ depending on the relative strengths of these signals. This might explain why we observed some variability in the effects of infliximab on IL-10 production when analysing multiple donors. (Figure 2B) Overall, IL-10 production was always higher in infliximab treated cultures compared to certolizumab.

Analysis by qPCR showed that infliximab inhibited the expression ofIL-12B mRNA, whereas certolizumab did not. In line with figure 1, expression ofIL-12A and IL-23A was not inhibited by infliximab or certolizumab. (Figure 2C)

We recently showed that activating FcγR are indispensable for the therapeutic efficacy of anti-TNF in experimental colitis.2 Activating FcγR have been shown to signal via spleen tyrosine kinase (Syk) and ligation of FcγR results in the phosphorylation of Syk.15 To study this further in a human setting, we measured phosphorylated Syk in human M1 macrophages. Indeed, only in infliximab treated M1 macrophages the presence of phosphorylated Syk was detected (Figure 2D). Furthermore, inhibition of Syk completely reversed the infliximab mediated decrease in IL-12p40 as well as the increase in IL-10 production (Figure 2E).

TNF-anti-TNF complex formation is crucial for IL-23 inhibition Our previous data indicate an important role for FcγR engagement in the mode of action of IgG1 anti-TNFs in IBD, however, etanercept also contains an Fc-region but failed to show clinical efficacy in Crohn’s disease.16 Interestingly, when M1 macrophages were cultured with etanercept, the production of IL-12p40 and IL-23 was not inhibited. (Figure 3A) Also, IL-10 production was decreased to levels comparable to certolizumab treated macrophages. This suggests that the mere presence of an IgG1 Fc-region is not sufficient to increase the IL-10 to IL-12/IL-23 ratio. We hypothesized that anti-TNF antibodies like infliximab and adalimumab differ from etanercept in their ability to cross-link FcγR. Antibodies like infliximab and adalimumab have two Fab’ arms that can simultaneously bind two TNF molecules. Importantly, biologically active TNF is a homotrimer and it has been shown that TNF can be bound by three infliximab molecules at once. These features will enable the formation of TNF- anti-TNF immune complexes that can cross-link FcγR. In contrast, etanercept contains only one TNF binding domain and has been shown to bind soluble TNF in a 1:1 ratio.17 Importantly, cross-linking of FcγR by immune complexes affects downstream FcγR effector functions to a much greater extent than Fc engagement by monomeric antibodies.

62 Anti-TNF immune complexes inhibit IL-12/IL-23 secretion

3

Figure 2. Inhibition of IL12/IL-23 is Fc-dependent M1 macrophages were reseeded with LPS 100 ng/ml and infliximab or certolizumab 10 µg/ml for 48 hours unless otherwise stated. (A) IL-12p40 and IL-23 production by M1 macrophages measured by ELISA. ANOVA followed by Bonferroni post-test. (B) Normalized IL-10 production by M1 macrophages from multiple donors. Wilcoxon signed rank test. (C) mRNA expression of indicated genes normalized to GAPDH expression measured by RT-PCR. (D) Macrophages were fixed 30 minutes after the addition of LPS and anti-TNF, stained with anti-pSYK-PE or isotype control and visualized by immunofluorescence and flow cytometry. (E) Syk-inhibitor R406 was added prior to the addition of LPS and anti-TNF (1 µg/ml), IL-12p40 and IL-10 production was measured by ELISA after 24 hours. ANOVA followed by Bonferroni post-test. Data are representative of at least 3 independent experiments with different donors, bars show means, error bars indicate standard error of the mean, ** p<0.01, *** p<0.001.

63 Chapter 3

To confirm these differences in TNF-anti-TNF immune complex formation for multiple anti- TNF compounds, we performed size-exclusion chromatography. (Figure 3B) Infliximab, adalimumab and golimumab formed a complex that roughly corresponds in size to a complex where 3 antibodies are bound to 1 soluble TNF molecule. Etanercept formed a complex with soluble TNF almost equal in size to unbound etanercept, indicating that one etanercept molecule was bound to one TNF molecule. (Figure 3B). In line with these data, we have previously shown that in the presence of soluble TNF, FcγR binding capacity becomes much higher for adalimumab than for etanercept.2 Thus, although etanercept has an Fc-region, it does not bind FcγR as potently as monoclonal antibodies infliximab or adalimumab in the presence of soluble TNF and this is reflected by the inability of etanercept to increase the IL-10 to IL-12/IL-23 ratio.

Finally, to further investigate the need for immune complex formation to achieve IL-12/ IL-23 inhibition, we compared non-specific monomeric IgG1 to complexed IgG1. Indeed, complexed IgG1 increased IL-10 and inhibited both IL-12p40 and IL-23 compared to monomeric IgG1, confirming their functional difference. (Figure 3C) Furthermore, when complexed IgG1 was added to certolizumab treatment (to add TNF blockade to the FcγR crosslinking by the complexed IgG), the production of IL-10, IL-12p40 and IL-23 became equal to infliximab treated M1 macrophages. (Figure 3D) Thus, the increase in IL-10 versus IL-12/IL23 ratio mediated by infliximab can be mimicked by a combination of certolizumab (TNF blockade) and IgG1 immune complexes (FcγR cross-linking).

64 Anti-TNF immune complexes inhibit IL-12/IL-23 secretion

3

Figure 3. Anti-TNF immune complexes mediate alterations in IL-10 and IL-23 pro- duction. (A) M1 macrophages stimulated with LPS 100 ng/ml and different anti-TNF compounds 10 μg/ml. Cytokine production after 48 hrs. ANOVA followed by Bonferroni post test (B) Complex formation of TNF with anti-TNF. TNF 2 µg/ml was incubated with anti-TNF 60 µg/ml and analysed on a Superdex 200 column. For each sample a representative elution pattern is shown out of at least two measurements. Arrows indicate elution of TNF-anti-TNF complexes. (C) M1 macrophages stimulated with IgG1 or complexed IgG1 (both 10 μg/ml) in the presence of LPS 100 ng/ml. Cytokine production measured by ELISA after 48 hrs. Unpaired T-test (D) M1 macrophages stimulated with anti-TNF and complexed IgG1 where indicated in the presence of LPS 100 ng/ml. Cytokine production measured by ELISA after 48 hrs. ANOVA followed by Bonferroni post-test. Data are representative at least three experiments with different donors, *** p<0.001.

65 Chapter 3

IgG1 anti-TNF inhibits IL-12/IL-23 in mucosal CD14+ cells from IBD patients In an inflamed intestine, mucosal macrophages are constantly exposed to a wide range of microbial products. To explore if our findings could be reproduced using an equivalent stimulation, we cultured M1 macrophages with heat-killed bacteria. Indeed, when M1 macrophages were cultured with a combination of heat-killed E. coli and heat-killed E. faecalis, we found that infliximab inhibits IL-12p40 in an Fc-dependent manner also in this setting. (Figure 4A) Additionally, this was also the case when M1 macrophages were cultured with either heat-killed E. coli or heat-killed E. faecalis separately (Supplementary Figure 3).

Next, we isolated lamina propria mononuclear cells (LMPC) from IBD patients. Cells were stimulated overnight with a combination of heat-killed E. coli and E. faecalis in the presence of infliximab, certolizumab or an IgG1 isotype control antibody. Analyses by flow cytometry revealed that infliximab inhibited IL-12p40 protein production in the CD14+ lamina propria myeloid population to levels comparable with unstimulated cultures. (Figure 4B) Furthermore, IL-12p40 was significantly decreased in infliximab treated CD14+ lamina propria cells compared to certolizumab or IgG1 treated LPMC. (Figure 4C)

Finally, mucosal mRNA expression of the gene encoding IL-12/IL-23p40 (IL-12B) was analyzed in a cohort of patients with Crohn’s disease treated with infliximab or adalimumab. We included 7 endoscopic responders and 8 endoscopic non-responders. (Patient characteristics shown in Supplementary Table 1) The expression of IL-12B decreased significantly in responders upon anti-TNF treatment, but remained stable in non-responders. (Figure 4D)

66 Anti-TNF immune complexes inhibit IL-12/IL-23 secretion

3

Figure 4. IgG1 anti-TNF inhibits mucosal IL-12/IL-23 in IBD patients. (A) M1 macrophages were stimulated with heat-killed E.coli and E. faecalis (both 1 × 108 cells/ml) and treated with anti-TNF (10 µg/ml) for 48 hours. IL-12p40 was measured by ELISA. ANOVA followed by Bonferroni post-test. (B/C) Representative histograms and graph of the CD14+ lamina propria myeloid compartment stimulated with heat-killed E. coli and E. faecalis (both 1 × 108 cells/ml) overnight in the presence of IgG1, certolizumab or infliximab (all 10 µg/ml). IL-12p40 was measured in CD14+ macrophages by flow cytometry. ANOVA followed by Bonferroni post-test. (D) mRNA expression of IL-12B measured by RT-PCR in intestinal biopsy specimens from patients with Crohn’s disease treated with infliximab or adalimumab, 7 endoscopic responders and 8 endoscopic non-responders. Each dot represents an individual patient. Mann Whitney U-test. Data are representative of at least 3 experiments with different donors or different IBD patients and bars show means, error bars indicate standard error of the mean.* p<0.05. ** p<0.01.*** p<0.001.

67 Chapter 3

DISCUSSION

The mode of action of anti-TNFs in IBD is still a matter of investigation.1 Here we find that the trimeric cytokine TNF forms immune complexes with IgG1 anti-TNF monoclonal antibodies that potently inhibit IL-12/IL-23 secretion at the level of IL12B mRNA in an Fc dependent manner.

Only a few years after the introduction of infliximab, it became clear that the mode of action of anti-TNFs in IBD was less certain than anticipated when etanercept completely failed to achieve either clinical response or remission in patients with Crohn’s disease, with patients on etanercept doing numerically worse than placebo.16 Also, treatment with etanercept has been associated with paradoxical development of IBD in patients with rheumatic diseases.1, 18 Certolizumab is significantly better than placebo for maintenance treatment but did not reach the primary endpoint for induction of clinical remission in three independent trials.19-22 In addition, there are no placebo-controlled data that support induction of mucosal healing by certolizumab. One uncontrolled endoscopic study showed that only 4% of patients with Crohn’s disease achieved mucosal healing after 10 weeks of certolizumab treatment.23 By comparison, complete mucosal healing was observed in 31% of patients on infliximab and 28% of patients on adalimumab in the same time frame.24, 25

We have previously shown that one of the potential mechanisms of action that differentiates the IgG1 antibodies infliximab and adalimumab from other anti-TNF blocking strategies is dependent on the antibody Fc region. We found that IgG1 antibodies but not certolizumab or etanercept generate CD206 positive macrophages with regulatory and wound healing properties in a mixed lymphocyte reaction in vitro.5 We found that responders but not non- responders to infliximab show accumulation of such CD206+ macrophages in the lamina propria in vivo.4 In a preclinical model we used two anti-TNF antibodies with the same variable region that differed in their ability to bind FcγR. In a head-to-head comparison we found that both antibodies were equally effective in the collagen induced model of arthritis but that FcγR interaction was required for mucosal healing in the CD45RBhigh transfer model of colitis.3 More recently we demonstrated that mice lacking activating FcγR failed to generate CD206+ regulatory macrophages and were completely unresponsive to anti-TNF therapy in the same model of colitis. Conversely, a hypo-fucosylated anti-TNF with increased affinity for FcγR showed increased CD206+ macrophage formation as well as improved therapeutic efficacy.2 In all our previous work the anti-TNF induced CD206+ macrophages were likely formed from monocyte precursors.

68 Anti-TNF immune complexes inhibit IL-12/IL-23 secretion

Here, we studied the effects of Fc-engagement on pro-inflammatory human macrophages with an M1-like phenotype. We found that in addition to generating CD206+ macrophages, anti-TNF monoclonal antibodies specifically inhibit IL-12/IL-23 production by M1-like macrophages. The inhibition of IL-12/IL-23 was mediated via Syk and was only achieved with anti-TNF compounds that have an Fc-region and were capable of forming immune complexes. TNF is a homotrimeric molecule containing three binding sites for anti-TNF antibodies. For antibodies like infliximab and adalimumab this promotes the formation of TNF-anti-TNF complexes. Complexed antibodies affect FcγR effector functions to a much greater extent than monomeric antibodies, by virtue of their ability to cross-link multiple FcγR. Although we have confirmed the primary finding of inhibition of IL-12/IL-23 in primary intestinal macrophages, it would be interesting to further evaluate the effects of immune complexes in this particular cell type as well. 3

In vivo, we observed a decrease in IL-12B expression in patients responding to anti-TNF therapy, but not in non-responders. It should be noted that our cohort was a clinical follow up cohort without hard criteria for response. However, a sub-analysis of the ACT-1 trial, a prospective study using well defined endoscopic criteria showed similar results with decreased expression of IL12B in responders after 8 weeks of therapy.26

As other current targets in IBD do not have a homomultimeric structure, it seems unlikely that other monoclonal antibodies used in IBD share a similar mode of action. However, the inhibition of IL12/IL23 by TNF-anti-TNF immune complexes does pose questions around the differentiation of therapies blocking IL-12/IL-23 directly from anti-TNF monoclonal antibodies. Our data suggest that that the mode of action of anti-TNFs in IBD may be partially overlapping with ustekinumab (anti-IL12/IL23) and IL-23 specific therapies such as risankizumab. No controlled endoscopic data are currently available for ustekinumab. In a recent phase 2 study with risankuzimab, that involved mainly anti-TNF exposed patients, the delta between placebo and the top dose was only 5% in terms of mucosal healing.10 Thus, although we need to await further data, the available endoscopic data support the idea that this therapeutic class may not be as differentiated from anti-TNFs as previously hoped.

Collectively, our results point out a previously unrecognized effector function of anti-TNF therapy that involves inhibition of the IL-12/IL-23 axis. These findings could contribute to our understanding about the failure of multiple biologicals in therapy refractory patients, but also suggest a new immunosuppressive mechanism for therapeutic antibodies capable of immune complex formation.

69 Chapter 3

REFERENCES

1. Levin AD, Wildenberg ME, van den Brink GR. Mechanism of Action of Anti-TNF Therapy in Inflammatory Bowel Disease. J Crohns Colitis 2016;10:989-97.

2. Bloemendaal FM, Levin AD, Wildenberg ME, et al. Anti-Tumor Necrosis Factor with a Glyco-engineered Fc Region Has Increased Efficacy in Mice with Colitis. Gastroenterology 2017.

3. McRae BL, Levin AD, Wildenberg ME, et al. Fc Receptor-mediated Effector Function Contributes to the Therapeutic Response of Anti-TNF Monoclonal Antibodies in a Mouse Model of Inflammatory Bowel Disease. J Crohns Colitis 2016;10:69-76.

4. Vos AC, Wildenberg ME, Arijs I, et al. Regulatory macrophages induced by infliximab are involved in healing in vivo and in vitro. Inflamm Bowel Dis 2012;18:401-8.

5. Vos AC, Wildenberg ME, Duijvestein M, et al. Anti-tumor necrosis factor-alpha antibodies induce regulatory macrophages in an Fc region-dependent manner. Gastroenterology 2011;140:221-30.

6. Gallo P, Goncalves R, Mosser DM. The influence of IgG density and macrophage Fc (gamma) receptor cross-linking on phagocytosis and IL-10 production. Immunol Lett 2010;133:70-7.

7. Sutterwala FS, Noel GJ, Clynes R, et al. Selective suppression of interleukin-12 induction after macrophage receptor ligation. J Exp Med 1997;185:1977-85.

8. Sutterwala FS, Noel GJ, Salgame P, et al. Reversal of proinflammatory responses by ligating the macrophage Fcgamma receptor type I. J Exp Med 1998;188:217-22.

9. Kamada N, Hisamatsu T, Okamoto S, et al. Unique CD14 intestinal macrophages contribute to the pathogenesis of Crohn disease via IL-23/IFN-gamma axis. J Clin Invest 2008;118:2269-80.

10. Feagan BG, Sandborn WJ, D’Haens G, et al. Induction therapy with the selective interleukin-23 inhibitor risankizumab in patients with moderate-to-severe Crohn’s disease: a randomised, double-blind, placebo-controlled phase 2 study. Lancet 2017;389:1699-1709.

11. Feagan BG, Sandborn WJ, Gasink C, et al. Ustekinumab as Induction and Maintenance Therapy for Crohn’s Disease. N Engl J Med 2016;375:1946-1960.

12. Sandborn WJ, Gasink C, Gao LL, et al. Ustekinumab induction and maintenance therapy in refractory Crohn’s disease. N Engl J Med 2012;367:1519-28.

13. Sands BE, Chen J, Feagan BG, et al. Efficacy and Safety of MEDI2070, an Antibody Against Interleukin 23, in Patients With Moderate to Severe Crohn’s Disease: A Phase 2a Study. Gastroenterology 2017;153:77-86 e6.

14. Repnik U, Knezevic M, Jeras M. Simple and cost-effective isolation of monocytes from buffy coats. J Immunol Methods 2003;278:283-92.

15. Nimmerjahn F, Ravetch JV. Fcgamma receptors as regulators of immune responses. Nat Rev Immunol 2008;8:34-47.

16. Sandborn WJ, Hanauer SB, Katz S, et al. Etanercept for active Crohn’s disease: a randomized, double- blind, placebo-controlled trial. Gastroenterology 2001;121:1088-94.

17. Scallon B, Cai A, Solowski N, et al. Binding and functional comparisons of two types of tumor necrosis factor antagonists. J Pharmacol Exp Ther 2002;301:418-26.

18. O’Toole A, Lucci M, Korzenik J. Inflammatory Bowel Disease Provoked by Etanercept: Report of 443 Possible Cases Combined from an IBD Referral Center and the FDA. Dig Dis Sci 2016;61:1772-4.

19. Sandborn WJ, Feagan BG, Stoinov S, et al. Certolizumab pegol for the treatment of Crohn’s disease. N Engl J Med 2007;357:228-38.

20. Sandborn WJ, Schreiber S, Feagan BG, et al. Certolizumab pegol for active Crohn’s disease: a placebo- controlled, randomized trial. Clin Gastroenterol Hepatol 2011;9:670-678 e3.

70 Anti-TNF immune complexes inhibit IL-12/IL-23 secretion

21. Schreiber S, Khaliq-Kareemi M, Lawrance IC, et al. Maintenance therapy with certolizumab pegol for Crohn’s disease. N Engl J Med 2007;357:239-50.

22. Schreiber S, Rutgeerts P, Fedorak RN, et al. A randomized, placebo-controlled trial of certolizumab pegol (CDP870) for treatment of Crohn’s disease. Gastroenterology 2005;129:807-818.

23. Hebuterne X, Lemann M, Bouhnik Y, et al. Endoscopic improvement of mucosal lesions in patients with moderate to severe ileocolonic Crohn’s disease following treatment with certolizumab pegol. Gut 2013;62:201-8.

24. Rutgeerts P, Diamond RH, Bala M, et al. Scheduled maintenance treatment with infliximab is superior to episodic treatment for the healing of mucosal ulceration associated with Crohn’s disease. Gastrointest Endosc 2006;63:433-42; quiz 464.

25. Rutgeerts P, Van Assche G, Sandborn WJ, et al. Adalimumab induces and maintains mucosal healing in patients with Crohn’s disease: data from the EXTEND trial. Gastroenterology 2012;142:1102-1111 e2.

26. Toedter G, Li K, Marano C, et al. Gene Expression Profiling and Response Signatures Associated With Differential Responses to Infliximab Treatment in Ulcerative Colitis. Am J Gastroenterology 2011;106:1272- 1280. 3

Funding This project was supported by Health Holland, Top Sector Life Sciences and Health

Disclosures G.R. van den Brink is currently an employee of GlaxoSmithKline; G.R.A.M. D’Haens received consulting and/or lecture fees from AbbVie, ActoGeniX, AM Pharma, Boehringer Ingelheim GmbH, Centocor, ChemoCentryx, Cosmo Technologies, Elan Pharmaceuticals, Engene, Dr Falk Pharma, Ferring, Galapagos, Giuliani SpA, Given Imaging, GlaxoSmithKline, Jansen Biologics, Merck Sharp and Dohme Corp, Millennium Pharmaceuticals, Inc (now Takeda), Neovacs, Novonordisk, Otsuka, PDL Biopharma, Pfizer, Receptos, Salix, Setpoint, Shire Pharmaceuticals, Schering-Plough, Tillotts Pharma, UCB Pharma, Versant, and Vifor Pharma and reports receiving research grants from Abbott Laboratories, Jansen Biologics, Given Imaging, MSD, DrFalk Pharma, and Photopill; and speaking honoraria from Abbott Laboratories, Tillotts, Tramedico, Ferring, MSD, UCB, Norgine, and Shire; M.E. Wildenberg reports personal fees from Takeda, personal fees and non-financial support from Janssen Parmaceuticals, grants from GlaxoSmithKline, grants from Tillotts Pharma, grants from Immunic, outside the submitted work; C.J. Buskens reports being on the Advisory board Johnson & Johnson (energy devices).

71 Chapter 3

SUPPLEMENTARY FIGURES

Supplementary Figure 1. TNF is equally neutralized by both certolizumab and infliximab. M1 macrophages were stimulated with LPS 100 ng/ml and anti-TNF 10 μg/ml. TNF was measured by Cytometric Bead Array.

Supplementary Figure 2. Infliximab Fab fragments do not inhibit IL-12p40 pro- duction by inflammatory macrophages. Infliximab was digested using papain and purified using a protein A column. (A) Total digest, purified Fab and remaining Fc fragments were run on a non-reducing SDS-PAGE gel, size indicated by marker in kD. (B) M1 macrophages were stimulated with LPS (100 ng/ml) and treated with infliximab Fab fragments (10 µg/ml) for 48 hours. IL-12p40 was measured by ELISA. Bars show means, error bars represent standard error of the mean.

72 Anti-TNF immune complexes inhibit IL-12/IL-23 secretion

Supplementary Figure 3. Anti-TNF inhibits IL-12p40 in Fc-dependent manner 3 during heat-killed E. coli or E. faecalis stimulation. M1 macrophages were stimulated with heat-killed E.coli or heat-killed E. faecalis (both 1 ×10E8 cells/ml) and treated with anti-TNF (10 µg/ml) for 48 hours. IL-12p40 was measured by ELISA. Data are representative of 2 independent experiments with different donors. ANOVA followed by Bonferroni post-test, bars show means, error bars represent standard error of the mean. ** p<0.01.*** p<0.001.

73 Chapter 3

Supplementary Table 1. Patient characteristics

Montreal classification* Gender A1/A2/A3 anti-TNF ** L1/L2/L3/L4 B1/B2/B3/+p Age at inclusion (years) weeks between biopsies

Patient 1 F 2 3 1 8 23 IFX Patient 2 M 1 3 3 8 39 ADA Patient 3 M 2 3 2 183 25 ADA Patient 4 F 2 3 2 17 20 ADA Patient 5 F 2 3 2p 129 56 ADA Patient 6 F 2 3 2 66 35 ADA

Responders Patient 7 F 2 3 1 13 40 ADA Patient 8 F 2 1 1 8 31 ADA Patient 9 F 2 3 2 8 23 IFX Patient 10 F 2 2 3/+P 8 42 ADA Patient 11 M 2 3 2 8 61 ADA Patient 12 M 2 1 2 8 20 ADA Patient 13 M 2 3+4 3p 29 39 ADA Patient 14 M 2 3 3 20 48 ADA

Non-responders Patient 15 F 2 3 3p 54 51 ADA

* Montreal classification: Age at diagnosis: A1: <17 years, A2: 17–40 years, A3: >40 years; Maximal location of disease: L1: ileal, L2: colonic, L3: ileocolonic, L4: upper disease; Maximal disease behaviour: B1: non-stricturing, non-penetrating, B2: stricturing, B3: penetrating, p: perianal disease. ** IFX: Infliximab, ADA: Adalimumab

74