Introduction

Inflammatory bowel disease (IBD) is a chronic intestinal disease with unknown causes. Symptoms include pain and bloody diarrhoea. Little is known about environmental factors that trigger flare-ups, including the effects of commonly used medications. Recent studies suggest a link between microbial factors and IBD, making it important to investigate the predictive level of antibiotics on flare-ups. New machine learning methods may be able predict IBD flare-ups by analysing patient data to identify contributing factors.

Aims & Methods

We aimed to investigate whether certain antibiotics increase the risk of flares in patients with IBD. We utilized a previously described¹ cohort of all patients with IBD in Denmark between 1974 and 2018. The cohort is based on the Danish Nationwide Patient Registry, which registers all healthcare related visits, treatments and procedures on an individual level. We utilized data between 1994 and 2018, and two distinct cohorts were created for the purpose of identifying flares. The first cohort was characterized by having an IBD-related hospital stay. The second cohort was characterized by experiencing a flare of IBD in need of systemic steroids. Patients were matched 1:5 using K-nearest neighbour. Antibiotics were grouped into the 3rd level of the Anatomical Therapeutic Chemical (ATC) classification system, with anthelmintics grouped into the 2nd level due to low numbers. All were given within 60 days before the outcome.
Cohorts were divided into an 80/20 training/testing set and trained using a 5-fold cross-validation with eXtreme Gradient Boosted decision tree (GBDT) framework with variables including year of remission/flare, sex, age at diagnosis, year of diagnosis, type of IBD, socioeconomic status², subtypes of IBD-medication, subtypes of biologics and subtypes of antibiotics. The final model was chosen based on the area under the receiver operating characteristic curve (AUROC). The final model was evaluated on the test set (i.e. unseen data for the model). Evaluation of the models was reported as accuracy (ACC), positive predictive value (PPV), and negative predictive value (NPV).

Results

There were in total 69,908 IBD patients in the original cohort. After applying the abovementioned algorithm, a total of 15,636 and 5,178 patients were included in the cohort for hospitalisation and steroid use, respectively.
The models achieved an AUROC of 0.71 (SD: 0.03) and 0.85 (SD: 0.008) for predicting a course of steroids and need for IBD-related hospitalization, respectively, on the training set. On the test-set, the models achieved an ACC of 82.72%, with a PPV of 36.36% and a NPV of 83.73% on predicting a course of steroids, and an ACC of 85.23%, with a PPV of 60.14% and a NPV of 87.80% on predicting IBD-related hospitalization. Quinolones and agents against amoebiasis and protozoal were identified as top 10 most important variables for making accurate predictions and for splitting decision trees. Other antibiotics, such as tetracyclines, betalactam, and intestinal anti-infectives, were also useful as splitting variables.

Conclusion

This study demonstrated the importance of antibiotics, particularly quinolones and agents against amoebiasis and protozoal, in predicting an IBD-related flare. The GDBT-trained models achieved a high accuracy in predicting hospitalization and steroid courses, with AUROC of 0.85 and 0.71, respectively. These findings highlight the impact antibiotics on the subsequent risk for a severer flare-up of IBD flare and may guide future research on the role of microbial factors in IBD.

References

¹ Lo B, Zhao M, Burisch J. Identifying patients with inflammatory bowel disease in the Danish National Patient Register. Dan Med J. 2023 Mar 27;70(4):A07220458. PMID: 36999818.
² https://www.dst.dk/en/Statistik/dokumentation/nomenklaturer/socio

Disclosure

An unrestricted grant was received from the Aage og Johanne Louis-Hansens Fond to support this study.

Introduction

Inflammatory bowel disease (IBD) is an idiopathic and chronic disorder that includes ulcerative colitis (UC) and Crohn’s disease (CD). Both diseases are different but show an uncontrolled intestinal immune response that generates tissue inflammation. Dendritic cells (DC) are phagocytic professional antigen presenting cells that link the innate and adaptive immune systems. Indeed, they are the only cell type able to stimulate naïve T cells generating an antigen-specific immune response. In humans, DC can be divided into plasmacytoid (pDC) and c conventional type 1 and 2 (cDC1, cDC2). Although cDC migration towards the GI mucosa is enhanced in IBD, the mechanisms underlying these migrations and whether biological drugs can modulate them are currently unknown.

Aims & Methods

This study aimed to analyze DC migratory capacity towards the GI mucosa, and whether biological drugs with different mechanisms of action can modulate this process in health and IBD.This is a cross-sectional observational postmarketing study where the modulatory effect of different biological treatments on samples from patients with an endoscopic diagnosis of IBD and healthy controls was studied ex vivo. Peripheral blood mononuclear cells (PBMC) were obtained from 15 patients per group and were cultured in presence of golimumab or ustekinumab. Following culture, surface markers expression was determined, assessing the differential effect of biological drugs (if any) over the DC subsets in the different groups. After that, the migratory capacity of circulating DC towards CCL2, CCL25, and MadCam1 from the different study groups was determined. Finally, we evaluated the DC migratory capacities in presence of biological drugs or with culture media alone.

Results

The cDC2 subset showed higher expression levels of β7 and CCR5 compared to cDC1. Golimumab and ustekinumab did not modify the expression of the homing markers in any studied subset in pDC and cDC2.On the contrary, in cDC1 both golimumab and ustekinumab increased CCR6 expression in qCD patients and CCR9 in aUC patients. Regarding to the migratory capacities of the different DC subsets toward intestinal chemoattractants ustekinumab decreased the migratory capacity of cDC2 from CD patients towards a medium supplemented with MadCam1 and increases toward CCL2 in case of cDC1 from CU. In the case of migratory capacities of cDC2 from HC were increased by ustekinumab and golimumab toward CCL2 and CCL25 just in case of golimumab.

Conclusion

cDC2 showed more migratory profile markers than other DC populations. Golimumab and ustekinumab modify the expression of the homing markers just in cDC1 subset. Nevertheless, ustekinumab and golimumab modify the migratory capacities of cDC in IBD patients and HC, but do not exert any effect on pDC. These effects occur in a different manner in CD and UC indicates that the behavior of cDC is different in CD and UC, and that the migration capacity of this subset is not dependent exclusively on homing markers.

Introduction

Therapeutic progress in inflammatory bowel disease (IBD) is limited by the lack of understanding of the pathophysiology and the mechanisms involved in response or lack/loss of response to treatments. IBD patients are treated with anti-inflammatory drugs, including biologics such as anti-tumor necrosis factor (TNF) agents, which have been crucial in achieving steroid-free remission and prevention of long-term complications. However, a significant subgroup of patients becomes refractory to these biologicals due to the development of immunogenicity, characterized by the presence of anti-drug antibodies (ADAs). The antibody production and peripheral tolerance are initiated and maintained within lymphoid organs, such as the peripheral lymph nodes, but the exact role of immune cells in these lymphoid structures and whether they are involved in anti-drug antibody (ADA) formation is unknown.

Aims & Methods

We aimed to unravel the pathophysiology behind the development of anti-drug antibodies. Therefore, we performed ultrasound-guided inguinal lymph node needle biopsies from which we isolated leukocytes along with PBMCs from the same patients. Leukocyte populations were analyzed at single cell level via high dimensional phenotyping through mass cytometry (CyTOF).

Results

Using this unique approach, we have so far results from 4 Crohn’s Disease (CD) patients with ADAs (75% females, 50% active disease) and 4 without ADAs (50% females, 75% active disease). Active disease was defined, as an observation of an elevated CRP (c-reactive protein) (>5 mg/L) and/or calprotectin >250 (mg/kg), including symptoms (Harvey-Bradshaw index score > 4 or disease activity index score-CDAI >150).
The proportion of CD4⁺ TEMRA (terminally differentiated effector memory T-cells), CD4⁺ effector and central memory T-cells were increased in the peripheral lymph nodes of patients that developed immunogenicity. Mature NK cells, classical and non-classical monocytes were also increased. Furthermore, the levels of the alpha-4 integrin expression were higher in the lymph node T-lymphocytes, as well as, in the lymph node naïve and memory B-cell subsets in the patient group with ADAs.

Table: Cytometry and Biochemical Characteristics.

Conclusion

Lymph node T-helper cells, monocytes and NK cells potentially play a role in the development ADAs in IBD patients treated with biologics. Although, there was no significant difference in the proportion of naïve and memory B-cells, our findings concerning the alpha-4 integrin expression levels also indicate that both lymph node T and B-lymphocytes, are possibly characterized by a higher migratory potential during immunogenicity conditions.

References

1. Stein RE, Lee DY, Leonard MB, et al. The association between drug levels, anti-drug antibodies, and therapeutic response during infliximab therapy in pediatric Crohn’s disease. J Crohns Colitis 2014; 8: S434.
2. Bodini G, Savarino V, Dulbecco P, et al. The influence of anti-adalimumab antibodies on adalimumab trough levels, TNF-alpha levels and clinical outcome. J Crohns Colitis 2014; 8: S42
3. Stefan S, Sandborn WJ, Choi J, et al. Anti-drug antibodies separate responses of markers of inflammation (c-reactive protein, fecal calprotectin) in Crohn’s disease patients treated with certolizumab pegol. Inflamm Bowel Dis 2014; 20: S68.
4. Strand V, Balsa A, Al-Saleh J, et al. Immunogenicity of Biologics in Chronic Inflammatory Diseases: A Systematic Review. BioDrugs. 2017;31(4):299-316. doi:10.1007/s40259-017-0231-8.
5. Brun MK, Goll GL, Jørgensen KK, et al. Risk factors for anti-drug antibody formation to infliximab: Secondary analyses of a randomised controlled trial. J Intern Med. 2022;292(3):477-491. doi:10.1111/joim.13495.

Disclosure

The authors state no conflict of interest.

Introduction

Inflammatory Bowel Diseases (IBD) are chronic inflammatory disorders characterised by alterations of acute relapses and remission. The past decades have brought substantial advances in the pharmacological management of IBD by the introduction of immunosuppressive agents, biologics and lately small molecules. Although, the sustained efficacy of the currently available anti-inflammatory therapies is still far from optimal. Therefore, the development of novel therapies represents a major unmet clinical need in this field. Recently the enrichment of the coagulation pathway genes and among these, plasminogen activator inhibitor 1 (PAI-1) was described in IBD, whereas PAI-1 emerged as a link between the epithelium and inflammation. Moreover, the genetic deletion of PAI-1 impaired the severity of experimental colitis in mice.

Aims & Methods

Based on these, our aim was to develop and test novel PAI-1 inhibitors for the treatment of IBD. Novel, peptide-based PAI-1 inhibitors were designed and screened in silico, then the selected ones were synthetized using the ixPeptide® platform. The inhibitory potential of the peptides was determined by an in vitro PAI-1 activity assay. Cytotoxicity was monitored in vitro in Hela cells with the CellTiter Glo assay. Based on the in vitro results LDN-071 was selected for further analysis. Experimental colitis was induced in mice with 3% DSS and after the induction of colitis 10 mg/bwkg LDN-071 was administrated per os for four consecutive days. The changes of body weight, spleen weight and colon length were measured. Haematoxylin eosin (HE) and Periodic acid-Schiff (PAS) staining were performed for histology. ELISA and qRT-PCR measurements were applied to analyse the mucosal cytokine expressions. In vivo toxicity of LDN-071 was determined in mice.

Results

LDN-071 significantly inhibited PAI-1 in vitro without decreasing the cell viability in a concentration range of 1µM-1mM. The administration of 10 mg/bwkg LDN-071 in the DSS-treated mice prevented the body weight loss, the development of bloody diarrhoea, the decrease of colon length and increase of the spleen weight, which were displayed by the DSS+vehicle-treated animals. The histological analysis of the colon revealed that DSS treatment induced a severe transmural inflammation, ulceration, loss of crypt structures and goblet cell depletion. Importantly, all of these parameters were significantly impaired by the administration of 10 mg/bwkg LDN-071. Additionally, mucosal gene and protein expressions of TNF-α, IL-1β and IL-6 were significantly impaired by PAI-1 inhibition. Importantly, no signs of toxicity were LDN-071 observed in mice exposed to 100 mg/bwkg LDN-071 for 7 days.

Conclusion

Our results showed that LDN-071, a novel peptide-based inhibitor of PAI-1. It significantly reduces the severity of experimental colitis in mice. We propose that the inhibition of PAI-1 could be a potential novel therapeutic strategy in IBD.

Disclosure

LDN-071 PAI-1 inhibitor is owned by Ladon Therapeutics Ltd.

Introduction

Ulcerative colitis (UC) is a refractory inflammatory bowel disease of the colon with repeated relapses and remissions that requires lifelong treatment. Previous studies indicated that proinflammatory cytokines involved the onsets and developments of UC. New treatments for targeting these cytokines, such as TNF-α, IL-12 and IL-23 have been developed in a recent decade, underlying that immune abnormalities are the core for the etiology of UC.
Platelets play a role not only in hemostatic coagulation but in inflammatory mediator by interacting with various cells in the innate and adaptive immunity. Chronic inflammatory response confers platelets activation leading to elevated platelet count and hypercoagulability. Moreover, platelets aggregates with leukocyte in active inflammatory milieu. In fact, thrombocytosis, thromboembolism and upregulation of platelet-monocyte complex (PMC) in UC patients have been reported, implying that platelets could be novel biomarker, therapeutic target and clues for elucidating the pathology of UC. However, the clinical significance of platelet activation and PMC formation in UC is not fully elucidated and few studies were conducted to assess how platelet activation contributes to disease aggravations of UC. Therefore, we comprehensively analyzed platelet activation and platelet-monocyte complex by focusing on a correlation with clinical severity of UC.

Aims & Methods

Patients with UC (n=70) including MES0-2 group (n=34) and MES3 group (n=27), diagnosed according to a combination of symptoms, endoscopic findings, histologic findings, and the absence of alternative diagnoses, and age and sex-matched healthy controls (HCs) (n =19) were enrolled for this study. We analyzed the surface markers such as CD62P, CD40L, CD63, PAC-1 and Annexin V on CD61+ platelets by flowcytometry (FCM) in patients with UC in comparison with those in healthy adults. We also analyzed the proportion of CD61+CD14+ platelet-monocyte complexes and quantified the activation of PMCs by FCM. Moreover, we also investigated the correlation among surface markers, proportion of platelet-monocyte aggregates and Mayo endoscopic score (MES) in UC patients.

Results

The expression of CD62P but not other surface markers on platelets was significantly upregulated in UC when compared to HCs (p<0.01), and was significantly upregulated in in MES3 group when compared to the MES0-2 group (P<0.001). The proportion of PMC was also higher in MES3 groups when compared to MES0-2 group (P<0.05), and was markedly reduced by treatment (P<0.05). In addition, CD16+ monocytes, phenotypically mature and pro-inflammatory cell population, were increased in patients with UC when compared to HCs (P<0.01). The expression of CD16 on monocytes aggregated with platelets was significantly higher when compered to that without platelet (P<0.001).

Conclusion

Our study suggests that platelet activation accelerates mucosal inflammation via adhesion with monocytes in UC and contributes to the aggravation of disease.

References

Semple JW, et al. Platelets and the immune continuum.
Nat Rev Immunol. 2011 Apr;11(4):264-74.
Collins CE, et al. Platelets circulate in an activated state in inflammatory bowel disease.
Gastroenterology. 1994 Apr;106(4):840-5.

Introduction

Crohn's disease (CD) location would influence the risk of complications and therapeutic strategies. The objective of this study was to compare the clinical presentation at diagnosis and the natural history of colonic CD in comparison to ileal CD and ulcerative colitis (UC) in pediatric-onset IBD.

Aims & Methods

All children (< 17 years) with a diagnosis of CD or UC made between 1988 and 2011 in a population-based registry were followed retrospectively until 2013. The presentation at diagnosis, the risks of complications, surgery, hospitalization and exposure to different treatments in ileal CD (L1), colonic CD (L2) and UC were compared. Disease location was defined at diagnosis. In case of extension of the disease (L1 or L2 to L3), follow-up was censored.

Results

A total of 215 L1 patients, 234 L2 patients and 337 UC patients were included. At diagnosis, the proportion of male (L1 53.5%; L2 53.4%; UC 42.7%; p=0.012), age at diagnosis (15.0; 13.7; 14.0 years; p=0.003), family history of IBD (12.6%; 12.7%; 5.0%; p=0.005), diagnostic delay (3.0; 3.0; 2.0 months; p=0.001) and smoking prevalence (11.6%; 8.1%; 5.6%; p=0.041) were different between the three groups. Over the study period, incidence of L1, L2 and UC were 0.65 [0.57-0.74], 0.71 [0.62-0.81] and 1.02 [0.92-1.14] per 100,000 person-years, respectively. Bloody stool and diarrhea at diagnosis was observed in 14.9% respectively; 44.4%; and 91.1% (p < 0.001) and 47.0%; 71.8%; and 65.0% (p < 0.001) of L1, L2 and UC. At diagnosis, granuloma was identified in 12.6% of L1 patients and 31.2% of L2 patients (p<0.001). The risk of extension to L3 was significantly higher in the L2 group than in the L1 group (at 5 years - 37% vs. 14%, p<0.001). L2 location was associated with a lower risk of luminal fistula (HR 0.41 [0.21-0.79], p=0.008) but a higher risk of anoperineal lesion (HR 2.11 [1.29-3.45], p=0.003). The prevalence of extra-intestinal manifestations, articular (p<0.001) and cutaneous (p<0.001) was higher in L2 CD. While the risk of surgery was significantly higher in case of L1 location (at 5 years - 37%, 13% and 13%; p<0.001), the risk of hospitalization (at 5 years - 22%, 39% and 33%; p=0.01), exposure to corticosteroids (at 5 years - 55%, 69% and 67%; p<0.001), immunosuppressants (at 5 years - 47%, 61% and 42% - p<0.001) and anti-TNF (at 5 years - 16%, 35% and 22%; p<0.001) were higher in case of L2 location.

Conclusion

The clinical presentation and evolution of ileal and colonic CD at diagnosis differ significantly. Colonic location is associated with a high risk of perianal CD, extra-intestinal manifestations, hospitalization and exposure to steroids, immunosuppressants and anti-TNFs. These differences could justify different therapeutic approaches in the future.

Introduction

Inflammatory bowel diseases are chronic disabling conditions with a complex and multifactorial aetiology, which is still not completely understood. OCTN1, an organic cation transporter could have a role in modulating the inflammatory response, whose genetic polymorphisms were associated to increased risk of inflammatory bowel diseases. Up to now, limited information exists on its role in predicting/modulating response to therapies.

Aims & Methods

The aim of this study was to evaluate the role of OCTN1 in modifying gut microbiota and immune composition to Infliximab therapy in murine colitis. Dextran sodium sulphate model of colitis on C57BL/6 mice and ocnt1 gene knockout mice was used to assess clinical efficacy of infliximab administered intravenously. Stool, colon and mesenteric lymph node samples were collected to assess differences in gut microbiota composition, histology for Rachmilewitz score and T cell populations respectively.

Results

Octn1^-/-influences microbiota profile and is associated with a worse dysbiosis in colitis mice. Infliximab treatment attenuates colitis-associated dysbiosis with increase of bacterial richness and evenness in both strains. In comparison with wild type, octn1^-/- mice have a higher baseline percentage of Treg, Tmemory, Th2 and Th17.

Conclusion

Our data support murine model to study OCTN1 genetic asset in inflammatory bowel diseases. This could be the first step to consider this membrane transporter as a biomarker in inflammatory conditions and as a predictor of response to therapies.