Scandinavian Journal of Gastroenterology. 2015; Early Online, 1–9
ORIGINAL ARTICLE
Scand J Gastroenterol Downloaded from informahealthcare.com by East Carolina University on 04/20/15 For personal use only.
Global mucosal and serum cytokine profile in patients with ulcerative colitis undergoing anti-TNF therapy
RAHIL DAHLÉN1,2, MARIA K MAGNUSSON1,2, ANTAL BAJOR1, ANDERS LASSON3, KJELL-ARNE UNG4, HANS STRID1 & LENA ÖHMAN1,2 1
Department of Microbioloy and Immunology, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden, Department of Microbiology and Immunology, Sahlgrenska Academy, Gothenburg, Sweden, 3Södra Älvsborg Hospital, Borås, Sweden, and 4Skaraborgs Hospital, Skovde, Sweden
2
Abstract Background and objective. The knowledge of the effects of anti-tumour necrosis factor (TNF) treatment on the global cytokine profile in patients with ulcerative colitis (UC) is limited. A better understanding of these mechanisms could improve the ability to select patients that should undergo the therapy. Therefore, the aim was to determine the global mucosal and serum cytokine profile before and during induction therapy with anti-TNF in UC patients. Materials and methods. In total, mucosal biopsies (n = 28) and serum samples (n = 42) were collected from UC patients (total n = 48) before anti-TNF therapy. At week 14 response to the therapy was evaluated and again mucosal biopsies (n = 14) and serum samples (n = 42) were collected. Quantitative real-time PCR was used to determine mucosal cytokine mRNA expression and the MSD MULTI-ARRAY assay system platform was used for analysis of cytokines in serum. The global cytokine profile was evaluated by multivariate factor analysis. Results. At baseline, the global profile of mucosal cytokine mRNA expression and serum cytokines discriminated therapy responders from non-responders. Responders had lower mucosal mRNA expression of interleukin 1b (IL-1b), IL-17A, IL-6 and interferon g (IFN-g) than non-responders. Fourteen weeks after therapy start mucosal IL-1b and IL-6 were down-regulated in therapy responders but not in non-responders. At week 14, serum levels of IL-6 were decreased in therapy responders whereas IFN-g and IL-12p70 were increased in non-responders. Conclusions. Our data suggest that patients with a therapy failure have a more severe pro-inflammatory cytokine profile before start of anti-TNF treatment, which is less well suppressed by the treatment as compared to therapy responders.
Key Words: adalimumab, biomarker, cytokine, IBD, infliximab, prediction, therapy response, ulcerative colitis
Introduction Treatment of ulcerative colitis (UC) has been improved by the use of anti-TNF (tumor necrosis factor) therapy. However, at least 30% of the patients do not respond to anti-TNF therapy [1], and it is so far not possible to identify these individuals before treatment start [2]. Moreover, the immunological mechanisms of anti-TNF therapy linked to therapy response are not fully understood. Few studies have compared the immunopathology of therapy responders and non-responders of UC patients undergoing
anti-TNF therapy. Although still incompletely characterized, the mucosal cytokine profile in patients with active UC seems to involve a variety of cytokines, as demonstrated by increased mRNA expression of interferon gamma (IFN-g), interleukin (IL)-13, IL-17A, IL-1b, IL-6, TNF and IL-8 [3–5]. Additionally, UC patients with active disease have higher serum levels of IL-6, IL-8 and IL-10 as compared with healthy [6]. Thus, current data indicates that UC is associated with an increased expression of cytokines reflecting innate immune responses such as TNF, IL-1b, IL-6 and IL-8. Moreover, pathogenesis of
Correspondence: Lena Öhman, Department Microbiology and Immunology, University of Gothenburg, Box 435, 405 30 Gothenburg, Sweden. Tel: +46 31 786 6214. Fax: +46 31 786 6210. E-mail: [email protected]
(Received 2 February 2015; revised 11 March 2015; accepted 15 March 2015) ISSN 0036-5521 print/ISSN 1502-7708 online 2015 Informa Healthcare DOI: 10.3109/00365521.2015.1031167
Scand J Gastroenterol Downloaded from informahealthcare.com by East Carolina University on 04/20/15 For personal use only.
2
R. Dahlén et al.
UC also includes adaptive T helper (Th) cell responses of all kinds such as Th1 (IFN-g), Th2 (IL-4, IL-5, IL-13), Th17 (IL-17A) and T regulatory cells (IL-10, TGF-b) [7]. TNF is known to activate pro-inflammatory cytokines including IL-1b and IL-6 [8], and has wideranging effects in the body. The effects of anti-TNF therapy on immune cells and their effector cytokines are incompletely known, although a down-regulation of mucosal TNF and IFN-g mRNA expression in UC patients undergoing infliximab treatment has been shown [9,10]. It has been demonstrated that infliximab therapy influences the expression of several genes associated with adaptive immune responses differently in therapy responders and non-responders [11]. Additionally, infliximab has been reported to have the potential to down-regulate the gene expression of IFN-g, IL-6 and IL-2 [12], and the secretion of cytokines such as IFN-g, IL-17A and IL-13 [13] in vitro of blood T cells isolated from UC patients. Moreover, our group has demonstrated that infliximab treatment decreases serum levels of IL-5, IL-8 and TNF in UC patients responding to the therapy, but not in therapy failures, 2 weeks after treatment start [14]. Potentially, the patient’s global cytokine profile before start of treatment could predict the response to anti-TNF therapy. Interestingly, it has been demonstrated that high pre-treatment mucosal TNF expression was associated with insufficient response to anti-TNF therapy in UC patients. Thus, the infliximab treatment was less effective in patients with high mucosal expression of TNF mRNA, indicating a more severe disease in therapy non-responders than responders [15]. Still, the knowledge of the effects of anti-TNF treatment on cytokines associated with innate as well as adaptive T-cell immune responses in UC patients is limited and needs to be expanded to better understand the mechanisms of the therapy and
to guide the decision on which patients that are suited to undergo the therapy or perhaps benefit from an escalated dosing. In the present study of UC patients, we hypothesized that the global cytokine profile, representative of innate as well as adaptive T cell responses, in the intestinal mucosa and serum before start of treatment was different in responders as compared with nonresponders. We also postulated that anti-TNF therapy affected the cytokine profile differently in responders and non-responders. Therefore, the aim was to determine the global cytokine profile, including mucosal tissue and serum, in UC patients before start of anti-TNF therapy and at evaluation of therapy response at week 14, in therapy responders and nonresponders, respectively. Methods Patient cohort and sample collection Patients with UC (n = 48) were consecutively included in the study at the outpatient clinics at Sahlgrenska University Hospital, Gothenburg, Kärnsjukhuset in Skövde, and Södra Älvsborg Hospital, Borås, Sweden. All patients had an active disease at the time of inclusion, had failed increased dosing of 5-aminosalicylic acid and/or thiopurines and were either corticosteroidrefractory or dependent. Mucosal biopsies from sigmoid colon obtained during colonoscopy and serum samples were collected at baseline, i.e. before start of anti-TNF therapy and at week 14. From 6 of the patients biopsies alone were obtained, from 22 of the patients both serum and biopsies were obtained whereas only serum samples were obtained from the remaining 20 patients (in total 48). The clinical response was evaluated at week 14, and the patients were classified as therapy responders or non-responders, respectively (Figure 1). Primary therapy response
Mucosal biopsies +/- serum: 28 Serum alone: 20 Total: 48
14 28 42* Responders (n = 26)
Anti-TNF treatment Baseline disease assessment
Week 14 Therapy evaluation
Non-responders (n = 22)
*6 patients in early termination = no samples Figure 1. Study design and patient treatment response. Patients with ulcerative colitis (n = 48) were included in the study before start of antiTNF therapy. Disease activity assessment was performed at inclusion, termed baseline, and at week 14, or at the termination of treatment if it ended during therapy induction (six patients experienced early termination). Serum samples and biopsies from sigmoid colon, or serum alone were obtained at start of anti-TNF therapy and at week 14. Patients were defined as responders (n = 26) or non-responders (n = 22) at week 14. Abbreviation: TNF = Tumour necrosis factor.
Cytokine profiles in ulcerative colitis
Scand J Gastroenterol Downloaded from informahealthcare.com by East Carolina University on 04/20/15 For personal use only.
was defined as a decrease in Mayo score of ‡3 points compared with baseline. Patients not achieving this decrease in the Mayo score were considered as nonresponders. If the criteria for a full Mayo score evaluation was not fulfilled (n = 4), therapy response was defined by one or several of the following parameters: C-reactive protein (CRP), fecal calprotectin, and a partial Mayo score. The study was performed after receiving written informed consent from all subjects and the protocol was approved by the Regional Ethical Review Board at the University of Gothenburg. PCR analysis of mucosal biopsies Mucosal biopsies from sigmoid colon were collected for RT-PCR analysis. Biopsies were obtained from 28 patients at baseline (therapy responders n = 16 and non-responders n = 12). At week 14, biopsies from the evaluation visit were obtained from 14 patients (therapy responders n = 7 and non-responders n = 7). Total mRNA from mucosal biopsies was extracted using Nucleospin DNA, RNA, and protein purification kit (Macherey-Nagel, Düren, Germany) and cDNA was prepared from 1 mg RNA using the QuantiTect Reverse Transcription kit (Qiagen, Hilden, Germany) according to the manufacturers’ protocols. cDNA was then used for real-time RT-PCR using Taqman Universal PCR Master Mix (Applied Biosystems, Foster City, CA) and Taqman Gene Expression assays (Applied Biosystems). Expression of TNF, IL-6, IL-1b, IFN-g, IL-17A, IL-13 and FoxP3 was determined. The results were normalized to the expression level of GAPDH and expressed as 2-(Target-GAPDH). Cytokine analysis and trough level analysis of serum samples Serum samples were collected for analysis of cytokine content. Serum was obtained from 42 patients at baseline and at week 14 (therapy responders n = 26 and non-responders n = 16). The cytokines IL-1b, IL-4, IL-6, IL-8, IL-10, IL-12p70, IL-13, IL-17A, TNF and IFN-g were analyzed with MSD MULTI-ARRAY Assay System platform (MSD SCALE DISCOVERY, Rockville, USA). Levels of infliximab in serum samples was determined by the TNF a-Blocker monitoring, infliximab drug level (e.g. REMICADE), ELISA (Immundiagnostik AG, Bensheim, Germany). All assays were performed according to the manufacturers’ instructions. Statistical analysis Multivariate factor analysis (SIMCA-P+ software; Umetrics, Umeå, Sweden) was used to examine the
3
relation between therapy outcome and the examined cytokines. Orthogonal partial least squares discriminant analyses (OPLS-DA) were implemented to examine whether therapy outcome could be discriminated based on the totality of X-variables. The quality of the OPLS-DA was based on the parameter R2, that is, the goodness of the fit of the model (R2 = 1 represents best possible fit, whereas R2 ‡0.05 represents fairly good discrimination). In the OPLS-DA loading column plots, the importance of each X-variable (cytokine) to Y (therapy outcome) is represented by column bars. The variables positioned in the same direction as the bar representing therapy response are positively associated, whereas variables in the opposite direction are inversely related to therapy response. The larger the bar and the smaller the error bar, the stronger and more reliable is the contribution to the model. Univariate statistical analyses were performed with the GraphPad Prism 6.0 (GraphPad Software, La Jolla, CA, USA). The Wilcoxon signed rank test was used to evaluate differences between two sets of paired samples, and the Mann Whitney U test was performed to evaluate differences between two groups. p-Values < 0.05 were considered as statistically significant. Data given in the text and Tables I and II are demonstrated as median (range) and the horizontal lines in the figures show the median. Data given in Supplementary Tables I and II are demonstrated as median (25–75 percentiles).
Results Demographics and clinical parameters in therapy responders and non-responders In total, 48 patients with UC were included in the study, whereof 44 patients were treated with infliximab and 4 patients were treated with adalimumab (Table I). At week 14, 26 patients were classified as therapy responders and 22 patients as nonresponders. Therapy responders and non-responders showed no major differences in gender distribution, age, disease duration, Mayo score, CRP or fecal calprotectin before start of anti-TNF treatment, hereafter termed baseline (Table I). All the patients had active disease at baseline as determined by Mayo score. At week 14, treatment responders showed a reduction in endoscopic Mayo score, CRP and calprotectin whereas no changes were detected for non-responders (Table II). Trough levels at week 14 in patients treated with infliximab tended to be higher in treatment responders vs. non-responders (7700 ng/ml (1000–42200) vs. 3800 ng/ml (0– 14400), p = 0.07).
4
R. Dahlén et al.
Table I. Demographics of anti-TNF therapy naïve UC patients before anti-TNF therapy start.
Scand J Gastroenterol Downloaded from informahealthcare.com by East Carolina University on 04/20/15 For personal use only.
Responders Total number of patients Male/female Age, years (range) Smoking habit (never/ex smoker/active) Disease duration, years (range) Mayo score, median (range) Endoscopic Mayo score, median (range) Colonic area involved, left side/extensive CRP (mg/l), median (range) Calprotectin (mg/g), median (range) Anti-TNF treatment (IFX/ADA) Other treatments than anti-TNF: Corticosteroids, 5-ASA, thiopurines Corticosteroids, 5-ASA Thiopurines, 5-ASA Corticosteroids, thiopurines 5-ASA Corticosteroids Thiopurines No treatment
30 5 8 2 6 1100
p1
Non-responders
26 20/6 (19–54) 20/5/1 (1–37) (3–11) (1–3) 5/21 (0.99 0.10 0.98 0.25 >0.99 0.77 0.57 >0.99 0.20
1 4 5 1 6 1 2 2
Abbreviations: 5-ASA = 5-aminosalicylic acid; ADA = Adalimumab; CRP = C-reactive protein; IFX = Infliximab; TNF = Tumour necrosis factor; UC = Ulcerative colitis. 1 Mann Whitney U test. 2 Data from one patient missing.
Global mucosal and serum cytokine profile at baseline in responders and non-responders A multivariate analysis combining mucosal mRNA expression of cytokines (n = 7) and serum cytokine levels (n = 10) from patients whom provided both biopsy and serum (n = 28) at baseline was performed. The multivariate analysis (OPLS-DA) demonstrated a discrimination between responders and non-responders based on the 17 variables (R2 = 0.5). The majority of the responders were found in the two left quadrants, whereas the majority of non-responders were plotted in the two right quadrants (Figure 2A). The association between the cytokines with therapy outcome was demonstrated in the OPLS-DA column plots. Bars pointing in the same direction as response or non-response,
respectively, are positively associated with that particular therapy outcome. The length of the bars indicates the relative importance of the different cytokines to the separation between the two patient groups. Therapy failure was positively associated with high mucosal mRNA expression of IL6, IL-1b, IL-17A and IFN-g (Figure 2B). In contrast, none of the investigated cytokines were strongly associated with therapy response. The multivariate findings were corroborated with univariate analyses. Patients responding to the therapy showed lower mucosal mRNA expression of IL-6, IL-1b, IL-17A and IFN-g as compared with nonresponding patients (Figure 2C). Therapy responders also tended to have lower mRNA expression of TNF than non-responders (Figure 2C). No differences between the groups were detected for IL-13,
Table II. Changes in endoscopic Mayo score, calprotectin and CRP from baseline to week 14. Baseline Endoscopic Mayo score, median (range) Responders Non-responders Calprotectin (mg/g), median (range) Responders Non-responders CRP (mg/l), median (range) Responders Non-responders 1
Wilcoxon signed rank test.
2 (1–3) 2 (1–3) 1200 (180–18000) 1400 (270–15400) 5 (
ORIGINAL ARTICLE
Scand J Gastroenterol Downloaded from informahealthcare.com by East Carolina University on 04/20/15 For personal use only.
Global mucosal and serum cytokine profile in patients with ulcerative colitis undergoing anti-TNF therapy
RAHIL DAHLÉN1,2, MARIA K MAGNUSSON1,2, ANTAL BAJOR1, ANDERS LASSON3, KJELL-ARNE UNG4, HANS STRID1 & LENA ÖHMAN1,2 1
Department of Microbioloy and Immunology, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden, Department of Microbiology and Immunology, Sahlgrenska Academy, Gothenburg, Sweden, 3Södra Älvsborg Hospital, Borås, Sweden, and 4Skaraborgs Hospital, Skovde, Sweden
2
Abstract Background and objective. The knowledge of the effects of anti-tumour necrosis factor (TNF) treatment on the global cytokine profile in patients with ulcerative colitis (UC) is limited. A better understanding of these mechanisms could improve the ability to select patients that should undergo the therapy. Therefore, the aim was to determine the global mucosal and serum cytokine profile before and during induction therapy with anti-TNF in UC patients. Materials and methods. In total, mucosal biopsies (n = 28) and serum samples (n = 42) were collected from UC patients (total n = 48) before anti-TNF therapy. At week 14 response to the therapy was evaluated and again mucosal biopsies (n = 14) and serum samples (n = 42) were collected. Quantitative real-time PCR was used to determine mucosal cytokine mRNA expression and the MSD MULTI-ARRAY assay system platform was used for analysis of cytokines in serum. The global cytokine profile was evaluated by multivariate factor analysis. Results. At baseline, the global profile of mucosal cytokine mRNA expression and serum cytokines discriminated therapy responders from non-responders. Responders had lower mucosal mRNA expression of interleukin 1b (IL-1b), IL-17A, IL-6 and interferon g (IFN-g) than non-responders. Fourteen weeks after therapy start mucosal IL-1b and IL-6 were down-regulated in therapy responders but not in non-responders. At week 14, serum levels of IL-6 were decreased in therapy responders whereas IFN-g and IL-12p70 were increased in non-responders. Conclusions. Our data suggest that patients with a therapy failure have a more severe pro-inflammatory cytokine profile before start of anti-TNF treatment, which is less well suppressed by the treatment as compared to therapy responders.
Key Words: adalimumab, biomarker, cytokine, IBD, infliximab, prediction, therapy response, ulcerative colitis
Introduction Treatment of ulcerative colitis (UC) has been improved by the use of anti-TNF (tumor necrosis factor) therapy. However, at least 30% of the patients do not respond to anti-TNF therapy [1], and it is so far not possible to identify these individuals before treatment start [2]. Moreover, the immunological mechanisms of anti-TNF therapy linked to therapy response are not fully understood. Few studies have compared the immunopathology of therapy responders and non-responders of UC patients undergoing
anti-TNF therapy. Although still incompletely characterized, the mucosal cytokine profile in patients with active UC seems to involve a variety of cytokines, as demonstrated by increased mRNA expression of interferon gamma (IFN-g), interleukin (IL)-13, IL-17A, IL-1b, IL-6, TNF and IL-8 [3–5]. Additionally, UC patients with active disease have higher serum levels of IL-6, IL-8 and IL-10 as compared with healthy [6]. Thus, current data indicates that UC is associated with an increased expression of cytokines reflecting innate immune responses such as TNF, IL-1b, IL-6 and IL-8. Moreover, pathogenesis of
Correspondence: Lena Öhman, Department Microbiology and Immunology, University of Gothenburg, Box 435, 405 30 Gothenburg, Sweden. Tel: +46 31 786 6214. Fax: +46 31 786 6210. E-mail: [email protected]
(Received 2 February 2015; revised 11 March 2015; accepted 15 March 2015) ISSN 0036-5521 print/ISSN 1502-7708 online 2015 Informa Healthcare DOI: 10.3109/00365521.2015.1031167
Scand J Gastroenterol Downloaded from informahealthcare.com by East Carolina University on 04/20/15 For personal use only.
2
R. Dahlén et al.
UC also includes adaptive T helper (Th) cell responses of all kinds such as Th1 (IFN-g), Th2 (IL-4, IL-5, IL-13), Th17 (IL-17A) and T regulatory cells (IL-10, TGF-b) [7]. TNF is known to activate pro-inflammatory cytokines including IL-1b and IL-6 [8], and has wideranging effects in the body. The effects of anti-TNF therapy on immune cells and their effector cytokines are incompletely known, although a down-regulation of mucosal TNF and IFN-g mRNA expression in UC patients undergoing infliximab treatment has been shown [9,10]. It has been demonstrated that infliximab therapy influences the expression of several genes associated with adaptive immune responses differently in therapy responders and non-responders [11]. Additionally, infliximab has been reported to have the potential to down-regulate the gene expression of IFN-g, IL-6 and IL-2 [12], and the secretion of cytokines such as IFN-g, IL-17A and IL-13 [13] in vitro of blood T cells isolated from UC patients. Moreover, our group has demonstrated that infliximab treatment decreases serum levels of IL-5, IL-8 and TNF in UC patients responding to the therapy, but not in therapy failures, 2 weeks after treatment start [14]. Potentially, the patient’s global cytokine profile before start of treatment could predict the response to anti-TNF therapy. Interestingly, it has been demonstrated that high pre-treatment mucosal TNF expression was associated with insufficient response to anti-TNF therapy in UC patients. Thus, the infliximab treatment was less effective in patients with high mucosal expression of TNF mRNA, indicating a more severe disease in therapy non-responders than responders [15]. Still, the knowledge of the effects of anti-TNF treatment on cytokines associated with innate as well as adaptive T-cell immune responses in UC patients is limited and needs to be expanded to better understand the mechanisms of the therapy and
to guide the decision on which patients that are suited to undergo the therapy or perhaps benefit from an escalated dosing. In the present study of UC patients, we hypothesized that the global cytokine profile, representative of innate as well as adaptive T cell responses, in the intestinal mucosa and serum before start of treatment was different in responders as compared with nonresponders. We also postulated that anti-TNF therapy affected the cytokine profile differently in responders and non-responders. Therefore, the aim was to determine the global cytokine profile, including mucosal tissue and serum, in UC patients before start of anti-TNF therapy and at evaluation of therapy response at week 14, in therapy responders and nonresponders, respectively. Methods Patient cohort and sample collection Patients with UC (n = 48) were consecutively included in the study at the outpatient clinics at Sahlgrenska University Hospital, Gothenburg, Kärnsjukhuset in Skövde, and Södra Älvsborg Hospital, Borås, Sweden. All patients had an active disease at the time of inclusion, had failed increased dosing of 5-aminosalicylic acid and/or thiopurines and were either corticosteroidrefractory or dependent. Mucosal biopsies from sigmoid colon obtained during colonoscopy and serum samples were collected at baseline, i.e. before start of anti-TNF therapy and at week 14. From 6 of the patients biopsies alone were obtained, from 22 of the patients both serum and biopsies were obtained whereas only serum samples were obtained from the remaining 20 patients (in total 48). The clinical response was evaluated at week 14, and the patients were classified as therapy responders or non-responders, respectively (Figure 1). Primary therapy response
Mucosal biopsies +/- serum: 28 Serum alone: 20 Total: 48
14 28 42* Responders (n = 26)
Anti-TNF treatment Baseline disease assessment
Week 14 Therapy evaluation
Non-responders (n = 22)
*6 patients in early termination = no samples Figure 1. Study design and patient treatment response. Patients with ulcerative colitis (n = 48) were included in the study before start of antiTNF therapy. Disease activity assessment was performed at inclusion, termed baseline, and at week 14, or at the termination of treatment if it ended during therapy induction (six patients experienced early termination). Serum samples and biopsies from sigmoid colon, or serum alone were obtained at start of anti-TNF therapy and at week 14. Patients were defined as responders (n = 26) or non-responders (n = 22) at week 14. Abbreviation: TNF = Tumour necrosis factor.
Cytokine profiles in ulcerative colitis
Scand J Gastroenterol Downloaded from informahealthcare.com by East Carolina University on 04/20/15 For personal use only.
was defined as a decrease in Mayo score of ‡3 points compared with baseline. Patients not achieving this decrease in the Mayo score were considered as nonresponders. If the criteria for a full Mayo score evaluation was not fulfilled (n = 4), therapy response was defined by one or several of the following parameters: C-reactive protein (CRP), fecal calprotectin, and a partial Mayo score. The study was performed after receiving written informed consent from all subjects and the protocol was approved by the Regional Ethical Review Board at the University of Gothenburg. PCR analysis of mucosal biopsies Mucosal biopsies from sigmoid colon were collected for RT-PCR analysis. Biopsies were obtained from 28 patients at baseline (therapy responders n = 16 and non-responders n = 12). At week 14, biopsies from the evaluation visit were obtained from 14 patients (therapy responders n = 7 and non-responders n = 7). Total mRNA from mucosal biopsies was extracted using Nucleospin DNA, RNA, and protein purification kit (Macherey-Nagel, Düren, Germany) and cDNA was prepared from 1 mg RNA using the QuantiTect Reverse Transcription kit (Qiagen, Hilden, Germany) according to the manufacturers’ protocols. cDNA was then used for real-time RT-PCR using Taqman Universal PCR Master Mix (Applied Biosystems, Foster City, CA) and Taqman Gene Expression assays (Applied Biosystems). Expression of TNF, IL-6, IL-1b, IFN-g, IL-17A, IL-13 and FoxP3 was determined. The results were normalized to the expression level of GAPDH and expressed as 2-(Target-GAPDH). Cytokine analysis and trough level analysis of serum samples Serum samples were collected for analysis of cytokine content. Serum was obtained from 42 patients at baseline and at week 14 (therapy responders n = 26 and non-responders n = 16). The cytokines IL-1b, IL-4, IL-6, IL-8, IL-10, IL-12p70, IL-13, IL-17A, TNF and IFN-g were analyzed with MSD MULTI-ARRAY Assay System platform (MSD SCALE DISCOVERY, Rockville, USA). Levels of infliximab in serum samples was determined by the TNF a-Blocker monitoring, infliximab drug level (e.g. REMICADE), ELISA (Immundiagnostik AG, Bensheim, Germany). All assays were performed according to the manufacturers’ instructions. Statistical analysis Multivariate factor analysis (SIMCA-P+ software; Umetrics, Umeå, Sweden) was used to examine the
3
relation between therapy outcome and the examined cytokines. Orthogonal partial least squares discriminant analyses (OPLS-DA) were implemented to examine whether therapy outcome could be discriminated based on the totality of X-variables. The quality of the OPLS-DA was based on the parameter R2, that is, the goodness of the fit of the model (R2 = 1 represents best possible fit, whereas R2 ‡0.05 represents fairly good discrimination). In the OPLS-DA loading column plots, the importance of each X-variable (cytokine) to Y (therapy outcome) is represented by column bars. The variables positioned in the same direction as the bar representing therapy response are positively associated, whereas variables in the opposite direction are inversely related to therapy response. The larger the bar and the smaller the error bar, the stronger and more reliable is the contribution to the model. Univariate statistical analyses were performed with the GraphPad Prism 6.0 (GraphPad Software, La Jolla, CA, USA). The Wilcoxon signed rank test was used to evaluate differences between two sets of paired samples, and the Mann Whitney U test was performed to evaluate differences between two groups. p-Values < 0.05 were considered as statistically significant. Data given in the text and Tables I and II are demonstrated as median (range) and the horizontal lines in the figures show the median. Data given in Supplementary Tables I and II are demonstrated as median (25–75 percentiles).
Results Demographics and clinical parameters in therapy responders and non-responders In total, 48 patients with UC were included in the study, whereof 44 patients were treated with infliximab and 4 patients were treated with adalimumab (Table I). At week 14, 26 patients were classified as therapy responders and 22 patients as nonresponders. Therapy responders and non-responders showed no major differences in gender distribution, age, disease duration, Mayo score, CRP or fecal calprotectin before start of anti-TNF treatment, hereafter termed baseline (Table I). All the patients had active disease at baseline as determined by Mayo score. At week 14, treatment responders showed a reduction in endoscopic Mayo score, CRP and calprotectin whereas no changes were detected for non-responders (Table II). Trough levels at week 14 in patients treated with infliximab tended to be higher in treatment responders vs. non-responders (7700 ng/ml (1000–42200) vs. 3800 ng/ml (0– 14400), p = 0.07).
4
R. Dahlén et al.
Table I. Demographics of anti-TNF therapy naïve UC patients before anti-TNF therapy start.
Scand J Gastroenterol Downloaded from informahealthcare.com by East Carolina University on 04/20/15 For personal use only.
Responders Total number of patients Male/female Age, years (range) Smoking habit (never/ex smoker/active) Disease duration, years (range) Mayo score, median (range) Endoscopic Mayo score, median (range) Colonic area involved, left side/extensive CRP (mg/l), median (range) Calprotectin (mg/g), median (range) Anti-TNF treatment (IFX/ADA) Other treatments than anti-TNF: Corticosteroids, 5-ASA, thiopurines Corticosteroids, 5-ASA Thiopurines, 5-ASA Corticosteroids, thiopurines 5-ASA Corticosteroids Thiopurines No treatment
30 5 8 2 6 1100
p1
Non-responders
26 20/6 (19–54) 20/5/1 (1–37) (3–11) (1–3) 5/21 (0.99 0.10 0.98 0.25 >0.99 0.77 0.57 >0.99 0.20
1 4 5 1 6 1 2 2
Abbreviations: 5-ASA = 5-aminosalicylic acid; ADA = Adalimumab; CRP = C-reactive protein; IFX = Infliximab; TNF = Tumour necrosis factor; UC = Ulcerative colitis. 1 Mann Whitney U test. 2 Data from one patient missing.
Global mucosal and serum cytokine profile at baseline in responders and non-responders A multivariate analysis combining mucosal mRNA expression of cytokines (n = 7) and serum cytokine levels (n = 10) from patients whom provided both biopsy and serum (n = 28) at baseline was performed. The multivariate analysis (OPLS-DA) demonstrated a discrimination between responders and non-responders based on the 17 variables (R2 = 0.5). The majority of the responders were found in the two left quadrants, whereas the majority of non-responders were plotted in the two right quadrants (Figure 2A). The association between the cytokines with therapy outcome was demonstrated in the OPLS-DA column plots. Bars pointing in the same direction as response or non-response,
respectively, are positively associated with that particular therapy outcome. The length of the bars indicates the relative importance of the different cytokines to the separation between the two patient groups. Therapy failure was positively associated with high mucosal mRNA expression of IL6, IL-1b, IL-17A and IFN-g (Figure 2B). In contrast, none of the investigated cytokines were strongly associated with therapy response. The multivariate findings were corroborated with univariate analyses. Patients responding to the therapy showed lower mucosal mRNA expression of IL-6, IL-1b, IL-17A and IFN-g as compared with nonresponding patients (Figure 2C). Therapy responders also tended to have lower mRNA expression of TNF than non-responders (Figure 2C). No differences between the groups were detected for IL-13,
Table II. Changes in endoscopic Mayo score, calprotectin and CRP from baseline to week 14. Baseline Endoscopic Mayo score, median (range) Responders Non-responders Calprotectin (mg/g), median (range) Responders Non-responders CRP (mg/l), median (range) Responders Non-responders 1
Wilcoxon signed rank test.
2 (1–3) 2 (1–3) 1200 (180–18000) 1400 (270–15400) 5 (
