Cancer Immunol Immunother (2015) 64:41–50 DOI 10.1007/s00262-014-1615-y
ORIGINAL ARTICLE
Evaluation of complement proteins as screening markers for colorectal cancer Line Storm · Ib J. Christensen · Jens C. Jensenius · Hans J. Nielsen · Steffen Thiel · the Danish Study Group on Early Detection of Colorectal Cancer
Received: 10 April 2014 / Accepted: 17 September 2014 / Published online: 27 September 2014 © Springer-Verlag Berlin Heidelberg 2014
Abstract Background Colorectal cancer (CRC) is a leading cause of cancer death worldwide. Lack of symptoms results in late detection and increased mortality. Inflammation, including complement activation, plays an important role in tumorigenesis. Experimental design The concentrations of nine proteins of the lectin pathway of the complement system were determined using time-resolved immunofluorometric assays. The first cohort investigated comprised a matched case– control study of 95 patients with CRC, 48 patients with adenomas and 48 individuals without neoplastic findings. Based on the results, Collectin-liver 1 (CL-L1), M-ficolin and MAp44 were determined as the most promising biomarkers and were subsequently evaluated in a case–control study of 99 CRC patients, 196 patients with adenomas and 696 individuals without neoplastic bowel lesions. Results Using logistic regression, we found that CL-L1, M-ficolin and MAp44 levels could significantly distinguish between patients with CRC, patients with adenomas and individuals without neoplastic bowel lesions. Higher levels of CL-L1 or MAp44 were associated with lower odds of CRC (OR 0.42 (0.25–0.70) p = 0.0003 and OR 0.39 L. Storm · J. C. Jensenius · S. Thiel (*) Department of Biomedicine, Faculty of Health Sciences, Aarhus University, Wilhelm Meyers Allé 4, Aarhus C, Denmark e-mail: [email protected] I. J. Christensen Finsen Laboratory, Rigshospitalet and Biotech Research and Innovation Center (BRIC), University of Copenhagen, Copenhagen, Denmark H. J. Nielsen Department of Surgical Gastroenterology, Hvidovre Hospital, University of Copenhagen, Hvidovre, Denmark
(0.23–0.65) p = 0.0003, respectively), whereas higher levels of M-ficolin were associated with higher odds of CRC compared to individuals without CRC (OR 1.94 (1.46– 2.59) p 15 as compared to a collection of previously measured values. If the concentration of the samples exceeded the highest value of the standard curve, the assays were repeated at higher dilution. The determinations were performed at Department of Biomedicine, Aarhus University, Denmark. Statistics Associations between the variables were evaluated by the Spearman rank correlation. CRC patients with recurrence were compared to those without using a two-sample t test with equal variances. Wilcoxon rank sum test was used to compare basic characteristics between the groups in study 2. Logistic regression analysis was used in both studies modeling the probability for CRC. Conditional logistic regression analysis was used for study 1. Marker levels were log-transformed using log base 2, implying that odds ratios (OR) are for twofold differences in marker levels. Sensitivities and specificities were calculated based on the multivariate logistic regression for study 2. Goodness of fit was assessed using the Hosmer–Lemeshow test. P values
13
44
Cancer Immunol Immunother (2015) 64:41–50
less than 0.05 were considered significant. All calculations have been done with SAS (v9.2, SAS Institute, Cary, N.C., USA).
Results Study 1 The concentration of nine different proteins of the lectin pathway in patients with CRC, patients with adenoma and healthy individuals was measured. Using logistic regression, we found that the concentrations of several of the proteins differed significantly between CRC and non-CRC, i.e., adenomas and healthy individuals. The distribution of the concentrations of CL-L1, M-ficolin and MAp44 for CRC patients, patients with adenomas and healthy individuals is shown in Fig. 1. Higher levels of CL-L1 were associated with significantly lower odds of CRC when comparing CRC to both non-CRC (OR 0.22 (0.09–0.57) p = 0.002), healthy individuals (OR 0.23 (0.08–0.70) p = 0.009) and adenomas (OR 0.25 (0.09–0.75) p = 0.013) (Table 1). Higher levels of MAp44 were associated with significantly lower odds of CRC with respect to both nonCRC (OR 0.24 (0.10–0.57) p = 0.001) and adenomas (OR 0.14 (0.05–0.39) p = 0.0002), while higher levels of MAp44 were associated with higher odds of adenomas compared to healthy individuals (OR 5.29 (1.42–19.64) p = 0.013). On the other hand, higher levels of M-ficolin were significantly associated with both higher odds of CRC (OR 1.64 (1.05–2.57) p = 0.03) and higher odds of adenomas when compared to healthy individuals (OR 1.64 (1.05–2.57) p = 0.03). Logistic regression analyses of the six other lectin pathway proteins, H-ficolin, MBL, MAp19, MASP-1, MASP-2 and MASP-3, were also conducted in study 1 and are presented in Table 1. Only marginal significance for any comparisons was seen for H-ficolin, MAp19 and MASP-1 (Table 1). We found no significant difference of the tested protein concentrations between CRC patients with or without recurrence (data not shown).
Fig. 1 The concentration of CL-L1 (a), M-ficolin (b) and MAp44 (c) in serum of individuals in study 1. Scatter plots of the concentrations divided into separate groups of CRC patients (n = 95), patients with adenomas (n = 48) and healthy individuals (n = 48). Lines indicate mean and interquartile range (25th–75th percentiles). Each dot represents an individual
Study 2 The characteristics of the patients in study 2 are shown in Table 2. A highly significant age difference (p
ORIGINAL ARTICLE
Evaluation of complement proteins as screening markers for colorectal cancer Line Storm · Ib J. Christensen · Jens C. Jensenius · Hans J. Nielsen · Steffen Thiel · the Danish Study Group on Early Detection of Colorectal Cancer
Received: 10 April 2014 / Accepted: 17 September 2014 / Published online: 27 September 2014 © Springer-Verlag Berlin Heidelberg 2014
Abstract Background Colorectal cancer (CRC) is a leading cause of cancer death worldwide. Lack of symptoms results in late detection and increased mortality. Inflammation, including complement activation, plays an important role in tumorigenesis. Experimental design The concentrations of nine proteins of the lectin pathway of the complement system were determined using time-resolved immunofluorometric assays. The first cohort investigated comprised a matched case– control study of 95 patients with CRC, 48 patients with adenomas and 48 individuals without neoplastic findings. Based on the results, Collectin-liver 1 (CL-L1), M-ficolin and MAp44 were determined as the most promising biomarkers and were subsequently evaluated in a case–control study of 99 CRC patients, 196 patients with adenomas and 696 individuals without neoplastic bowel lesions. Results Using logistic regression, we found that CL-L1, M-ficolin and MAp44 levels could significantly distinguish between patients with CRC, patients with adenomas and individuals without neoplastic bowel lesions. Higher levels of CL-L1 or MAp44 were associated with lower odds of CRC (OR 0.42 (0.25–0.70) p = 0.0003 and OR 0.39 L. Storm · J. C. Jensenius · S. Thiel (*) Department of Biomedicine, Faculty of Health Sciences, Aarhus University, Wilhelm Meyers Allé 4, Aarhus C, Denmark e-mail: [email protected] I. J. Christensen Finsen Laboratory, Rigshospitalet and Biotech Research and Innovation Center (BRIC), University of Copenhagen, Copenhagen, Denmark H. J. Nielsen Department of Surgical Gastroenterology, Hvidovre Hospital, University of Copenhagen, Hvidovre, Denmark
(0.23–0.65) p = 0.0003, respectively), whereas higher levels of M-ficolin were associated with higher odds of CRC compared to individuals without CRC (OR 1.94 (1.46– 2.59) p 15 as compared to a collection of previously measured values. If the concentration of the samples exceeded the highest value of the standard curve, the assays were repeated at higher dilution. The determinations were performed at Department of Biomedicine, Aarhus University, Denmark. Statistics Associations between the variables were evaluated by the Spearman rank correlation. CRC patients with recurrence were compared to those without using a two-sample t test with equal variances. Wilcoxon rank sum test was used to compare basic characteristics between the groups in study 2. Logistic regression analysis was used in both studies modeling the probability for CRC. Conditional logistic regression analysis was used for study 1. Marker levels were log-transformed using log base 2, implying that odds ratios (OR) are for twofold differences in marker levels. Sensitivities and specificities were calculated based on the multivariate logistic regression for study 2. Goodness of fit was assessed using the Hosmer–Lemeshow test. P values
13
44
Cancer Immunol Immunother (2015) 64:41–50
less than 0.05 were considered significant. All calculations have been done with SAS (v9.2, SAS Institute, Cary, N.C., USA).
Results Study 1 The concentration of nine different proteins of the lectin pathway in patients with CRC, patients with adenoma and healthy individuals was measured. Using logistic regression, we found that the concentrations of several of the proteins differed significantly between CRC and non-CRC, i.e., adenomas and healthy individuals. The distribution of the concentrations of CL-L1, M-ficolin and MAp44 for CRC patients, patients with adenomas and healthy individuals is shown in Fig. 1. Higher levels of CL-L1 were associated with significantly lower odds of CRC when comparing CRC to both non-CRC (OR 0.22 (0.09–0.57) p = 0.002), healthy individuals (OR 0.23 (0.08–0.70) p = 0.009) and adenomas (OR 0.25 (0.09–0.75) p = 0.013) (Table 1). Higher levels of MAp44 were associated with significantly lower odds of CRC with respect to both nonCRC (OR 0.24 (0.10–0.57) p = 0.001) and adenomas (OR 0.14 (0.05–0.39) p = 0.0002), while higher levels of MAp44 were associated with higher odds of adenomas compared to healthy individuals (OR 5.29 (1.42–19.64) p = 0.013). On the other hand, higher levels of M-ficolin were significantly associated with both higher odds of CRC (OR 1.64 (1.05–2.57) p = 0.03) and higher odds of adenomas when compared to healthy individuals (OR 1.64 (1.05–2.57) p = 0.03). Logistic regression analyses of the six other lectin pathway proteins, H-ficolin, MBL, MAp19, MASP-1, MASP-2 and MASP-3, were also conducted in study 1 and are presented in Table 1. Only marginal significance for any comparisons was seen for H-ficolin, MAp19 and MASP-1 (Table 1). We found no significant difference of the tested protein concentrations between CRC patients with or without recurrence (data not shown).
Fig. 1 The concentration of CL-L1 (a), M-ficolin (b) and MAp44 (c) in serum of individuals in study 1. Scatter plots of the concentrations divided into separate groups of CRC patients (n = 95), patients with adenomas (n = 48) and healthy individuals (n = 48). Lines indicate mean and interquartile range (25th–75th percentiles). Each dot represents an individual
Study 2 The characteristics of the patients in study 2 are shown in Table 2. A highly significant age difference (p
