MIF and AIMP1 in the early detection of colorectal
carcinoma in PSC-IBD patients.
Exploring possibilities for the use of cytokines MIF and AIMP1 as biomarkers to
detect colorectal carcinoma in patients with primary sclerosing cholangitis and
concomitant inflammatory bowel disease.
Scientific report
Date of publication: 1/07/2017 Name: Mariska Emily van Kan Student number: 10753249
Education: Bèta-gamma - Biomedische Wetenschappen (Track: Patiëntgericht onderzoek) Educational institution: Universiteit van Amsterdam
Research institution: Tytgat Institute for liver and intestinal research Daily supervisor: Manon de Krijger
2
Preface
Every medical student of the University of Amsterdam will follow the same Anatomy and Physiology course at the Academical Medical Centre. This course is divided in several subjects such as respiration, electrophysiology and heart and circulation. The intestinal tract was only discussed half a lecture before moving on to the next subject. I asked why the organ that is - in my opinion- one, if not thé, most important role player in human wellbeing was subordinated so quickly. The tutor answered: “I do not know much about the intestinal tract. Whenever you’ll ask a group of physicians to give an anatomy lecture, they will queue up for hearts, brains and lungs. But talking about the intestines is very unpopular. The few physicians who do have substantial knowledge of matters are in general too busy to teach and inspire new students about the intestines.”
This matter gained my interest and a little research brought me very fast via Dr. C.I.J. Ponsioen to the Tytgat Institute for Liver and Intestinal research (TI). I hoped to have at least the smallest little contribution to intestinal research during my internship. TI proved to be the best place for this.
I was given a starting point for this research, namely PSC-IBD CRC, with availability to an exceptional database, methods and materials that I could only dream of. This thesis proposes to supplement in Kirsten Boonstra’s research and to provide input in the direction of further investigation of this matter. Unfortunately, I have not been able to take full advantage of the opportunities due to lack of time and inevitable research issues. But after all it was an enriching experience in which I could take a valuable look behind the scenes of biomedical research.
I would like to thank Manon de Krijger for her mentoring during my internship and for guiding my research and thesis into the right direction. Also my fellow colleagues of the Wouter group who were always very willing to help, guide and advice deserve a special thanks. Lastly I am very grateful for this opportunity provided by Dr. C.I.J. Ponsioen and Prof. Dr. W.J. de Jonge.
Mariska van Kan Amsterdam, 1/7/2017
3
Abstract
The risk of developing colorectal carcinoma (CRC) for patients with primary sclerosing cholangitis with concomitant inflammatory bowel disease (PSC-IBD) is 50% after 25 years since disease onset. Currently, early adenomas can only be detected using colonoscopy, and there are no sufficient strategies yet to determine on beforehand which of the PSC-IBD patients will develop colorectal carcinoma. In order to take away the healthcare and patient disease burden, biomarkers for early detection of CRC are needed. In a PCR array conducted at the Tytgat Institute (TI) by Kirsten Boonstra in 2013, significant upregulation of MIF and AIMP1 was found. Therefore in this research the possibilities for macrophage migration inhibitory factor (MIF) and Aminoacyl TRNA Synthetase Complex Interacting Multifunctional Protein 1 (AIMP1) as biomarkers in CRC for PSC-IBD patients are explored.It is hypothesized that MIF expression is upregulated in colon tissue of PSC-IBD patients compared to IBD-patients and healthy controls. Furthermore, it is hypothesized that PSC-IBD patients who will develop CRC have a significant MIF upregulation in non-cancerous tissue compared to non-CRC PSC-IBD patients. Due to time considerations only the first hypothesis will be tested in this research.
In order to test whether MIF and AIMP1 could be efficient biomarkers for early CRC detection, their operating features should be defined. First, a literature study was conducted to determine the biological functions of these cytokines. Based on found literature expectations towards the immunohistochemical stainings conducted in this research were formulated. Thereafter IHC protocol was refined for several conditions. Eventually the conducted stainings were analysed in various ways.
During the immunohistochemical stainings formalin fixed paraffin embedded (FFPE) colon biopsies from a previously prospectively conducted cohort were stained. The tissues used correspond to the samples used in the PCR array conducted by Boonstra (2013). Eight PSC-IBD patients, 8 IBD patients and 4 healthy controls (HCs) were included. The same patients were used as for the PCR array conducted by Kirsten Boonstra in 2013.
The stainings were qualified manually and quantified digitally. No significant results were found. Due to lack of significance, specificity and reproducibility this research could not be validated to conduct hard conclusions. Therefore, it cannot be said whether MIF is upregulated or not on protein level in PSC-IBD patients compared to PSC-IBD patients and HCs. In further research a more robust IHC protocol should be developed with other antibodies. Based on the found literature not only MIF and AIMP1, but also IL22 and mast cells could be interesting starting points in the search for biomarkers in early detection of CRC.
4 CONTENT Preface 2 Abstract 3 Abbreviations 6 1. Introduction 6
1.1 PSC as independent risk factor for CRC 7
1.2 Biomarker selection 7
1.3 Research design 8
2. Biomarkers in PSC-IBD CRC 8
2.1 MIF 8
2.2 AIMP1 9
3. Methods and materials 9
3.1 TIssue Immunohistochemistry 9
3.2 Immunohistochemistry 9
3.3 Pathological staining evaluation 9
3.4 Digital staining evaluation 10
3.5 Statistical analysis 12 4. Results 12 4.1 Dataset 12 4.2 Morphology 13 4.3 Staining intensity 15 4.4 Digital evaluation 15 5. Conclusion 17 5.1 Biomarker potential 17 5.2 Staining analysation 17 6. Discussion 17 6.1 Validity 17 6.2 Expectations 17 6.3 Digital methods 18 6.4 Research design 18 6.5 Insights 19 6.6 Research recommendations 19 7. Literature 20
APPENDIX A: PCR array results Kirsten Boonstra 24
5
APPENDIX C: Results 33
APPENDIX D: Illustration and explanation insights 34
7
Abbreviations
AB Antibody
AIMP1 Aminoacyl TRNA synthetase complex Interacting Multifunctional Protein 1 (p54)
BSA Bovine Serum Albumin
CA Colon ascendens
CD Crohn’s disease
CRC Colorectal carcinoma
FFPE Formalin Fixed Paraffin Embedded
HC Healthy controls
IBD Inflammatory Bowel Disease
IHC Immunohistochemistry
(M)MIF Macrophage Migration Inhibitory Factor
MC Mast Cell
OD Optical Density
PBS Phosphate Buffered Saline PBT PBS, BSA, Triton solution PSC Primary Sclerosing Cholangitis SOP Standard Operating Procedure UC Ulcerative Colitis
1. Introduction
Primary sclerosing cholangitis (PSC) is a rare, cholestatic liver disease that is occurring at an increasing extent (Torres, 2011). It involves a chronic fibro inflammatory syndrome of the biliary tract. Also, PSC is often accompanied by inflammatory bowel disease (IBD) expressed by progressive, more right sided inflammation of the colon (Razumilava, 2011; Torres, 2011). Additionally, PSC-IBD patients have a significantly increased risk for developing colorectal cancer (CRC), which is the second most common cause of cancer death in the Western world (He et al., 2008). Factors such as lethality, the lack of efficient therapies, the high amount of complications and the fact that PSC patients are relatively young make the disease a substantial healthcare burden (Shetty, 1999; Torres, 2011). However, despite a large amount of literature and experimental research, the mechanisms by which patients with PSC-IBD develop CRC remain unidentified.
1.1 PSC as independent risk factor for CRC
In former research, inflammatory CRC development is considered subsequently different and more destructive from sporadic CRC (Vagefi & Longo, 2005). IBD patients are in general at higher risk of developing inflammatory CRC (Soetikno, 2002), which again suggests a relationship between
8 inflammation and CRC (Triantafillidis, 2009). The mortality rate in inflammatory CRC shows to be significantly higher than in sporadic CRC, accounting for 15% of all IBD deaths (Vagefi & Longo, 2005).
In the normal population there is a 2%-7.5% IBD occurrence contrasting with the PSC patients of which 80% suffer from IBD (Torres, 2011). Despite having mild inflammation compared to IBD patients, PSC-IBD patients have a significantly increased risk for CRC compared to IBD patients. After 25 years since onset of the disease, PSC-IBD patients will have a compelling 50% risk of CRC, which is 9 times higher than for IBD patients (Maaser et al., 2002). One peculiar difference between the two patient groups is that PSC-IBD patients are predominated with 65% of colonic neoplasia in the colon ascendens (Torres, 2011; Razumilava, 2011). Concluding can be stated that the presence of PSC is an independent risk factor for CRC development (Torres, 2011).
To limit the mortality of CRC, PSC-patients conduct a more intense colonoscopic surveillance examination than usually recommended, which in can result in complications (Soetikno, 2002; Razumilava, 2011). Since patients with PSC-IBD are already disadvantaged with a sensitive system, it is important to develop an cancer surveillance strategy that is as least stressful as possible. In order to do this, insight in the mechanisms of colitis associated colorectal cancer occurrence and the increased risk of CRC in PSC-IBD patients is urgently required. Clarification of the molecular mechanisms underlying the maturation and metastasis of PSC-IBD CRC is of great interest for developing sufficient medical therapies. One strategy could be early detection of CRC using biomarkers. Detecting these specific markers in colon biopsies could classify the patients based on their risk of developing CRC. In this way personalized therapies could be formed, relieving both patient and healthcare.
1.2 Biomarker selection
From, 2008 -, 2011, the EpiPSCPBC cohort was formed of 695 patients with PSC, of which 483 had concomitant IBD. Of twenty four of these patients who developed CRC during the course of the surveillance biopsy tissue over the years of the surveillance is available. A multiplex PCR (Qiagen) of colonic tissue of PSC-IBD patients versus IBD patients and healthy controls (HC) from this cohort was conducted by Boonstra (2013, unpublished). The PCR array showed significantly higher RNA expression of several genes, among which macrophage migration inhibitory factor (MIF) and aminoacyl tRNA synthetase complex interacting multifunctional protein 1 (AIMP1) were most abundant (App. A).
The pro-inflammatory cytokine MIF is known for its upregulated expression in all sorts of cancers (Nihihira, 2009). MIF expression is found to be significantly higher in cancerous colorectal mucosa compared to healthy mucosa (He et al., 2008). Also its intestinal tumorigenesis promoting properties (Wilson, 2005) are proven, making it an interesting target for CRC therapy as a chemoprevention target (Torres, 2011) and a biomarker in CRC diagnosis (He et al., 2008).
The cytokine AIMP1 is specifically induced by apoptosis, and is known for its angiogenesis and immune response controlling properties and anti-tumor activities (Lee et al., 2005). Less is known about AIMP1 as an anti-tumor agent but its upregulation in PSC makes it an interesting research target.
1.3 Research design
In order to limit mortality due to CRC among PSC-IBD patients, biomarkers for early detection of inflammatory CRC are wanted. Therefore, in this research the potential of MIF and AIMP1 as biomarkers in the detection of early CRC in PSC-IBD were explored. First, the biological functions of MIF as a pro-inflammatory cytokine were investigated via literature study. Based on this, predictions were made for the immunohistochemistry morphology. The same setup was carried out for the anti-tumorigenesis cytokine AIMP1. Based on Boonstra’s PCR array, an upregulation of MIF and AIMP1 on protein level in PSC-IBD compared to IBD tissue and HC tissue was hypothesized. To test the hypothesis IHC stainings were conducted on a selection of samples from the aforementioned PSColon PCR array.
9 The IHC results were analysed in three approaches. First a qualitative analysis of the stainings was carried out by a pathologist (Verheij, J., personal communication, Amsterdam Medical Centre, may 2017) Secondly an automated digital IHC image analysis was conducted using ImageJ with IHC specific plug-ins. Lastly the generated data were aalysed statistically using GraphPad.
2. Biomarkers in PSC-IBD CRC
To determine the relevance of MIF and AIMP1 expression in the development of CRC in PSC-IBD, literature research into the biological functions of the proteins is needed. In the next section the biological functions of the two proteins are explained and expectations towards protein expression in colon are given.
2.1 MIF
2.1.1 Biological properties
MIF has several functions as a biological, pro-inflammatory cytokine (Hoi et al., 2007). It is known to be a key role player in regulating the inflammatory response on various levels, differing from altering gene expression to immune-cell activation and other damage mechanisms. Firstly, one of these mechanisms is the inhibition of anti-inflammatory inhibitor glucocorticoids. This overriding anti-inflammatory property makes MIF an inflammation induces, which is a rather singular property.
Secondly, upon MIF expression macrophages are withheld from migration. Through this mechanism epithelial cells can use MIF in first order immune reactions in the lamina propria (Maaser et
al., 2002). Due to MIFs interference with inflammation it is a big player in immune inflammatory diseases
pathogenesis like IBD (Hoi et al., 2007).
Thirdly, MIF is a noted cancer inducer and is linked inadmissible to cell proliferation, tumor differentiation, angiogenesis, tumor progression and cancer metastasis. (He et al., 2008, Bernhagen 1998). These properties together with its ubiquitous expression in colorectal epithelium, makes it an accountable role-player for early colorectal carcinogenesis and sporadic colorectal adenomas (Wilson, 2005).
2.1.2 Expected protein expression
MIF is a cytoplasmic cytokine expressed in response to stress in practically all pituitary and peripherally cells in human tissues (Bernhagen 1998; Nobre et al., 2016). These tissues include the entire gastrointestinal tract where expression can especially be found in the epithelium (Wilson, 2005; Nishihira & Mitsuyama, 2009). MIF is expressed in the intracellular environment in both immune and non-immune cells like eosinophils, neutrophils, granulocytes, monocytes, macrophages, B and T-lymphocytes, endocrine cells, endothelial cells, epithelial cells and neuronal cells. (Nobre et al., 2016; Maaser et al., 2002; Nishisira & Mitsuyama, 2009). Especially corticotropic pituitary cells (Bernhagen 1998), macrophages (Hoi et al., 2007) human intestinal cells and activated T-cells (Maaser et al., 2002) are a major source for MIF. So far little is known about the induction of this production and expression regulation in these cells.
In epithelial cells MIF is constitutively present in crypt and villus of colonic epithelial cells (Maaser et al., 2002). MIF is expressed with an increasing gradient from basolateral to apical that is seen throughout the epithelium. MIF expression should also be seen consistently throughout the basolateral region of epithelial cells (Wilson, 2005). Positive MIF expression in the lamina propria points towards active T-cells or macrophages
10 is upregulated in colonic tissue during the development of inflammatory CRC. This makes MIF a potential biomarker for early detection of CRC in PSC-IBD.
2.2 AIMP1
2.2.1 Biological properties
AIMP1 is, in contrary to MIF, described as an anti-tumorigenesis cytokine, responsible for endothelial cell apoptosis and angiogenetic activity inhibition. Based on earlier studies Lee et al. (2006) concluded that AIMP1 indeed has anti-colontumorous activity. Therefore it could be useful in targeted therapy as primary antitumor agent or as a supplement to primary cytotoxic anticancer drugs. Due to AIMP1s uknown expression mechanisms in CRC, there seems to be a paradox in how its functioning is describes. However, besides AIMP1s anti-angiogenic activities there are also immune-stimulating activities of AIMP1 described creating a link to inflammatory CRC.
2.2.2 Expected protein expression
Kim et al. (2011) conducted IHC on CRC tissue and found a significant difference of AIMP1 expression in cancerous compared to normal tissue. They stated that a loss of AIMP1 was a contributor to CRC development. On the contrary, Boonstra et al. also found a significant difference between PSC-IBD, IBD and HC tissues with a significant higher level of AIMP1 in PSC tissue. In the light of these opposite statements, it is interesting to investigate whether and in what direction AIMP1 plays a contributing role in the development of CRC in PSC-IBD. The expectation in this research is based on Boonstra’s (2013) findings that there was an upregulation of RNA in PSC-IBD patients. Because the same biopsies were used, a correlated upregulation on protein level was expected.
To determine the exact role of MIF and AIMP1 in the onset of PSC-IBD CRC, the expression of these genes on protein level should be researched. In this experiment this is done using IHC staining, as will be described in the next section.
3. Materials and Methods
The practical experiments are conducted at the Tytgat Institute for liver and intestinal research. The experiments are conducted from 1/4/2017 until 1/6/2017 under surveillance of Manon de Krijger, under supervision of Dr. C.I.J. Ponsioen. All experiment protocols and results can be found in laboratory notebook no. 679 retained by the Tytgat Institute. The used tissue samples were from the PSColon database and are selected to be the same samples as Kirsten Boonstra used in her PCR array (App. A).
3.1 Tissue
Of all patients, tissue was available of the colon ascendens, colon transversum and recto sigmoid. Because of the predominance of occurrence of CRC in the right sided colon, colon ascendens biopsies were used (Claessens, 2010).
3.2 Immunohistochemistry
Immunohistochemistry (IHC) is a well-known method for biomarker identif and disease classification (Varghese, 2014). In IHC, protein expression is detected in tissues using biomarker-specific antibodies (AB). The primary AB binds to MIF and AIMP1. A secondary AB will gain an AB-AB reaction with the primary AB. Upon secondary AB binding, a brown color-producing reaction in presence of DAB will be catalysed, resulting in biomarker specific tissue staining.
11
3.2.1 IHC protocol
Formalin-fixed, paraffin-embedded tissues were immunohistochemically stained for MIF according to SOP: TI-An-O40: Detection of proteins (antigens) in tissue sections by immunohistochemistry (App. B). Blocks were cut into 4.5µm sections and dried overnight at 37oC. After deparaffinisation, endogenous
peroxidase activity was blocked by treating the deparaffinised sections with 0.3% H2O2 in 100% MeOH
for the duration of, 20 min, after which the sections were washed two times 5’ with 10 x PBS (pH:7.4). The epitopes were disclosed for antigen retrieval in 10 ml 0.1M citrate buffer dissolved in 90 ml ddH2O. In
a half closed tray the buffer was brought to the boil during 2’ on 700 Watt in a microwave and kept hot during, 20’ on 80 Watt. After, 20’ cooling on ice the sections were washed twice with 10 x PBS again and blocked with 100 µl PBT. PBT was made from 50 ml 1 x PBS and 50 µl 10 x triton X with phosphate buffered saline (Sigma life science, >96% pH: 6.5-7.5). The tissues were incubated overnight at 4oC with
100 µl antigen dilution of 1:500 with either a polyclonal rabbit AIMP1 antibody (GTX 105175, 0.54 mg/ml, GeneTEx) or a polyclonal mouse MIF antibody (MCA3263z, 0.1mg/ml, Bio-RAD). After two washings with 10 x PBS the sections were incubated 60 min with secondary antibody either polyclonal mouse antibody for MIF (poly HRP-ANTI-Mouse IgG, Bright Vision, Duiven, The Netherlands) or a polyclonal rabbit antibody for AIMP1 (poly HRP-ANTI-Rabbit IgG, Bright Vision, Duiven, The Netherlands). After additional 5’ 10 x PBS washing, 100 µl DAB (DAB substrate buffer and DAB chromogen, Chromogen system) was added and the tissues were incubated during 8’ in the dark. The staining was stopped in ddH2O. Finally
the sections were counterstained with 4’ hematoxylin, 1’ washed in ddH2O and mounted with glycerol
gelatine.
3.3 Pathological staining evaluation
The obtained tissue samples were evaluated on quality, specificity and reproducibility by a trained pathologist. Pathological analysis of tissue samples is a time consuming and subjective procedure, wherein the intensity and specificity of antibody staining is visually judged. If the quality of the stainings is found to be sufficient, the pathologist counts the samples manually. Manual scoring decisions are directly influenced by visual bias. This instigated the design of a method for automated digital IHC image analysis for an unbiased, quantitative assessment of antibody staining intensity in tissue sections.
3.4 Digital staining evaluation
For digital staining evaluation a color deconvolution was conducted on images of the stained tissues using the deconvolution plugin in ImageJ. This results in isolation of the DAB, hematoxylin and complementary background staining spectra, making it possible to analyse the DAB staining (Figure 1). There were different digital strategies applied on the DAB-images. After image acquisition strategy the methods for optical density vector determination, threshold counting and IHV profiler scoring are explained.
3.4.1 Image acquisition
Images of the stainings were captured with a standard light microscope (Olympus BX 51) equipped with an U-CMAD-3T7 lens (Tokyo, Japan). Objective lenses 10X, 20X and 40X were used. Snapshots were made using Olympus cellsens entry software on light preset:8.
3.4.2 Optical density vector determination
The optical density (OD) protocol of Ruifrok et al. was used as described by Varghese et al. (2014). Varghese defined OD as being proportional to the concentration of the stain. The OD was calculated from DAB images using the ImageJ measurements plugin. The measurements were given based on a pixel-by-pixel analysis and OD was calculated manually following:OD = log (Imax/Imean), in which Imax was 255 for all samples.
12
3.4.3 Threshold counting
After color deconvolution the images were converted to 8-bit images (grayscale) and the intensity-threshold was manually set for each tissue. To limit subjective bias as much as possible, the darkest positive cell was chosen and the threshold was set on the point that this cell is fully selected (Figure 2.a). The ImageJ scale was set on 1.585 pixels/µm. The size-threshold for counting was set on a distance of 80-1000 pixels for every tissue. The selected particles were counted using the count particles plugin in ImageJ.
3.4.4 IHC profiler scoring
An automated digital IHC image analysis was conducted using the ImageJ IHC profiler plugin. This plugin is designed by Varghese (2014) for IHCs and applicable for cytoplasmic or nuclear protein stainings. The IHC profiler plugin operates a pixel-by-pixel analysis on the DAB image and gives a score in a four tier system. Each image was corrected for darkness of the staining using a pixel intensity range. This pixel intensity range was set from 0 to 255 in which 255 was the lightest and 0 is the darkest shade (note: this
differs from the OD measurement method where 255 is the darkest shade). Subsequently a histogram of the image was generated (Figure 2.b). Based on where the peak of the histogram lies in the reference bar a score was assigned to the tissue (Figure 3).
a b c d
Figure 1. Deconvolution of an IHC-DAB stained tissue sample a; IHC stained colon tissue. b; DAB color image.c; Hematoxylin color image d; Background color image
Figure 2. a; Manually set threshold for MIF positive stained cells on an 8-bit grayscale DAB image. b; IHC profiler histogram of pixel intensity of DAB image
13
3.5 Statistical analysis
GraphPad Prism 7 software (GraphPad Software, La Jolla California USA, 2017) was used to statistically analyse and visualize the gained data. All datasets were tested for normality. Each dataset required the non-parametric Kruskall-Wallis test for comparing means between groups. Significance level was set on p-value <0.05.
4. Results
4.1 Dataset
A total of 24 samples was included, among which were one negative control, 2 positive controls and 4 HCs. Four CD patients and 4 UC patients formed the IBD group. In the MIF-batch 3 PSC-CDs and 5 PSC-UCs made up the PSC-IBD group. In the AIMP1-batch 1 PSC-CD patient and 6 PSC-UC patients were taken into account. This difference in PSC-IBD setup emerged from tissue lost during IHC staining and an error in experiment setup.
Figure 3. (Varghese, 2014)
Different zones assigned for the scoring of the DAB stained image. A: Shows the reference bar distributing the various zones ranging from 0 to 235. 235 to 255 pixel values are generally found to represent fatty tissues or blank areas and thus kept out of range for zone considerations. B: High positive (3+) image with its corresponding reference bar. C: Positive (2+) stained image with its corresponding reference bar. D: Low positive (1+) stained image with its corresponding reference bar. E: Negative (0) stained image with its corresponding reference bar
14
4.2 Morphology
In the MIF batch the negative control was blanco and the positive control stained very weak. Of the MIF batch 10 samples appeared to have chiefly background staining, 11 looked according to protein atlas (Figure 4.) and 1 sample appeared to be blanco. In the AIMP1 batch 11 stainings looked like protein atlas, 8 stainings appeared aspecific and 3 stainings appeared blanco. The appearance of the samples do not show similarities between or within the different batches or patient groups.
=
Figure 4. IHC staining expectation in healthy colon tissue based on protein atlas. a; weak staining intensity but specific positive cells. b; strong staining intensity with higher background staining and specific positive cells. c; d; AIMP1 expectation.
b
a c
15 a b c d e f g h i j
Figure 5. IHC stainings. a; Negative control on healthy colon tissue. b; Positive control MIF staining on colon tissue. c; MIF staining on CA tissue of a PSC patient. d; MIF staining on CA tissue an IBD patient. e; MIF staining on CA tissue of a HC, f; Negative control on (necrotic) tonsil tissue. g; Positive control AIMP1 staining on tonsil tissue. h; AIMP1 staining on CA tissue of a PSC patient. i; AIMP1 staining on CA tissue of an IBD patient. j; AIMP1 staining on CA tissue of a HC.
16 The consulted pathologist, J. Verheij, stated too little contrast between positive and negative cells was visible to determine staining specificity. However, some colored specifically stained mast cells. Also the IHC stainings lacked reproducibility. A further optimization of the protocol was advised by comparing different protein blocks and ABs from different companies.
4.3 Staining intensity
It was expected that positive colored cells could be distinguished from none or less-protein expressing cells based on staining intensity. By counting a significant higher or lower amount of positive cells in tissues of different patient groups, conclusions were intended to be drawn. In the conducted stainings the same tissues between and within IHC batches differed in color intensity. Looking to earlier literature and pictures of stainings of for instance protein atlas, the specificity of the antibody seems less questionable than the reproducibility. There are batches of stainings that were countable, but since there was a lack of reproducibility statistical analysis remains untrustworthy.
4.4 Digital evaluation
For each cytokine three datasets were generated. Among each dataset there were 3 groups: PSC-IBD, IBD and HC. For each dataset statistical analysis was conducted using GraphPad 7. The results are presented per strategy following optical density, positive cell counts and IHC profiler scoring and for the groups IHC, IBD and PSC-IBD.
4.2.1 Optical Density
MIF: Optical Density scores of MIF samples. The one-sample t-test for each group was significant. Only the PSC group did not met assumption for normality. The Kruskall Wallis test compared PSC with IBD, PSC with HC and IBD with HC. None were significant.
AIMP1: Optical Density scores of AIMP samples. The one-sample t-test for each group was significant. Every group met assumption of normality. The Kruskall Wallis test compared PSC with IBD, PSC with HC and IBD with HC . None were significant.
d
f
17
4.2.2 Positive cell count
MIF: The one-sample t-test for each group was not significant for the PSC group. The PSC group did not meet assumption of normality. The Kruskall Wallis test compared PSC with IBD, PSC with HC and IBD with HC. None were significant.
AIMP1: The one-sample t-test for each group was significant. All the groups met the assumption for normality. The Kruskall Wallis test compared PSC with IBD, PSC with HC and IBD with HC. None were significant.
4.2.3 IHC profiler score
MIF: The one-sample t-test for each group was not significant for the PSC group. This group also wasn’t normal. The Kruskall Wallis test should compare PSC with IBD, PSC with HC and IBD with HC. None were significant.
AIMP1: The one-sample t-test for each group was significant. Only PSC-IBD group met normality test. The Kruskall Wallis test should compare PSC with IBD, PSC with HC and IBD with HC. None were significant.
18
5. Conclusion
To explore whether MIF and AIMP1 can be useful biomarkers in the early detection of colorectal carcinoma in PSC-IBD patients, IHC stainings were conducted on CA tissues. To be an useful biomarker the proteins should be significantly differently expressed in PSC-IBD tissues of patients of whom is known they’ll develop CRC compared to non CRC-PSC-IBD patients, IBD patients and HC’s.
5.1 Biomarker potential
MIF: Based on literature MIF has a potential as biomarker for early CRC detection due to its pro-inflammatory properties. Its link with IBD and sporadic CRC make it a suspect in PSC-IBD CRC conduction. Expected was an upregulation of MIF in PSC-IBD patients compared to IBD patients and HCs. Also, MIF was significantly upregulated in Boonstra’s (2013, unpublished) PCR array.
AIMP1: Based on literature AIMP1 has a potential as biomarker for early CRC detection due to its anti-tumorigenesis properties. A downregulation of AIMP1-expression in PSC-IBD patients compared to IBD patients and HCs was expected. Contradictionairy to this expectation a significant upregulation of AIMP1 was seen in Boonstra’s (2013) PCR array. Upregulation of AIMP1 was expected in the stainings of the used tissue samples.
5.2 Staining analysation
The stainings could not be validated by a pathologist due to lack of specificity and reproducibility. The conducted statistical analysis on generated data gave without exception insignificant results. Based on this no hard conclusions regarding the main question can be made.
6. Discussion
6.1 Validity
For a research to be valid and is related results to be reliable the conducted experiments should measure what they intend to measure. Exactly the validity of the IHC stainings gave problems during this research. It cannot be said if the IHC outcomes were results or nonspecific stainings.
6.2 Expectations
The expectations point of this research were based on the PCR array of Kirsten Boonstra with significance difference in expression of close to 0.05. This significance is only seen when the non-parametric Kruskall Wallis test is conducted on certain groups. This makes it questionable whether the results were significant enough to distinguish the different patient groups from each other.
In this research the expectation for AIMP1 is made based on the assumption that RNA upregulation will cause protein expression upregulation. In reality however, RNA expression does not necessarily result in protein upregulation.
6.2.1 MIF
MIF is a well-known pro-inflammatory cytokine in carcinogenesis. Its expression is expected in nearly every cell in nearly every tissue. Due to MIFs widespread functioning and expression, a specificity bias is prevailing. In further research with MIF this should be taken into account.
19
6.2.2 AIMP1
AIMP1 is a rather unknown protein. Its RNA upregulation is an interesting starting point. However differences in expectations based on literature and Boonstra’s (2013) PCR array make it a hazardous target. On the other hand this could be even more interesting for further research towards AIMP1s precise functioning.
6.2.3 Staining morphology
The demarcation between aspecific staining or a result remains a point of discussion. The analysed stainings were disapproved based on the impression of one single pathologist. Similarity between obtained stainings, and expectations based on literature and protein atlas is seen from others perspective. This asks for a second opinion. However, executed statistics remain insignificant.
6.3 Digital methods
6.3.1 Image acquisition
The used microscope and software for image acquisition were of lesser quality than desired. Looking at the same staining through microscopes of higher quality could give different insights.
6.3.2 ImageJ
There are a few points of discussion about the conducted ImageJ analysis. Firstly for the obtained OD did not correct for the cell density of the tissues is executed. Also with IHC profiler software the analysis is very dependent on which part of the tissue is photographed and analysed. To reduce this bias analysation on 10x magnification can be done. Also within-patient comparison by analysing for instance 5 images of the same tissue sample can diminish this bias.
Secondly the manually set intensity threshold for each image can give subjective bias in counting the positive cells. However, the size threshold will diminish aspecific counting as much as possible. This counting method remains more reliable than counting manually. A more accurate size threshold based on expectations is desirable.
6.4 Research design
The sample size might be too small to effectuate sufficient statistical analysis. Also IHC staining might not be the best strategy to answer the main question. A western blot could beof compelling added value in further research.
6.5 insights
A lot of literature research was conducted in order to make a proper expectation. In the literature research IL22 was also taken into account, however, not mentioned in this scientific report since no IL22 stainings were executed. During the literature research and meetings with the pathologist also the importance of mast cells in CRC development became clear. Mast cells seemed to be the only cells that colored sufficient positive, different from the expectation that T-cells and Macrophages would color positive in the lamina basalis. A connection between MIF, AIMP1, IL22, mast cells, T-cells, PSC, IBD and inflammatory CRC was found as schematized in figure 9. In Appendix D the connections are briefly explained briefly with an appointment to the article where the connection was found.
20
6.6 Research recommendations
For follow-up research to MIF and AIMP1 expression on this sample set the aforementioned suggestions should be taken into account. When a valid research design is formed, PSC-IBD CRC samples can be used to further explore the relevance of these biomarkers in CRC development. Also different techniques, such
as a Western Blot could be conducted on this sample set
Follow-up research towards other biomarkers in the early detection of CRC in PSC-IBD patients should be done. Based on evaluated literature in this research IL22 and mast-cells could be relevant starting points.
21
7. Literature
1. Baron, T. H., Koehler, R. E., Rodgers, W. H., Fallon, M. B., & Ferguson, S. M. (1995). Mast cell cholangiopathy: another cause of sclerosing cholangitis. Gastroenterology, 109(5), 1677-1681.
2. Background Correction. Requested on 19/5/2017 from
https://imagej.nih.gov/IJ/plugins/background.html
3. Blatner, N. R., Bonertz, A., Beckhove, P., Cheon, E. C., Krantz, S. B., Strouch, M., ... & Khazaie, K. (2010). In colorectal cancer mast cells contribute to systemic regulatory T-cell dysfunction.
Proceedings of the National Academy of Sciences, 107(14), 6430-6435.
4. Bernhagen, J., Calandra, T., & Bucala, R. (1998). Regulation of the immune response by macrophage migration inhibitory factor: biological and structural features. Journal of Molecular
Medicine, 76(3), 151-161.
5. Brand, S., Beigel, F., Olszak, T., Zitzmann, K., Eichhorst, S. T., Otte, J. M., ... & Leclair, S. (2006). IL-22 is increased in active Crohn’s disease and promotes proinflammatory gene expression and intestinal epithelial cell migration. American Journal of physiology-Gastrointestinal and Liver
Physiology, 290(4), G827-G838.
6. Chen, W. S., Wei, S. J., Liu, J. M., Hsiao, M., Kou‐Lin, J., & Yang, W. K. (2001). Tumor invasiveness and liver metastasis of colon cancer cells correlated with cyclooxygenase‐2 (COX‐2) expression and inhibited by a COX‐2–selective inhibitor, etodolac. International journal of cancer, 91(6), 894-899.
7. Chichlowski, M., Westwood, G. S., Abraham, S. N., & Hale, L. P. (2010). Role of mast cells in inflammatory bowel disease and inflammation-associated colorectal neoplasia in IL-10-deficient mice. PloS one, 5(8), e12220.
8. Claessen, M. M. H. (2010). Colorectal carcinogenesis in patients with primary sclerosing
cholangitis and inflammatory bowel disease (Doctoral dissertation, Utrecht University).
9. Claessen, M. M. H., Lutgens, M. W. M. D., van Buuren, H. R., Oldenburg, B., Stokkers, P. C. F., van der Woude, C. J., ... & Siersema, P. D. (2009). More right‐sided IBD‐associated colorectal cancer in patients with primary sclerosing cholangitis. Inflammatory bowel diseases, 15(9), 1331-1336. 10. Cohen, M. C., Zeschke, R., Bigazzi, P. E., Yoshida, T., & Cohen, S. (1975). Mastocytoma cell
migration in vitro: Inhibition by MIF-containing supernatants. The Journal of Immunology, 114(5), 1641-1645.
11. Dictonary-normal:Colon-The Human Protein Atlas. Requested on 8/5/2017 from http://proteinatlas.org/learn/dictionary/normal/colon
12. Dictonary-normal:Pancreas-The Human Protein Atlas. Requested on 8/5/2017 from http://proteinatlas.org/learn/dictionary/normal/Pancreas
13. Dictonary-normal:Tonsil-The Human Protein Atlas. Requested on 8/5/2017 from http://proteinatlas.org/learn/dictionary/normal/Tonsil
14. Gaudenzio, N., Laurent, C., Valitutti, S., & Espinosa, E. (2013). Human mast cells drive memory CD4+ T cells toward an inflammatory IL-22+ phenotype. Journal of Allergy and Clinical
Immunology, 131(5), 1400-1407.
15. D'Haens, G. R., Lashner, B. A., & Hanauer, S. B. (1993). Pericholangitis and sclerosing cholangitis are risk factors for dysplasia and cancer in ulcerative colitis. American Journal of
Gastroenterology, 88(8).
16. He, X. X., Chen, K., Yang, J., Li, X. Y., Gan, H. Y., Lin, C. Y., ... & Al-Abed, Y. (2008). Macrophage Migration Inhibitory Factor (MIF) Promotes Coloarectal Cancer. Mol Med.
17. Heinrich, P. C., Behrmann, I., Serge, H., Hermanns, H. M., Müller-Newen, G., & Schaper, F. (2003). Principles of interleukin (IL)-6-type cytokine signalling and its regulation. Biochemical journal,
22 18. Hoi, A. Y., Iskander, M. N., & Morand, E. F. (2007). Macrophage migration inhibitory factor: a
therapeutic target across inflammatory diseases. Inflammation & Allergy-Drug Targets (Formerly
Current Drug Targets-Inflammation & Allergy), 6(3), 183-190.
19. Kim, S. S., Hur, S. Y., Kim, Y. R., Yoo, N. J., & Lee, S. H. (2011). Expression of AIMP1, 2 and 3, the scaffolds for the multi-tRNA synthetase complex, is downregulated in gastric and colorectal cancer. Korea, 2010, 0021159.
20. Kirchberger, S., Royston, D. J., Boulard, O., Thornton, E., Franchini, F., Szabady, R. L., ... & Powrie, F. (2013). Innate lymphoid cells sustain colon cancer through production of interleukin-22 in a mouse model. Journal of Experimental Medicine, 210(5), 917-931.
21. Lakatos, P. L., & Lakatos, L. (2008). Risk for colorectal cancer in ulcerative colitis: changes, causes and management strategies. World Journal of Gastroenterology, 14(25), 3937-3947.
22. Le, P. T., Pearce, M. M., Zhang, S., Campbell, E. M., Fok, C. S., Mueller, E. R., ... & Brubaker, L. (2014). IL22 regulates human urothelial cell sensory and innate functions through modulation of the acetylcholine response, immunoregulatory cytokines and antimicrobial peptides: assessment of an in vitro model. PloS one, 9(10), e111375.
23. Lee, Y. S., Han, J. M., Kang, T., Park, Y. I., Kim, H. M., & Kim, S. (2006). Antitumor activity of the novel human cytokine AIMP1 in an in vivo tumor model. Molecules & Cells (Springer Science &
Business Media BV), 21(2).
24. McCormack, G., Moriarty, D., O'Donoghue, D. P., McCormick, P. A., Sheahan, K., & Baird, A. W. (2001). Tissue cytokine and chemokine expression in inflammatory bowel disease. Inflammation
Research, 50(10), 491-495.
25. Maaser, C., Eckmann, L., Paesold, G., Kim, H. S., & Kagnoff, M. F. (2002). Ubiquitous production of macrophage migration inhibitory factor by human gastric and intestinal epithelium.
Gastroenterology, 122(3), 667-680.
26. Nishihira, J., & Mitsuyama, K. (2009). Overview of the role of macrophage migration inhibitory factor (MIF) in inflammatory bowel disease. Current pharmaceutical design, 15(18), 2104-2109. 27. Nobre, C. C. G., de Araújo, J. M. G., de Medeiros Fernandes, T. A. A., Cobucci, R. N. O., Lanza, D.
C. F., Andrade, V. S., & Fernandes, J. V. (2016). Macrophage Migration Inhibitory Factor (MIF): Biological Activities and Relation with Cancer. Pathology & Oncology Research, 1-10.
28. OD calibration. Requested on 19/5/2017 from
https://imagej.nih.gov.ij/docs/examples/calibration/index.html
29. Põlajeva, J., Bergström, T., Edqvist, P. H., Lundequist, A., Sjösten, A., Nilsson, G., ... & Pejler, G. (2014). Glioma‐derived macrophage migration inhibitory factor (MIF) promotes mast cell recruitment in a STAT5‐dependent manner. Molecular oncology, 8(1), 50-58.
30. Rao, C. V., & Yamada, H. Y. (2013). Genomic instability and colon carcinogenesis: from the perspective of genes. Frontiers in oncology, 3, 130.
31. Razumilava, N., Gores, G. J., & Lindor, K. D. (2011). Cancer surveillance in patients with primary sclerosing cholangitis. Hepatology, 54(5), 1842-1852.
32. Ren, X., & Colletti, L. M. (2009). IL-22 is involved in liver regeneration after hepatectomy.
American Journal of Physiology-Gastrointestinal and Liver Physiology.
33. Sekikawa, A., Fukui, H., Suzuki, K., Karibe, T., Fujii, S., Ichikawa, K., ... & Fujimori, T. (2010). Involvement of the IL-22/REG Iα axis in ulcerative colitis. Laboratory Investigation, 90(3), 496-505.
34. Shetty, K., Rybicki, L., Brzezinski, A., Carey, W. D., & Lashner, B. A. (1999). The risk for cancer or dysplasia in ulcerative colitis patients with primary sclerosing cholangitis. The American journal of
23 35. Shkolnik, T., Livni, E., Reshef, R., Lachter, J., & Eidelman, S. (1987). The macrophage migration
inhibition (MIF) assay as a marker of colorectal cancer. Diseases of the colon & rectum, 30(2), 101-105.
36. Shu, J., Qiu, G., Ilyas, M., & Kaye, P. (2010). Biomarker detection in whole slide imaging based on statistical color models. In MICCAI, 2010 Workshop on Computational Imaging Biomarkers for
Tumors: From Qualitative to Quantitative. 13th International Conference on Medical Image Computing and Computer Assisted Intervention, Beijing, China.
37. Shu, J., Qiu, G., & Mohammad, I. (2013, July). A semi-automatic image analysis tool for biomarker detection in immunohistochemistry analysis. In Image and Graphics (ICIG), 2013 Seventh
International Conference on (pp. 937-942).
38. Soetikno, R. M., Lin, O. S., Heidenreich, P. A., Young, H. S., & Blackstone, M. O. (2002). Increased risk of colorectal neoplasia in patients with primary sclerosing cholangitis and ulcerative colitis: a meta-analysis. Gastrointestinal endoscopy, 56(1), 48-54.
39. Sokol, H., Cosnes, J., Chazouilleres, O., Beaugerie, L., Tiret, E., Poupon, R., & Seksik, P. (2008). Disease activity and cancer risk in inflammatory bowel disease associated with primary sclerosing cholangitis. World J Gastroenterol, 14(22), 3497-503.
40. Sonnenberg, G. F., Fouser, L. A., & Artis, D. (2010). 1 Functional Biology of the IL-22-IL-22R Pathway in Regulating Immunity and Inflammation at Barrier Surfaces. Advances in immunology,
107, 1.
41. Stoyanova, I. I., & Gulubova, M. V. (2002). Mast cells and inflammatory mediators in chronic ulcerative colitis. Acta histochemica, 104(2), 185-192.
42. Talwalkar, J. A., & Lindor, K. D. (2005). Primary sclerosing cholangitis. Inflammatory bowel
diseases, 11(1), 62-72.
43. Tanaka, T., & Ishikawa, H. (2013, March). Mast cells and inflammation-associated colorectal carcinogenesis. In Seminars in immunopathology (Vol. 35, No. 2, pp. 245-254). Springer-Verlag. 44. Tissue expression of AIMP 1 – Summary – The Human Protein Atlas. Requested on 8/5/2017
from http://www.proteinatlas.org/ENSG00000164022-AIMP1/tissue
45. Tissue expression of IL22 – Summary – The Human Protein Atlas. Requested on 8/5/2017 from http://www.proteinatlas.org/ENSG00000127318-IL22/tissue
46. Tissue expression of MIF – Staining in colon – The Human Protein Atlas. Requested on 8/5/2017 from http://www.proteinatlas.org/ENSG00000240972-MIF/tissue/colon
47. Triantafillidis, J. K., Nasioulas, G., & Kosmidis, P. A. (2009). Colorectal cancer and inflammatory bowel disease: epidemiology, risk factors, mechanisms of carcinogenesis and prevention strategies. Anticancer research, 29(7), 2727-2737.
48. Torres, J., de Chambrun, G. P., Itzkowitz, S., Sachar, D. B., & Colombel, J. F. (2011). Review article: colorectal neoplasia in patients with primary sclerosing cholangitis and inflammatory bowel disease. Alimentary pharmacology & therapeutics, 34(5), 497-508.
49. Tsaitas, C., Semertzidou, A., & Sinakos, E. (2014). Update on inflammatory bowel disease in patients with primary sclerosing cholangitis. World journal of hepatology, 6(4), 178.
50. Vagefi, P. A., & Longo, W. E. (2005). Colorectal cancer in patients with inflammatory bowel disease. Clinical colorectal cancer, 4(5), 313-319.
51. Varghese, F., Bukhari, A. B., Malhotra, R., & De, A. (2014). IHC Profiler: an open source plugin for the quantitative evaluation and automated scoring of immunohistochemistry images of human tissue samples. PloS one, 9(5), e96801.
52. Väyrynen, J. P., Tuomisto, A., Klintrup, K., Mäkelä, J., Karttunen, T. J., & Mäkinen, M. J. (2013). Detailed analysis of inflammatory cell infiltration in colorectal cancer. British journal of cancer,
24 53. Wilson, J. M., Coletta, P. L., Cuthbert, R. J., Scott, N., MacLennan, K., Hawcroft, G., ... & Medina, F.
(2005). Macrophage migration inhibitory factor promotes intestinal tumorigenesis.
Gastroenterology, 129(5), 1485-1503.
54. Wolk, K., Witte, E., Hoffmann, U., Doecke, W. D., Endesfelder, S., Asadullah, K., ... & Sabat, R. (2007). IL-22 induces lipopolysaccharide-binding protein in hepatocytes: a potential systemic role of IL-22 in Crohn’s disease. The Journal of Immunology, 178(9), 5973-5981.
55. Quantifying Stained Liver Tissue. Requested on 19/5/2017 from
https://imagej.nih.gov/ij/docs/examples/stained-sections/index.html
56. Xu, L., Yi, H. G., Wu, Z., Han, W., Chen, K., Zang, M., ... & Qu, C. (2015). Activation of mucosal mast cells promotes inflammation-related colon cancer development through recruiting and modulating inflammatory CD11b+ Gr1+ cells. Cancer letters, 364(2), 173-180.
25
APPENDIX A: PCR array Results Kirsten Boonstra
PCR-array colon Qiagen results of Kirsten Boonstra (2013 unpublished). Different groups
were compared using the Kruskall-Wallis test or Mann Whitney-U test. Significant results
were seen in A, B, E and F.
27
APPENDIX B: SOP: TI-AN-O40
34
APPENDIX C: Results
35
APPENDIX D: Illustration and explanation insights
1. IBD – PSC: 60-80% of PSC patients have IBD (Torres, 2011).
2. IBD – UC-PSC: PSC is an independent risk factor for CRC development for UC patients. 70-80% of PSC patients have UC (Shetty, 1999)
3. IBD – CD-PSC: 10 % of PSC patients have CD (Torres, 2011).
4. IBD – MIF: MIF is multiple-level involved in the immune-inflammatory response and therefore potentially involved in immune-mediated inflammatory diseases like IBD. (Hoi et al., 2007) 5. PSC – Hepatoctemy.
6. PSC – CRC The risk of colorectal carcinogenesis in PSC patients is 50% after 25 years since disease ofset (Shetty, 1999)
7. IBD – CRC /inflammatory CRC: 8. IBD – CRC /inflammatory CRC:
9. CD – IL22: IL22 is produced by activated T-cells and is significantly more expressed in blood of patients with active Crohns disease (Wolk, 2007). Also, IL-22 mRNA expression is in general increased in inflamed colonic mucosa of patients with UC or CD (Brand et al.., 2005).
10. MIF – CRC: MIF expression is known to drive tumorgenic behaviour and is seen increasingly in sporadic human CRC (Wilson, 2005). It is already stated that a MIF assay can be useful as a clinical CRC marker (Shkolnik, 1986).
11. MIF – IL22: IL-22 is responsible for pro inflammatory cytokine upregulation and originates from activeated T-cell sites (Brand et al., 2005).
12. MIF – T-cells: MIF regulates macrophage, T and B-cell response (Brand et al., 2005). 13. PSC – MCs: MCs are potential role players in PSC diagnosis (Baron et al., 1995)
14. MIF – MCs: MC accumulation correlates positive with level of MIF (Polajeva et al., 2013). 15. AIMP1 – CRC: Decreased expression of AIMP1 in CRC tissue is observed. This suggests a role for
36 16. IL-22 – CRC: In colon cancer IL-22 induces cascades involved in inflammation(Ren & Colletti,
2009).
17. IL22 – T-cells: IL-22 is produced by activated T-cells (Brand et al., 2005).
18. IL22 – Hepatectomy: IL22 is connected in liver regeneration after 70% hepatectomy (Wolk, 2007).
19. MC – CRC: MC MCs are increased in inflamed colon of IBD patients. MCs stimulate tumor growth in inflammation-mediated CRC (Takana, 2013).
20. MC – IL22: MCs shape T-cells toward IL-22 production (Gaudenzio et al., 2013).
21. MC – T-cells: Interaction between T-cells and MCs promote immune suppression loss in autoimmunity diseases and stimulate inflammation in CRC. The cross talk between T-cells and MCs could be promising targets in CRC therapy (Blatner, 2010).
37
