Endothelial Fas-Ligand in Inflammatory Bowel Diseases and in Acute Appendicitis

Abstract

Fas-mediated induction of apoptosis is a major factor in the selection of lymphocytes and downregulation of immunological processes. In the present study, we have assessed endothelial Fas-ligand (FasL) expression in normal human ileum, appendix, and colon, and compared the expression levels with that in inflammatory bowel disease and in acute appendicitis. In a normal appendix, endothelial FasL levels were constant in almost half of the mucosal vessels; but, in the normal ileum and colon, endothelial FasL was practically restricted to areas in close proximity to lymphatic follicles, and was expressed mainly in the submucosal aspect of the follicles in the vessels with high endothelium. In samples from subjects with either Crohn’s disease or ulcerative colitis, the extent of endothelial FasL expression was elevated in the submucosa and associated with an elevated number of lymphoid follicles. In inflammatory bowel disease, ulcers and areas with a high density of mononuclear cells expressing FasL also showed an elevated density of blood vessels with endothelial FasL expression. Although the function of endothelial FasL remains unclear, such a specific expression pattern suggests that endothelial FasL expression has a role in the regulation of lymphocyte access to the peripheral lymphoid tissues, including the intestinal mucosa.

Introduction

The lymphoid system of the human gut is composed of an extensive set lymphoid tissue formations including Peyer’s patches (PP), appendix, isolated lymphoid follicles, and the intraepithelial compartment. In the intestine, antigens pass from the intestinal lumen via villus epithelium or M-cells. Within the organized structure of PPs, lymphocytes meet their cognate antigens and activate effector cells. During the activation and proliferation process, local dendritic cells mark freshly activated lymphoid cells to home back to the gut after a journey via the PP’s efferent lymphatics to mesenteric lymph nodes. From these nodes, cells eventually flow through the thoracic duct into the systemic blood circulation and finally back to the gut lamina propria (LP). Migration from the blood stream into gut tissue goes through the specialized high endothelial venules (HEVs). Migration into tissue requires guidance from both chemokines and addressins from the endothelial cells (Garside et al. 2004; Salmi and Jalkanen 2005).

HEVs are specialized post-capillary vessels that function as ports for lymphocytes to access tissue via the blood stream (Girard and Springer 1995; Tohya et al. 2010). Recently, we demonstrated that, in reactive human lymph nodes, these vessels express Fas-Ligand (FasL), a TNF-related inducer of apoptosis of Fas-wearing cells (Kokkonen et al. 2004; Kokkonen and Karttunen 2009). FasL is mainly found in lymphatic cells after activation (Krammer et al. 2007; Tohya et al. 2010) and in some epithelial cells of the eyes, testis and placenta (Pinkoski et al. 2002; Ferguson and Griffith 2006; Ferguson and Griffith 2007). The Fas/FasL system is considered to be focused on maintaining cellular homeostasis (Strasser et al. 2009). FasL in HEVs is hypothesized to induce apoptosis in circulating, activated lymphocytes and thereby restrict them from re-entering the lymph node (Sata and Walsh 1998; Kokkonen et al. 2004). In support of this, we found high numbers of Fas-positive apoptotic cells (including both CD3- and CD20-positive cells) near FasL-positive HEVs.

In the gut mucosa, FasL has been seen in the subsets of LP cells, including cytotoxic T lymphocytes (Pinkoski et al. 2000), plasma cells (Sträter et al. 1999), NK-cells (Nagata and Suda 1995; Nagata and Golstein 1995), macrophages and neutrophils (Griffith et al. 1995). However, no endothelial expression has been reported in these studies. Of the intestinal epithelial cells, Paneth cells express FasL (Moller et al. 1996) but no expression has been reported in the normal villus epithelium. Denning and colleagues (2002) showed that mouse villus epithelial cells express FasL after exposure to oxidative stress.

During inflammation, homing of lymphocytes is altered by a change in the hemodynamics and permeability after activation of local microvascular endothelial cells. In intestinal inflammatory conditions, such as inflammatory bowel disease (IBD) and acute appendicitis, the expression levels of addressins are modified by cytokine levels and specifically regulate the migration of lymphocytes to LP (Salmi and Jalkanen 2005; Hachim and Ahmed 2006). The main categories of inflammatory bowel disease are Crohn’s disease (CD) and ulcerative colitis (UC) (Yantiss and Odze 2006). In CD, FasL gene polymorphism determines whether or not a patient will respond to treatment with a TNF-alpha analogue (Hlavaty et al. 2005), suggesting that the Fas/FasL system has an important role in the pathogenesis of CD. Several studies have analyzed FasL expression in IBD, but none have described endothelial expression of FasL (Ueyama et al. 1998; Di Sabatino et al. 2003; Melgar et al. 2004; Souza et al. 2005).

Recently, we demonstrated FasL expression in HEV vessels adjacent to histologically normal or hyperplastic lymphoid follicles of ileal and duodenal mucosal biopsies of children (Kokkonen et al. 2011), suggesting the expression of endothelial FasL in intestinal lymphoid tissue. In the present study, we have systematically assessed endothelial FasL expression in the different layers of the intestinal wall. By using immunohistochemical methods with proven, reliable antibodies (G247-4) (Sträter et al. 2001), we explore here FasL endothelial expression in the wall of ileum, colon and appendix. We have also assessed endothelial expression of FasL in IBD and in acute appendicitis to evaluate potential changes in the expression related to chronic inflammation in autoimmune diseases and in acute inflammation, respectively. Since FasL expression is predominantly seen in vessels with morphology of HEVs, we used HEV markers (MAdCAM-1 and MECA-79) to confirm the functional identity of these FasL-expressing vessels.

Materials & Methods

Patient Samples

Tissue samples representing normal and inflamed ileum, colon and appendix were selected for evaluation from the files of the Department of Pathology, Oulu University Hospital. For each diagnostic group and anatomical region, 10 specimens were selected. Patients had undergone partial resection of the gut due to severe inflammation or intestinal cancer, or they had undergone an appendectomy. Specimens for Crohn’s disease or ulcerative colitis groups were selected on the basis of a diagnosis confirmed by both the clinician and pathologist. Specimens used in both normal groups (ileum and colon) had been resected due to intestinal cancer and a sample from a healthy area, distant from the tumor, was included. Samples in the appendicitis and normal appendix groups had come from patients who had displayed symptoms of appendicitis and their appendixes had been surgically removed and diagnosis was made by a pathologist. Table 1 summarizes the demographic data of the subjects.

Table 1.

Study Samples.

	Total (n)	Men (n) (%)	Age Mean (SD)
Normal Intestine
Ileum	8	4 (50%)	65.9 (9.69)
Colon	12	3 (25%)	66.5 (14.0)
Appendix	11	4 (36%)	34.5 (14.8)
Crohn’s Disease
Ileum	5	2 (40%)	44.6 (15.1)
Colon	5	3 (60%)	33.2 (18.1)
Ulcerative Colitis	10	7 (70%)	44.6 (10.5)
Acute Appendicitis	10	5 (50%)	38.4 (18.9)
Total	61	28 (46%)	47.9 (19.3)

Immunohistochemistry

Immunohistochemistry was performed using Dako Cytomation Envision™ kit according to manufacturer’s instructions (Glostrup, Denmark). Primary antibodies used were Fas-Ligand (G247-4; BD Pharmingen; San Diego, CA) at 1:1000 overnight incubation; von Willebrand Factor VIII (rabbit anti-human, A 0082; DAKO) at 1:5000; Podoplanin (Acris Andibodies GmbH, Herford; Germany) at 1:200; MECA-79 (BD Pharmingen) at 1:200; and MAdCAM-1 (17F5:sc-59790; Santa Cruz Biotechnology, Dallas, TX) at 1:100. FasL antibodies were incubated overnight at room temperature in a humidified chamber, and all others were incubated for 60 min at 37°C in a humidified chamber.

The specificity of the antibodies has been reported previously for Factor VIII (Sehested and Hou-Jensen 1981), podoplanin (Wendemagegn et al. 2015), FasL (Sträter et al. 2001), MECA-79 (Streeter et al. 1988), and MAdCAM-1 (Leung et al. 2004). For negative control staining, the primary antibody was omitted.

Assessment of Immunohistochemical Staining

Positively stained vessels in sections stained for Factor VIII and FasL were counted with a microscope using a 10× objective and 10× oculars, one with a 10×10 square grid. With 100× magnification, the total area of the whole 10×10 grid was 1 mm². We counted as many fields of view as possible in each anatomical compartment in each slide. In most cases, we could reach the aim of 10 fields. However, there are few slides for which this was not possible due to small tissue samples. We normalized counting results to indicate positive vessels per mm². Vessels in each of the following anatomical areas of the bowel wall were separately counted: mucosa, mucosal lymphoid follicles, submucosa, submucosal lymphoid follicles, subserosa, and subserosa lymphoid follicles. In cases representing IBD or appendicitis, ulcers were separately evaluated if present. Ulcers were defined as regions where mucosal structures, including the crypts, were missing.

Vessels were classified as HEV or non-HEV vessels. Characteristic to HEVs is that the endothelial cell height is the same or greater than the length. When counting vessels in the mucosa, one side of the grid was placed on the muscularis mucosae and the height of the mucosae determined the counting area. Lymphoid follicles were defined as a dense, roundish collection of lymphocytes. Vessel density was separately counted in lymphoid follicles with and without germinal centers. In addition to vessel counts, the number of FasL-positive, non-endothelial cells in the LP was counted and is expressed as cells/mm².

Statistical Analysis

Analysis was performed with SPSS Statistics 16.0 (SPSS, Chicago, IL). Normality of distribution was assessed with the Kolmogorov–Smirnov test and Shapiro–Wilk test, as well as with histograms and normal Q-Q Plots. For those few variables that did not show a normal distribution, the Mann–Whitney test was used for a comparison of differences between the groups; otherwise, a Student’s t-test was used. Correlation analyses were performed using Spearman’s rank correlation.

Results

Endothelial FasL Expression in Normal Ileum, Colon and Appendix

FasL expression was seen in occasional endothelial cells and in single cells (Figs. 1, 2). To assess whether endothelial FasL expression has any association with the functionally and immunologically different regions of the intestinal wall or with specific vessel types, FasL expression was separately analyzed in the mucosa, submucosa and subserosa. Additionally, each layer was examined separately in areas unrelated to lymphoid follicles but in the vicinity of lymphoid follicles. Expression in HEV vessels, as well as in those with flat endothelium (Fig. 1), was separately assessed. The number of Factor VIII-expressing vessels was similarly counted to obtain a density estimate for all vessels.

Figure 1.

Immunohistochemical staining of FasL. (A) Sample from a patient with Crohn’s disease (CD) in the colon, showing different vessel types expressing FasL in the submucosa (×100). FasL is seen in HEVs (H), veins (v) and venules, arterioles and sometimes in arteries (not shown). (B) Sample from a patient with ulcerative colitis (UC) showing damaged mucosa lamina propria (×100). Many FasL-positive HEVs are accompanied with many FasL-positive mononuclear cells. (C) Sample from a patient with CD in the subserosa (×100). FasL-positive HEVs are seen at the sides or just around the follicles without germinal centers. Scale, 100 µm.

Figure 2.

Scatterplots of FasL vascular expression correlation with single FasL cell expression in different anatomical areas. Data includes all of the patients.

Areal densities of different FasL-expressing vessels not related with lymphoid follicles are summarized in Table 2, and the proportions (%) of FasL-positive vessels of all vessels (as seen in Factor VIII staining) in Table 3. In the mucosa outside of the lymphoid follicles, FasL expression was rare both among HEV and non-HEV vessels in both the normal ileum and colon (Table 2), with only 13% of subjects showing vascular FasL in the ileum and 17% in the colon. Correspondingly, the proportions of FasL-positive vessels were low, always less than 3% of all vessels, with ileal HEV vessels showing the highest positivity rate. Similarly, in the submucosa, vascular FasL expression was absent from the normal ileum and only rarely present (8%) in the normal colon. The maximum proportion of vessels expressing FasL of all vessels in normal colon was about 3% (Table 3). In contrast, in the ileum, FasL expression was commonly present in the subserosal layer (43% of subjects), being detected here in about half of the vessels (Table 3). In the subserosa of the colon, FasL-expressing vessels were rare (11% of subjects), and the expression showed no association with vessels of the HEV type (Table 3).

Table 2.

Areal Density (vessels/mm²) of Fas-L-Positive Vessels in Different Anatomical Areas of the Gut.

		Structure
Diagnosis/ Organ	Vessel Type	Mucosa	Ulcers	Submucosa	Subserosa
Normal Ileum	HEV	0 (0–0.2)	np	0 (0–0)	0 (0–1.43)
n=8	nHEV	0 (0–0)	np	0 (0–0)	0 (0–10.83)
	all	0 (0–0.2)	np	0 (0–0)	0 (0–12.26)
Normal Colon	HEV	0 (0–0)	np	0 (0–0)	0 (0–0)
n=12	nHEV	0 (0–0.16)	np	0 (0–0.54)	0 (0–12)
	all	0 (0–0.16)	np	0 (0–0.54)	0 (0–12)
Normal Appendix	HEV	0.62 (0–7.36)*I,C	np	0.2 (0–1.29)*I,C	0 (0–0.74)
n=10	nHEV	0.3 (0–2.33)*I,C	np	0.15 (0–3.55)	0.16 (0–1.11)
	all	0.91 (0–7.87)*I,C	np	0.43 (0–4.84)*I,C	0.16 (0–1.85)
Crohn’s Disease Colon	HEV	0.15 (0–2.78)#	4.63 (1.55–4.67)*	0.12 (0–1.79)#	0 (0–1.79)
n=5	nHEV	0 (0–33.33)	4.14 (1.29–24.67)*	0.27 (0–5.36) *	1.53 (0–13.39)
	all	0.15 (0–36.11)	5.91 (5.69–29.33) *	0.53 (0–7.14) *	1.53 (0–15.18)
Crohn’s Disease Ileum	HEV	0.26 (0–1.8)	5.54 (0–8.7)	0 (0–8.37)	0.56 (0–3.51)
n=5	nHEV	0 (0–1.64)	2.97 (0–10.77)	0.7 (0–4.57)	0.31 (0–7.37)
	all	0.48 (0–3.44)	10.81 (0–18.7)	0.7 (0–12.93)	0.87 (0–10.88)
Ulcerative Colitis	HEV	0 (0–0)	1.56 (0–2.69)*	0 (0–4.56)	0 (0–0.56)
n=10	nHEV	0 (0–4.4)	8.75 (0–14.62)*	0.15 (0–9.22) t	0 (0–8.81)
	all	0 (0–4.4)	9.17 (0–16.92)*	0.15 (0–13.78) t	0 (0–8.81)
Appendicitis	HEV	0.65 (0–5.19)	0.38 (0.16–4.76)	0.52 (0–6.74)	0 (0–0.83)
n=9	nHEV	1.08 (0–2.38)	0.07 (0–3.11)	0 (0–2.92)	0.33 (0–6.25)
	all	1.2 (0–6.4)	1.46 (0.16–5.35)	0.61 (0–7.77)	0.33 (0–7.08)

The areal densities of HEVs and non-HEVs (nHEV), and all vessels are separately shown. Values are the median densities with minimum and maximum values in parentheses. Significant difference: *p<0.05; trend for a difference: #p value between 0.0501–0.058. Disease groups were compared with the corresponding normal group. Normal appendix was compared with normal ileum and normal colon; in these cases, a significant difference is marked as I (for ileum) or C (for colon). np, not present.

Table 3.

Proportion (%) of Areal Density of Fas-L Positive Vessels of the Areal Density of FVIII Positive Vessels in Different Anatomical Areas of the Gut.

		Structure
Diagnosis /Organ	Vessel Type	Mucosa	Ulcers	Submucosa	Subserosa
Normal Ileum	HEV	0 (0–2.61)	np	0 (0–0)	0 (0–152.82)
n=8	nHEV	0 (0–0)	np	0 (0–0)	0 (0–45.53)
	all	0 (0–0.52)	np	0 (0–0)	0 (0–41.46)
Normal Colon	HEV	0 (0–0)	np	0 (0–0)	0 (0–0)
n=12	nHEV	0 (0–0.5)	np	0 (0–2.87)	0 (0–28.47)
	all	0 (0–0.41)	np	0 (0–2.87)	0 (0–24.95)
Normal Appendix	HEV	10.65 (0–70.52)*I, *C	np	3.12 (0–32.61)*I, *C	0 (0–0)
n=10	nHEV	1.05 (0–12.88)	np	0.53 (0–12.7)	0 (0–3.32)
	all	2.35 (0–25.52)*I, *C	np	1.73 (0–14.74)*I, #C	0 (0–2.9)
Crohn’s Disease Colon	HEV	2.65 (0–14.29)#	17.3 (10.77–29.54)	2.69 (0–18.6)#	0 (0–9.44)
n=5	nHEV	0 (0–21)	5.74 (2.47–34.02)	0.87 (0–15.09)*	6.1 (0–26.31)
	all	0.36 (0–19.69)	8.74 (7.02–25.32)	1.45 (0–15.84)*	5.12 (0–21.74)
Crohn’s Disease Ileum	HEV	5.79 (0–26.88)	35.14 (0–49.5)	0 (0–150.88)	23.43 (0–38.38)
n=5	nHEV	0 (0–5.94)	5.6 (0–14.92)	1.42 (0–25.66)	1.2 (0–42.25)
	all	1.56 (0–10.04)	15.69 (0–18.54)	1.14 (0–55.42)	4.35 (1.01–33.55)
Ulcerative Colitis	HEV	0 (0–0)	14.43 (0–30.13)	0 (0–91.11)	0 (0–5.56)
n=10	nHEV	0 (0–5.95)	10.47 (0–22)	0 (0–30.24)	0 (0–34.41)
	all	0 (0–5.76)	9.82 (0–22.25)	0 (0–38.81)	0 (0–22.02)
Appendicitis	HEV	6.74 (0–48.18)	3.21 (1.18–38.56)	11.08 (0–84.93)	0 (0–4.66)
n=9	nHEV	2.91 (0–5.62)	1.16 (0–8.49)	2.01 (0–13.89)	0.93 (0–56)
	all	4.39 (0–13.83)	3.94 (0.49–15.69)	2.94 (0–34.15)	0.72 (0–56)

The areal densities of HEVs and non-HEVs (nHEV), and all vessels are separately shown. Values are the median densities with minimum and maximum values in parentheses. Significant difference: *p<0.05; trend for a difference: #p value between 0.0501–0.058. Disease groups were compared with the corresponding normal group. Normal appendix was compared with normal ileum and normal colon; in these cases, a significant difference is marked as *I (for ileum) or *C (for colon); a trend as #I or #C. np, not present.

In the normal appendix, endothelial FasL expression was significantly more common than in the ileum or colon in the submucosa [70% vs. ileum 0% (p=0.012) and vs. colon 8 % (p=0.011)] and tended to be present more often in the mucosa (60%) and subserosa (50%; p=not significant). Proportions of FasL-expressing vessels were significantly higher in the appendiceal mucosa (median 2.4%) and submucosa (median 1.7%) than in the corresponding regions of the ileum (median values 0%, 0%; p<0.005) or colon (0%, 0%; p<0.05; Table 3). The expression was common in the HEV-type vessels (median, mucosa 10.7%; submucosa 3.12%; Table 3).

In the areas in the vicinity of lymphoid follicles, endothelial FasL expression was elevated in the submucosa but not in the mucosa (Fig. 1; Fig. 3). Proportions of FasL-positive HEVs of all FasL-positive vessels were higher near or in lymphatic follicles of submucosa (median, 66.7%; min to max 0% to 100%) compared to other regions of submucosa (median, 25.8%; min to max, 0% to 100%; p=0.006, all groups included).

Figure 3.

Boxplot of areal density vessels expressing FasL in intestinal lamina propria and submucosa. Vessels were separately counted within lymphoid follicles and in lamina propria or submucosa outside of lymphoid follicles. HEVs and non-HEVs are shown separately, and significant differences are shown between lamina propria and follicles. The y-axis shows positive vessel density per mm². Boxplots show median, quartiles and 5% to 95% intervals. Data include all of the patients.

The type of lymphoid follicle was related with FasL expression in the adjacent vessels. In the mucosa of the normal appendix, there were significantly more FasL-positive vessels associated with follicles without germinal centers (GC0 median 12.16/mm², range 2.55–23.27/mm²) than with those with germinal centers (GC1 median 0.88/mm², range 0.0–4.39 /mm²; p=0.018). Similarly, submucosal GC0 showed a higher vascular FasL density compared with GC1, both among HEVs (4.17/mm², range 0.0–23.9 /mm² vs 0.0/mm², range 0.0–4.39/mm²; p<0.001) and non-HEVs (1.43/mm², range 0.0–12.5/mm² vs 0.0/mm², range, 6.25/mm²; p<0.001).

Endothelial FasL Expression in IBD and in Acute Appendicitis

FasL expression around lymphoid follicles was similar in the disease groups and controls (data not shown). In the colon mucosa outside the lymphoid follicles, FasL expression in the HEVs was seen significantly more frequent in subjects with CD than in the controls (60% vs 0%; p=0.015, Fischer’s exact test). Patients with CD in the colon also showed an elevated vascular FasL expression in the submucosa in the non-HEV vessels, as both areal density (p=0.027) and proportion of FasL-expressing vessels were elevated (p=0.027) as compared with those values in normal colon submucosa (Table 3). Similarly, in colon affected by CD, the density of FasL-expressing HEVs in both mucosa and submucosa showed a trend of increment as compared with that in the normal colon (p=0.064, p=0.064, respectively; Table 3). Subjects with CD in the ileum had similar mucosal vascular FasL expression densities and percentages of Factor VIII as that found in normal ileum (Table 3). However, in the submucosa, FasL non-HEVs were present more often in subjects with CD (60%) than in those with normal ileum (0%; p=0.035).

FasL-expressing non-HEVs tended to be more common in patients with UC (50%) than in the subjects with a normal colon (0%; p=0.056). Patients with UC also showed elevated areal densities of FasL-expressing HEVs and non-HEVs (p=0.009 and p=0.009, respectively) as compared with that in normal submucosa of the colon (Table 3).

In subjects with appendicitis, vascular FasL density, the proportion of all vessels and number of patients that had vascular FasL did not differ from those values measured in samples from patients with normal appendices (Tables 2 and 3).

Vascular density in the ulcer base was compared with the anatomically corresponding layer of the intestinal wall; i.e., the submucosa. Ulcers in CD of the colon showed an increment in the density of both FasL-expressing HEVs (p=0.004) and non-HEVs (p=0.004), and FasL expression was significantly more common in both HEVs (100%; p=0.002) and non-HEVs (100%; p=0.009) as compared with normal colon submucosa (8%; Table 2). Similarly, in ulcers of subjects with CD in the ileum, FasL was seen more often in both HEVs and non-HEVs (both 60%), than in normal ileum submucosa (both 0%; both p=0.035). Also, in the ulcer bases in UC, both FasL HEVs (80%; p=0.002) and FasL non-HEVs (80%; p=0.010) were more common than that observed in the normal colon submucosa (8% both). In appendiceal ulcers, FasL expression was statistically similar in both HEVs and non-HEVs (100% both) as those in the submucosa of normal appendix (70%).

Lymphoid Follicle Density and Vascular and Non-endothelial FasL Expression

As expected, patients with CD had more lymphoid follicles in the ileal submucosa as compared with the normal submucosa (median 4.1, range 1.2–5.0 vs 0.25, 0.1–3.9; p=0.006). A similar difference was seen in the ileal subserosa (CD median 3.8, range 0.0–4.7 vs controls 0.0, 0.0–1.0; p=0.019), and colon submucosa (CD, median 3.4, range 3.2–4.6 vs controls 1.7, 0.2–3.8; p=0.006). A similar incremental trend was seen in the subserosa of the colon (CD, median 0.8, range 0.0–4.1 vs 0.0, 0.0–0.0; p=0.064). UC tissue did not differ from normal colon tissue in numbers of lymphoid follicles. Patients with appendicitis had a significantly elevated number of follicles in the subserosa as compared with those with a normal appendix (median 2.5, range 0.2–5.0 vs 0.0, 0.0–0.8; p=0.001). Numbers of lymphoid follicles did not correlate to endothelial FasL density in mucosa or submucosa, but a significant correlation was present in the subserosa (c = 0.403; p=0.003).

FasL Expression in Non-endothelial Cells

FasL was also expressed in single non-endothelial cells, mostly in the mononuclear cells. Areal densities of these cells are shown in Table 4. Elevated densities of FasL-positive single cells were seen in UC and CD as compared with the controls. Patients with CD in the colon had significantly elevated densities of single FasL-positive cells in the submucosa (p=0.004), subserosa (p=0.05) and ulcers (p=0.004) as compared with that in normal colon. In ileal CD, the density of single FasL-positive cells was elevated as compared with normal ileum in the mucosa (p=0.030), submucosa (p=0.002) and in the ulcers (p=0.019; compared to normal ileum submucosa) but not in the subserosa. UC patients showed an elevated density of single FasL-positive cells in the submucosa (p=0.009) and in ulcers (p=0.019 compared with normal colon submucosa) but no difference was seen in the mucosa or subserosa. In acute appendicitis, the density of FasL-positive single cells was decreased in the mucosa (p=0.002) and in ulcers (p=0.011), but in submucosa and subserosa no differences in comparison to normal appendix was seen.

Table 4.

Areal Densities (cells/mm²) of Fas-Ligand-Expressing, Single Non-Endothelial Cells in Different Anatomical Areas of Normal and Inflamed Gut.

	Structure
Diagnosis/Organ	Mucosa (Lamina Propria)	Ulcers	Submucosa	Subserosa
Normal Ileum n=8	2.9 (0.2–12.2)	np	0 (0–0)	0 (0–2.6)
Normal Colon n=12	17.6 (9.6–38.5)	np	0 (0–0.7)	0 (0–0.4)
Normal Appendix n=10	54 (18–93.5)	64.1 (15.6–151.7)	2.7 (0–8.8)	0 (0–1.7)
Crohn’s Disease Colon n=5	25 (3.9–333.3)	24 (17.7–308) p=0.004	2.7 (0.1–148.2) p=0.004	2.7 (0–118.8) p=0.050
Crohn’s Disease Ileum n=5	14.4 (6.8–21.6) p=0.030	13.2 (0–16.4) p=0.019	5.3 (1.5–35.6) p=0.002	5 (0–32)
Ulcerative Colitis n=10	16.5 (5.2–46.1)	19.7 (0–31.3) p=0.019	1.7 (0–6.9) p=0.009	0 (0–3.6)
Appendicitis n=9	21.1 (7.8–47) p=0.002	4.6 (0.6–70.2) p=0.011	1.7 (0–13.8)	0.4 (0–8.8)

Values are the median, with the range in parentheses. Significant differences between diseases and controls are shown. Ulcers were compared with normal submucosa. np, not present.

FasL single cell density in the mucosa correlated with the density of mucosal vessels expressing FasL (HEVs: c = 0.349, p=0.007; non-HEV: c = 0.484, p<0.001). Similar correlations were seen in the submucosa (HEV: c = 0.916, p<0.001; non-HEV: c = 0.628, p<0.001) and subserosa (HEV: c = 0.588, p<0.001; non-HEV: c = 0.583, p<0.001) but not in the ulcers (Figure 2).

Correlation of Endothelial FasL Expression with Other HEV Markers (MECA-79 and MAdCAM-1)

MECA-79 expression was present in vessels consisting mostly of HEV-sized venules. Common locations for MECA-79 venules were near submucosal parts of lymphoid follicles with corresponding FasL expression in the HEV vessels (Fig. 4) and near muscularis mucosae. MECA-79-expressing veins were especially seen in the appendices between follicles at the border of the mucosa and submucosa. Interestingly, some veins near follicles showed MECA-79 positivity and higher endothelial cells on the follicle side but flat and negative staining endothelium on the other side of vessel. MAdCAM-1 expression in vessels was seen more on non-HEVs than on HEVs. MAdCAM-1 vessels were mostly found in the mucosa and somewhat less in submucosa near the muscularis mucosae but none in the subserosa.

Figure 4.

Immunohistochemical staining of FasL, MECA-79, MadCAM-1, and Podoplanin in a sample of histologically normal colon. (A) In FasL staining, HEVs (arrows) at the outer aspect of a lymphoid follicle show a weak positive reaction whereas other vessels are negative. (B) In MECA-79 staining, HEVs in the outer parts of a lymphoid follicle, corresponding vessels decorated by FasL positivity in (A), show a positive reaction (arrows). In addition, vessels with low endothelium at the inferior (submucosal) aspect of the lymphoid follicle show a positive reaction for MECA-79 mainly at the side against the lymphoid follicle (arrowheads). (C) In MadCAM-1 staining, a positive reaction is present in HEVs (arrows) and in mucosal vessels with low endothelium (arrowheads). (D) Podoplanin antibodies label only vessels with low endothelium in the very basal part of mucosa (arrow) and submucosa (arrowheads), consistent with lymphatic vessels. Scale, 50 µm.

Densities of Factor VIII HEVs correlated with FasL HEVs in mucosa (p<0.001) and submucosa (p=0.002) but not in ulcers or subserosa. Factor VIII non-HEVs and FasL non-HEVs instead have correlations in mucosa (p=0.047) and ulcers (p=0.002) but not in submucosa or subserosa. Despite these correlations, total Factor VIII HEV and non-HEV densities are greater than that of total FasL HEV and non-HEV densities in all areas (p<0.001 in mucosa, ulcers, submucosa and subserosa; Table 2). Areal density of MECA-79 HEVs instead correlate in all areas with FasL HEVs (mucosa, p<0.001; ulcers, p=0.001; submucosa, p=0.006; and subserosa, p=0.006). MECA-79 non-HEVs instead correlate with FasL non-HEVs only in mucosa (p<0.001). Correlations between MAdCAM-1 HEVs and non-HEVs and FasL HEVs and non-HEVs are found only in the mucosa (HEVs p=0.001; non-HEVs p=0.004). Figure 5 illustrates FasL correlations with Factor VIII, MECA-79 or MAdCAM-1 as scatterplots (includes both HEV and non-HEV densities).

Figure 5.

Scatterplots showing endothelial (all vessel positivity) FasL correlations with Factor VIII (upper row), MECA-79 (middle lane) and MAdCAM-1 (lower lane). Correlations are calculated in different anatomical areas mucosa (first column from left), ulcers (second from left), submucosa (third from left) and subserosa (last column).

Discussion

The main finding in the present study is that, in the intestinal wall, FasL is expressed in the endothelium of a subset of HEVs, and in flat-walled venules, arterioles and arteries (Fig. 1). In normal mucosa and submucosa of the ileum, colon and appendix, vascular FasL expression was mostly restricted to the vicinity of lymphoid follicles, where about 0.4% to 4.5% of non-HEV and HEV vessels, respectively, showed expression (Tables 2, 3; Figs. 1, 2). Overall, the density of FasL-expressing vessels was higher in the appendix, likely related to the abundance of lymphoid follicles in this organ. In IBD, an elevated density of FasL-expressing vessels was seen at the ulcer base and in the submucosa of samples from patients with CD of the colon (Tables 2, 3). Our findings indicate that endothelial FasL expression in the intestinal wall shows a specific pattern related to immunologically different regions of the intestinal wall. Furthermore, a characteristic increase in submucosal FasL expression in CD suggests that abnormalities in endothelial function are involved in the pathogenesis of this disease.

Our analysis with HEV markers (MECA-79, MAdCAM-1) indicated that FasL-positive vessels with high endothelial cells show functional features of HEV vessels. However, only a minor proportion of HEVs (according to HEV markers) showed FasL expression in the normal mucosa. The finding of FasL expression in HEVs and near the lymphoid follicles is in concordance with our previous findings of its expression in lymph nodes, where vascular FasL was seen mostly in the HEV vessels in the paracortical areas of the activated lymph nodes (Kokkonen et al. 2004; Kokkonen and Karttunen 2009). Although some previous studies have reported a selective expression of FasL in endothelial cells, such as an upregulation in the endothelium overlying atherosclerosis (Imanishi et al. 2001), none of the previous studies on FasL expression in the intestinal mucosa comment on endothelial expression (Nagata and Golstein 1995; Nagata and Suda 1995; Pinkoski et al. 2000). There are several potential explanations for this discrepancy, including, but not limited to, unreliable FasL antibodies used in a majority of these studies. Many of the available FasL antibodies were assessed by Sträter et al. (1999). The group showed that, among the 12 tested antibodies, only the FasL antibody (G247-4) used in the present work was reliable. Since the expression intensity in the endothelial cells was mostly less than that in the positive lymphoid cells in the LP, methodological reasons, such as sensitivity of the detection system, and optimization of pretreatment and antibody dilution, may be additional explanations for this discrepancy. Finally, we presume that some of the previous studies may have ignored endothelial expression of FasL due to their focus on immune cells.

Blood vessels in the gut form a complex network that is capable of changing according to the needs of tissue nutrition, immunological activity, and other functional needs. Vascular density is higher around the lymphoid follicles, with vessels forming a basket-like network around the follicles. This basket-like network, consisting largely of HEV-like vessels, has been shown to be a location of lymphocyte extravasation (Azzali et al. 2008). Interestingly, here we demonstrate FasL expression in these vessels in the submucosa, but only rarely in the mucosa. Lymphoid follicles usually overlay the muscularis mucosae extending from the mucosa to the submucosa. We speculate that endothelial FasL expression in the submucosal side of these follicles is related to the main area of extravasation of lymphoid cells and even to the polar structure of the follicles, with overall migration of the cells from the contraluminal dark zone to the luminal light zone of the germinal centers (Hauser et al. 2007).

We show in this study, for first time, that endothelial FasL expression is elevated in IBD, but not in an acute form of mucosal inflammation, as in acute appendicitis, suggesting that the increase is a feature of chronic inflammation with an immunological background. Elevation of FasL-expressing vessel density was mostly noted in ulcers in samples from patients with CD and UC. Some patients showed extensive patterns of vascular FasL expression in all layers of the gut, as seen in the supplemental video (CD patient, FasL staining). We speculate that this is due to ulcers being specific spots of severe immunological activation, as well as sites of tissue healing. Vascular FasL was also significantly elevated in the submucosa in samples from patients with CD without any ulceration. This could be related to the well-documented deep or transmural inflammation characteristic of CD (Yantiss and Odze 2006). Interestingly, Cheng et al. (2010) showed that cultured human umbilical endothelial cells expressed Fas and FasL after treatment with nicotine. Since smoking is known to exacerbate the course of CD by increasing the risk of developing fistulas and strictures, we speculate that one potential mechanism for the effects of nicotine could be in promoting immunological activation by the enhancement of endothelial FasL expression. Another speculative mechanism potentially linking smoking with endothelial FasL and aggravation of the disease could be the activation of endothelial cell apoptosis by Fas/FasL-mediated mechanisms. Studies of polymorphisms of the FasL gene provide additional evidence concerning the role of FasL in the pathogenesis of IBD. Hlavaty et al. (2005) showed that, of patients carrying the FasL gene -843 CC/CT polymorphism, 75% responded to infliximab treatment, whereas for those with the FasL gene TT genotype, only 38% responded.

Elevated endothelial FasL was associated with elevated FasL single cell expression at the sites of chronic inflammation in samples from patients with CD and UC. Upon activation in the clonal expansion phase, T-cells express the Fas receptor, and FasL is expressed on the cell membrane. Although, resistance to Fas-mediated apoptosis remains for at least 48 hours following activation (Peter et al. 2007). These freshly activated lymphocytes exit lymph nodes via the lymphatics and spread into the circulation. These activated cells, which manage to drift to the site of inflammation within 48 hours, are resistant to Fas-mediated apoptosis and are not eliminated. We interpret our finding to fit with the concept that FasL-positive cells likely represent these activated cells.

Vascular FasL has been suggested to have a functional role in the regulation of lymphocyte extravasation of the vascular wall (Sata et al. 2001) and of lymph nodes by inducing apoptosis of activated Fas-expressing leukocytes (Kokkonen et al. 2004; Kokkonen and Karttunen 2009), which is similar to FasL on epithelial cells in immune-privilege organs. However, currently the concept of immune privilege and FasL’s role is controversial (Ferguson and Griffith 2006; Strasser et al. 2009). The activation of Fas in certain conditions leads to survival and a pro-inflammatory effect on T- and B-cells and not result in apoptosis (Peter et al. 2007). However, autoimmune lymphoproliferative syndrome (ALPS) caused by mutation in the gene coding for Fas or FasL emphasizes the importance of Fas/FasL-mediated apoptosis in homeostasis of the immune system. Patients with ALPS have a greatly elevated risk for autoimmune diseases and they have lymphadenopathy and splenomegaly caused by defects in Fas/FasL-mediated elimination of lymphoid cells.

In summary, we demonstrate here that endothelial FasL expression is present in functionally specific areas of the intestinal wall. FasL expression was rarely present in mucosa but was detected in a significant proportion of HEV and non-HEV vessels of the submucosal aspect of lymphoid follicles. This suggests that endothelial FasL has a role in lymphocyte transmigration from the blood to mucosal lymphoid tissue. Inflammation of the ileum and colon in samples from IBD patients was associated with an increase of endothelial FasL expression, but acute appendicitis did not affect endothelial FasL expression. Regulation of endothelial FasL expression and the biological role of such expression warrant mechanistic studies. Based on the location of specific vessels and its characteristic alteration in mucosal autoimmune diseases, we suggest that endothelial FasL expression in endothelial cells may play an important role in the regulation of mucosal immunity and in the pathogenesis of IBD.