Altered Expression of the Epithelial Mucin MUC1 Accompanies Endoscopic Recurrence of Post-operative Crohn’s disease

Abstract

Background:

MUC1-glycoprotein is expressed at low levels and in fully glycosylated form on epithelial cells. Inflammation causes MUC1 overexpression and hypoglycosylation. We hypothesized that overexpression of hypoglycosylated MUC1 would be found in postoperative Crohn’s disease (CD) recurrence and could be considered an additional biomarker of recurrence severity.

Methods:

We examined archived neo-terminal ileum biopsies from patients with prior ileocecal resection who had postoperative endoscopic assessment of CD recurrence and given a Rutgeerts ileal recurrence score. Consecutive tissue sections were stained using two different anti-MUC1 antibodies, HMPV that recognizes all forms of MUC1 and 4H5 that recognizes only inflammation-associated hypoglycosylated MUC1.

Results:

71 postoperative CD patients were evaluated. There was significant increase in MUC1 expression of both glycosylated/normal(p<0.0001) and hypoglycosylated/abnormal(p<0.0001) forms in patients with severe endoscopic CD recurrence(i3+i4), ileal score i2, compared to patients in endoscopic remission(i0+i1). Results were similar regardless of anti-TNF-α use. While MUC1 expression and Rutgeerts scores were in agreement when characterizing the majority of cases, there were a few exceptions where MUC1 expression was characteristic of more severe recurrence than implied by Rutgeerts score.

Conclusions:

MUC1 is overexpressed and hypoglycosylated in neo-terminal ileum tissue of patients with postoperative CD recurrence. Increased levels are associated with more severe endoscopic recurrence scores, and this is not influenced by anti-TNF-α use. Discrepancies found between Rutgeerts scores and MUC1 expression suggest that addition of MUC1 as a biomarker of severity of postoperative CD recurrence may improve categorization of recurrence status and consequently treatment decisions.

Introduction

Crohn’s disease (CD) is an incurable inflammatory bowel disease (IBD) that affects around 780,000 individuals in the USA [1, 2]. CD is characterized by chronic inflammation and progressive damage to the gastrointestinal tract. Over half of CD patients require ileocecal resection for complications of their disease [3]. Although intestinal resection may induce a surgical remission, histologic recurrence has been found to occur as early as 1 week postoperatively and most patients subsequently develop clinical recurrence [4]. Recurrence most commonly starts at the ileocolonic anastomosis and neo-terminal ileum, just proximal to the anastomosis [5]. Current guidelines recommend performing an ileocolonoscopy 6–12 months postoperatively to diagnose endoscopic recurrence [6, 7]. Subsequent colonoscopies are scheduled at 1 to 3 year intervals, depending on symptoms and results of most recent endoscopic evaluation. Rutgeerts et al. developed an endoscopic recurrence score that has been widely utilized by gastroenterologists to predict rates of future clinical and surgical recurrences [8]. This score aids in making therapy decisions. At time of ileocolonoscopy, patients are given a score between i0 and i4 based on the appearance of their neo-terminal ileum [8]. Most studies have designated patients with score of i0 or i1 as being in endoscopic remission. A score of i2, i3, or i4 implies endoscopic recurrence of respectively increasing severity. Patients in endoscopic remission (i0 and i1) have a lower likelihood of progressing to clinical recurrence or requiring a second operation within the subsequent 8 years, while those with endoscopic recurrence (i2-i4) will most likely require a repeat operation within 3–4 years [8–11].

MUC1 is a large transmembrane glycoprotein and a member of the mucin family of molecules [12]. Normal epithelial cells express the fully glycosylated form of MUC1 at low levels on the apical/luminal side of various epithelial cells. The extracellular domain of MUC1 is dominated by the variable number of tandem repeats region (VNTR), comprised of 20 amino acids that repeat on an average of 80–200 times per molecule [13]. Each repeat contains five potential O-linked glycosylation sites that in healthy epithelial cells are occupied by long branched sugars. When epithelial cells undergo malignant transformation or in cases of chronic inflammation such as IBD, there is a quantitative (overexpression) as well as a qualitative (hypoglycosylation) alteration in MUC1 expression [13–16]. The hypoglycosylated form of MUC1 exposes the polypeptide backbone which may either be lacking sugars completely or be glycosylated with short monosaccharide Tn or disaccharide T antigen or their sialylated forms, sTn and sT. Overexpression and changes in glycosylation terminate the state of immune ignorance and the abnormal form becomes immunogenic attracting cells of both innate and adaptive immune system. Expression of the abnormal form of MUC1 is characteristic of chronically inflamed epithelia, such as colonic and intestinal epithelium in IBD [14, 16, 17]. Once MUC1 expression changes (overexpression and hypoglycosylation), it continues to be a main driver of chronic inflammation and IBD progression [18, 19].

It has been demonstrated that abnormal MUC1 expression (overexpression and hypoglycosylation) is detected with early intestinal inflammation. Not only is abnormal MUC1 expressed as a consequence of inflammation, it also contributes to inflammation itself and has been shown to lead to colitis associated colon cancer in IBD mouse models [14]. Intervention with a vaccine against abnormal hypoglycosylated MUC1 administered early in life ameliorated subsequent progression to IBD and cancer in mice that spontaneously develop IBD [17, 20].

Our hypothesis is that this scenario similarly plays out in human IBD and in CD not only at initiation and progression of disease but also in disease recurrence. Since changes in MUC1 expression mark very early inflammation-related changes in the epithelium, hypoglycosylated MUC1 has the potential to be a highly sensitive diagnostic biomarker of early CD recurrence. Additionally, because hypoglycosylated MUC1 is also a promoter of inflammation, its presence would be expected to predict more severe recurrences. The goal of this study was to examine MUC1 expression in CD biopsies with pre-assigned endoscopic postoperative CD recurrence scores, which had not been done before, and determine if intensity and amount of expression and hypoglycosylation correlate with the severity of the endoscopic score. Our results support prospective studies to validate MUC1 expression as an additional and potentially more accurate diagnostic marker of early CD recurrence and progression. Additionally, our results support MUC1 as a potential target for immunotherapy of IBD.

Methods

Case Selection

Electronic medical records of adult patients who were evaluated at The University of Pittsburgh Medical Center outpatient gastroenterology clinic or ileocolonoscopy appointment between 6/2016 until 6/2017 were reviewed. Inclusion criteria selected patients with isolated terminal ileal CD who had a prior curative ileocolonic resection with primary ileocolonic anastomosis. Ileocolonoscopy using high-definition white light Olympus colonoscopes had to have been performed with punch biopsies from the neo-terminal ileum. Additionally, all patients had to have an assigned postoperative endoscopic Rutgeerts ileal score (i0 to i4) at the time of ileocolonoscopy.

Once the patients were identified, their archived neo-terminal ileum specimen paraffin blocks were obtained in order to stain for MUC1 expression.

Immunohistochemistry for MUC1 expression

Tissue paraffin sections (4–5 um thick) were deparaffinized by baking overnight at 59 degrees Celsius. Endogenous peroxidase activity was eliminated by treatment with 30% H₂O₂ for 15 minutes at room temperature. Antigen retrieval was performed by microwave heating in 0.1% citrate buffer. Nonspecific binding sites were blocked with 1% bovine serum albumin (BSA) for 20 minutes at room temperature. The anti-MUC1 antibody HMPV, which recognizes all forms of MUC1 by binding the epitope APDTR in the VNTR region in a glycosylation-independent manner, was purchased from BD PharMingen, San Diego, CA. The anti-MUC1 antibody VU-4H5 that recognizes the epitope APDTRPAP in the VNTR region of hypoglycosylated MUC1, was purchased from Santa Cruz Biotechnology, Santa Cruz, CA. The purified mouse IgG1 k isotype control was purchased from BD PharMingen, San Diego, CA. Anti-MUC1 antibody HMPV was diluted 1:200, anti-MUC1 antibody VU-4H5 was diluted 1:100, purified mouse IgG1 k isotype control was diluted 1:200, using 1% BSA in a protein-free blocking buffer (PBS, phosphate-buffered saline, ThermoFisher Scientific, Waltham, MA). Staining was performed by the avidin-biotin peroxidase complex method with a commercial kit (Vectastain ABC kit; Vector Laboratories, Burlingame, CA). Color development was performed using a 3,3’-diaminobenzidine kit (BD PharMingen, San Diego, CA). Quantitating MUC1 staining intensity and assigning a relative score was done by two investigators blinded to the patients’ clinical status and endoscopic Rutgeerts score. Each case was first semiquantitatively scored from 0 to 4 based on the percent of cells staining positive relative to the total number of cells in the area scored: 0 = no stain, 1 = 1–10%, 2 = 11–25%, 3 = 26–40%, 4 ≥ 41%. Each case was then scored again for intensity of staining from 1(light brown) to 4(dark brown). The average of these two values was used to calculate the total MUC1 expression score. To date there are no established guidelines for reporting MUC1 immunohistochemistry results. We are following a method that reflects percent of positive cells as well as intensity of staining and our method is similar to several reports in the literature [21, 22].

Statistical Analysis

Data were entered into Statistical Package for Social Sciences (SPSS) version 23, which was used for data cleaning, management, and analyses. Categorical variables were presented as number and percent, whereas continuous variables were presented as mean and standard deviation. Association between Rutgeerts ileal score and continuous variables was assessed using ANOVA test, whereas association with categorical variables was assessed using Chi-square test or Fishers exact test, as deemed appropriate. A p-value of <0.05 was used to indicate statistical significance.

Ethical Considerations

All study material and data were kept confidential. This study was approved by the University of Pittsburgh’s Institutional Review Board (PRO16030106).

Results

Sample Characteristics

Seventy-one postoperative CD patients were include in this study (Table 1). Mean age of the patients was 43.01±13.11 years and 60.6% were female. Twenty-six patients (36.6%) were receiving anti-tumor necrosis factor (TNF)-α medications at time of neo-terminal ileum biopsy and 32.4% were receiving immunomodulatory drugs (thiopurines or methotrexate).

Table 1.

Patient characteristics and histologic grading of postoperative CD neo-terminal ileum biopsies

		Endoscopic Recurrence Score (Rutgeerts ileal score)
		i0 (n=21)	i1 (n=14)	i2 (n=17)	i3 (n=9)	i4 (n=10)	p-value
Gender	Male	11 (52.4)	8 (57.1)	3 (17.6)	3 (33.3)	3 (30.0)	0.12
	Female	10 (47.6)	6 (42.9)	14 (82.4)	6 (66.7)	7 (70.0)
Age	Mean (±SD)	42.38 ± 12.70	43.14 ± 15.41	42.24 ± 12.22	50.44 ± 15.24	38.80 ± 9.31	0.49
IMM use	n (%)	10 (47.6)	4 (28.6)	4 (23.5)	3 (33.3)	2 (20.0)	0.50
Anti-TNF-α use	n (%)	14 (66.7)	3 (21.4)	4 (23.5)	0 (0.0)	5 (50.0)	0.002
Steroid use	n (%)	0 (0.0)	1 (7.1)	0 (0.0)	0 (0.0)	2 (20.0)	0.054

SD: standard deviation; ICR: ileocecal resection; IMM; immunomodulator; Anti-TNF-α; anti-tumor necrosis factor-α

Patients were classified into three groups: (1) endoscopic remission with a Rutgeerts ileal score of i0 or i1 (n=35); (2) mild endoscopic recurrence with a Rutgeerts ileal score of i2 (n=17); and (3) more severe recurrence with a Rutgeerts ileal score of i3 or i4 (n=19).

MUC1 expression

Figure 1 shows representative images of different degrees of MUC1 expression, both normal glycosylated and abnormal hypoglycosylated forms, in neo-terminal ileum samples from patients with indicated Rutgeerts endoscopic recurrence scores. MUC1 expression score was assigned to all samples as described in the Methods section and compared between different endoscopic scores (Figure 2). Patients with severe endoscopic recurrence (i3 and i4) showed statistically significantly higher levels of total MUC1 expression as detected by HMPV antibody compared to patients with mild endoscopic recurrence (i2) or in endoscopic remission (3.32±0.82 vs 2.44±1.01 vs 1.93±1.06, respectively; p<0.0001). Similarly, there was a significantly higher expression of the hypoglycosylated MUC1 as detected by 4H5 antibody in patients with more severe recurrence compared to those with mild recurrence or those in endoscopic remission (2.16±1.15 vs 1.41±1.21 vs 0.56±1.00, respectively; p<0.0001). Amongst these three groups of patients, there was no significant difference between the patients mean age (p=0.89), the time period between ileocecal resection and time of colonoscopy for neo-terminal ileum biopsy (p=0.16), steroid use (p=0.31), immunomodulator use (p=0.43), or use of anti-TNF-α (p=0.43). There was, however, a difference in gender distribution amongst the 3 groups (p=0.03), the significance for which is not clear. Females accounted for 45.7% of patients in remission (i0+i1), compared to 82.4% of patients with an i2 recurrence score, and 68.4% of those with more severe recurrence (i3+i4). This difference was not significant when patients were compared based on their postoperative recurrence score, i0 vs i1 vs i2 vs i3 vs i4, as demonstrated in Table 1.

Figure 1:

MUC1 expression on neo-terminal ileum biopsies representing each postoperative Rutgeerts score. HMPV antibody stains normal glycosylated MUC1 while the 4H5 antibody stains the abnormal hypoglycosylated MUC1. Scores of i0 and i1 designate patients in endoscopic remission, i2 with mild endoscopic recurrence, while scores of i3 and i4 reflect more severe endoscopic recurrence. Images are taken with 10x and 40x objectives.

Figure 2:

Expression of normal glycosylated and abnormal hypoglycosylated MUC1, as measured by HMPV and 4H5 antibodies, respectively, on neo-terminal ileum tissue of CD patients in endoscopic remission (i0 + i1), mild endoscopic recurrence (i2), and those with more severe endoscopic recurrence (i3 + i4). Each dot represents a single patient.

When MUC1 expression was compared between each Rutgeerts score separately, there still was a statistically significant difference amongst the groups. MUC1 staining using HMPV was 1.67±1.13 for i0, 2.32±0.85 for i1, 2.44±1.01 for i2, 3.72±0.44 for i3, and 2.95±0.93 for i4; p< 0.0001. Similarly, there was a significant difference between the groups when stained with 4H5: 0.50±0.92 for i0, 0.64±1.13 for i1, 1.41±1.21 for i2, 2.44±1.21 for i3, and 1.90±1.10 for i4; p<0.0001.

Anti-TNF-α use at the time of biopsy

Because MUC1 expression is known to be regulated by TNF-α [23–25], we evaluated MUC1 expression separately in patients receiving anti-TNF-α agents at time of their neo-terminal ileum biopsies (Figure 3) versus not (Figure 4). Of 26/71 patients on anti-TNF-α, 14 had Rutgeerts ileal score of i0, 3 had a score of i1, 4 a score of i2, none had a score of i3, and 5 had a score of i4. When samples from these 26 patients were analyzed for MUC1 according to their endoscopic recurrence score (endoscopic remission vs i2 vs more severe endoscopic recurrence), the statistically significant difference in the average MUC1 score was still there, both as detected by HMPV (p=0.02) and 4H5 (p=0.04), with higher MUC1 scores associated with more severe endoscopic scores. Of 45 patients not on anti-TNF-α therapy (Figure 4), 7 had Rutgeerts ileal score of i0, 11 had a score of i1, 13 a score of i2, 9 a score of i3, and 5 had a score of i4. In this group as well, the statistically significant difference in the average MUC1 expression remained, both as detected by HMPV (p<0.0001) and 4H5 (<0.0001), with higher levels of expression in more severe scores. Amongst the different groups, there was no difference between the mean age of patients (p=0.19), time between ileocecal resection and time of neo-terminal ileum biopsy (p=0.10), steroid use (p=0.35) or use of other immunosuppressive therapy (p=0.12). When comparing neo-terminal ileum tissue of patients with an endoscopic score of i2 with patients who had a more severe endoscopic recurrence score (only i4 patients in this subgroup), there was no statistically significant difference detected between the HMPV-stained and 4H5-stained MUC1 levels. This subgroup analysis is somewhat compromised because of the small sample size and the lack of i3 patients in this sample.

Figure 3:

Expression of normal glycosylated and abnormal hypoglycosylated MUC1, as measured by HMPV and 4H5 antibodies, respectively, on neo-terminal ileum tissue of CD patients receiving anti-TNF-α therapy. Patients were grouped into those in endoscopic remission (i0 + i1), those with mild endoscopic recurrence (i2), and those with more severe endoscopic recurrence (i3 + i4). Each dot represents a single patient.

Figure 4:

Expression of normal glycosylated and abnormal hypoglycosylated MUC1, as measured by HMPV and 4H5 antibodies, respectively, on neo-terminal ileum tissue of CD patients not receiving anti-TNF-α therapy. Patients were grouped into those in endoscopic remission (i0 + i1), those with mild endoscopic recurrence (i2), and those with more severe endoscopic recurrence (i3 + i4). Each dot represents a single patient.

As above, there was no difference in the mean age of patients (p=0.69), time between ileocecal resection and time of neo-terminal ileum biopsy (p=0.75), steroid use (p=1.00) or with the use of other immunosuppressive therapy (p=1.00), amongst the three groups of patients. Here, there was no difference in gender distribution amongst the three groups of patients (p=0.13).

Discrepancy between MUC1 expression and endoscopic postoperative score

In addition to the majority of biopsies with higher Rutgeerts scores showing higher MUC1 expression and hypoglycosylation, we found in several cases a potentially diagnostically and prognostically important discrepancy between MUC1 expression and the endoscopic score. This discrepancy was more apparent for the hypoglycosylated MUC1 (4H5 antibody staining), where the MUC1 staining would suggest a more advanced disease recurrence that would have warranted a higher endoscopic score. Figure 5 displays the neo-terminal ileum tissue from Patients 2 and 3 that were both given endoscopic score i2. Based on the MUC1 expression, however, much higher staining intensity of Patient 3 tissue suggests a more severe recurrence, more consistent with higher endoscopic scores. In fact, Figure 5 shows that expression of hypoglycosylated MUC1 (stained by 4H5) in the neo-terminal ileum tissue of Patient 3 (i2) is at the level seen in patients with a Rutgeerts score of i3. This discrepancy is also displayed in Figure 2 amongst the patients in the endoscopic remission group. In that group, there were 4 patients who had overexpression of the hypoglycosylated MUC1 higher than seen or expected for patients in remission and more consistent with a score between i2 and i4. Similarly, there were 3 patients in the i2 group whose MUC1 expression pattern was similar to patients with a more severe recurrence score (i3/i4).

Figure 5:

Representative examples of discrepancies in MUC1 expression scores and endoscopic scores amongst postoperative CD patients. Representative immunostaining of neo-terminal ileum biopsies of patients with different endoscopic scores i1, i2, i3, and i4. HMPV antibody stains the normal fully glycosylated MUC1; 4H5 antibody stains abnormal (hypoglycosylated) MUC1. Images are taken with 20x objective.

Discussion

In this first observational study of MUC1 expression in human postoperative CD, we show that the abnormal hypoglysylated form of MUC1, previously associated with colon cancer and other adenocarcinomas and more recently with IBD, is expressed in the neo-terminal ileum of CD patients with postoperative recurrence and might be considered for future evaluation as an additional biomarker of recurrence. We show that MUC1 expression increases and glycosylation decreases with increasing severity of recurrence as diagnosed by the widely utilized Rutgeerts postoperative endoscopic ileal recurrence score. There were, however, cases that we consider noteworthy, where MUC1 expression did not directly match the endoscopic score (Figures 2 and 5). These cases suggest that hypoglycosylated MUC1 overexpression might be more sensitive than the postoperative CD recurrence endoscopic score for diagnosing recurrence. Currently, endoscopic recurrence scores dictate a patient’s treatment regimen and frequency of surveillance with the goal of prevention of recurrence rather than treatment of recurrent active disease. Our results suggest that abnormal MUC1 expression scores in combination with the Rutgeerts postoperative endoscopic score could allow development of a more accurate model for prediction of post-operative CD recurrence.

The highest expression of both normal glycosylated MUC1 and abnormal hypoglycosylated MUC1 was found in biopsies scored i3 while patients with an endoscopic score of i4 showed lower MUC1 expression. This is most likely due to the severe epithelial tissue damage in more severe recurrence stage (ie score i4) that lowers its ability to produce MUC1.

Alterations in MUC1 glycosylation and level of expression affect its cellular localization and function, and have a significant impact on its role in perpetuating inflammation. The tandem repeat region of MUC1 upregulates the expression of the inflammatory cytokines IL-6 and TNF-α by binding to their promoter regions in an NFkB p65-dependent manner [18]. In addition, TNF-α has been shown to modulate the expression of MUC1 through NF-κB [23–25]. This positive feedback loop between MUC1 and TNF-α is especially important in CD and postoperative recurrence of CD because anti-TNF-α therapies (such as infliximab, adalimumab, and certolizumab pegol) have shown promise for prevention of postoperative CD recurrence and treatment of CD. It is likely that some of the benefit seen from anti-TNF-α therapy is indirect, due to its ability to prevent overexpression of the hypoglycosylated (abnormal form) MUC1 as long as it is applied early in recurrence. Continued high-level expression of MUC1 in patients on anti-TNF-α therapy may be an early biomarker of resistance to anti-TNF-α therapy, alerting in a more timely manner to the need to switch to another treatment for optimal disease control in these patients. Prospective studies will be needed to determine if MUC1 expression can identify patients destined to fail anti-TNF-α therapy.

In addition, if further studies validate these preliminary observations, MUC1 could convert from a biomarker of postoperative CD recurrence to a target for prevention of recurrence. In our preclinical mouse models of IBD, overexpression of hypoglycosylated MUC1 induces migration of neutrophils to chronically inflamed colon and neutrophils where they produce pro-inflammatory cytokines that in turn cause increase in MUC1 expression. This feedback loop suggests that although MUC1 might not be an initial inducer of IBD, it perpetuates chronic inflammation. Similarly, MUC1 expression in CD, if not prevented or lowered would be expected to promote inflammation and speed up recurrence. In preclinical IBD models, MUC1 expression could be prevented by a hypoglycosylated MUC1 vaccine given early in disease, which led to slowing down of IBD and prevention of progression to colitis associated colon cancer by elimination of hypoglycosylated MUC1 expressing (diseased) cells [17, 20]. One could invision a similar vaccine given to CD patients postoperatively to elicit immunity against hypoglycosylated MUC1 expressing cells that would signal CD recurrence.

This study is an observational study for the initial proof-of-concept of our hypothesis and as such it has the usual caveats including the retrospective nature and relatively small sample size. One specific limitation of all studies employing tissue biopsies is the potential that the disease site was not adequately sampled, especially important for patients with a postoperative ileal score of i1 or i2, who tend to have patchy disease. Although it is our practice to biopsy non-ulcerated tissue when possible, given the retrospective nature of this study, uniform sampling amongst patients remains a limitation. Other caveats include the heterogeneous population based on immunosuppressive medications and time from ileocecal resection. In spite of these limitations, this first study achieved its goal to show abnormal MUC1 expression in recurrent CD and a positive association between abnormal MUC1 expression and presence of endoscopic recurrence in postoperative CD patients. Additionally, discrepancies between postoperative endoscopic CD scores and abnormal MUC1 expression support further studies to validate hypoglycosylated MUC1 as a diagnostic marker of early postoperative recurrence and prognostic marker for progression of recurrent disease, and a future target for prevention of recurrence.