Pattern of MUC6 expression across 119 different tumor types: A tissue microarray study on 15 412 tumors

Mucin 6 (MUC6) is a secreted gel‐forming mucin covering the surfaces of gastrointestinal and other tissues. Published work demonstrates that MUC6 can also be expressed in several cancer types and can aid in the distinction of different tumor entities. To systematically analyze MUC6 expression in normal and cancerous tissues, a tissue microarray containing 15 412 samples from 119 different tumor types and subtypes as well as 608 samples of 76 different normal tissue types was analyzed by immunohistochemistry. At least a weak MUC6 positivity was seen in 50 of 119 (42%) tumor entities. Thirty‐three tumor entities included tumors with strong positivity. MUC6 immunostaining was most frequent in mucinous carcinomas of the breast (44%), adenocarcinomas of the stomach (30%–40%) and esophagus (35%), and neuroendocrine carcinomas of the colon. Strong MUC6 staining was linked to advanced pT stage (p = 0.0464), defective mismatch repair status and right‐sided tumor location (p < 0.0001 each) in colorectal cancer, as well as to high tumor grade (p = 0.0291), nodal metastasis (p = 0.0485), erb‐b2 receptor tyrosine kinase 2 positivity (p < 0.0001) and negative estrogen receptor (p = 0.0332)/progesterone receptor (p = 0.0257) status in breast carcinomas of no special type. The broad range of tumor types with MUC6 expression limits the utility of MUC6 immunohistochemistry for the distinction of different tumor types.


INTRODUCTION
Mucin 6 (MUC6) is one of five secreted gel-forming mucins (MUC2, MUC5AC, MUC5B, MUC6, and MUC19) which are expressed from a gene cluster at chromosome 11p15. Their characteristic cysteine-rich regions enable oligomerization and formation of the mucin layer which protects the epithelial surfaces from chemical and mechanical stress and microbial pathogens. MUC6 is normally expressed by mucus-forming cells in the corpus and by the pyloric glands in the antrum of the stomach, by Brunner glands in the duodenum, the gallbladder, parts of the pancreas, and in male and female reproductive organs. 1 Abnormal MUC6 expression has also been reported from a variety of cancers, and there is growing evidence for a role in tumor development and progression. [2][3][4][5] For example, MUC6 alterations have been linked to tumor aggressiveness in cancers of the ovaries, 6 breast, 7 stomach, 8 esophagus, 9 colon, 10 pancreas, 11 and bile ducts. 12,13 In routine pathology, MUC6 immunohistochemistry (IHC) can offer additional diagnostic information for the distinction between gastric and intestinal metaplasia of the Barrett esophagus, 14 or between intraductal papillarymucinous, intraductal tubulo-papillary and intraductal oncocytic-papillary neoplasias of the pancreas. 15 Published studies on MUC6 expression in cancers have reported highly variable data. For example, immunohistochemical MUC6 positivity ranges from 7% to 85% in esophageal adenocarcinoma, 16,17 from 4% to 93% in cholangiocellular carcinoma, 18,19 from 24% to 64% in pancreatic ductal adenocarcinoma, 11,20 from 20% to 100% in cervical adenocarcinoma, 21,22 and from 0% to 92% in lobular breast carcinoma. 23,24 These discrepant data are likely due to the use of different antibodies and staining protocols and make it impossible to assess the potential diagnostic significance of MUC6 for the distinction of tumor entities.
This study was designed to provide a comprehensive analysis of MUC6 expression across a broad range of different tumor entities. For this purpose, >15 000 tissue samples from 119 different tumor types and subtypes, and 76 non-neoplastic tissues were evaluated by IHC in a tissue microarray (TMA) format under highly standardized conditions.

TMAs
TMAs composed of normal and tumorous tissues were employed for this study. The normal tissue TMA contained eight samples from eight different donors from each of 76 different normal tissue types. Normal tissues were obtained from donors who underwent surgery for reasons other than cancer. The cancer TMAs contained a total of 15 412 primary tumors from 119 tumor types and subtypes. Histopathological data including grade, pT and pN status was available from 524 ovarian cancers, 259 endometrial cancers, 598 pancreatic cancers, 327 gastric cancers, and 1475 breast cancers. The breast cancer dataset also included molecular information on estrogen receptor (ER), progesterone receptor (PR), erb-b2 receptor tyrosine kinase 2 (HER2) as well as follow-up information on a subset of 877 patients with a median follow-up time of 49 months (range, 1-88). The colon cancer TMA included data on lymphatic infiltration (L), mismatch repair (MMR) protein status, and rat sarcoma virus (RAS) mutation status. The composition of both normal and cancer TMAs is described in detail in the results section. All samples were from the archives of the Institutes of Pathology, University Medical Center Hamburg-Eppendorf, Germany, the Institute of Pathology, Clinical Center Osnabrueck, Germany, and Department of Pathology, Academic Hospital Fuerth, Germany. Tissues were fixed in 4% buffered formalin (10% dilution of saturated [38%] formalin solution) and then embedded in paraffin. The TMA manufacturing process was described earlier in detail. 25,26 TMA tissue spot diameter was 0.6 mm. The use of archived remnants of diagnostic tissues for manufacturing of TMAs and their analysis for research purposes as well as patient data analysis has been approved by local laws (HmbKHG, §12) and by the local ethics committee (Ethics Commission Hamburg, WF-049/09). All work has been carried out in compliance with the Helsinki Declaration.

IHC
Freshly cut TMA sections were immunostained on one day and in one experiment. Slides were deparaffinized and exposed to heat-induced antigen retrieval for 5 min in an autoclave at 121°C in pH 7.8 buffer. Primary antibody (MSVA-806R, rabbit recombinant, MS Validated Antibodies) was used for MUC6 detection. The antibody was applied at 37°C for 60 min at a dilution of 1:150. For the purpose of antibody validation, the normal TMA was also analyzed with the anti-MUC6 mouse monoclonal antibody CLH5 (Bio SB, cat. #BSB6171, ready-to-use) after antigen retrieval at pH 9 (Agilent/Dako Omnis Target Retrieval Solution High pH, cat. #K8004). Bound antibody was then visualized using the EnVision Flex reagent (Agilent #52023) according to the manufacturer's directions. For tumor tissues, the percentage of positive neoplastic cells was estimated, and the staining intensity was semiquantitatively recorded (0, 1+, 2+, 3+). For statistical analyses, the staining results were categorized into four groups. Tumors without any staining were considered negative. Tumors with 1+ staining intensity in ≤70% of tumor cells or 2+ intensity in ≤30% of tumor cells were considered weakly positive. Tumors with 1+ staining intensity in >70% of tumor cells, 2+ intensity in 31%-70%, or 3+ intensity in ≤30% of tumor cells were considered moderately positive. Tumors with 2+ intensity in >70% or 3+ intensity in >30% of tumor cells were considered strongly positive.

Statistics
Statistical calculations were performed with JMP 16 software (SAS Institute Inc.). Contingency tables and the χ 2 test were performed to search for associations between MUC6 immunostaining and tumor phenotype. Survival curves were calculated according to Kaplan-Meier. The log-rank test was applied to detect significant differences between groups. A p-value of ≤0.05 was considered as statistically significant.

Technical issues
A total of 11 685 (75.8%) of 15 412 tumor samples were interpretable in our TMA analysis. The remaining 3727 (24.2%) samples were not interpretable due to the lack of unequivocal tumor cells or a lack of the entire tissue spot. On the normal tissue TMA, enough samples (≥4) were always analyzable per tissue type to determine MUC6 staining patterns in individual cell types.

MUC6 in normal tissue
Strongest cytoplasmic MUC6 staining was seen in epithelial cells of seminal vesicles, Brunner glands of the duodenum, and the mucous secreting glands of the stomach. Positive staining was also consistently found in intercalated and interlobular ducts of the pancreas, in small juxtaportal bile ducts (but not in the portal bile ducts), in gallbladder surface epithelium (not in all samples), in the cauda (more intense) and caput (less intense and more focal) of the epididymis, in scattered cells of the fallopian tube, and (weakly) in endocervical glands. A few scattered MUC6 positive cells were also seen at least in some samples of collecting ducts of the kidney, breast glands, endometrium in pregnancy, and in the trophoblast of the first trimenon placenta. Representative images are shown in Figure 1. All these findings were observed by both MSVA-806R and CLH5 antibodies (Supporting Information: Supplementary Figure 1). Normal tissues without detectable MUC6 staining included skin and nonkeratinizing squamous epithelium of oral cavity, lip, tonsil surface and crypts, ectocervix, esophagus, stomach surface epithelium, colon mucosa, corpuscles of Hassall's in the thymus, mesenchymal tissues (myometrium, lymphatic and hematopoietic cells), hepatocytes, proximal and distal tubuli as well as glomeruli of the kidney, salivary glands, prostate, testis, respiratory epithelium, lung, ovary including corpus luteum and follicular cysts, mature placenta, adrenal gland, thyroid, cerebrum, cerebellum, adenoand neurohypophysis.  Table 1). The tumor categories with the highest rate of positive staining included mucinous carcinomas of the breast (44%), gastric adenocarcinomas (30%-40%), esophageal adenocarcinomas (35%), neuroendocrine carcinomas of the colon and gallbladder (33% each), endometrioid carcinomas of the ovary (29%), tubular carcinomas of the breast (29%), ductal adenocarcinomas of the pancreas (28%), endometrioid endometrial carcinomas (24%), carcinosarcomas of the ovary (24%), and lobular carcinoma of the breast (21%). A graphical representation of a ranking order of MUC6 positive and strongly positive cancers is given in Figure 3. The relationship between MUC6 positivity and histopathological tumor phenotype is shown in Table 2. Strong MUC6 staining was linked to advanced pT stage (p = 0.0464), defective MMR status, and right-sided tumor location (p < 0.0001 each) in colorectal cancer, as well as to high tumor grade (p = 0.0291), nodal metastasis (p = 0.0485), HER2 positivity (p < 0.0001) and a negative ER (p = 0.0332)/PR (p = 0.0257) status in breast carcinomas of no special type. MUC6 expression was unrelated to clinico-pathological features in serous ovarian carcinomas, endometrioid endometrial carcinomas, ductal adenocarcinomas of the pancreas, and in gastric adenocarcinomas.

DISCUSSION
More than 11 600 samples from 119 different tumor types and subtypes were successfully analyzed for MUC6 expression in our study. The analysis identified 51 tumor entities with at least occasional MUC6 MUC6 EXPRESSION ACROSS 119 DIFFERENT TUMOR TYPES | 283 expression. Among the 16 tumor entities for which MUC6 expression had not been previously reported there were several clinically relevant tumor types such as for example serous and endometrioid ovarian carcinomas, neuroendocrine tumors of the pancreas, small cell neuroendocrine carcinomas of the prostate and the urinary bladder, as well as testicular yolk sac tumors. Most MUC6 positive cancers showed a glandular differentiation. The highest prevalences of MUC6 expression were seen in carcinomas of the breast, adenocarcinomas of the stomach, esophagus, colorectum, and the pancreas, as well as in carcinomas of the endometrium and the ovary. It is of note that a considerable fraction of these tumor types is derived from epithelial tissues (i.e., colon mucosa, endometrium) that do not regularly express MUC6.
The ranking list of tumors according to their MUC6 positivity rate is an important result of this study. Given the maximal standardization of our experimental procedure and the rigorous validation of our reagents, we assume that this list reflects the relative importance of MUC6 expression for these tumors. Although the absolute positivity rates obtained in our study are specific for our assay, we expect that other protocols or the use of other highly specific antibodies would result in a comparable ranking order. The comparison of our data with previous data collected from the literature (Figure 4) demonstrates that a comparable ranking order could not be generated from existing literature data because of the very high variability of published results. These are reflective of the range of technical issues connected to such studies, including the use of different antibodies, IHC protocols and scoring strategies which all can massively impact the outcome of immunohistochenmical analyses. These observations also highlight the important role of large-scale studies involving many different tumor categories for assessing the utility of diagnostic immunohistochemical markers.
Our findings do not suggest a major role of MUC6 IHC for the distinction of tumor entities. The fact that none of the analyzed tumor categories showed MUC6 positivity in more than 40% of cases makes it clear that the absence of MUC6 immunostaining cannot exclude any tumor entity. In principle, MUC6 IHC would provide the highest level of diagnostic information with respect to tumor types that are always MUC6 negative. A positive MUC6 immunostaining would literally exclude such tumor entities from diagnostic considerations. In the present analysis, 59 of 119 analyzed tumor types and subtypes did not show MUC6 immunostaining. These for example included all different categories of squamous cell carcinomas, mesotheliomas, and hepatocellular carcinomas. Also, renal cell and thyroidal carcinomas showed MUC6 positivity in <1% of cases and are therefore considered unlikely sources for MUC6 positive metastases.
The existence of sizable subgroups of MUC6 positive and negative tumors within many clinically relevant cancer types raises the question of a possible prognostic or predictive role of MUC6 expression. The case numbers analyzed in this study were large enough to search for associations between MUC6 staining and features of tumor aggressiveness in six different tumor entities. That MUC6 positivity was linked to advanced stage in colorectal adenocarcinoma and to high Bloom-Richardson-Elston grade, nodal metastasis and molecular features that are related to aggressive tumor phenotype (ER negative, PR negative, and HER2 positive) in breast cancers of no special type demonstrates that high MUC6 expression can be linked to a more aggressive tumor behavior. As MUC6 is usually not expressed in normal colorectal epithelial cells, our findings suggest that a neo-expression of MUC6 can parallel tumor progression in at least some tumor types. Of note, Betge et al. 10 had previously analyzed MUC6 in 381 colorectal adenocarcinomas and found that high MUC6 expression was linked to long progression-free (p = 0.024) and cancer-specific survival (p = 0.043). Other authors had addressed the prognostic role of MUC6 expression in 225 gastric cancers, 8 101 pancreatic cancers, 11 85-100 cholangiocarcinomas, 12,13 73 salivary gland carcinomas, 27 and 36 ovarian cancers. 6 Most of these studies had suggested a better prognosis in patients with high MUC6 expression. Considering these controversial results and that we could not find any associations between MUC6 expression and histological features of aggressive tumor behavior in four further tumor    29 Given the large scale of our study, emphasis was placed on an adequate validation of our assay. According to the recommendations of the international working group for antibody validation (IW-GAV), we validated our approach by comparing our IHC findings in normal tissues with data obtained by another independent anti-MUC6 antibody and RNA data derived from two different publicly accessible databases. [30][31][32][33] The use of 76 different normal tissues for antibody validation ensured that a very large fraction of the proteins expressed in cells of adult humans were exposed to our antibody. Such a broad tissue validation increases the likelihood for detecting possible cross-reactivities. Validity of our assay was supported by the detection of MUC6 immunostaining in all organs with unequivocal MUC6 RNA expression (stomach, duodenum, seminal vesicles, pancreas, cervix uteri). The additional MUC6 stains obtained in epithelial cells of the epididymis, juxtaportal bile ducts of the liver, gallbladder epithelium, breast epithelium, endometrium of pregnancy, scattered cells of the fallopian tube, some collecting ducts of the kidney, and placental trophoblastic cells were confirmed by the use of the independent second antibody CLH5. That MUC6 RNA expression had not been described for these organs is probably due to the small number of F I G U R E 3 Ranking order of mucin 6 (MUC6) immunostaining in tumors. Both the frequency of positive cases (blue dots) and the frequency of strongly positive cases (orange dots) are shown.