In experimental models, mucin-depleted foci (MDF), formed by dysplastic crypts devoid of mucin production have been recognized to be correlated with colorectal carcinogenesis and to serve as preneoplastic lesions of colorectal cancer (CRC). In humans, there is only one report of identification of MDF in patients with familial adenomatous polyposis and CRC; however, the histological characteristics of human MDF are not discussed extensively in the report. In the present study, colonic samples from 53 patients with sporadic CRC were stained with Alcian blue and examined for the presence of MDF. Subsequently, the samples were examined for the presence of aberrant crypt foci (ACF) by methylene blue staining. We classified MDF into two categories: flat-MDF and protruded-MDF (having the characteristics of both ACF and MDF). We found a total of 354, 41 and 19 colonic mucosal lesions with a mean multiplicity of 44, 38.9 and 66.9 crypts (ACF, flat-MDF and protruded-MDF, respectively). The density of MDF was 0.0082 lesions/cm2. The ACF identified in sporadic CRC patients corresponded to hyperplastic or non-dysplasic lesions. However, MDF identified in these patients corresponded to low-grade dysplasia. In addition, we found that Paneth cell metaplasia and inflammatory cell infiltration were specific histological features of MDF. These histological characteristics are reported to be associated with the development of CRC. Therefore, our results indicate that MDF might represent preneoplastic lesions in human colorectal carcinogenesis. (Cancer Sci 2012; 103: 144–149)
Although endoscopic screening and removal of adenomas can reduce the incidence of colorectal cancer (CRC),(1) CRC remains one of the most common causes of cancer-related death in developed countries.(2) Recently, the strategy of cancer prevention using non-toxic chemical entities, termed chemoprevention, has attracted much attention. Adenomas have been established as premalignant lesions of CRC and are characterized by the presence of genetic and histological changes; they have thus been used as the efficacy endpoint in chemoprevention trials.(3) However, such an endpoint requires hundreds of subjects and a very long observation period; therefore, more useful surrogate biomarkers are needed.
In experiments on colorectal carcinogenesis, the earliest identifiable precursors of malignant lesions are aberrant crypt foci (ACF). ACF were discovered as the earliest microscopic lesions to appear in the colonic mucosa of mice treated with azoxymethane.(4) Since they were initially described, numerous studies of ACF in animal models have demonstrated that ACF are precursors of CRC and have led to their widespread use as surrogate biomarkers of colorectal carcinogenesis.(5–9) Thus, in rodent models, ACF have been established as precursor lesions of CRC. In humans also, ACF have been identified in the colonic mucosa, in vivo, using methylene blue staining.(10,11) Although some researchers have reported a strong correlation between human ACF and the presence both of synchronous adenomas and/or carcinomas,(11–13) others have reported a lack of such correlation.(14) The validity of ACF as surrogate endpoints and/or precursors of CRC has not been established in humans; therefore, additional biomarkers that would show a robust correlation with CRC are needed.
Caderni et al.(15) identify mucin-depleted foci (MDF), formed by dysplastic crypts devoid of mucin production, in the colons of carcinogen-treated rodents. Some chemopreventive studies in rodent models demonstrate a correlation between the presence of MDF and the development of CRC, and suggest that MDF might, therefore, serve as surrogate biomarkers of CRC.(16–18) Although MDF is reported to be easily identifiable in unsectioned colons of rodents treated with colon-specific carcinogens, there has been only one report of identification of MDF in the human colon. Femia et al.(19) identify MDF in the unsectioned colons of patients with familial adenomatous polyposis (FAP), and at a much lower density in patients with sporadic CRC. In their study, only one MDF is identified among all the sporadic CRC patients examined; therefore, the histological characteristics of these lesions could not be adequately analyzed. Additional study is needed to determine the relationship between MDF and colorectal carcinogenesis in humans. We examined the surrounding colonic tissues in resected specimens of CRC, because patients with resected CRC have been reported to be at a higher risk of a second CRC,(20) and it is easily conceivable that they carry preneoplastic lesions in the colonic mucosa surrounding the resected CRC. In the present study, we attempted to identify MDF in a large sample of patients with sporadic CRC and analyzed the histological characteristics of these colonic mucosal lesions. Moreover, we also examined the histological characteristics of MDF identified in the colon of 1 FAP patient in comparison with those identified in the colons of the sporadic CRC patients, just as in the report by Femia et al.(19) The main aim of the present study was to evaluate whether MDF might serve as preneoplastic lesions of CRC in humans. Although it is difficult to validate MDF as preneoplastic lesions of CRC by only histological investigation, this study might represent the first step in this process.
Materials and Methods
Clinical materials. This study was approved by the institutional review board for research ethics at the University of Ryukyus and the Yokohama City University School of Medicine. Colonic samples from 53 patients who underwent resection for CRC were obtained from the University of Ryukyus Hospital in 2006–2009. Colonic tissues at least 5-cm away from the original neoplastic lesion were subjected to histopathological examination. None of the patients had any familial risk factors for this disease based on the clinical manifestations, age at onset or family history. A summary of the clinical data of the patients enrolled in the study is presented in Table 1. A surgically resected colonic specimen was also obtained from one FAP patient (36 year-old male) at the Yokohama City University Hospital. Written informed consent for use in the research was obtained from all the patients.
Table 1. Characteristics of the studied sporadic colorectal cancer patients
Mean age* (range)
Location of colon evaluated
Area examined cm2/patient* (range)
Prevalence of ACF/flat-MDF/protruded-MDF, %
A, ascending colon; D, descending colon; R, rectum; S, sigmoid colon; T, transverse colon. All colonic samples evaluated were from patients with sporadic CRC who had no familial risk factors for this disease. *Values are means ± SD.
63.8 ± 12.3 (33–88)
Right colon (17): A (9), T(8)
139.1 ± 81.7 (42–363)
Left colon (36): D (3), S (9), R (24)
Identification of aberrant crypt foci and mucin-depleted foci in the unsectioned colon. Using scissors, microscopically normal mucosa was gently peeled from the muscularis propria. Subsequently, the peeled mucosa was stretched for the staining. The area (cm2) of each sample was calculated using a ruler after it was pinned onto a polystyrene board. The staining for identification of the colonic mucosal lesions was performed as previously described.(21) We previously reported that Alcian blue (AB) staining is useful and convenient for detecting both ACF and MDF.(21) In brief, the removed colonic mucosa was stained with a 1% solution of AB, pH 2.5 (Sigma Chemical, St Louis, MO, USA), in 3% acetic acid for 5 min and immediately washed with distilled water. Subsequently, after MDF detection, the colonic mucosa was stained with 0.2% methylene blue solution to identify ACF. The stained colonic mucosal segments containing both ACF and MDF were photographed at low magnification under a light microscope. The number and multiplicity (number of crypts/mucosal lesion) ACF and MDF were determined.
Mucin-depleted foci were identified as focal lesions characterized by loss of AB staining, attributable to the loss or reduction of mucin, as compared with the surrounding normal crypts (Fig. 1). ACF were defined as lesions with larger crypts and showing darker staining with methylene blue than normal crypts, often with oval or slit-like lumens, and a thicker epithelial lining (Fig. 1).
Histological examination of the identified colonic mucosal lesions. After identification and measurement of the MDF and ACF, each lesion was dissected and embedded in paraffin for histological analysis. Each section was basically examined using an en face preparation, and, in addition, some lesions were also examined in longitudinal preparations. The 4-μm-thick sections were stained with H&E and evaluated by at least two pathologists. Immunohistochemical staining for MUC2 protein (Santa Cruz Biotechnology, Santa Cruz, CA, USA) was also performed on each of the serial HE-stained sections, using the universal immuno-enzyme polymer method with microwave activation (Simple Stain MAX PO M kit; Nichirei, Tokyo, Japan).
Identification of aberrant crypt foci and mucin-depleted foci in the colon of sporadic colorectal cancer patients. Colonic samples were stained with AB, pH 2.5, to identify the presence of MDF, and subsequently stained with methylene blue to identify the presence of ACF. As expected, both ACF and MDF were easily visualized. MDF could be distinguished into two categories: flat-MDF and protruded-MDF (Fig. 1F and K, respectively). The characteristics of these colonic lesions are shown in Table 2. We found a total of 354, 41 and 19 colonic mucosal lesions with a mean multiplicity of 44, 38.9 and 66.9 crypts (ACF, flat-MDF and protruded-MDF, respectively). In this study, the density of MDF (flat-MDF and protruded-MDF) was 0.0082 lesions/cm2, which is higher than that reported in sporadic CRC patients (0.0006 lesions/cm2).(19) The multiplicity of crypts in the protruded-MDF was significantly higher than that in any of the other lesions (P = 0.02). In regard to the locations of these colonic lesions, the density of ACF was significantly higher in the left colon than that in the right colon (P = 0.011). This trend was similar to the epidemiology of CRC.(2) In contrast, the density of flat-MDF and protruded-MDF were higher in the left colon than in the right colon, although the differences were not statistically significant (P = 0.063 and P = 0.61, respectively).
Table 2. Characteristics of the colonic mucosal lesions in colorectal cancer patients
Right colon: ascending colon and transverse colon. Left colon: descending colon, sigmoid colon, and rectum. Values are mean ± SD. *P = 0.02 calculated by Kruskal–Wallis test. †The differences in the density of the colonic mucosal lesions between the right colon and left colon was calculated by the Mann–Whitney U-test (P =0.011, P =0.063 and P =0.61, ACF, flat-MDF and protruded-MDF, respectively).
6.7 ± 6.9
0.8 ± 1.5
0.4 ± 0.9
44.0 ± 55.7
38.9 ± 62.7
66.9 ± 82.6
0.054 ± 0.058
0.007 ± 0.017
0.003 ± 0.009
0.023 ± 0.034
0.002 ± 0.006
0.001 ± 0.003
0.067 ± 0.061
0.009 ± 0.019
0.004 ± 0.010
Histological analysis of the identified aberrant crypt foci and mucin-depleted foci in the colons of sporadic colorectal cancer patients. The identified flat-MDF (41 lesions), protruded-MDF (19 lesions) and the representative sample of ACF (20 lesions) were sectioned for histopathology. Three different types of crypt foci were identified in the HE-stained sections. ACF were characterized by larger crypts and wider lumens, as compared with the surrounding crypts, corresponding to hyperplastic or non-dysplasic lesions (Fig. 1C–E). In contrast, the flat-MDF did not show nuclear stratification or loss of polarity, but showed Paneth cell metaplasia and decrease/loss of goblet cells, indicative of low-grade dysplasia (Fig. 1H–J). Protruded-MDF showed the features of both ACF and MDF, also corresponding to low-grade dysplasia (Fig. 1M–O). We also detected inflammatory cell infiltration as a specific histological feature of MDF (Fig. 1J,O). Immunohistochemical staining for MUC2 revealed that MDF were depleted of MUC2 expression, the most abundant type of mucin in the colon (Fig. 2).
Examination of mucin-depleted foci in the familial adenomatous polyposis patient. The density of MDF in the FAP patient was 0.068 lesions/cm2, which was much higher than that in the sporadic CRC patients in this study, similar to Femia’s report. All the MDF (eight lesions) from the FAP patient were flat-MDF, and showed slight nuclear stratification and loss of polarity, and Paneth cell metaplasia, corresponding to moderate-grade dysplasia (Fig. 3).
Although previous published studies reveal that some ACF show dysplastic changes and might serve as surrogate biomarkers of colorectal carcinogenesis, the majority of ACF in CRC patients are not dysplastic.(11,22–27) Consistent with these reports, we confirmed that representative samples of ACF from sporadic CRC patients corresponded to hyperplastic or non-dysplastic lesions. In contrast, protruded-MDF, which have the characteristics of both ACF and MDF, showed changes corresponding to low-grade dysplasia. In addition, we revealed that protruded-MDF have a larger number of crypts than ACF. ACF with a large number of crypts have been reported to show a stronger correlation with the presence of CRC than those with a small number of crypts.(11) Thus, the results in the present study suggest that protruded-MDF might, at least in part, correspond to dysplastic ACF and probably represent more advanced lesions in colorectal carcinogenesis.
In our study, the density of MDF in sporadic CRC patients was much higher than that reported from a previous study.(19) We analyzed larger areas of the colorectum, which could partially explain our finding a higher number of MDF. However, this is not sufficient to explain the discrepancy; another reason might be differences in the characteristics of the study participants, such as race and age. For example, ACF are reported to be detected at a higher incidence in Japanese patients than in American patients.(11,14,27,28) Yet another reason might be the difference in the method of detection used between the two studies. MDF were originally identified using a stain for more stringent mucin staining; namely, high iron diamine-AB.(19) In experimental models, we previously showed that approximately one-quarter of the MDF identified with AB staining were actually normal-like crypts, and approximately 10% of MDF identified with AB staining corresponded to classical ACF with dysplasia.(21) MDF identified in the present study might contain these lesions, and this could explain the higher density of MDF in our study.
Concerning the histological characteristics of MDF, MDF identified in sporadic CRC patients showed changes corresponding to low-grade dysplasia. Paneth cell metaplasia, which has been shown to be correlated with the presence of CRC,(29) was also observed in both flat-MDF and protruded-MDF; however, no nuclear stratification or loss of polarity was observed. In contrast, MDF identified in the FAP patient showed slight nuclear stratification and loss of polarity, corresponding to moderate-grade dysplasia. Indeed, our sample size was too small to conclude the histopathological characteristics of MDF in FAP patients; FAP patients carry germ-line mutations in the APC, and, therefore, accumulated gene mutations might promote progression from MDF to an advanced stage of colorectal carcinogenesis in humans. Furthermore, we found inflammatory cell infiltration as a specific histological feature of MDF. Although the significance of this inflammatory cell infiltration is not clear, it might be related to immune defenses or recognition of the new altered cell population. The most abundant mucin, MUC2, has been reported to be depleted in the colonic MDF in carcinogen-treated rats,(30) and it has been suggested that the lack of MUC2 protein activates inflammatory markers in MDF.(31) It has also been reported that genetically MUC2-deficient mice develop CRC spontaneously, as well as colitis.(32,33) In the case of humans, several studies have revealed diminished MUC2 mRNA expression in CRC patients.(34–36) These results indicate that the lack of a protective layer of mucus might activate local inflammation, and contribute to further progression of MDF to a more advanced stage of colorectal carcinogenesis in humans.
Although our study indicates that MDF might serve as preneoplastic lesions not only in experimental models, but also in humans, there ae some limitations. AB staining with acidic solution is unsuitable for in vivo endoscopic identification of MDF in humans. In addition, the low prevalence of MDF in human sporadic CRC patients means that it is difficult to use them as surrogate biomarkers of CRC. Therefore, we could not show clinical usefulness of MDF immediately. However, MDF might be useful for select high-risk subjects who need frequent surveillance. In the future, to evaluate whether MDF have a robust correlation with future CRC development, we need to establish a method of endoscopic identification of these lesions with future technical progress of in vivo visualization or particular mucosal staining that enables the detection of MDF. Moreover, in this study, we did not evaluate the genetic and molecular alterations in human MDF. In experimental models, MDF has been reported to show nuclear stratification and loss of polarity,(21) and activation of Wnt signaling, driven in part by mutations of the β-catenin and Apc genes.(17,37) In contrast, human MDF identified in sporadic CRC patients does not show nuclear stratification and loss of polarity. The primary mechanisms of colorectal carcinogenesis in carcinogen-treated rodent models are the frequent gene mutations of β-catenin and altered cellular localization of the protein.(38,39) However, stabilizing mutations of β-catenin in the absence of mutations of the APC have been described in only approximately 7% of sporadic CRC patients.(40) This might have an important implication in the different histological characteristics of MDF between humans and experimental rodent models. To elucidate the relevance of human MDF in colorectal carcinogenesis, further study of the underlying genetic and molecular alterations is needed.
In conclusion, we confirmed that MDF can also be identified in sporadic CRC patients just as in carcinogen-treated rodent models. We demonstrated that these lesions show histological features corresponding to low-grade dysplasia and are characterized by the specific histopathological characteristics of Paneth cell metaplasia and inflammatory cell infiltration. These characteristics have been demonstrated to be associated with an increased risk of CRC development. Thus, MDF in humans might represent an early-step histological change in the process of colorectal carcinogenesis, and could probably serve as preneoplastic lesions of CRC.
This work was supported in part by a Grant-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan and a Grant-in-Aid from the Ministry of Health, Labour and Welfare.