|Home | About | Journals | Submit | Contact Us | Français|
In certain regions of the body, transition zones exist where stratified squamous epithelia directly abut against other types of epithelia. Certain transition zones are especially prone to tumorigenesis an example being the anorectal junction, although the reason for this is not known. One possibility is that the abrupt transition of the simple columnar epithelium of the colon to the stratified squamous epithelium of the proximal portion of the anal canal may contain a unique stem cell niche. We investigated whether the anorectal region contained cells with stem cell properties relative to the adjacent epithelium. We utilized a tetracycline-regulatable histone H2B-GFP transgenic mice model, previously used to identify hair follicle stem cells, to fluorescently label slow-cycling anal epithelial cells (e.g., prospective stem cells) in combination with a panel of putative stem cell markers. We identified a population of long-term GFP label-retaining cells concentrated at the junction between the anal canal and the rectum. These cells are BrdU-retaining cells and expressed the stem cell marker CD34. Moreover, tracking the fate of the anal label-retaining cells in vivo revealed that the slow-cycling cells only gave rise to progeny of the anal epithelium. In conclusion, we identified a unique population of cells at the anorectal junction which can be separated from the other basal anal epithelial cells based upon the expression of the stem cell marker CD34 and integrin α6, and thus represent a putative anal stem cell population.
Tissue stem cells are thought to proliferate, self-renew and differentiate throughout the entire life of an animal. Their cell progeny participate in tissue homeostasis and repair.1,2 Being long-term residents in an epithelium, stem cells are uniquely susceptible to the accumulation of multiple oncogenic changes by repeated divisions giving rise to tumors. One explanation of this apparent paradox is that tissue stem cells are relatively quiescent cells, which divide infrequently and give rise to another stem cell daughter and to a rapidly proliferating, transient amplifying (TA) daughter cell.3,4 TA cells, unlike stem cells, have a limited growth potential. Yet, for a short period of time they divide repeatedly to make a large number of progeny, thereby fulfilling the need to maintain adequate number of cells of the tissue of origin. In the absence of specific molecular markers, the slow-cycling property of stem cells has been used to identify stem cells in their tissues of origin.5–9 Slowly cycling cells are identified by a repeated “pulse” with bromodeoxyuridine (BrdU) to label all the proliferating cells in a tissue, followed by a “chase” period. The rapidly dividing TA cells divide and dilute the label, while infrequently dividing cells retain the label (label-retaining cells or LRCs). Therefore, label retention reflects the growth history of the cells. Pulse-chase labeling schemes are highly variable and depend on the growth kinetics of the tissue being studied.
LRCs can be found in many tissues including the oral mucosa and the skin epidermis,10 the hair follicle,11 the eccrine glands,12 the bone marrow,13–15 the mammary gland,16 the kidney,17 the liver,18 the trachea,19 the esophagus,20 the tibia,21 the pancreas,8 the heart,22 the bladder,23 the limbus of the eye,5,24,25 and the mouse ovary.26
Stem cells are critical for wound healing, due to their potent ability to regenerate their own tissue. However, their extensive capacity for proliferation also implicates them in tumorigenesis.27 Transitional epithelia are defined by the abrupt change from one type of epithelium to another and have been shown to exist in the human eye, esophagus, stomach, bladder, cervix and anus. Interestingly, these transition zones have been reported to be susceptible to tumor formation, which raises the possibility that a stem cell niche exists within the transition zone.28–32 For instance, in the cervix, cancers arise exclusively in the vaginal-cervical squamocolumnar junction.33 In anal cancers, tumors can develop in the perianal skin, anal margin and anal canal. Interestingly, tumors of the anal canal develop at the transition zone between the stratified squamous epithelium of the anal canal and the columnar epithelium of the rectum. These tumors are more frequent than those at the anal margin and the perianal skin and their prognosis is less favorable.34 In the absence of TGFβRII, mouse anal transitional epithelia spontaneously generate squamous cell carcinomas.35 Similarly, mice with a targeted disruption of BMPR1A develop polyps in the intestinal epithelium, but carcinomas result in the gastrointestinal transitional zone.36 Transitional epithelia are poorly characterized and the presence of putative slow-cycling cells has not previously been investigated.
In this study, we utilized a previously developed strategy to detect cells in anal epithelium based on their proliferation history.37 Specifically, we used tetracycline-inducible mice driving histone H2B-GFP to follow cell proliferation through the dilution of GFP label. Compared to BrdU label retention, the histone H2B-GFP pulse-chase system is more sensitive, ensures initial uniform labeling of cells within tissues, and affords enhanced sensitivity in monitoring the infrequent division behavior of stem cells.38 We have detected a minor population of LRCs in the basal layer of the anal epithelium at the transition zone that meet the simple epithelium of the rectum. The stem cell surface marker CD34, also expressed in hair follicles39 and esophageal stem cells,20 has been found to colocalize within the anal LRCs.
We began by analyzing how cells connect at the dentate line where squamous anal cells abut columnar cells from the rectum (Fig. 1A). We analyzed four adult CD-1 mice from four to six weeks-old. Ultrastructural analyses revealed that cells at the anorectal transition zone show fewer desmosomes than anal cells that connect between each other in the basal layer (Figs. 1B-B′ and S1). Hemidesmosomes can be detected between the basement membrane and the anal keratinocytes (Fig. 1B″). We next examined how cells differentiate in the anal epithelium. As the electron microscopy images indicated, the mouse anal epithelium is composed of a basal layer of cells that differentiate into a spinous, granular and terminally differentiated stratum corneum layer (Fig. S1A). The markers of anal differentiation that we examined were Keratin 6, Keratin 10, Loricrin and Filagrin. Keratin 6, a keratin naturally expressed in hair follicles, but aberrantly induced in epidermis under hyperproliferative conditions, is expressed in the suprabasal layers of the anal epithelium (Fig. 2A). Similarly, Keratin 10 (a marker of the spinous layer of the epidermis) is highly expressed in the suprabasal layers of the anal epithelium (Fig. 2B). These results were not surprising as we have previously shown that the anal epithelium is naturally more proliferative than the epidermis of the skin.35 As in the epidermis, Loricrin marks the granular layer (Fig. 2C) and Filagrin marks the terminally differentiated stratum corneum layer (Fig. 2D). The anal epithelium expresses typical markers for stratified squamous epithelium and therefore may provide a niche for a population of stem cells similar to the stratified squamous epithelium of the skin and the esophagus.
To determine whether slow-cycling cells were present in the anal epithelium, we used the in vivo pulse-chase experiments previously employed for labeling adult hair follicle bulge cells with histone H2B-GFP.37 In this system, double transgenic tetracycline-inducible mice express H2B-GFP driven by the keratin 5 promoter (K5-TetVP16xTRE-H2B-GFP) specifically in skin. H2B-GFP expression is activated upon tetR-VP16 protein binding to the tetracycline responsive element DNA fragment, and can be turned off by addition of a tetracycline analogue (doxycyline, doxy) to the mouse diet. Upon H2B-GFP repression in the adult and embryo, the brightest label-retaining cells have been found in the hair follicle bulge.37,40 Here, we used this tet-off system to examine the presence of H2B-GFP LRCs in the anal epithelium. In unchased mice, all perianal skin and anal epithelial cells displayed H2B-GFP epifluorescence, consistent with strong keratin 5 promoter activity in these tissues (Fig. 3A and B). No H2B-GFP expression was detected in the rectum of H2B-GFP founder mice or in any double transgenic mice expressing the K5-driven tetracycline-regulatable transactivator mouse strain confirming that there is no leaky expression of GFP in this organ (Fig. 3B). After three to four weeks of doxycycline chase, H2B-GFP LRCs clustered prominently within the anal transition zone at the junction with the dentate line (Fig. 3C and D). This pattern of expression was consistent in all the mice analyzed (n = 10). After five weeks of chase almost no LRCs were detected in the anal transition zone, in contrast to the hair follicle bulge which can retain label for ten weeks of chase.37,38
To verify the presence of the LRCs in the anal transitional epithelium, we utilized a second pulse-chase technique. Slow-cycling cells have been detected in the past by a repeated “pulse” of the nucleotide analog bromodeoxyuridine (BrdU) to label all the proliferating cells in a tissue, followed by a “chase” period.11,41 We applied the BrdU pulse and chase strategy to the double transgenic mice K5-TetVP16xTRE-H2B-GFP to analyze the presence of H2B-GFP and BrdU LRCs in the anal epithelium. We labeled the tissue with BrdU for five days (P22-P26) and chased until 43 or 56 days of age. Simultaneously, we turned off the H2B-GFP expression by feeding the mice with doxycycline (Fig. 4A). Before the chase, immunofluorescence staining of perianal skin and anorectal frozen sections indicated that most of the cells were labeled with BrdU (Fig. 4B and C). After three weeks of chase, BrdU and H2B-GFP LRCs were found in the bulge area of the perianal hair follicle (Fig. 4D) and in the anorectal transition zone (Fig. 4E). This pattern of expression was consistent in all mice analyzed (n = 6). Due to the requirement that cells be actively undergoing DNA synthesis to incorporate BrdU,42 not all of the H2B-GFP LRCs colocalized with the BrdU positive cells, as the K5-TetVP16xTRE-H2B-GFP pulse stage is cell cycle independent. These results clearly indicate the presence of a population of LRCs in the mouse anal transition zone.
To determine if the anal H2B-GFP LRCs might express stem cell markers, we conducted immunofluorescence microscopy on anorectal frozen sections with antibodies directed against a number of adult stem cell markers. Similarly to the hair follicle stem cell niche,39,43,44 we found that the surface marker CD34 is expressed in a subset of H2B-GFP LRCs (Fig. 5A). p63, a transcription factor involved in maintenance of the stratified squamous epithelia and essential for the proliferative potential of their stem cells,45–47 is expressed throughout the entire anal epithelium and in the anal glands, but not in the rectum (Fig. 5B–D). In contrast to the hair follicle bulge,48 Keratin 19 marks the simple epithelium of the rectum and is absent in the stratified epithelium of the anal canal (Fig. 5C). Other bulge stem cell markers, such as Keratin 15, Lhx2 and Sox9,40,49,50 are not detected by immunofluorescence in the anal epithelium (data not shown). Sox2, a transcription factor involved in embryonic stem cell maintenance51,52 and in stratified squamous epithelium such as the esophagus53 is strongly expressed in the basal layer of the anal epithelium and its expression is decreased upon differentiation. In contrast to p63, Sox2 is not expressed in the anal gland (Fig. 5D and E).
To track the fate of anal LRCs during the differentiation of the anal epithelium, we monitored GFP fluorescence intensities relative to those of the differentiation-associated markers Keratin 6 (K6), Filagrin and Keratin 10 (K10) (Fig. 6). Over-exposure of the anorectal sections, from the K5-TetVP16xTRE-H2B-GFP double transgenic mice chased for four weeks (P56), verified that suprabasal and differentiated cells were largely GFP-positive, deriving from anal LRCs. We conclude that there is a small population of anal LRCs at the anorectal transitional epithelium that may give rise to progeny of the anal epithelium.
The identification and characterization of anal stem cells would significantly advance our understanding of the biology of the anal epithelium in health and disease. However, the proliferative compartment within the mouse anal epithelium remains poorly characterized despite the clinical importance of this tissue. The mouse anal epithelium can be divided into two compartments: a superficial stratified layer consisting of large, differentiated and keratinizing squamous cells, and a basal layer composed of densely packed columnar cells. Transitional epithelia are defined by the abrupt change from one type of epithelium to another. Whether or not there are stem cells present in the transition zone responsible for maintaining the anal epithelium has never been investigated. In other stratified epithelia (such as the epidermis of the skin and esophagus) stem cells have been shown to reside in the basal layer.20,54 We have identified a rare population of slow-cycling cells that retains both H2B-GFP and BrdU for prolonged chase times (3–4 weeks) and resides in the basal layer of the anal epithelium, at the transition between the stratified anal epithelium and the simple epithelium of the rectum. The anal epithelium is naturally more proliferative than the perianal skin; therefore, it is not surprising that H2B-GFP LRCs may be detected following shorter chase periods than their counterparts in the hair follicle bulge. No LRCs were found in epithelial cells of the anal canal close to the perianal skin.
Epithelial stem cells are frequently described as part of a population of slow-cycling cells or LRCs, based upon the fact that stem cells divide less frequently than other differentiated cells.10,24 Therefore, the anatomical finding of LRC distribution in the anal epithelium is useful in predicting stem cell distribution because it is likely that many of the LRCs are in fact stem cells.6 For example, slow-cycling cells detected in the limbus at the transition zone of the cornea with the conjunctiva24,25 have been largely demonstrated to be stem cells.55
We found that the anal slow-cycling cell population expresses the surface marker CD34, which is expressed by a variety of pluripotent cells and tissue stem cells56 including hair follicle bulge stem cells,39 esophageal stem cells20 and muscle satellite cells.57 Moreover anal slow-cycling cells also expressed several other markers consistent with known stem cells.46,51,52,56 As in other stem cell niches, we found heterogeneity in the LRC and the presence of stem cell marker such as CD34.7
Further analysis will determine the properties of the anal LRCs. Mouse anal epithelial cell culture studies and anorectal injury models will be required to directly address the stem cell potential of the anal LRCs.
Anorectal regions from adult CD-1 mice were dissected and fixed in 2% glutaraldehyde, 4% paraformaldehyde and 2 mM CaCl2 in 0.05 M sodium cacodylate buffer, pH 7.2 at 4°C for one hour. The samples were then post fixed in 1% osmium tetroxide in 0.2 M sodium cacodylate buffer, processed through a graded series of alcohols, infiltrated and embedded in LX-112 resin. After polymerization at 60°C for three days, ultrathin sections (100 nm) were cut using a Reichert-Jung Ultracut E microtome and counterstained in 2% aqueous uranyl acetate and Reynolds lead citrate. Images were taken with a transmission electron microscope (Hitachi H-6750) equipped with a digital camera (AMT 2k × 2K tem CCD).
One percent toluidine blue O was prepared in 1% sodium borate solution and used to stain 1 µm anorectal sections. Cryostat sections (10 µm) of mouse anorectal regions were fixed for 10 min in freshly prepared 4% paraformaldehyde in 1X phosphate-buffered saline solution (1X PBS) and washed 3 times for 10 min in 1X PBS at room temperature. Sections were then permeabilized in 0.1% Triton X 100 for 10 minutes and non-specific staining was eliminated via the following blocking solution: 2.5% normal goat serum, 2.5% normal donkey serum, 2% gelatin, 0.1% Triton X 100 and 1% bovine serum albumin in 1X PBS. Whenever mouse monoclonal antibodies were utilized, the MOM kit (Vector Laboratories, Burlingame, CA) was used.
Primary antibodies against the following proteins were used at the dilution indicated: bromodeoxyuridine (Abcam, 1/200), Keratin 5 (Seven Hills Bioreagents, Cincinnati, OH, 1/2,000), Keratin 17 (a generous gift from Dr. Pierre Coulombe, 1/5,000), Keratin 6 (a generous gift from Dr. Elaine Fuchs, 1/500), Keratin 10 (Covance, Emeryville, CA, 1/1,000), Keratin 8 and Keratin 19 (these antibodies, developed by Dr. Brulet and Dr. Kemler, were obtained from the NICHD Developmental Studies Hybridoma Bank maintained by the University of Iowa), Loricrin (Covance, Emeryville, CA, 1/500), Filagrin (Covance, Emeryville, CA, 1/1,000), CD34 (eBiosciences Inc., San Diego, CA, 1/50), P-cadherin (R&D Antibodies, North Las Vegas, NV, 1/100), CD49f (BD Biosciences, San Jose, CA, 1/100), p63 (Santa-Cruz Biotechnology Inc., Santa Cruz, CA, 4A4, 1/50), Sox2 (a generous gift from Dr. Jeffrey Whittsett, 1/5,000). 4′,6-diamidino-2-phenylindole (DAPI) was utilized as a marker of cell nuclei (Sigma Chemical Co., St. Louis, MO, 1/5,000). Secondary antibodies conjugated with FITC or Rhodamine (Jackson ImmunoResearch Laboratories Inc., West Grove, PA) were used at a dilution of 1/250.
Immunostained sections were analyzed using a fluorescent microscope AxioImager M1 (Zeiss) and pictures were taken with an axioCam MRm camera (Zeiss). Images in different focal planes were combined using the Extended Focus Module within the Axiovision software suite (Zeiss).
All experiments were approved by the Cincinnati Children's Hospital Research Foundation IACUC and carried out using standard procedures. We generated Tet-off H2B-GFP mice by crossing heterozygous K5tTA (FVB) mice with pTREH2B-GFP (CD1) mice37 and identified animals expressing GFP with a blue led flashlight (Tektite) and amber eyeglasses (EyeSave Sunglasses and Readers). Mice were fed with 1 g/Kg of Doxycycline food (Harlan) to repress H2B-GFP expression. For label retention studies, 5-bromo-2-deoxyuridine (BrdU, Sigma-Aldrich, St. Louis, MO) was injected intraperitoneally to a final concentration of 50 µg/g of body weight starting at postnatal days P22-P26 at 12 h intervals, and was subsequently added to the drinking water (0.8 mg/ml) during the pulse period. A total of six mice were analyzed.
We thank Dr. James Lessard for valuable discussions, Drs. Elaine Fuchs, Pierre Coulombe and Jeffrey Whittsett for providing their antibodies. We thank Dr. Adrian McNairn for valuable criticism of the manuscript. We also thank Dr. Amalia Pasolli for her advice on the electron microscopy procedures. This work was funded by CCHMC Trustee Grant Award and part by PHS Grant P30 DK 078392 (G.G.).
Previously published online: www.landesbioscience.com/journals/cc/article/12437