- Email: [email protected]
PARACRINE ACTION OF TRANSFORMING GROWTH FACTOR-ALPHA IN RECTAL CRYPT EPITHELIUM OF HUMANS
Cell Biology International 2002, Vol. 26, No. 12, 1029–1034 doi:10.1006/cbir.2002.0962, available online at http://www.idealibrary.com on
PARACRINE ACTION OF TRANSFORMING GROWTH FACTOR-ALPHA IN RECTAL CRYPT EPITHELIUM OF HUMANS IVAN L. CAMERON1* and W. ELAINE HARDMAN2 1
Department of Cellular and Structural Biology, The University of Texas Health Science Center at San Antonio, San Antonio, TX 78229-3900, U.S.A., 2Pennington Biomedical Research Center, Baton Rouge, LA 70808, U.S.A. Received 22 January 2002; accepted 21 August 2002
Colon and rectal mucosal crypt epithelium is a rapidly renewing cell population, where cell proliferation is normally balanced by cell loss. This report concerns the putative paracrine action of transforming growth factor (TGF-) in this homeostatic process. Immunohistochemical staining for proliferating cell nuclear antigen (PCNA) and TGF- was performed on biopsy specimens of rectal mucosa taken from consenting patients. The height of the proliferative compartment in mid-axially sectioned crypts in each individual was determined from the distribution of PCNA stained cells. The number of TGF- stained cells that exhibited intense positive staining in a continuous column from the mouth down the side of the crypt was also scored in each individual patient. There was a significant positive correlation (P=0.05, n=22 patients) between the height of the proliferative compartment and the number of cells staining for TGF-. Noncellular TGF- reactivity was also observed in the lamina propria adjacent to the TGF- reactive epithelial cells, indicating secretion of TGF- by these epithelial cells. These findings suggest that TGF- is released from epithelial cells in the upper compartment of the crypt into the adjacent lamina propria and then diffuses to the epithelial cells in the lower part of the crypt, resulting in 2002 Elsevier Science Ltd. All rights reserved. expansion of the proliferative compartment. K: transforming growth factor-alpha (TGF-); rectal mucosa; immunohistochemistry (IHC); cell proliferation.
INTRODUCTION Colon and rectal mucosal crypt epithelium is a rapidly renewing cell population. Stem and progenitor cells capable of cell proliferation are located towards the base of the crypt and mature nonproliferating epithelial cells are located in a column of cells above the zone of cell proliferation. Exogenous factors, including damaging agents (i.e. carcinogens, chemicals and gamma irradiation), inflammatory agents, surgical resection or dietary modification, are all known to cause adaptive changes in crypt column height and in the size of the zone of cell proliferation (Cameron et al, 1990, 2000; Deschner 1990; Beauchamp and Townsend, Jr., 1990; Egger, et al., 1998). These exogenous factors work by modulation of multiple endogen*To whom correspondence should be addressed: Fax: +1 (210) 5673803; E-mail: [email protected] 1065–6995/02/$-see front matter
ous factors that are either known to be or are suspected of being involved in the adaptive crypt changes. Endogenous factors may include neuronally derived factors (Bjerknes and Chang, 2001), endocrine factors (Beauchamps and Townsend, 1990), or various epithelial derived growth factors. Among these growth factors are peptides of the epidermal growth factor (EGF) family which display intracrine, autocrine, juxtacrine and paracrine action in the gastrointestinal mucosa (Derynck, 1988). One member of the epidermal growth factor family thought to be a key mediator of mucosal homeostasis is TGF- (Egger et al., 1998). TGF- and other ligands of the EGF family, i.e. epidermal growth factor (EGF), betacellulin, heparin-binding EGF, heregulin, CRIPTO and amphiregulin, are expressed in mucosal crypts in the colon and rectum (Barnard et al., 1995; Hardman et al., 1997; 2002 Elsevier Science Ltd. All rights reserved.
1030
Normanno et al., 2001). All of these ligands share a stretch of six cysteines spaced at specific intervals within a 40 amino acid sequence. TGF- is present at the lumenal surface and along the crypt epithelium (Xian et al., 1999). The intensity of immuno-reactivity of TGF- in the zone of cell proliferation in large bowel mucosal crypts appears less intense and more variable than in the upper crypt epithelium (Hardman et al., 1997). TGF- is made in the crypt epithelium and can exist as a membrane bound pro-TGF-, a diffusible mesoTGF- (with the pro-region present), a membranebound active TGF- (without the pro-region), or the diffusible 50 amino acid functional TGF- (Massague, 1990). The receptor for TGF- and the other ligands of the EGF family is the epidermal growth factor receptor (EGFR) that is found on the basolateral surface of epithelial cells in the proliferative zone (Huang et al., 1995; Sauma et al., 1995). In this study, we investigated the hypothesis that TGF- secreted at the base of epithelial cells located in the upper fraction of human rectal mucosal crypts influences the height of the column of proliferatively active epithelial cells. The findings support the hypothesis that TGF- secreted from epithelial cells in the upper region of the crypt diffuses through the lamina propria, acting to expand the number of cells present in the column of active proliferative epithelial cells. MATERIALS AND METHODS The human rectal biopsy specimens used for this study were obtained from previous studies (Hardman et al., 1997; Barnes et al., 1999). The protocols were approved by the Institutional Review board of the University of Texas Health Science Center. Informed consent was obtained from each volunteer. Subjects scheduled for routine endoscopy for detection and removal of colorectal polyps were screened for entry into the protocol. Specimens from 12 males and 10 females (total of 22 subjects), mean age 57.5 years (range 47 to 73 years) were used. None of the subjects displayed evidence of large bowel disease. A Go-Lytely enema (Braintree Labs, Braintree, MA, U.S.A.) was given prior to colonoscopy. Biopsy specimens from normal appearing flat rectal mucosa were taken about 6 cm above the anal verge using 5 mm pinch forceps. Two or more biopsies from each subject were fixed in Omnifix (An-Con Genetics, Melville, NY, U.S.A.), an alcohol-based fixative that does not cross-link
Cell Biology International, Vol. 26, No. 12, 2002
antigen epitopes. The selection of this particular fixative for immunohistochemical localization of TGF- in intestinal crypts is based on previous studies (Hardman et al., 1997; Hardman and Cameron, 1998; Barnes et al., 1999). Only rectal mucosa from subjects found to be without large bowel polyps were used in this study. The treatment of biopsy specimens for histology was as previously described (Hardman et al., 1997). Immunohistochemistry Routine immunohistochemical (IHC) techniques were used to localize PCNA or TGF- proteins on the deparaffinized serial sections of rectal biopsies. PCNA primary antibody (clone PC10, 1:25 dilution) was from Signet Laboratories, Dedham, MA, U.S.A. TGF- primary antibody (Ab-2, clone 2314.4, 1:10 dilution) was from Oncogene Science, Cambridge, MA, U.S.A. Secondary antibodies (Super Sensitive link (anti-IgG) and label (streptavidin-horseradish peroxidase)) and liquid diaminobenzidine were from Biogenex, San Ramon, CA, U.S.A. A single primary antibody was applied to each of two tissue sections on each slide. The third tissue section on each slide was used as a negative control, with normal mouse serum substituted for the primary antibody and incubated with all other reagents. Determination of proliferative zone height and crypt height Only complete mid-axial, longitudinally sectioned crypts on PCNA stained slides were used to determine the proliferative parameters. Crypt height was taken as the mean number of cells in a single column from the center of the base of the crypt to the mouth of the crypt. The top cell counted was the last cell that was more a part of the crypt wall than of the lumenal surface. Seven or more complete crypts were scored for each patient. The upper limit of the proliferative zone was calculated from data on the position (in number of cells from the base of the crypt) of each PCNA stained cell. Analyses of the distribution of stained cells in the crypt column showed that the distribution was significantly different from normal. A square root transformation of the position of each stained cell was performed and resulted in a normal distribution of cell heights, so that parametric statistical analyses could be used to help determine the proliferative zone. The transformed mean of the heights of stained cells and the SD from that mean were calculated for each patient. The upper
Cell Biology International, Vol. 26, No. 12, 2002
Fig. 1. Photomicrographs of 4 m thick histological sections of human rectal biopsies, lightly counterstained with Meyer’s hematoxylin. (A) The immunohistochemical localization of PCNA in nuclei of epithelial cells near the base of the colonic crypt (two examples at arrowheads) is shown by the intense dark deposit of diaminobenzidine. (B) The immunohistochemical localization of TGF- is shown by the intense dark deposit of diaminobenzidine in the upper half of the colonic crypt. Arrowheads approximately denote the end of continuous staining in the upper half of the crypt.
limit of the proliferative zone (PZ) was defined as the mean of the square root-transformed heights of the stained cells plus 1 SD from that mean. This was calculated for each patient. Calculation of TGF- parameters Complete mid-axially sectioned, longitudinal crypts were selected to determine the extent of TGF- stain (Barnes et al., 1999). TGF- staining was recorded as the number of cells that exhibited positive staining in a continuous column from the mouth down the side of the crypt. Solitary TGF- positive cells and cells at the bottom of the crypt were not included in the count. Statistical analyses Prism (Graph Pad Software, San Diego, CA, U.S.A.) and SAS for the PC were used for statistical analyses.
RESULTS Figure 1A is a photomicrograph of PCNA stained cells in the proliferative zone. Figure 1B illustrates that all the epithelial cells in the upper part of the crypts of rectal biopsies stain positive for TGF-. A
1031
Fig. 2. Linear regression showing the significant positive correlation (P=0.05) between the number of cells that are stained intensely positive for TGF- (as defined in the text) and the number of cells in the zone of cell proliferation as measured in mid-axially sectioned crypts of rectal mucosal biopsies from humans (n=22).
few cells at the very base of the crypt showed positive nuclear and cytoplasmic staining for TGF-. The mean (SEM) total column height of the rectal mucosal crypts expressed in number of cells for this patient population was 65.60.93, and the untransformed upper limit of the proliferative zone ranged from 16 to 48 cells with a mean (SEM) of 21.51.15 cells. Figure 2 summarizes the immunohistochemistry results from the rectal mucosa of 22 humans. The untransformed upper limit of the proliferative zone, as expressed in number of cells from the crypt base, is plotted against the number of cells at the top of the crypt stained for TGF-. In 3 out of 22 (14%) of the rectal specimens, TGF- staining occurred along the entire column of epithelial cells. However, most individuals had a lower fraction of intensely TGF- reactive cells. Least squares linear regression analysis of the data in Figure 2 demonstrated a significant positive correlation (P=0.05) between the number of crypt cells that stained intensely positive for TGF- and the height of the proliferative zone. TGF- immunohistochemical staining was also localized within the lamina propria between adjacent crypts (Fig. 3), and was seen in the noncellular areas of the loose connective tissue, but not in the cells of the lamina propria. There was a readily observable gradient of TGF- reactivity, with more TGF- staining in the lamina propria adjacent to the upper part of the crypt than towards the base (Fig. 3). The epithelial cells at the lumen surface retained TGF- reactivity, but the intensity appeared somewhat less than in those lining the upper compartment of the crypt, particularly in the basal region of the surface epithelial cells. The negative controls, using an irrelevant
1032
Fig. 3. The IHC localization of TGF- in the lamina propria between the crypts reveals the intensity of TGF- staining that is higher near the lumen surface and lower adjacent to the zone of cell proliferation towards the bottom of the crypts.
mouse MAb of the same isotype (anti-trp E, type IgG2a, Oncogene Science), revealed no staining in the crypts.
DISCUSSION TGF- is normally expressed in the mucosa throughout the gastrointestinal tract (Thomas et al., 1992; Alison et al., 1993; Montaner et al., 1999), where it is thought to play a role in stimulating epithelial cell proliferation and migration (Podolsky, 1994; Barnard et al., 1995; Barnes et al., 1999; Wilson and Gibson, 1999). TGF- has been localized in the colon and rectal mucosal crypts of mouse, rat and human using the mouse monoclonal IgG (GHF10) antibody raised against human TGF- C-terminal peptide 34-50 (Oncogene Science, Cambridge, MA, U.S.A.). However, different immunohistochemical studies using this particular MAb have given variable localization patterns in colon crypts. Such variability has been attributed to the method of tissue preparation, i.e. fresh frozen vs fixed (Xian et al., 1999), and to the type of fixative used, i.e. formalin or Methacarn (Xian et al., 1999), or alcohol, formalin or Omnifix (Hardman and Cameron, 1998). Xian et al. (1999) found that two TGF- polyclonal antibodies and the TGF- MAb-GF10 gave the same localization pattern in fresh frozen
Cell Biology International, Vol. 26, No. 12, 2002
sec-tions of human colon, and all three antibodies co-localized with TGF- mRNA, as revealed by in situ hybridization, suggesting that TGF- was being identified by the antibodies. However, the specificity of the TGF- MAb-GF10 was questioned when colon crypts of TGF- knockout mice gave the same localization as those of wild type mice. Xian et al. (1999) suggested that MAb-GF10 may crossreact with other members of the EGF family that are also expressed in mouse colon crypts. Thus, we are left with some uncertainty as to the exact identity of the EGF family members that MAb-GF10 reacts with in mouse colon crypts, and cannot therefore be certain of the specificity for TGF- in human rectal crypts. If the MAb-GF10 antibody does not specifically identify TGF-, then it is probably crossreacting with a related EGF family member. However, our method of tissue preparation, using Omnifix fixation and paraffin embedding of human rectal mucosal crypts and MAb-GF10, appears to give the same pattern of immunohistochemical localization as the polyclonal antibodies against TGF- (Xian et al., 1999). The latter study revealed strong TGF- staining reactivity on the surface epithelium, and distinct but somewhat weaker staining reactivity in epithelial cells progressing from the crypt mouth to the lower region of the crypt, where cell proliferation is known to occur. There was no evidence of TGF- reactivity in cells of the lamina propria. In addition, our results reveal modest but definite TGF- reactivity in the non-cellular areas of lamina propria, especially noticeable adjacent to the epithelium in the upper region of the crypt (Fig. 3). This finding suggests a paracrine model of TGF- action in rectal crypts. In this model, mucosal crypt epithelial cells divide in the zone of cell proliferation towards the bottom of the crypt. Most daughter cells of these newly divided cells migrate up the crypt wall, cease cell proliferation and differentiate into functionally mature cells as they move towards the crypt mouth, and eventually move onto the lumen surface where they exfoliate. Total transit time is about 4–8 days in humans (Lipkin and Deschner, 1976). A few cells in the proliferative zone migrate towards the base of the crypt. It is established that the zone of cell proliferation changes in size due to various exogenous factors without a concomitant increase in the zone of differentiated cells (Cameron et al., 2000). Indeed, Deschner (1990) and others demonstrate a disproportionate expansion of the zone of cell proliferation as a risk factor for colorectal adenocarcinoma. It is proposed that TGF- plays a
Cell Biology International, Vol. 26, No. 12, 2002
role in regulating the size of the zone of cell proliferation, as described below. Cells in the zone of cell proliferation are sometimes observed to have TGF- reactivity in the supranuclear area where the Golgi apparatus resides. As these cells move out of the zone of cell proliferation they begin to accumulate stores of TGF- and as the cells migrate into the zone of differentiated cells they all appear to have a large store of TGF-. The loose lamina propria connective tissue in the regions adjacent to the lumenal surface and upper crypt epithelium demonstrates a gradient of TGF- reactivity, decreasing from the mucosal lumen surface to the crypt proliferative zone region. Presumably, pro-TGF- is transported to the basal membrane, where it is enzymatically cleaved to release the diffusible 50-amino acid functional TGF-. The method of fixation and use of the MAb against this functional TGF- apparently allowed detection of the diffusible and functional TGF- in the lamina propria, as would be expected if epithelial cells were releasing TGF-. Given that the receptor for TGF- is reported to be located on the basolateral surface of epithelial cells in the zone of cell proliferation (Huang et al., 1995; Sauma et al., 1995), it is hypothesized that functional TGF- diffuses through the lamina propria to the EGF receptors on the epithelial cells in the proliferative zone, and stimulates these cells to proliferate. The more TGF-, the greater the number of cells in the zone of cell proliferation. This model of paracrine action fits with the results of several reports that implicate dysregulation of TGF- as a contributory factor in the development of colorectal cancer. For example, Tanaka et al. (1992) reported that 24% of colon adenomas and 81% of colon adenocarcinomas overexpress TGF-. Moskal et al. (1995) report that most patients diagnosed with colorectal adenocarcinomas have higher than normal levels of TGF- in their peripheral blood. The latter findings account for an observed increase in the size of the zone of cell proliferation in large bowel crypts located far away from a colorectal adenocarcinoma that is secreting large amounts of TGF- (Barnes et al., 1999). It is also reported that the mitogenic activity of TGF- can be blocked by antibodies to the EGFR or by a TGF- antisense construct (Sizeland and Burgess, 1992; Ciardiello et al., 1993; Barnard et al., 1995). Given that TGF- is a mitogenic factor involved in the promotion of colorectal cancer, it is not surprising to learn that the extent of TGF- immunohistochemical staining in the crypts of rectal mucosa has proven to be a
1033
useful biomarker of the risk of colorectal cancer (Hardman et al., 1997; Barnes et al., 1999).
ACKNOWLEDGEMENTS This work was partially supported by grants from the National Institute of Cancer, the Susan G. Komen Breast Cancer Foundation and the American Institute for Cancer Research. REFERENCES A MR, N MM, A TB, S CE, 1993. Transforming growth factor- immunoreactivity in a variety of epithelial tissues. Cell Prolif 26: 449–460. B JA, B RD, R WE, D RN, C RJ, 1995. Epidermal growth factor-related peptides and their relevance to gastrointestinal pathophysiology. Gastroenterology 108: 564–580. B CJ, H-M RL, H WE, C IL, S KV, L M, 1999. Effect of aspirin on prostaglandin E2 formation and transforming growth factor alpha expression in human rectal mucosa from individuals with a history of adenomatous polyps of the colon. Cancer Epidemiol Biomarkers Prev 8: 311–315. B CJ, H WE, C IL, 1999. Presence of well-differentiated distal, but not poorly differentiated proximal, rat colon carcinomas is correlated with increased cell proliferation in and lengthening of colon crypts. Int J Cancer 80: 68–71. B RD, T CM J, 1990. Growth factors. In: Moyer MP, Poste GH, eds. Colon Cancer Cells. San Diego, Academic Press, Inc. B M, C H, 2001. Modulation of specific intestinal epithelial progenitors by enteric neurons. Proc Natl Acad Sci 98: 12497–12502. C IL, O VA, H KE, N MV, P GM, H DW, 1990. Quantitative contribution of factors regulating rat colonie crypt epthelium: role of parenteral and enteral feeding, caloric intake, dietary cellulose level and the colon carcinogen, DMH. Cell Tissue Kinet 23: 227–235. C IL, H WE, H DW, C JW, 2000. Dietary fibre on cell proliferation in large bowel mucosal crypts near or away from lymphoid nodules and on mineral bioavailability. Cell Prolif 33: 367–379. C F, B C, N N, B G, P S, T G, B AR, S DS, 1993. Infection with a transforming growth factor a anti-sense retroviral expression vector reduces the in vitro growth and transformation of a human colon cancer cell line. Int J Cancer 54: 952–958. D R, 1988. Transforming growth factor . Cell 54: 593–595. D EE, 1990. Kinetics of normal, preneoplastic, and neoplastic colonic epithelium. In: Moyer MP, Poste GH, eds. Colon Cancer Cells. San Diego, Academic Press, Inc. E B, C HV, P F, C NN, S EP, L J, B VS, F SW, B
1034
MW, E VE, 1998. Reduced susceptibility of mice overexpressing transforming growth factor alpha to dextran sodium sulphate induced colitis. Gut 43: 64–70. H WE, C IL, B WH, S KV, K SC, L KA, 1997. Transforming growth factor a distribution in rectal crypts as a biomarker of decreased colon cancer risk in patients consuming cellulose. Cancer Epidemiol Biomarkers Prev 6: 633–637. H WE, C IL, 1998. The type of chemical fixative affects the immunohistochemical localization of transforming growth factor a in rat colon crypts. Histochem J 30: 301–305. H F, S S, Y Z, F E, 1995. Colon absorptive epithelial cells lose proliferative response to TGF- as they differentiate. Exp Cell Res 219: 8–14. L M, D E, 1976. Early proliferative changes in intestinal cells. Cancer Res 36: 2665–2668. M J, 1990. Transforming growth factor-: A model for membrane-anchored growth factors. J Biol Chem 265: 21393–21396. M B, A M, P-T R, 1999. Immunolocalization of transforming growth factor-alpha and epidermal growth factor receptor in the rat gastroduodenal area. Dig Dis Sci 44: 1408–1416. M TL, H S, E LM, F HA J, C S, 1995. Serum levels of transforming growth factor alpha in gastrointestinal cancer patients. Cancer Epidemiol Biomarkers Prev 4: 127–131.
Cell Biology International, Vol. 26, No. 12, 2002
N N, B C, DL A, S DS, 2001. The role of EGF-related peptides in tumor growth. Front Biosci 6: 685–707. P DK, 1994. Peptide growth factors and regulation of growth in the intestine. In: Walsh JH, Dockray GJ, eds. Gut Peptides: Biochemistry and Physiology. New York, Raven Press. 803–823. S S, H F, S S, F E, 1995. Colon goblet cells lose proliferative response to TGF- as they differentiate. Int J Cancer 61: 848–853. S AM, B AW, 1992. Anti-sense transforming growth factor or oligonucleotides inhibit autocrine stimulated proliferation of a colon carcinoma cell line. Mol Biol Cell 3: 1235–1243. T S, I K, H K, T T, Y M, S K, K G, 1992. Immunoreactive transforming growth factor- and epidermal growth factor in colorectal adenomas and carcinomas. Oncology 49: 381–385. T DM, N MM, G WJ, A MR, 1992. Immunoreactivity of transforming growth factor alpha in the normal adult gastrointestinal tract. Gut 33: 628–631. W AJ, G PR, 1999. Role of epidermal growth factor receptor in basal and stimulated colonic epithelial cell migration in vitro. Exp Cell Res 250: 187–196. X CJ, M CE, R LC, 1999. Specificity of the localization of transforming growth factor-alpha immunoreactivity in colon mucosa. J Histochem Cytochem 47: 949–958.
PARACRINE ACTION OF TRANSFORMING GROWTH FACTOR-ALPHA IN RECTAL CRYPT EPITHELIUM OF HUMANS IVAN L. CAMERON1* and W. ELAINE HARDMAN2 1
Department of Cellular and Structural Biology, The University of Texas Health Science Center at San Antonio, San Antonio, TX 78229-3900, U.S.A., 2Pennington Biomedical Research Center, Baton Rouge, LA 70808, U.S.A. Received 22 January 2002; accepted 21 August 2002
Colon and rectal mucosal crypt epithelium is a rapidly renewing cell population, where cell proliferation is normally balanced by cell loss. This report concerns the putative paracrine action of transforming growth factor (TGF-) in this homeostatic process. Immunohistochemical staining for proliferating cell nuclear antigen (PCNA) and TGF- was performed on biopsy specimens of rectal mucosa taken from consenting patients. The height of the proliferative compartment in mid-axially sectioned crypts in each individual was determined from the distribution of PCNA stained cells. The number of TGF- stained cells that exhibited intense positive staining in a continuous column from the mouth down the side of the crypt was also scored in each individual patient. There was a significant positive correlation (P=0.05, n=22 patients) between the height of the proliferative compartment and the number of cells staining for TGF-. Noncellular TGF- reactivity was also observed in the lamina propria adjacent to the TGF- reactive epithelial cells, indicating secretion of TGF- by these epithelial cells. These findings suggest that TGF- is released from epithelial cells in the upper compartment of the crypt into the adjacent lamina propria and then diffuses to the epithelial cells in the lower part of the crypt, resulting in 2002 Elsevier Science Ltd. All rights reserved. expansion of the proliferative compartment. K: transforming growth factor-alpha (TGF-); rectal mucosa; immunohistochemistry (IHC); cell proliferation.
INTRODUCTION Colon and rectal mucosal crypt epithelium is a rapidly renewing cell population. Stem and progenitor cells capable of cell proliferation are located towards the base of the crypt and mature nonproliferating epithelial cells are located in a column of cells above the zone of cell proliferation. Exogenous factors, including damaging agents (i.e. carcinogens, chemicals and gamma irradiation), inflammatory agents, surgical resection or dietary modification, are all known to cause adaptive changes in crypt column height and in the size of the zone of cell proliferation (Cameron et al, 1990, 2000; Deschner 1990; Beauchamp and Townsend, Jr., 1990; Egger, et al., 1998). These exogenous factors work by modulation of multiple endogen*To whom correspondence should be addressed: Fax: +1 (210) 5673803; E-mail: [email protected] 1065–6995/02/$-see front matter
ous factors that are either known to be or are suspected of being involved in the adaptive crypt changes. Endogenous factors may include neuronally derived factors (Bjerknes and Chang, 2001), endocrine factors (Beauchamps and Townsend, 1990), or various epithelial derived growth factors. Among these growth factors are peptides of the epidermal growth factor (EGF) family which display intracrine, autocrine, juxtacrine and paracrine action in the gastrointestinal mucosa (Derynck, 1988). One member of the epidermal growth factor family thought to be a key mediator of mucosal homeostasis is TGF- (Egger et al., 1998). TGF- and other ligands of the EGF family, i.e. epidermal growth factor (EGF), betacellulin, heparin-binding EGF, heregulin, CRIPTO and amphiregulin, are expressed in mucosal crypts in the colon and rectum (Barnard et al., 1995; Hardman et al., 1997; 2002 Elsevier Science Ltd. All rights reserved.
1030
Normanno et al., 2001). All of these ligands share a stretch of six cysteines spaced at specific intervals within a 40 amino acid sequence. TGF- is present at the lumenal surface and along the crypt epithelium (Xian et al., 1999). The intensity of immuno-reactivity of TGF- in the zone of cell proliferation in large bowel mucosal crypts appears less intense and more variable than in the upper crypt epithelium (Hardman et al., 1997). TGF- is made in the crypt epithelium and can exist as a membrane bound pro-TGF-, a diffusible mesoTGF- (with the pro-region present), a membranebound active TGF- (without the pro-region), or the diffusible 50 amino acid functional TGF- (Massague, 1990). The receptor for TGF- and the other ligands of the EGF family is the epidermal growth factor receptor (EGFR) that is found on the basolateral surface of epithelial cells in the proliferative zone (Huang et al., 1995; Sauma et al., 1995). In this study, we investigated the hypothesis that TGF- secreted at the base of epithelial cells located in the upper fraction of human rectal mucosal crypts influences the height of the column of proliferatively active epithelial cells. The findings support the hypothesis that TGF- secreted from epithelial cells in the upper region of the crypt diffuses through the lamina propria, acting to expand the number of cells present in the column of active proliferative epithelial cells. MATERIALS AND METHODS The human rectal biopsy specimens used for this study were obtained from previous studies (Hardman et al., 1997; Barnes et al., 1999). The protocols were approved by the Institutional Review board of the University of Texas Health Science Center. Informed consent was obtained from each volunteer. Subjects scheduled for routine endoscopy for detection and removal of colorectal polyps were screened for entry into the protocol. Specimens from 12 males and 10 females (total of 22 subjects), mean age 57.5 years (range 47 to 73 years) were used. None of the subjects displayed evidence of large bowel disease. A Go-Lytely enema (Braintree Labs, Braintree, MA, U.S.A.) was given prior to colonoscopy. Biopsy specimens from normal appearing flat rectal mucosa were taken about 6 cm above the anal verge using 5 mm pinch forceps. Two or more biopsies from each subject were fixed in Omnifix (An-Con Genetics, Melville, NY, U.S.A.), an alcohol-based fixative that does not cross-link
Cell Biology International, Vol. 26, No. 12, 2002
antigen epitopes. The selection of this particular fixative for immunohistochemical localization of TGF- in intestinal crypts is based on previous studies (Hardman et al., 1997; Hardman and Cameron, 1998; Barnes et al., 1999). Only rectal mucosa from subjects found to be without large bowel polyps were used in this study. The treatment of biopsy specimens for histology was as previously described (Hardman et al., 1997). Immunohistochemistry Routine immunohistochemical (IHC) techniques were used to localize PCNA or TGF- proteins on the deparaffinized serial sections of rectal biopsies. PCNA primary antibody (clone PC10, 1:25 dilution) was from Signet Laboratories, Dedham, MA, U.S.A. TGF- primary antibody (Ab-2, clone 2314.4, 1:10 dilution) was from Oncogene Science, Cambridge, MA, U.S.A. Secondary antibodies (Super Sensitive link (anti-IgG) and label (streptavidin-horseradish peroxidase)) and liquid diaminobenzidine were from Biogenex, San Ramon, CA, U.S.A. A single primary antibody was applied to each of two tissue sections on each slide. The third tissue section on each slide was used as a negative control, with normal mouse serum substituted for the primary antibody and incubated with all other reagents. Determination of proliferative zone height and crypt height Only complete mid-axial, longitudinally sectioned crypts on PCNA stained slides were used to determine the proliferative parameters. Crypt height was taken as the mean number of cells in a single column from the center of the base of the crypt to the mouth of the crypt. The top cell counted was the last cell that was more a part of the crypt wall than of the lumenal surface. Seven or more complete crypts were scored for each patient. The upper limit of the proliferative zone was calculated from data on the position (in number of cells from the base of the crypt) of each PCNA stained cell. Analyses of the distribution of stained cells in the crypt column showed that the distribution was significantly different from normal. A square root transformation of the position of each stained cell was performed and resulted in a normal distribution of cell heights, so that parametric statistical analyses could be used to help determine the proliferative zone. The transformed mean of the heights of stained cells and the SD from that mean were calculated for each patient. The upper
Cell Biology International, Vol. 26, No. 12, 2002
Fig. 1. Photomicrographs of 4 m thick histological sections of human rectal biopsies, lightly counterstained with Meyer’s hematoxylin. (A) The immunohistochemical localization of PCNA in nuclei of epithelial cells near the base of the colonic crypt (two examples at arrowheads) is shown by the intense dark deposit of diaminobenzidine. (B) The immunohistochemical localization of TGF- is shown by the intense dark deposit of diaminobenzidine in the upper half of the colonic crypt. Arrowheads approximately denote the end of continuous staining in the upper half of the crypt.
limit of the proliferative zone (PZ) was defined as the mean of the square root-transformed heights of the stained cells plus 1 SD from that mean. This was calculated for each patient. Calculation of TGF- parameters Complete mid-axially sectioned, longitudinal crypts were selected to determine the extent of TGF- stain (Barnes et al., 1999). TGF- staining was recorded as the number of cells that exhibited positive staining in a continuous column from the mouth down the side of the crypt. Solitary TGF- positive cells and cells at the bottom of the crypt were not included in the count. Statistical analyses Prism (Graph Pad Software, San Diego, CA, U.S.A.) and SAS for the PC were used for statistical analyses.
RESULTS Figure 1A is a photomicrograph of PCNA stained cells in the proliferative zone. Figure 1B illustrates that all the epithelial cells in the upper part of the crypts of rectal biopsies stain positive for TGF-. A
1031
Fig. 2. Linear regression showing the significant positive correlation (P=0.05) between the number of cells that are stained intensely positive for TGF- (as defined in the text) and the number of cells in the zone of cell proliferation as measured in mid-axially sectioned crypts of rectal mucosal biopsies from humans (n=22).
few cells at the very base of the crypt showed positive nuclear and cytoplasmic staining for TGF-. The mean (SEM) total column height of the rectal mucosal crypts expressed in number of cells for this patient population was 65.60.93, and the untransformed upper limit of the proliferative zone ranged from 16 to 48 cells with a mean (SEM) of 21.51.15 cells. Figure 2 summarizes the immunohistochemistry results from the rectal mucosa of 22 humans. The untransformed upper limit of the proliferative zone, as expressed in number of cells from the crypt base, is plotted against the number of cells at the top of the crypt stained for TGF-. In 3 out of 22 (14%) of the rectal specimens, TGF- staining occurred along the entire column of epithelial cells. However, most individuals had a lower fraction of intensely TGF- reactive cells. Least squares linear regression analysis of the data in Figure 2 demonstrated a significant positive correlation (P=0.05) between the number of crypt cells that stained intensely positive for TGF- and the height of the proliferative zone. TGF- immunohistochemical staining was also localized within the lamina propria between adjacent crypts (Fig. 3), and was seen in the noncellular areas of the loose connective tissue, but not in the cells of the lamina propria. There was a readily observable gradient of TGF- reactivity, with more TGF- staining in the lamina propria adjacent to the upper part of the crypt than towards the base (Fig. 3). The epithelial cells at the lumen surface retained TGF- reactivity, but the intensity appeared somewhat less than in those lining the upper compartment of the crypt, particularly in the basal region of the surface epithelial cells. The negative controls, using an irrelevant
1032
Fig. 3. The IHC localization of TGF- in the lamina propria between the crypts reveals the intensity of TGF- staining that is higher near the lumen surface and lower adjacent to the zone of cell proliferation towards the bottom of the crypts.
mouse MAb of the same isotype (anti-trp E, type IgG2a, Oncogene Science), revealed no staining in the crypts.
DISCUSSION TGF- is normally expressed in the mucosa throughout the gastrointestinal tract (Thomas et al., 1992; Alison et al., 1993; Montaner et al., 1999), where it is thought to play a role in stimulating epithelial cell proliferation and migration (Podolsky, 1994; Barnard et al., 1995; Barnes et al., 1999; Wilson and Gibson, 1999). TGF- has been localized in the colon and rectal mucosal crypts of mouse, rat and human using the mouse monoclonal IgG (GHF10) antibody raised against human TGF- C-terminal peptide 34-50 (Oncogene Science, Cambridge, MA, U.S.A.). However, different immunohistochemical studies using this particular MAb have given variable localization patterns in colon crypts. Such variability has been attributed to the method of tissue preparation, i.e. fresh frozen vs fixed (Xian et al., 1999), and to the type of fixative used, i.e. formalin or Methacarn (Xian et al., 1999), or alcohol, formalin or Omnifix (Hardman and Cameron, 1998). Xian et al. (1999) found that two TGF- polyclonal antibodies and the TGF- MAb-GF10 gave the same localization pattern in fresh frozen
Cell Biology International, Vol. 26, No. 12, 2002
sec-tions of human colon, and all three antibodies co-localized with TGF- mRNA, as revealed by in situ hybridization, suggesting that TGF- was being identified by the antibodies. However, the specificity of the TGF- MAb-GF10 was questioned when colon crypts of TGF- knockout mice gave the same localization as those of wild type mice. Xian et al. (1999) suggested that MAb-GF10 may crossreact with other members of the EGF family that are also expressed in mouse colon crypts. Thus, we are left with some uncertainty as to the exact identity of the EGF family members that MAb-GF10 reacts with in mouse colon crypts, and cannot therefore be certain of the specificity for TGF- in human rectal crypts. If the MAb-GF10 antibody does not specifically identify TGF-, then it is probably crossreacting with a related EGF family member. However, our method of tissue preparation, using Omnifix fixation and paraffin embedding of human rectal mucosal crypts and MAb-GF10, appears to give the same pattern of immunohistochemical localization as the polyclonal antibodies against TGF- (Xian et al., 1999). The latter study revealed strong TGF- staining reactivity on the surface epithelium, and distinct but somewhat weaker staining reactivity in epithelial cells progressing from the crypt mouth to the lower region of the crypt, where cell proliferation is known to occur. There was no evidence of TGF- reactivity in cells of the lamina propria. In addition, our results reveal modest but definite TGF- reactivity in the non-cellular areas of lamina propria, especially noticeable adjacent to the epithelium in the upper region of the crypt (Fig. 3). This finding suggests a paracrine model of TGF- action in rectal crypts. In this model, mucosal crypt epithelial cells divide in the zone of cell proliferation towards the bottom of the crypt. Most daughter cells of these newly divided cells migrate up the crypt wall, cease cell proliferation and differentiate into functionally mature cells as they move towards the crypt mouth, and eventually move onto the lumen surface where they exfoliate. Total transit time is about 4–8 days in humans (Lipkin and Deschner, 1976). A few cells in the proliferative zone migrate towards the base of the crypt. It is established that the zone of cell proliferation changes in size due to various exogenous factors without a concomitant increase in the zone of differentiated cells (Cameron et al., 2000). Indeed, Deschner (1990) and others demonstrate a disproportionate expansion of the zone of cell proliferation as a risk factor for colorectal adenocarcinoma. It is proposed that TGF- plays a
Cell Biology International, Vol. 26, No. 12, 2002
role in regulating the size of the zone of cell proliferation, as described below. Cells in the zone of cell proliferation are sometimes observed to have TGF- reactivity in the supranuclear area where the Golgi apparatus resides. As these cells move out of the zone of cell proliferation they begin to accumulate stores of TGF- and as the cells migrate into the zone of differentiated cells they all appear to have a large store of TGF-. The loose lamina propria connective tissue in the regions adjacent to the lumenal surface and upper crypt epithelium demonstrates a gradient of TGF- reactivity, decreasing from the mucosal lumen surface to the crypt proliferative zone region. Presumably, pro-TGF- is transported to the basal membrane, where it is enzymatically cleaved to release the diffusible 50-amino acid functional TGF-. The method of fixation and use of the MAb against this functional TGF- apparently allowed detection of the diffusible and functional TGF- in the lamina propria, as would be expected if epithelial cells were releasing TGF-. Given that the receptor for TGF- is reported to be located on the basolateral surface of epithelial cells in the zone of cell proliferation (Huang et al., 1995; Sauma et al., 1995), it is hypothesized that functional TGF- diffuses through the lamina propria to the EGF receptors on the epithelial cells in the proliferative zone, and stimulates these cells to proliferate. The more TGF-, the greater the number of cells in the zone of cell proliferation. This model of paracrine action fits with the results of several reports that implicate dysregulation of TGF- as a contributory factor in the development of colorectal cancer. For example, Tanaka et al. (1992) reported that 24% of colon adenomas and 81% of colon adenocarcinomas overexpress TGF-. Moskal et al. (1995) report that most patients diagnosed with colorectal adenocarcinomas have higher than normal levels of TGF- in their peripheral blood. The latter findings account for an observed increase in the size of the zone of cell proliferation in large bowel crypts located far away from a colorectal adenocarcinoma that is secreting large amounts of TGF- (Barnes et al., 1999). It is also reported that the mitogenic activity of TGF- can be blocked by antibodies to the EGFR or by a TGF- antisense construct (Sizeland and Burgess, 1992; Ciardiello et al., 1993; Barnard et al., 1995). Given that TGF- is a mitogenic factor involved in the promotion of colorectal cancer, it is not surprising to learn that the extent of TGF- immunohistochemical staining in the crypts of rectal mucosa has proven to be a
1033
useful biomarker of the risk of colorectal cancer (Hardman et al., 1997; Barnes et al., 1999).
ACKNOWLEDGEMENTS This work was partially supported by grants from the National Institute of Cancer, the Susan G. Komen Breast Cancer Foundation and the American Institute for Cancer Research. REFERENCES A MR, N MM, A TB, S CE, 1993. Transforming growth factor- immunoreactivity in a variety of epithelial tissues. Cell Prolif 26: 449–460. B JA, B RD, R WE, D RN, C RJ, 1995. Epidermal growth factor-related peptides and their relevance to gastrointestinal pathophysiology. Gastroenterology 108: 564–580. B CJ, H-M RL, H WE, C IL, S KV, L M, 1999. Effect of aspirin on prostaglandin E2 formation and transforming growth factor alpha expression in human rectal mucosa from individuals with a history of adenomatous polyps of the colon. Cancer Epidemiol Biomarkers Prev 8: 311–315. B CJ, H WE, C IL, 1999. Presence of well-differentiated distal, but not poorly differentiated proximal, rat colon carcinomas is correlated with increased cell proliferation in and lengthening of colon crypts. Int J Cancer 80: 68–71. B RD, T CM J, 1990. Growth factors. In: Moyer MP, Poste GH, eds. Colon Cancer Cells. San Diego, Academic Press, Inc. B M, C H, 2001. Modulation of specific intestinal epithelial progenitors by enteric neurons. Proc Natl Acad Sci 98: 12497–12502. C IL, O VA, H KE, N MV, P GM, H DW, 1990. Quantitative contribution of factors regulating rat colonie crypt epthelium: role of parenteral and enteral feeding, caloric intake, dietary cellulose level and the colon carcinogen, DMH. Cell Tissue Kinet 23: 227–235. C IL, H WE, H DW, C JW, 2000. Dietary fibre on cell proliferation in large bowel mucosal crypts near or away from lymphoid nodules and on mineral bioavailability. Cell Prolif 33: 367–379. C F, B C, N N, B G, P S, T G, B AR, S DS, 1993. Infection with a transforming growth factor a anti-sense retroviral expression vector reduces the in vitro growth and transformation of a human colon cancer cell line. Int J Cancer 54: 952–958. D R, 1988. Transforming growth factor . Cell 54: 593–595. D EE, 1990. Kinetics of normal, preneoplastic, and neoplastic colonic epithelium. In: Moyer MP, Poste GH, eds. Colon Cancer Cells. San Diego, Academic Press, Inc. E B, C HV, P F, C NN, S EP, L J, B VS, F SW, B
1034
MW, E VE, 1998. Reduced susceptibility of mice overexpressing transforming growth factor alpha to dextran sodium sulphate induced colitis. Gut 43: 64–70. H WE, C IL, B WH, S KV, K SC, L KA, 1997. Transforming growth factor a distribution in rectal crypts as a biomarker of decreased colon cancer risk in patients consuming cellulose. Cancer Epidemiol Biomarkers Prev 6: 633–637. H WE, C IL, 1998. The type of chemical fixative affects the immunohistochemical localization of transforming growth factor a in rat colon crypts. Histochem J 30: 301–305. H F, S S, Y Z, F E, 1995. Colon absorptive epithelial cells lose proliferative response to TGF- as they differentiate. Exp Cell Res 219: 8–14. L M, D E, 1976. Early proliferative changes in intestinal cells. Cancer Res 36: 2665–2668. M J, 1990. Transforming growth factor-: A model for membrane-anchored growth factors. J Biol Chem 265: 21393–21396. M B, A M, P-T R, 1999. Immunolocalization of transforming growth factor-alpha and epidermal growth factor receptor in the rat gastroduodenal area. Dig Dis Sci 44: 1408–1416. M TL, H S, E LM, F HA J, C S, 1995. Serum levels of transforming growth factor alpha in gastrointestinal cancer patients. Cancer Epidemiol Biomarkers Prev 4: 127–131.
Cell Biology International, Vol. 26, No. 12, 2002
N N, B C, DL A, S DS, 2001. The role of EGF-related peptides in tumor growth. Front Biosci 6: 685–707. P DK, 1994. Peptide growth factors and regulation of growth in the intestine. In: Walsh JH, Dockray GJ, eds. Gut Peptides: Biochemistry and Physiology. New York, Raven Press. 803–823. S S, H F, S S, F E, 1995. Colon goblet cells lose proliferative response to TGF- as they differentiate. Int J Cancer 61: 848–853. S AM, B AW, 1992. Anti-sense transforming growth factor or oligonucleotides inhibit autocrine stimulated proliferation of a colon carcinoma cell line. Mol Biol Cell 3: 1235–1243. T S, I K, H K, T T, Y M, S K, K G, 1992. Immunoreactive transforming growth factor- and epidermal growth factor in colorectal adenomas and carcinomas. Oncology 49: 381–385. T DM, N MM, G WJ, A MR, 1992. Immunoreactivity of transforming growth factor alpha in the normal adult gastrointestinal tract. Gut 33: 628–631. W AJ, G PR, 1999. Role of epidermal growth factor receptor in basal and stimulated colonic epithelial cell migration in vitro. Exp Cell Res 250: 187–196. X CJ, M CE, R LC, 1999. Specificity of the localization of transforming growth factor-alpha immunoreactivity in colon mucosa. J Histochem Cytochem 47: 949–958.