Annals of Surgical Oncology Cite Track
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bernstein, H.
Right arrow Articles by Bernstein, C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bernstein, H.
Right arrow Articles by Bernstein, C.
Related Collections
Right arrow Other Laboratory Research
Annals of Surgical Oncology 9:505-517 (2002)
© 2002 Society of Surgical Oncology

ORIGINAL ARTICLES

Patchy Field Defects of Apoptosis Resistance and Dedifferentiation in Flat Mucosa of Colon Resections From Colon Cancer Patients

Harris Bernstein, PhD, Hana Holubec, MS, James A. Warneke, MD, Harinder Garewal, MD, PhD, David L. Earnest, MD, Claire M. Payne, PhD, Denise J. Roe, PhD, Haiyan Cui, PhD, Elaine L. Jacobson, PhD and Carol Bernstein, PhD

From the Departments of Microbiology and Immunology (HB, HH, CMP, CB), Surgery (JAW), Medicine (HG, DLE), and Pharmacology and Toxicology (ELJ), the College of Public Health (DJR), and the Arizona Cancer Center (HB, JAW, HG, DLE, CMP, ELJ, HC, DJR), University of Arizona, Tucson, Arizona; the Tucson Veterans Affairs Hospital (HG), Tucson, Arizona; and the Cancer Center (ELJ), University of Kentucky, Lexington, Kentucky.

Correspondence: Address correspondence and reprint requests to: Carol Bernstein, PhD, Department of Microbiology and Immunology, College of Medicine, University of Arizona, Tucson, AZ 85724; Fax: 520-626-2100; E-mail: bernstein3@ earthlink.net.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Background: Abnormal areas in normal-appearing flat colonic mucosa (field defects) may predispose individuals to colon cancer. Markers of field defects would indicate cancer risk.

Methods: We evaluated apoptosis capability, dedifferentiation, frequency of simple aberrant crypts, aberrant crypt foci, microadenomas, and total nicotinamide adenine dinucleotide levels at locations within normal-appearing flat mucosa obtained from colon resections.

Results: Among goblet cells from colonic mucosa samples of individuals without colonic neoplasia, there was a high mean deoxycholate-induced apoptotic index (AI) of 59.1% and high Dolichos biflorus agglutinin (DBA) lectin reactivity (differentiation) in 85.0% of samples. In contrast, flat mucosa samples from colon cancer patients had a significantly (P < .01) lower average AI of 37.4%, and a significantly (P = .03) lower percentage (40.5%) had high DBA reactivity. For colon cancer patients, AI and DBA reactivity values were patchy within a resection. Nicotinamide adenine dinucleotide levels were highly variable among individuals without neoplasia, and aberrant crypt foci and microadenomas were rare.

Conclusions: AI and aberrant DBA reactivity are promising indicators of colon cancer risk. Our results attest to the importance of obtaining multiple samples to assess colon cancer risk because of the patchy nature of field defects.

Key Words: Field defects • Apoptosis resistance • Differentiation • Colon cancer • Dolichos biflorus agglutinin


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Colon carcinogenesis is a multistep process that requires multiple mutations,1 somatically inheritable epigenetic changes (epimutations) caused by hypermethylation of CpG dinucleotide islands in DNA,2,3 or both. Numerous alterations are observed in the flat mucosa of patients with colon cancer.4 These alterations characterize field defects that seem to occur during progression to sporadic adenocarcinoma of the colon.

An early observed functional change is loss of the capacity to undergo apoptosis in response to damage.5–7 If a cell acquires mutations or epimutations that cause apoptosis resistance, this can lead to increased clonogenic survival and consequent clonal expansion.8 In addition, suppression of apoptosis leads to increased mutagenesis.9,10 Differentiation of colonic cells is associated with increased apoptosis susceptibility.11–14 Thus, loss of differentiation in colonic cells may be associated with resistance to apoptosis. Mutations or epimutations that produce a growth advantage or mutator phenotype are likely alterations during carcinogenesis.15 Environmental (e.g., smoking) and dietary (e.g., high fat) factors probably influence the incidence of these mutations and epimutations. Figure 1 schematically indicates alterations occurring during progression to colon cancer and the likely influence of diet and smoking. Other factors that may have similar effects include dietary alcohol use, low intakes of calcium or antioxidant vitamins, and the nondietary factors of obesity and low physical activity.



View larger version (17K):
[in this window]
[in a new window]
  		FIG. 1. Schematic diagram indicating alterations important in progression to colon cancer.

 
In a previous study7 we examined 68 patients for the single biomarker of apoptosis resistance (in the flat mucosa), evaluated in a few biopsies in widely separated regions of the colon. In this study, we examined 16 human colon resections, usually taking 24 tissue samples from each, thus examining a much higher number of samples than we took from each patient in the previous study. We had two goals with our more intensive study of individual patients. One goal was to use the established marker of apoptosis resistance to evaluate several different potential biomarkers for specificity and sensitivity in detecting cancer risk in the flat-appearing colonic mucosa. A second goal was to evaluate the patchiness of occurrence of field defects of any useful marker.

We measured the distribution of the previously established biomarker of apoptosis resistance along the flat mucosa and compared this distribution with loss of cellular differentiation (measured by reactivity with the lectin Dolichos biflorus agglutinin [DBA]). We also used multiple tissue samples from these colon resections to evaluate the frequency of preneoplastic lesions (including simple aberrant crypts [SAC], aberrant crypt foci [ACF], and microadenomas) as indicators of cancer risk and compared them with apoptosis resistance and lectin reactivity. Because we had previously determined that poly(adenosine diphosphate-ribose) polymerase (PARP), a nicotinamide adenine dinucleotide (NAD)-consuming enzyme, is protective against apoptosis,16 we also examined the correlation of total NAD levels with apoptosis resistance and cancer risk [where total NAD is measured as the sum of levels of NAD(P)+, NAD(P)H, NAD+, and NADH].

We confirmed that apoptosis resistance is specific to high-risk patients and have now established that this defect is patchy. We were able to characterize epithelial cell differentiation (assessed with lectin staining) into three categories: normal, sparse, and aberrant. We found that sparse epithelial cell differentiation of the flat mucosa is strongly associated with colonic neoplasia and is also patchy. In addition, a high level (more than 35%) of aberrant lectin staining at any tissue sample site was specific for high-risk patients and was also patchy. Although the group of patients was smaller in this study, the high number of tissue samples allowed us to show that apoptosis resistance and defective (sparse or aberrant) differentiation correlate with each other and are both correlated with colon cancer risk. There was no significant association with distance from the tumor itself, indicating that patients at risk for colon cancer have field defects that are not directly produced by an adjacent tumor. Our study emphasizes the patchy nature of field defects, and this has unveiled an important caveat for biomarker development. It is now clear that evaluation of single biopsies taken from one area of the colon is insufficient to assess cancer risk.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Patient Samples
Mucosal samples were obtained from patients undergoing colonic resection or colonoscopy for clinical indications at the University Medical Center of the University of Arizona (Table 1). Protocols approved by the Human Subjects Committee of the University of Arizona were used, and informed consent for use of surgically removed tissues for research was obtained from each subject. Mucosal samples were removed from the colonic resections within 20 minutes after surgical removal and placed on ice. Samples taken as colonic biopsies were treated similarly by being placed in ice-cold medium after removal from the colon.


View this table:
[in this window]
[in a new window]
  		TABLE 1. Patients, their colonic diagnoses, and locations of lesions
 
Quantitation of Bile Acid-Induced Apoptosis by Using an Ex Vivo Live Cell Bioassay
Incubation medium with or without 1.0 mM of sodium deoxycholate (NaDOC) was prepared as previously described.7 Mucosal samples that were excised from colon resections or obtained during colonoscopy were placed in incubation medium that had been prewarmed to 37°C and equilibrated with 5% CO2 for 30 minutes. These samples were then incubated for 3 hours at 37°C in a humidified CO2 incubator, after which the incubation medium was removed and 2 mL of cold, half-strength Karnovsky’s fixative was added. The samples were kept in the refrigerator at 4°C overnight and then transferred to .1 M of cacodylate buffer (pH 7.2). Samples were then postfixed in osmium tetroxide, dehydrated in a graded series of ethanols, and embedded in Spurr’s epoxy resin. One-micrometer sections were stained with a polychrome stain, and the frequency of apoptotic goblet cells was measured as previously described.7 Only goblet cells were scored with respect to the frequency of apoptosis because, under our conditions of treatment, only the goblet cells were susceptible to apoptosis. Apoptosis was confirmed with transmission electron microscopy as the gold standard (Fig. 2).17



View larger version (170K):
[in this window]
[in a new window]
  		FIG. 2. An electron micrograph showing a goblet cell in a colonic mucosa tissue sample undergoing classic apoptosis after treatment of the tissue sample with 1.0 mM of sodium deoxycholate in incubation media for 3 hours. The nucleus shows a condensation and margination of the chromatin and the loss of fibrillar centers in the nucleolus. The nucleoplasm and cytoplasm show increased electron density.

 
Measurement of DBA Reactivity
Tissue samples of colonic mucosa were fixed in 70% ethanol and embedded in paraffin, and 4-µm sections were obtained. DBA lectin histochemical staining was performed by using a modification of the method of Boland et al.18 The tissue sections were deparaffinized, rehydrated, endogenous peroxidase blocked with 2.0% H2O2 in methanol, and reacted for 30 minutes with biotinylated DBA (Vector Laboratories, Inc., Burlingame, CA) at 10 µg/mL, followed by avidin-biotin-peroxidase complex (Vector Laboratories) for 30 minutes. The sections were finally reacted with diaminobenzidine tetrahydrochloride (Sigma Biochemical Co., St. Louis, MO) at .5 mg/mL in phosphate-buffered saline (consisting of 8 g/L of NaCl, .2 g/L of KCl, .2 g/L of KH2PO4, and 1.15 g/L of Na2HPO4) in the presence of .015% H2O2 for 3 minutes. The sections were counterstained with hematoxylin.

DBA reactivity of tissue sections was assessed by using bright-field microscopy in conjunction with Image-Pro PlusTM software (Media Cybernetics, Silver Spring, MD). First, 28 colonic mucosal samples from 2 patients without any colon disease (normal group) and 2 patients with diverticulitis (low-risk group) were studied for the degree and pattern of DBA staining. This was performed to establish the degree of variability in lectin staining and to establish the normal pattern of staining along the crypts for comparison with patients with colonic neoplasms (higher-risk group). The observer then evaluated 82 samples from 12 patients with colonic neoplasms (tubulovillous adenomas [TVAs] or adenocarcinomas) and was blinded as to patient diagnosis and to percentage apoptosis. In addition, five of the biopsies from patients with colonic neoplasms were re-evaluated, with blinding as to previous value, diagnosis, and percentage apoptosis, to establish intraobserver variability. There was excellent concordance between both sets of measurements.

Representative fields showing the observed variation in DBA reactivity are presented in Fig. 3. An area of a tissue section was designated normal if most of its goblet cells exhibited DBA reactivity, and the frequency of goblet cells exhibiting DBA reactivity was much higher in the upper half of the crypts and within the associated surface epithelium than in the lower half of the crypts (Fig. 3A). An area was designated sparse if its goblet cells had a substantially lower frequency of DBA reactivity, within both the crypts and the associated surface epithelial area (Fig. 3B). In these sparse areas, when goblet cells exhibited DBA reactivity, the DBA reactivity was as strong as for goblet cells that were DBA positive in normal areas of the section; however, a much smaller fraction of goblet cells showed this DBA reactivity. An area was designated aberrant if the goblet cells were negative for DBA but nongoblet cells exhibited DBA reactivity, as shown in Fig. 3C and at higher magnification in Fig. 3D. To quantitate the percentage of a tissue sample that had normal, sparse, or aberrant staining with DBA, first, the length of the surface epithelium was determined from digitized images in conjunction with Image-Pro Plus software. Then, the percentage of that length occupied by each of the staining patterns was determined. The degree of DBA reactivity designated as normal was based on our observations of tissue samples from individuals with normal colons. The DBA reactivity of goblet cells was previously shown by Bresalier et al.19 to be primarily present in the upper portion of the crypts in the proximal colon and in both the upper and lower portions of the crypts in the distal colon. This pattern was also present in our study.



View larger version (121K):
[in this window]
[in a new window]
  		FIG. 3. Crypts with normal levels of Dolichos biflorus agglutinin (DBA) reactivity (A); crypts with sparse DBA reactivity (B); crypts with aberrant DBA reactivity (C); and an enlarged image of a portion of a crypt with aberrant DBA reactivity (D). In panel D, the arrows point to two goblet cells with nonstaining (white) mucin, whereas other nongoblet cells of the crypt have DBA reactivity and are stained brown.

 
Measurement of Preneoplastic Lesions
Tissue sections were screened for SACs, ACF, and microadenomas from formalin-fixed, paraffin-embedded tissue samples. The total square millimeters of mucosa, which includes the length of surface epithelium to the base of the crypts, was determined from digitized images in conjunction with Image-Pro Plus software. The frequency of SACs was then expressed as the number of SACs per square millimeter of mucosa. We used an established method of detecting these preneoplastic lesions by using sectioned tissue samples stained with hematoxylin and eosin.20– 22

Extraction and Assay of NAD+, NADP+, NADH, and NADPH
NAD+ and NADP+ were extracted from half of each frozen tissue sample by homogenization in .5 M of HClO4. After 10 minutes on ice, the extract was centrifuged at 200 x g for 10 minutes. The pellet was reserved for protein quantification. The supernatant was neutralized with 2 M of KOH, and the insoluble KClO4 was removed by centrifugation. The resulting supernatant was assayed for NAD+ and NADP+ as described previously.23 NADH and NADPH were extracted in the other half of each frozen tissue sample by homogenizing in .5 M of NaOH. The extract was heated to 60°C for 10 minutes to destroy oxidized pyridine nucleotides. The reduced pyridine nucleotides were then oxidized by adding a 1/10 volume of 2 mM phenazine ethosulfate (Sigma Biochemical Co.). The sample was then brought to .5 M of HClO4 and processed as described previously for NADP+ extraction and assay. The protein precipitated by HClO4 was dissolved in 1 M of NaOH, and the protein concentration was determined by the Bradford method. NADP+ values were normalized to protein concentration in each sample extracted.

Statistical Methods
For each patient, an apoptotic index (AI; percentage apoptosis of goblet cells) was measured by the same person (H.B.) using two samples at each of multiple different locations within the resected colon segment. Similarly, DBA reactivity was measured at the same distances from the tumor or at similar sites within a neoplasia-free colon resection as the samples taken for AI measurements. The mixed linear model was used to analyze the differences in AI among patients without colonic neoplasia, patients with TVA, and patients with cancer. For reactivity with the lectin DBA, a categorical variable was created, with 1 representing all locations aberrant; 2 representing a mixture of aberrant and sparse DBA reactivity and/or normal DBA reactivity; and 3 representing all locations normal. For this categorical variable, the generalized estimation equation model was used to analyze the differences among patients without neoplasia, patients with TVA, and patients with cancer. Variance components analysis was used to determine the proportion of variance between patients and within patients (location and resampling for AI and location for DBA reactivity). The Spearman correlation coefficient was used to measure the association between AI and DBA reactivity.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Tissue samples were taken from 16 colon resections (all those listed in Table 1, except for patients C99-40 and MS-091099, for whom only biopsies were taken during colonoscopies). Typical locations of the samples and subsequent treatments are indicated schematically in Fig. 4. All tissue samples were taken from the flat mucosa (the nontumorous mucosa), with samples taken adjacent to, 2 cm distant from, or maximally distant from the tumors in resected colon segments, or from flat mucosa from colon segments taken from colons without neoplasia.



View larger version (17K):
[in this window]
[in a new window]
  		FIG. 4. Schematic diagram showing locations and treatments of tissue samples taken from a typical colon resection from a patient with a tumor. MEM, minimum essential medium; NaDOC, sodium deoxycholate; Pmax, Dmax, samples taken maximally proximal or distal, respectively, from the tumor; P2, D2, samples taken 2-cm proximal or distal from the tumor; P0, D0, samples taken adjacent to the tumor on the proximal or distal side.

 
Patients Free of Colonic Neoplasia
Two colon resections were from patients with normal colons, where the patients had a recurrent liposarcoma external to, but not involving, the colonic wall. The colon was resected as part of the surgical procedure to remove the liposarcoma (and surrounding tissues). Two further colon resections were from patients with diverticulitis; one of these patients (C97-43) had considerable inflammation of the affected segment of the colon.

As shown in Fig. 5, these four resections were evaluated for (1) percentage apoptosis of goblet cells after incubation with 1.0 mM of NaDOC for 3 hours (also called AI); (2) differentiation measured by DBA reactivity; and (3) total NAD. The diagnosis for each patient, the location of the resection, and the age and sex of each patient are shown in the upper left of the graphs of AI.



View larger version (70K):
[in this window]
[in a new window]
  		FIG. 5. Data obtained from resections from patients free of colonic neoplasia. In all graphs, the abscissa shows the location of tissue samples that were evaluated, with distances indicated as centimeters (x) proximal (Px) or distal (Dx) to the center of resection (CR), a fistula (F0), or, in the case of patient C97-43, distal to an abscess. The bar graphs on the left indicate the apoptotic indices of goblet cells in pairs of tissue samples taken at the distances indicated; the bar graphs in the center show values of Dolichos biflorus agglutinin (DBA) reactivity at the distances indicated, where the fraction of each tissue section with normal DBA reactivity is shown by a filled area of a bar, the fraction of the tissue sample with sparse goblet cell–DBA reactivity is shown by a diagonally striped area, and the fraction of the tissue section with aberrant DBA reactivity is shown by an open area; the bar graphs on the right show total nicotinamide adenine dinucleotide (NAD) concentrations in tissue samples at the distances indicated.

 
In Fig. 5 (and in later figures where percentage apoptosis is plotted), the horizontal dashed line at 31.6% is 2 SD below the mean AI for biopsies treated with NaDOC. This value was obtained in a previous study of 14 neoplasia-free individuals by using the standard exposure to 1.0 mM of NaDOC for 3 hours.7 With a normal distribution of AI values for individuals without neoplasia, 97.7% of AI values are expected to be above this dashed line. The AI value 31.6% is designated AInor - 2SD. In this study, the AIs of the 27 tissue samples from resections of neoplasia-free colons all fell above AInor - 2SD, the dashed line. The AIs of pairs of tissue samples from these neoplasia-free individuals were generally close in value, as were the AIs at different locations within a resection. The coefficient of variation of AIs from patients without neoplasia was 16%.

Figure 2 shows an electron micrograph of a goblet cell in a tissue sample treated with NaDOC. The goblet cell is undergoing apoptosis, with classic condensation and margination of the chromatin.5,17 The images obtained by electron microscopy confirm that the goblet cells with dark-staining nuclei, counted as apoptotic in the light microscope, are indeed undergoing apoptosis.

The four graphs in the center column of Fig. 5 indicate DBA reactivity (i.e., level of differentiation) in samples at the positions shown on the abscissa. A black bar reaching 100% indicates that differentiation was entirely normal in the sample. Crypts with normal DBA staining are shown in Fig. 3A. The brown staining is associated with the mucin granules in the goblet cells of the surface epithelium and crypts.

A bar showing a fraction of its length with diagonal stripes indicates the fraction of the tissue sample with sparse DBA staining within the crypts. Crypts showing a sparse frequency of DBA-reactive (brown-staining) goblet cells are shown in Fig. 3B.

A bar (Fig. 5, center column) showing a fraction of its length as white (open) indicates the fraction of the sample having aberrant (non–goblet cell) staining with DBA. Crypts with all aberrant DBA staining are shown in Fig. 3C. A magnified image of a crypt with aberrant staining in shown in Fig. 3D. For the colon resections that were free of neoplasia (Fig. 5), 85.0% of the sample areas examined showed frequent DBA-reactive goblet cells, indicating normal differentiation of the goblet cells. Only 7% of the sampled area of the epithelium of the two normal colons showed a sparse frequency of goblet cells staining with DBA. One colon resection from a patient with diverticulitis had normal DBA staining in all regions examined. The other patient (C97-43) with diverticulitis had inflammation throughout the colon and also had the most samples with sparse or aberrant staining. The coefficient of variation of differentiation, as measured by reactivity with DBA, from patients without neoplasia was 30%.

The right column of four graphs in Fig. 5 shows the total concentration of NAD normalized to protein in the colonic epithelium. Two of these patients with colons without neoplasia (C98-2 and C97-42) had relatively high levels of NAD, and the other two (C99-59 and C97-43) had relatively low levels of NAD.

All of the tissues represented in Fig. 5 were from the left colon. Differentiation of goblet cells was generally normal as defined in Materials and Methods. However, resections from patients with colonic neoplasia included tissues from both the right and left colon. To determine whether DBA reactivity was simply a reflection of location within the colon, we obtained biopsies from five regions of the colon (cecum, transverse colon, descending colon [40 cm from the anal verge], sigmoid colon [20 cm from the anal verge], and rectum) of two patients free of colonic neoplasia (patients C99-40 and MS-091 of Table 1). Figure 6 shows that for these patients, DBA reactivity was normal in all regions and was similar in staining pattern and intensity to the 14 tissue samples taken at surgery from the two colon resections of patients without colonic disease.



View larger version (19K):
[in this window]
[in a new window]
  		FIG. 6. Dolichos biflorus agglutinin (DBA) reactivity observed at different locations in colons free of neoplasia. Values of DBA reactivity are indicated as in Fig. 5. Trans, biopsies taken from the transverse colon; 40 cm and 20 cm, biopsies taken at 40 and 20 cm from the anal verge, respectively.

 
Patients With a Colonic Adenocarcinoma
Figures 7 and 8 show the AIs, DBA reactivities, and NAD levels in the flat mucosa of colon resections from nine patients with colonic adenocarcinoma. Figure 7 shows data for flat mucosa of patients with frequent low AI values (40% to 100% of AIs below AInor - 2SD). Figure 8 shows data for flat mucosa; most AIs were in the range shown by patients without colonic neoplasia, although they were at low-normal values within this range, and <10% of AIs were below AInor - 2SD.



View larger version (48K):
[in this window]
[in a new window]
  		FIG. 7. Data from resections from patients with adenocarcinoma and with relatively low apoptotic index values. The abscissa shows distances proximal and distal to the tumor. The ordinate values are as indicated in Fig. 5. DBA, Dolichos biflorus agglutinin; NAD, nicotinamide adenine dinucleotide.

 


View larger version (55K):
[in this window]
[in a new window]
  		FIG. 8. Data from resections from patients with adenocarcinoma and low-normal apoptotic index values. The ordinate values are as indicated in Fig. 5. DBA, Dolichos biflorus agglutinin; NAD, nicotinamide adenine dinucleotide.

 
The entire colon of patient C99-49 was resected because this patient had ulcerative colitis. All values of AI were below AInor - 2SD at the 13 locations tested. Only a single bar for AI value is shown at each location. Although pairs of tissue samples were taken at each location, the number of goblet cells was small in each sample, and few of this low number of goblet cells were undergoing apoptosis. Therefore, the AIs of the pairs of tissue samples at each location were combined for this colon resection. This colon resection also had aberrant DBA reactivity at all 12 locations assessed. This colon resection was the only one, of the 17 evaluated, that had all AIs below AInor - 2SD and all DBA reactivities aberrant at all locations tested.

Total NAD levels were low in all five subjects with adenocarcinomas or TVA (Figs. 7–9). Although NAD is apparently not a good biomarker because of its variation in nonneoplastic colons, the data do show that all cancer and other high-risk patients tested had low NAD levels. This was also true of subjects with other types of cancer.24



View larger version (37K):
[in this window]
[in a new window]
  		FIG. 9. Data from resections from patients with tubulovillous adenomas (TVA). The ordinate values are as indicated in Fig. 5. DBA, Dolichos biflorus agglutinin; NAD, nicotinamide adenine dinucleotide.

 
From Figs. 5, 7, and 8, it is clear that individuals free of neoplasia have AIs and DBA reactivities with lower variability than individuals with colon cancer. In patients with colon cancer, AIs and DBA reactivities had considerable variability between samples taken at the same distances from the tumor and between locations at different distances from the tumor. AI and degree of differentiation varied in a patchy manner in colon cancer patients, and there was no apparent effect of distance from the tumor.

Patients With Large TVAs
The values obtained for AI, DBA reactivity, and total NAD for samples of colon resections from three patients with large TVAs are shown in Fig. 9. One patient (C98-23) had values of AI and DBA reactivity comparable to those obtained for patients without colonic neoplasia. A second patient (C98-1) had normal AI values but DBA reactivity similar to that shown among the group of resections with abnormal values shown in Fig. 7. The third patient (C97-4) had low AIs but relatively normal DBA reactivity.

All Patients
Data from all 16 resections are summarized in Fig. 10. The data are grouped in three columns: those from patients without neoplasia, those who had a TVA, and those who had an adenocarcinoma.



View larger version (18K):
[in this window]
[in a new window]
  		FIG. 10. Scatter diagram showing percentage apoptosis (AI) of goblet cells after treatment of colonic tissue samples with 1.0 mM of sodium deoxycholate in incubation medium for 3 hours. Tissue samples were from patients free of neoplasia (left), with tubulovillous adenomas (TVA; center), or with adenocarcinoma of the colon (right). (•) The AI obtained was near a paired tissue sample that had normal Dolichos biflorus agglutinin (DBA) reactivity (normal differentiation). ({triangleup}) Proximity to a tissue sample that had some fraction with sparse DBA reactivity. (*) The tissue sample was located near a paired sample that had some fraction with aberrant DBA reactivity.

 
The mean value of AIs from all samples from individuals without neoplasia was 59.1% ± 9.5%, for patients with TVAs was 54.8% ± 13.4%, and for individuals with adenocarcinoma was 37.4% ± 15.6%. The difference between the mean AI of patients without neoplasia and the mean AI of patients with colon cancer was significant (P < .01). Similarly, the difference between the mean AIs of patients with TVA and of patients with colon cancer was significant (P = .02). The difference between the mean AIs of patients without neoplasia and patients with a TVA was not significant (P = .08).

The differences between DBA reactivities of patients without neoplasia and of patients with colon cancer were significant (P = .03). However, the differences between DBA reactivities of patients with TVA and of patients with colon cancer were not significant (P = .69), and there was no significant difference between DBA reactivities of patients without neoplasia and patients with a TVA (P = .88).

By using components of variance analysis, we assessed potential sources of variation in AI between individuals, between locations within a resection, and between tissue samples taken at the same distance from either the tumor or the center of the resection (for individuals without colonic neoplasia). If there was substantial interindividual variability compared with intraindividual variability (because of location and samples taken at the same distances from the tumor or center of resection), this would suggest a greater homogeneity of AI values within each individual. However, if considerable intraindividual variability were found, this would reflect biologic differences from area to area.

The largest component of variation in AI in cancer patients was between individuals (50.6%). However, 34.9% of the variance was between samples taken at the same distance from the tumor, and 14.5% was due to different locations, indicating patchiness in susceptibility to bile acid-induced apoptosis in the colon of these patients at high risk for subsequent cancer.

In Fig. 10, 60.4% of samples with a normal AI were near a location with a normal pattern of differentiation, whereas 20.8% were near a location with sparse DBA staining (low-normal differentiation) and 18.9% were near a location with aberrant differentiation. By comparison, 21.6% of samples with low AI were associated with a nearby sample with sparse or low differentiation, and 67.6% were associated with a nearby sample with aberrant differentiation.

Low AIs are very frequent in patients with an adenocarcinoma and are not present in samples from individuals without neoplasia. Thus, low AI is specific as a marker for susceptibility to cancer. Dedifferentiation is not as specific, but it seems to be more sensitive as a marker of susceptibility to neoplasia. Overall, for the association between DBA reactivity and AI, the Spearman correlation was .61, and the association was significant (P = .01). The degree of aberrant DBA staining in nongoblet cells was, however, specific for the high-risk group, because no biopsy taken from the neoplasia-free patients exhibited >35% aberrant DBA staining.

SAC frequencies were determined in 23 samples obtained from 4 different individuals. These were from one individual without neoplasia, one individual with a TVA, and two individuals with an adenocarcinoma. The frequencies of SACs in 7 samples from the individual without neoplasia were .4 to 1.5 per millimeter, in 6 samples from the individual with a TVA were 0 to .5 per millimeter, and in 13 samples from the 2 individuals with adenocarcinomas were 0 to 1.0 per millimeter. Because the frequency of SACs in tissue samples from the normal individual tended to be higher than in samples from individuals with TVA or adenocarcinomas, the frequency of SACs does not seem promising as a biomarker of progression toward colon carcinogenesis.

In all of the 105 tissue samples evaluated for DBA staining, shown in Figs. 5 and 7–9, only one ACF and one microadenoma were found. Thus, ACF and microadenomas do not seem promising as biomarkers of progression toward colon carcinogenesis with use of routinely prepared biopsies.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
We performed a study of the variability of potential biomarkers of colon cancer risk, using freshly obtained human colonic resections. It was determined that resistance to the induction of apoptosis with a novel ex vivo bioassay was the most specific of the biomarkers and was present in 59% of the normal-appearing mucosal samples from patients with colon cancer. Apoptosis resistance highly correlated with a low level of differentiation, assessed with DBA lectin staining. A high fraction (>35%) of tissue showing aberrant lectin staining (of nongoblet cells) was present only in the group of patients with colonic neoplasms, also making aberrant lectin reactivity a specific biomarker. In one case of a TVA patient, aberrant lectin staining placed the patient in the high-risk group, although apoptosis resistance was not present. The patchy nature of the field defects associated with apoptosis resistance and DBA staining underscores the necessity for multiple biopsies to assess colon cancer risk. The most important aspect of our study is to point out this caveat of patchiness, because it will probably apply to other biomarkers as well.

In a previous study, we evaluated AIs of biopsies taken at the cecum, descending colon, and sigmoid colon. Those biopsies were from three groups of individuals: those who never had colonic neoplasia, those with previous or current adenomatous polyps (including TVAs), and those with a history of colonic adenocarcinoma.7 In Table 2 we see that the values of AI obtained in this study are similar to those found in our previous study.


View this table:
[in this window]
[in a new window]
  		TABLE 2. Comparison of mean AI values
 
The typical high-fat Western diet causes the colonic epithelium to be exposed to a relatively high level of bile acids. This exposure may cause epithelial cells to frequently undergo apoptosis. There would then be a selective pressure for cells with mutations causing apoptosis resistance to survive and repopulate the colonic epithelium. Selection for apoptosis-resistant cells, by high-fat meals, may thus give rise to patches of the colonic epithelium with reduced ability to undergo apoptosis. Such apoptosis-resistant cells would probably survive even when containing unrepaired DNA damages, and thus, on replication, would have an increased tendency to undergo mutation, including carcinogenic mutation. In fact, Cherbonnel-Lasserre et al.25 and Liu et al.26 have shown that when Bcl-2 is overexpressed, causing resistance to apoptosis, surviving cells have a higher incidence of mutation or retained DNA damage after either x-ray or UV irradiation than surviving control cells that are not apoptosis resistant. By the previous reasoning, apoptosis resistance, or low AI values in response to bile acid treatment, is expected as a biomarker of colon cancer risk.4,6,7

It was somewhat surprising to find that ACF and microadenomas, considered by many investigators to represent preneoplastic lesions, were rare in the 77 samples that we evaluated from colon cancer and TVA patients (one ACF and one microadenoma). Aberrant crypts have long been suggested to reflect early events in carcinogenesis.27 More recent work using carcinogen-treated rats has shown a correlation of cancer risk with large ACF with more than six crypts per focus but not with small ACF.28 Studies of human colon resections indicated that large ACF showed dysplasia when examined histologically, indicating that large ACF were microadenomas.29 However, with our methods we found that the frequency of SACs in flat mucosa did not correlate with colon cancer and that ACF or microadenomas are too rare to be useful as biomarkers of colon cancer risk. Therefore, the absence of these preneoplastic lesions from routine biopsies should not be taken as an indication of low colon cancer risk. The tissue samples we evaluated for ACF and microadenomas were similar in size to biopsies that would be available from patients undergoing colonoscopy. Although ACF and microadenomas are expected as biomarkers of colon cancer risk when an entire animal colon is examined with methylene blue staining, our data indicate that such lesions are not easily detected in biopsy-sized tissue samples taken from human colons.

The repair of DNA damages consumes NAD as a result of PARP activity to make adenosine diphosphate–ribose polymers.30 The marked consumption of NAD during inflammatory processes is underscored by the development of pellagra in a patient with ulcerative colitis.31 Patients with familial adenomatous polyposis also have decreased NAD levels and show reduced stimulation of PARP activity by DNA damage.32 However, we find that the total NAD content in the flat colonic mucosa of individuals without neoplasia varies over a wide range, as has been observed for blood cells23 and skin.24 This may reflect dietary differences between individuals. These data, therefore, preclude any conclusion about decreased NAD levels and colon cancer risk, even though all colonic mucosa samples tested from subjects with colon cancer and large adenomas had low NAD values.

Some agents thought to be protective against colon cancer— including butyrate (produced during the fermentation of fiber in the colon), dietary selenium, and fish oil—cause both increased differentiation of colon cells and an increased propensity to undergo protective apoptosis.11–14 The increased differentiation is correlated with alterations in the mucins produced by the goblet epithelial cells.14,33 Thus, deficient differentiation, marked by altered or deficient mucin production, could be expected to be a biomarker for colon cancer risk. Mucins are high– molecular-weight glycoproteins containing up to 85% oligosaccharides.34 Lectins are a class of polypeptides in which each specific type of lectin binds to a specific sugar moiety in a glycoprotein or glycolipid.35 Boland et al.33 showed that the plant lectin DBA binds preferentially to terminal N-acetyl-galactosamine of mucin and that terminal N-acetyl-galactosamine occurred in well-differentiated goblet cell mucin of the upper colonic crypt. On the basis of this, DBA can be used as a probe to assess the level of differentiation of goblet cells. We show that sparse or aberrant DBA reactivity is a sensitive biomarker for colon cancer risk and that a high level of aberrant lectin staining in nongoblet cells is specific for high-risk patients.

Overall, the results of our study suggest that the live-cell bioassay for AI and the more practical DBA staining assay on preserved tissue samples are promising biomarkers of colon cancer risk, but multiple samples must be obtained to give a valid indication of risk. The observed patchiness of field defects has important implications for future biomarker development.


   Acknowledgments
 
The authors thank Mandala V. Wilson for performing NAD(P) analyses. Supported by National Institutes of Health Program Project Grant CA72008, Arizona Disease Control Research Commission Grants 10016 and 6002, National Institutes of Health Institutional Core Grant CA23074, National Institute of Environmental Health Sciences Grant ES06694, National Institutes of Health Grants CA43894 and CA65579, and Veterans Affairs Hospital Merit Review Grant 2HG.

Received for publication November 7, 2001. Accepted for publication February 27, 2002.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell 1990; 61: 759–67.[CrossRef][Medline]
  2. Issa J-PJ, Vertino PM, Boehm CD, Newsham IF, Baylin SB. Switch from monoallelic to biallelic humanIGF2 promoter methylation during aging and carcinogenesis. Proc Natl Acad Sci U S A 1996; 93: 11757–62.[Abstract/Free Full Text]
  3. Baylin S. Tying it all together: epigenetics, genetics, cell cycle, and cancer. Science 1997; 277: 1948–9.[Free Full Text]
  4. Bernstein C, Bernstein H, Payne CM, Garewal H. Field defects in progression to adenocarcinoma of the colon and esophagus. Electronic J Biotechnol (online) 2000;3. Available at: http://www. ejb.org/content/vol3/issue3/full/1./ Accessed December 15, 2000.
  5. Payne CM, Bernstein H, Bernstein C, Garewal H. Role of apoptosis in biology and pathology: resistance to apoptosis in colon carcinogenesis. Ultrastruct Pathol 1995; 19: 221–48.[Medline]
  6. Garewal H, Bernstein H, Bernstein C, Sampliner R, Payne C. Reduced bile acid-induced apoptosis in "normal" colorectal mucosa: a potential biological marker for cancer risk. Cancer Res 1996; 56: 1480–3.[Abstract/Free Full Text]
  7. Bernstein C, Bernstein H, Garewal H, et al. A bile acid-induced apoptosis assay for colon cancer risk and associated quality control studies. Cancer Res 1999; 59: 2353–7.[Abstract/Free Full Text]
  8. Corbet SW, Clarke AR, Gledhill S, Wyllie AH. P53-dependent and -independent links between DNA-damage, apoptosis and mutation frequency in ES cells. Oncogene 1999; 18: 1537–44.[CrossRef][Medline]
  9. Kuo M-L, Shiah S-G, Wang C-J, Chuang S-E. Suppression of apoptosis by Bcl-2 to enhance benzene metabolites-induced oxidative DNA damage and mutagenesis: a possible mechanism of carcinogenesis. Mol Pharmacol 1999; 55: 894–901.[Abstract/Free Full Text]
  10. Williams GT, Critchlow MR, Hedge VL, O’Hare KB. Molecular failure of apoptosis: inappropriate cell survival and mutagenesis? Toxicol Lett 1998; 102–103: 485–9.
  11. Heerdt BG, Houston MA, Augenlicht LH. Potentiation by specific short-chain fatty acids of differentiation and apoptosis in human colonic carcinoma cell lines. Cancer Res 1994; 54: 3288–94.[Abstract/Free Full Text]
  12. Hague A, Manning AM, Hanlon KA, Huschtscha LI, Hart D, Paraskeva C. Sodium butyrate induces apoptosis in human colonic tumour cell lines in a p53-independent pathway: implications for the possible role of dietary fibre in the prevention of large-bowel cancer. Int J Cancer 1993; 55: 498–505.[Medline]
  13. Stewart MS, Davis RL, Walsh LP, Pence BC. Induction of differentiation and apoptosis by sodium selenite in human colonic carcinoma cells (HT29). Cancer Lett 1997; 117: 35–40.[CrossRef][Medline]
  14. Chang W-CL, Chapkin RS, Lupton JR. Fish oil blocks azoxymethane-induced rat colon tumorigenesis by increasing cell differentiation and apoptosis rather than decreasing cell proliferation. J Nutr 1997; 128: 491–7.[Abstract/Free Full Text]
  15. Jackson AL, Loeb LA. The mutation rate and cancer. Genetics 1998; 148: 1483–90.[Abstract/Free Full Text]
  16. Payne CM, Crowley C, Washo-Stultz D, et al. The stress-response proteins poly(ADP-ribose) polymerase and NF-{kappa}B protect against bile salt-induced apoptosis. Cell Death Differ 1998; 5: 623–36.[CrossRef][Medline]
  17. Payne CM, Bernstein C, Bernstein H. Apoptosis overview emphasizing the role of oxidative stress, DNA damage and signal transduction pathways. Leuk Lymphoma 1995; 19: 43–93.[Medline]
  18. Boland CR, Martin MA, Goldstein IJ. Lectin reactivities as intermediate biomarkers in premalignant colorectal epithelium. J Cell Biochem Suppl 1992; 16G: 103–9.
  19. Bresalier RS, Boland CR, Kim YS. Regional differences in normal and cancer-associated glycoconjugates of the human colon. J Natl Cancer Inst 1985; 75: 249–60.
  20. McLellan EA, Medline A, Bird RP. Sequential analyses of the growth and morphological characteristics of aberrant crypt foci: putative preneoplastic lesions. Cancer Res 1991; 51: 5270–4.[Abstract/Free Full Text]
  21. Pretlow TP, O’Riordan MA, Pretlow TG, Stellato TA. Aberrant crypts in human colonic mucosa: putative preneoplastic lesions. J Cell Biochem Suppl 1992; 16G: 55–62.
  22. Takayama T, Katsuki S, Takahashi Y, et al. Aberrant crypt foci of the colon as precursors of adenoma and cancer. N Engl J Med 1998; 339: 1277–84.[Abstract/Free Full Text]
  23. Jacobson EL, Jacobson KK. Tissue NAD as a biochemical measure of niacin status in humans. Methods Enzymol 1997; 280: 221–30.[CrossRef][Medline]
  24. Jacobson EL, Shieh WM, Huang AC. Mapping the role of NAD metabolism in cancer prevention and treatment. Mol Cell Biochem 1999; 193: 69–74.[CrossRef][Medline]
  25. Cherbonnel-Lasserre C, Gauny S, Kronenberg A. Suppression of apoptosis by Bcl-2 or Bcl-xL promotes susceptibility to mutagenesis. Oncogene 1996; 13: 1489–97.[Medline]
  26. Liu Y, Naumovski L, Hanawalt P. Nucleotide excision repair capacity is attenuated in human promyelocytic HL60 cells that overexpress Bcl2. Cancer Res 1997; 57: 1650–3.[Abstract/Free Full Text]
  27. Wargovich MJ, Medline A, Bruce WR. Early histopathologic events to evolution of colon cancer in C57BL/6 and CF1 mice treated with 1,2-dimethylhydrazine. J Natl Cancer Inst 1983; 71: 125–31.
  28. Davies MJ, Bowey EA, Adlercreutz H, Rowland IR, Rumsby PC. Effects of soy or rye supplementation of high-fat diets on colon tumour development in azoxymethane-treated rats. Carcinogenesis 1999; 20: 927–31.[Abstract/Free Full Text]
  29. Roncucci L, Medline A, Bruce WR. Classification of aberrant crypt foci and microadenomas in human colon. Cancer Epidemiol Biomarkers Prev 1991; 1: 57–60.[Abstract]
  30. Jacobson EL. Niacin deficiency and cancer in women. J Am Coll Nutr 1993; 12: 412–6.[Abstract]
  31. Jarrett P, Duffill M, Oakley A, Smith A. Pellagra, azathioprine and inflammatory bowel disease. Clin Exp Dermatol 1997; 22: 44–5.[CrossRef][Medline]
  32. Cristovao L, Lechner MC, Fidalgo P, Leitao CN, Mira FC, Rueff J. Absence of stimulation of poly(ADP-ribose) polymerase activity in patients predisposed to colon cancer. Br J Cancer 1998; 77: 1628–32.[Medline]
  33. Boland CR, Montgomery CK, Kim YS. Alterations in human colonic mucin occurring with cellular differentiation and malignant transformation. Proc Natl Acad Sci U S A 1982; 79: 2051–5.[Abstract/Free Full Text]
  34. Jass JR, Roberton AM. Colorectal mucin histochemistry in health and disease: a critical review. Pathol Int 1994; 44: 487–504.[Medline]
  35. Raedler A, Raedler E. The use of lectins to study normal differentiation and malignant transformation. J Cancer Res Clin Oncol 1985; 109: 245–51.[CrossRef][Medline]



This article has been cited by other articles:


Home page
 Cancer Epidemiol. Biomarkers Prev.Home page
C. M. Payne, H. Holubec, C. Bernstein, H. Bernstein, K. Dvorak, S. B. Green, M. Wilson, M. Dall'Agnol, B. Dvorakova, J. Warneke, et al.
Crypt-Restricted Loss and Decreased Protein Expression of Cytochrome c Oxidase Subunit I as Potential Hypothesis-Driven Biomarkers of Colon Cancer Risk
Cancer Epidemiol. Biomarkers Prev., September 1, 2005; 14(9): 2066 - 2075.
[Abstract] [Full Text] [PDF]

Home page
 J. Histochem. Cytochem.Home page
H. Holubec, C. M. Payne, H. Bernstein, K. Dvorakova, C. Bernstein, C. N. Waltmire, J. A. Warneke, and H. Garewal
Assessment of Apoptosis by Immunohistochemical Markers Compared to Cellular Morphology in Ex Vivo-stressed Colonic Mucosa
J. Histochem. Cytochem., February 1, 2005; 53(2): 229 - 235.
[Abstract] [Full Text] [PDF]

Home page
 CarcinogenesisHome page
H. Bernstein, C. M. Payne, K. Kunke, C. L. Crowley-Weber, C. N. Waltmire, K. Dvorakova, H. Holubec, C. Bernstein, R. R. Vaillancourt, D. A. Raynes, et al.
A proteomic study of resistance to deoxycholate-induced apoptosis
Carcinogenesis, May 1, 2004; 25(5): 681 - 692.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bernstein, H.
Right arrow Articles by Bernstein, C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bernstein, H.
Right arrow Articles by Bernstein, C.
Related Collections
Right arrow Other Laboratory Research

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS