THE AMERICAN JOURNAL OF GASTROENTEROLOGY
1, 2000
© 2000 by Am. Coll. of Gastroenterology
Published by Elsevier Science Inc.
Expression of Brain-Type
Glycogen Phosphorylase Is a
Potentially Novel Early Biomarker in the
Carcinogenesis of Human Colorectal Carcinomas
Satoshi Tashima, M.D., Shinya Shimada, M.D., Kenji Yamaguchi, M.D., Junji Tsuruta, M.D., and
Michio Ogawa, M.D.
Departments of Surgery II and Surgical Pathology, Kumamoto University School of Medicine,
Kumamoto, Japan
OBJECTIVE: Our previous studies have demonstrated the sig-
nificant role of brain-type glycogen phosphorylase (BGP) in
the carcinogenesis of gastric carcinoma. The aims of the
present study were to investigate the expression of BGP in
colorectal carcinoma as well as the timing of this expression
in the adenoma-carcinoma sequence (ACS), in comparison
with the overexpression of p53 protein. We also sought to
identify this marker in the particular colorectal mucosa
bearing de novo carcinoma.
METHODS: The expression of BGP and p53 protein in colo-
rectal carcinoma using affinity purified specific anti-human
BGP antibody (Ab) and anti-p53 Ab was studied using 96
resected specimens. Further investigation to examine the
timing of BGP expression in comparison with p53 overex-
pression was carried out using 13, 18, eight, and 16 speci-
mens of adenoma with mild, moderate, and severe dyspla-
sia, and carcinoma in adenoma, respectively. The BGP
immunohistochemistry in whole resected human colorectal
mucosa (two with carcinoma and one with ulcer) was car-
ried out using specific anti-BGP and anti-p53 Ab.
RESULTS: The BGP visualized by immunohistochemistry
was commonly present in colorectal carcinoma (83.3%).
The expression of this molecule during ACS showed excel-
lent correlation with the increased dysplasia and was found
before p53 overexpression, whereas no BGP expression was
seen in the normal human large intestine remote from the
cancer foci. Positive staining in overtly normal-looking co-
lonic mucosa was observed mainly around carcinomas with-
out any adenoma component.
CONCLUSIONS: The present study is the first to localize the
BGP molecule in colorectal carcinoma, adenoma, and nor-
mal mucosa. It is suggested that BGP is a novel biomarker
for carcinogenesis in both the pathways of ACS and the de
novo colorectal carcinoma. (Am J Gastroenterol 2000;95:
255–263. © 2000 by Am. Coll. of Gastroenterology)
Vol. 95, No.
ISSN 0002-9270/00/$20.00
PII S0002-9270(99)00751-0
INTRODUCTION
Adenoma-carcinoma sequence (ACS) has been widely ac-
cepted as the dominant pathway in the carcinogenesis of
colonic carcinoma (1–3). The scenario of ACS is persuasive
and the multistage progressive processes have been clearly
elucidated using various oncogenes and tumor suppresser
genes (2, 3). Utilizing these established multiple steps, tu-
mor markers expressed in carcinoma and precancerous le-
sions can be selected as biomarkers for cancer risk as well
as chemoprevention of colonic carcinoma. Furthermore,
such a marker may contribute to identify precancerous le-
sions leading to de novo carcinoma (4 –7), in which little
carcinogenesis has yet been elucidated, and to study the
direct link of genetic alterations.
Glycogen phosphorylase (GP) (EC 2.4.1.1) plays a cen-
tral role in the mobilization of carbohydrate reserves in a
wide variety of organs and tissues, and has a long history of
being investigated, starting in the 1930s (8 –10). Mamma-
lian GPs are found in three major isoforms, i.e., muscle,
liver, and brain (BGP), which can be distinguished by func-
tional and structural properties as well as by the tissues in
which they are predominantly expressed (10, 11). In recent
years, cDNAs encoding the three human GP isoforms have
been cloned and sequenced (10, 12, 13). The sequences have
been aligned and these data demonstrated a remarkable
conservation of primary sequence in these GPs. However,
chromosome mapping analyses have revealed that the genes
encoding muscle, liver, and brain-type GP are assigned to
chromosomes 11, 14, and 20, respectively, suggesting that
distinct cis-acting elements govern the differential expres-
sion of the phosphorylase isoforms in various tissues (13).
The physiological role of brain-type GP is poorly under-
stood, but it is generally thought to induce an emergency
glucose supply during a stressful period (10). In addition, it
has been suggested that the major isoform of GP found in
fetal tissue and tumor tissues is BGP (14 –18).
We have previously demonstrated, through enzyme and
immunohistochemical investigations, the abnormally ex-
256 Tashima et al.
pressed glycogen phosphorylase activity in gastric carci-
noma (15, 19 –21). The anti-BGP antibody (Ab) against an
immunogen peptide of BGP has demonstrated specific re-
activity on Western blot analysis and good reactivity in the
conventional paraffin section at a low Ab concentration
(21). This enabled us to localize BGP clearly in the carci-
noma tissues. Interestingly, immunohistochemical analyses
using this specific anti-BGP Ab revealed that BGP was
expressed in approximately 80% (58/72) of the intestinal-
type and 20% (14/72) of the diffuse-type carcinoma in the
cytoplasm. A strong correlation was seen between expres-
sion of BGP and both gastric cancer and intestinal metapla-
sia, whereas no positive staining was observed in the normal
gastric epithelial cells.
The p53 tumor suppresser gene has been considered to be
crucial in the demonstration of cell proliferation in cell cycle
phases (22) and is one of the most commonly mutated genes
in a variety of tumors (23–25). The genetic alteration of p53
has been investigated in the chain of ACS and reported to be
a late event in colonic carcinogenesis (26 –29).
Here, we describe for the first time the localization of
BGP in colonic normal mucosa, adenoma with varying
degrees of dysplasia, and carcinoma; the timing of expres-
sion of this molecule (when it expresses) in the ACS in
comparison with p53; and BGP-positive normal-appearing
mucosa adjacent to colonic carcinoma without adenoma as
potential precancerous foci of de novo carcinoma.
MATERIALS AND METHODS
Tissue Specimens
Human specimens were surgically (n ⫽ 96) or endoscopi-
cally (n ⫽ 55) resected from patients with colorectal carci-
noma or adenoma. To study the distribution of BGP foci in
the whole mucosa, and because a considerable number of
such cancerous lesions were assumed to be de novo carci-
nomas, we selected one resected colon and rectum with a
superficially depressed-type early carcinoma of small size
(1.1 cm and 1.5 cm, respectively), markedly invaded into the
submucosa without adenoma component (4, 6, 30 –33). One
with colonic ulcer (diagnosed as Behcet’s disease), resected
due to bleeding, was added as a control. Each specimen was
first cut into 40 ⫻ 10 mm blocks. Tissues were fixed in 10%
buffered formalin for 4 days, embedded in paraffin, and then
cut into five serial sections for histological and immunohis-
tochemical examinations.
Antibodies
Antibody against human BGP was raised as previously
reported by Ignacio et al. (34), with modification. Briefly, a
synthesized 13-residue peptide (CDLQIPPPNIPRD), corre-
sponding to cystein, coupled to the 12-carboxyterminal res-
idues of BGP was chosen for the immunogen. It had no
significant homology with other proteins including human
liver and muscle-type GP using Gene Works homology
search. The N-terminal cystein of the peptide was coupled
AJG – Vol. 95, No. 1, 2000
with the activated keyhole limpet hemocyanin (KLH)
(Pierce, Rockford, IL). Adult rabbits were immunized at
1-wk intervals by four subcutaneous (s.c.) injections of 100
␮g of the KLH peptide, with Freund’s complete or incom-
plete adjuvant. One week after the last injection the same
amount of KLH peptide was injected s.c. as a booster, and
1 wk later blood samples were collected from the jugular
vein. The IgG (IgG) fractions of pooled antisera (approxi-
mately 100 ml) were precipitated by adding saturated
(NH4)2SO4 (50% saturation), dissolved in 100 ml of PBS
and dialyzed extensively in the same buffer. A part of this
IgG fraction (2 ml) was then applied to a column of BGP-
coupled Sepharose (1 mg of the BGP peptide was coupled
with 1 ml of Hi Trap NHS-activated Sepharose) (Pharmacia
Biotech, Uppsala, Sweden) at a flow rate of 0.5 ml/min at
room temperature. The column was washed successively
with buffer A (0.5 mol/L ethanolamine, 0.5 mol/L NaCl, pH
8.3) and buffer B (0.1 mol/L acetate buffer, 0.5 mol/L NaCl,
pH 4.0). Then the antibodies bound in the column were
eluted with 0.1 mol/L glycine-HCl buffer, pH 2.5, contain-
ing 1 mol/L NaCl. Fractions containing antibodies, which
were detected by measuring the absorbance at 280 mn, were
collected, neutralized immediately with 1 mol/L NaOH, and
dialyzed against phosphate-buffered saline buffer. The spec-
ificity of the anti-human BGP Ab has been previously ex-
amined using Western blot analyses (21). The results
showed that affinity purified anti-BGP Ab against synthe-
sized peptide had specific reactivity (detected as a single
band at an expected size of approximately 10 kD) to the
extract of the U251 (the cDNA of BGP has been cloned
from this astrocytoma cell line) and AGS (gastric cancer)
cell lines, whereas none of the reactive bands were observed
in the homogenate from muscle, liver, and the AZ521 gas-
tric cancer cell line. In enzyme-histochemical and -cyto-
chemical analyses, the GP was positive in U251, AGS,
muscle, and liver but it was negative in AZ521. These
findings clearly indicated that anti-BGP Ab specifically re-
acts against the BGP, but not against the muscle or liver-
type GP.
Anti-p53 (BP53.12) Ab was purchased from Zymed Lab-
oratory Inc. (San Francisco, CA).
Immunohistochemistry
The ABC Elite kits (Vector Laboratories, Inc. Burlingame,
CA) for rabbit IgG (anti-BGP) and mouse IgG (anti-p53)
were used. Sections of formalin-fixed and paraffin-embed-
ded tissue (3-␮m thick) were deparaffinized and hydrated
through xylene and graded ethanols. Sections were incu-
bated with normal horse serum for 30 min after the blocking
of endogenous peroxidase activity with 0.3% H2O2 in meth-
anol for 30 min, incubated overnight at 4°C with optimally
diluted (see later section) primary Ab, and subsequently
incubated with biotinylated anti-rabbit and -mouse Ab and
avidin-biotin peroxidase complex for 30 min at room tem-
perature. They were washed in 10 mmol/L PBS (pH 7.2)
between each incubation step. The sites of peroxidase ac-
AJG – January, 2000
tivity were visualized by the diaminobenzidine method.
Sections were counterstained with hematoxylin for micro-
scopic examination. In p53 staining, heat antigen retrieval
was used (35). Sections were immersed in 1 L of citrate
buffer (2.1 g citric acid in 1 L distilled water, pH 6.0, with
2 mol/L NaOH) and boiled for 30 min before the immuno-
histochemical procedure. The working dilution of the pri-
mary antibody, affinity purified anti-BGP Ab BP53.12 (anti-
p53 Ab), was 1 ␮g/ml. BP 53.12 (anti-p53 Ab) was
prediluted by the manufacturer and applied directly without
additional dilution. As a negative control, nonimmunized
rabbit and mouse IgG2a were used instead of the primary
Ab. Furthermore, to confirm the specificity of the affinity
purified anti-BGP Ab, the supernatant of the Ab after ab-
sorption with the immunized peptide was used for the first
layer in the step of immunohistochemistry. Briefly, the
anti-BGP Ab (1 ␮g/ml) was incubated with the excess
amount of synthesized peptide (500 ng/ml; molecular ratio
of Ab and the peptide was 1:50) for 1 h at 37°C. After
centrifugation at 10,000 ⫻ g for 1 h, the supernatant was
used for the first layer of the immunohistochemistry process,
instead of the anti-BGP; the same procedures were followed
for anti-BGP Ab.
Staining Evaluation
The entire area of the tissue was examined by low-power (⫻
10) optical fields and the percentage of positively stained
cells was independently estimated in 10 fields for each case
by three of the authors (S.T., S.S., and J.T.). To evaluate the
correlation between positivity of BGP and that of p53, the
mean percentages (with range) were calculated and the
results were graded as follows: (⫺) for absent or ⬍10%,
(1⫹) for 10 –25%, (2⫹) for 25– 60%, and (3⫹) for ⬎60%
positive. The carcinomas with ⬎10% staining of BGP or
p53 were defined as positive.
Statistical Analysis
Statistical analyses were performed using Fisher’s exact
test.
RESULTS
BGP and p53 Expression in
Colorectal Carcinoma and Normal Mucosa
The reactivity of colonic carcinoma against anti-BGP Ab is
shown in Figure 1. Positive reactivity was observed in the
cytoplasm of cancer cells, whereas no staining was detected
when using the supernatant of the Ab after absorption with
the immunized peptide of BGP for the primary Ab. No
immunohistochemical staining of BGP was observed in
normal colonic and rectal mucosa remote from the cancer
foci, even in their proliferating zone (except in the particular
foci described later). The BGP positivity in the 96 cases of
colorectal carcinoma was 83.3% as a whole (Table 1). There
was no significant difference between the location of the
carcinoma and the BGP or p53 positivity. Further, the 96
BGP in Carcinogenesis of Colorectal Cancer 257
cases of resected colorectal carcinoma were classified with
Dukes’ classification. Neither the BGP nor p53 reactivity
showed significant correlation with the location of the car-
cinoma and the stage according to Dukes’ classification. As
shown in Table 1, the positivity of BGP in the colorectal
carcinoma (83.3%) was significantly higher than that of p53
(50%) (p ⬍ 0.001).
Timing of BGP and p53 Expression in the ACS
The polypectomized adenomas were classified into four
groups according to the UICC (36): adenoma with mild
dysplasia (n ⫽ 13), moderate dysplasia (n ⫽ 18), severe
dysplasia (n ⫽ 8), and carcinoma in adenoma (n ⫽ 16). The
representative reactivity of each group against anti-BGP Ab
is shown in Figure 2. The positive BGP staining was ob-
served in adenoma with moderate or severe dysplasia and
carcinoma in adenoma, whereas adenoma with mild dyspla-
sia had no apparent staining (Fig. 2A and 2D). The reactivity
and frequency of BGP positivity are shown in Table 2.
Interestingly, BGP was expressed abruptly at the stage of
the adenomas with moderate dysplasia, and the frequency
increased with the progress of the dysplasia. The percent-
ages of immunohistochemical positivity for anti-BGP were
7.7, 61.1, 87.5, and 93.8% for adenoma with mild dysplasia,
moderate dysplasia, severe dysplasia, and carcinoma in ad-
enoma, respectively. There were significant differences be-
tween adenoma with mild dysplasia and moderate (p ⫽
0.002) or severe (p ⫽ 0.0005) adenoma, and carcinoma in
adenoma (p ⬍ 0.0001), respectively. p53 overexpression in
ACS was also examined immunohistochemically on serial
sections of BGP (Fig. 2B and 2E). p53 protein was not
detected in either the adenoma with mild or moderate dys-
plasia, but was detected in only two (weakly positive) of six
adenomas with severe dysplasia and 43.8% of the carcinoma
in adenoma.
BGP-Positive Foci in Normal Mucosa
Adjacent to Carcinoma and in
Resected Colon With Carcinoma and Ulcer
During these studies, BGP-positive foci in the overtly nor-
mal-looking mucosa neighboring to the carcinoma were
occasionally observed (Fig. 3). BGP positivity in the normal
mucosa adjacent to carcinoma was examined using 66 spec-
imens with BGP-positive colorectal carcinoma involving
the overtly normal-looking mucosa. Interestingly, BGP-pos-
itive foci were frequently (90.9%) observed in the overtly
normal-looking mucosa adjacent to carcinoma (Table 3). To
investigate whether BGP is expressed in some focus in the
normal-appearing mucosa remote from cancer foci, resected
colon and rectum with small and early carcinoma (with
submucosal cancer invasion) without an adenoma compo-
nent and one with colonic ulcer resected because of bleeding
were studied. The BGP and p53 positivity was mapped on
the whole specimens (Fig. 4). The BGP and p53 expression
was found in both of the cancer foci, but not in the ulcers.
Interestingly, BGP-positive foci frequently existed in the
258 Tashima et al. AJG – Vol. 95, No. 1, 2000

Figure 1. Immunohistochemical staining of colonic carcinoma with anti-BGP antibody (A) and anti-p53 antibody (B). Strongly positive
reactivity was observed in the cytoplasm (A) and nuclei (B) of the cancer cells (negative in the neighboring normal colonic mucosa). None
of the staining was detected when using the supernatant of the anti-BGP antibody after absorption with the immunized peptide (C) or mouse
IgG2a instead of the primary antibody (D). Bar: 100 ␮m.

normal mucosa around carcinoma. There were no apparent DISCUSSION

morphological properties in the BGP-positive foci com-
pared with the BGP-negative normal mucosa (Figs. 3 and
5A). On the other hand, very few foci were observed in the
resected colon with ulcers. No p53 overexpression was
observed anywhere in the overtly normal-looking mucosa
examined.
It has been demonstrated that BGP is identical to GP in the
hepatocellular carcinomas or fetal liver, i.e., so-called “fe-
tal-type” (14, 37). Therefore, the expression of BGP in
colonic cancer can be considered to be an instance of on-
cofetal antigen reversion in cancer, in the same way that

Table 1. BGP and p53 Expression in Colorectal Carcinoma

Location Right Colon Left Colon Rectum Total
BGP Positivity (%) 22/26 (84.6) 38/42 (90.5) 20/28 (71.4) 80/96 (83.3)*
Dukes’ A 6/6 12/14 8/8 26/28
B 0/2 6/6 4/8 10/16
C 14/15 16/17 7/11 37/43
D 2/3 4/5 1/1 7/9
p53 Positivity (%) 11/26 (42.3) 25/42 (59.5) 12/28 (42.8) 48/96 (50.0)
Dukes’ A 2/6 10/14 4/8 16/28
B 1/2 1/6 4/8 6/16
C 6/15 11/17 3/11 20/43
D 2/3 3/5 1/1 6/9
*Significantly different from the value for p53 positivity at p ⬍ 0.001.
BGP ⫽ brain-type glycogen phosphorylase.
AJG – January, 2000 BGP in Carcinogenesis of Colorectal Cancer 259

Figure 2. Immunohistochemical analyses with anti-BGP antibody (A and D), anti-p53 Ab (B and E) and hematoxylin-eosin staining (C
and F) in adenoma (A–C) and carcinoma in adenoma (D–F). The adenoma consists of mild (center), moderate (right), and severe (left)
dysplasia components (A–C), and differential reactivity was observed in BGP (A) and p53 (B) staining. The positive BGP staining was
observed in adenoma with moderate, severe, and carcinoma in adenoma, whereas adenoma with mild dysplasia had no apparent staining.
Arrowhead: p53 reactive cells in adenoma with severe dysplasia. Bar: 100 ␮m.

␣-fetoprotein in hepatocellular carcinomas or carcinoem-
bryonic antigen in gastrointestinal cancer is. The present
study revealed that BGP was expressed in the colonic ade-
nomas with higher grades of dysplasia and cancers, and that
there was an excellent positive correlation between the BGP
expression and the ACS. Thus, our findings indicate that
BGP is a useful biomarker to detect the higher stages of
colonic carcinogenesis.
Matsuda et al. (38) have examined the ultrastructural
features of the adenomas in each step of the ACS, and have
260 Tashima et al.
Table 2. BGP and p53 Expression in Adenoma Carcinoma Sequence
Adenoma
Mild Moderate
Reactivity BGP p53 BGP
⫺ 12 13 7 18
⫹ 1 0 11
2⫹ 0 0 0 0
3⫹ 0 0 0 0
Positivity (%) 1/13 (7.7)* 0/13 (0) 11/18 (61.1)
*Significantly different from the value for moderate group at p ⬍ 0.05.
†
Significantly different from the value for mild group at p ⬍ 0.001.
‡
Significantly different from the value for moderate and mild group at p ⬍ 0.05.
BGP ⫽ brain-type glycogen phorphorylase.
demonstrated that the cellular characteristics of the adenoma
cells with mild and moderate atypia were similar to those in
normal epithelial cells and colonic carcinoma cells, respec-
tively. They concluded that the adenoma cells with moder-
ate atypia could be regarded as being at an important stage
in the process of becoming carcinoma cells. The present
evidence that BGP expressed abruptly at the stage of the
adenomas with moderate dysplasia corresponds well with
the ultrastructural observations. Therefore, this marker may
assist in the histological search for precancerous lesions
without any morphological change but with marked malig-
nant potential, in lesions such as those undergoing the de
novo carcinogenesis.
Concerning the carcinogenesis of sporadic colorectal car-
cinoma, attention has been focused on another pathway, i.e.,
de novo carcinogenesis (4 –7). A sizable number of colonic
carcinomas are assumed to arise directly from the colonic
mucosa, unlike ex-adenoma. If these assumptions are cor-
rect, it would be natural to consider that there are consid-
erable reserve troops undergoing genetic alterations toward
malignant transformation in the overtly normal-looking mu-
cosa bearing a carcinoma. In the present study, multifocal
expression of the BGP was observed mainly in the colorec-
Figure 3. Immunohistochemical staining of overt normal mucosa
adjacent to colonic carcinoma with anti-BGP antibody. Positive
reactivity was observed in the cytoplasm of cancer cells and the
neighboring normal colonic mucosa. Bar: 100 ␮m.
AJG – Vol. 95, No. 1, 2000
Microcarcinoma
Severe in Adenoma
p53 BGP p53 BGP p53
1 6 1 9
0 3 2 4 2
2 0 8 1
2 0 3 4
0/18 (0) 7/8 (87.5)† 2/8 (25.0) 15/16 (93.8)*† 7/16 (43.8)‡
tal mucosa adjacent to carcinoma (Fig. 4), which is, we
assume, de novo carcinoma (for the reasons described in the
Tissue Specimens section of Materials and Methods. As
shown in Figure 5, there were no apparent morphological
properties in the BGP-positive foci, compared with the
BGP-negative normal mucosa. Similar to the gastric ulcer,
as previously described (21), BGP was not expressed in the
regenerative atypical epithelia during wound healing (Fig.
4), and was barely expressed in adenoma with mild dyspla-
sia (Table 2). On the other hand, it was expressed in mor-
phologically normal mucosa adjacent to the de novo carci-
noma. As the timing of BGP expression was observed from
adenoma with moderate atypism to carcinoma in the present
ACS study, the cells in the BGP-positive foci might have
changes that are genetically equivalent to those of the cells
in the adenoma with moderate dysplasia. Therefore, these
findings suggest that the expression of BGP reflect the
abnormality that precedes morphological changes of neo-
plasia. Further, they suggest that BGP-positive foci can be
regarded as a promising candidate as a biomarker for ma-
lignant transformation.
The expression of particular molecules common in car-
cinoma may detect the malignant susceptibility of colonic
mucosa not only in the pathway of ACS but also in de novo
carcinoma. Shamsuddin et al. described the carbohydrate
moiety D-galactose-␤(1-3)-N-acetyl-D-galactosamine(Gal-
GalNAc) as a biomarker for colorectal carcinoma (39 – 41).
High Gal-GalNAc expression was observed in preneoplastic
lesions such as adenomatous polyps and chronic ulcerative
colitis, although they did not refer to a positive relationship
between the expression and the histological dysplasia. They
concluded that Gal-GalNAc is a biomarker that appears
during the early stages of progression of colonic carcino-
Table 3. BGP Expression in Adjacent Normal Mucosa to
Carcinoma
BGP Cases
Positive 60
Negative 6
Positivity (%) 60/66 (90.9)
BGP ⫽ brain-type glycogen phosphorylase.
AJG – January, 2000 BGP in Carcinogenesis of Colorectal Cancer 261

Figure 4. The distribution of BGP positive foci in the resected colon (case 1: colonic carcinoma; case 2: rectal carcinoma; case 3: ulcer).

genesis, and that its expression supports the field effect
theory of carcinogenesis and also explains the basis for mass
screening for colonic cancer and precancerous conditions
(41). BGP is the molecule commonly expressed in both
gastric (15, 19, 21) and colonic carcinomas and the preneo-
plastic lesions suggest this to be a broader maker of carci-
noma in the alimentary tract than Gal-GalNAc.
Aberrant crypt foci (ACF) may be another putative pre-
cursor in rat and human colonic adenoma and cancer (42–
45). ACF has been reported first by Bird (42) as lesions
consisting of large, thick crypts in methylene blue-stained
specimens of colon from mice treated with a carcinogen.
Patients with colon cancer had more ACF than those with
noncancerous lesions and the incidence of ACF was shown
to be modified by inhibitors (43– 45). Furthermore, in-
creased expression of c-fos and ras as well as K-ras muta-
tion in rats has been observed. Takayama et al. (45) sug-
gested that ACF, particularly those that are large and have
dysplastic features, are precursors of adenoma and cancer.
Our detailed histological analysis of BGP foci detected no
dysplastic change (Figs. 3, 5A) and our preliminary study
could not detect the K-ras mutation in the BGP-positive foci

Figure 5. Demonstration of a BGP positive focus in the normal appearing colonic mucosa from case 1 in Figure 4 (A) and hematoxylin-
eosin staining on the serial section (B). There are no apparent morphological properties in the BGP-positive focus compared with the
BGP-negative normal mucosa.
262 Tashima et al.
(unpublished data), suggesting that BGP foci are hardly
related to the ACF. Because it has been reported that de
novo cancers are involved in infrequent K-ras mutation in
carcinogenesis (5, 6), BGP foci may be related to the path-
way of de novo cancers.
Studies on preneoplastic lesions are essential for under-
standing the details of the molecular mechanism of colorec-
tal carcinogenesis. A major question about de novo carci-
nogenesis is why adenomatous change does not occur
despite the adenomatous polyposis coli (APC) gene muta-
tion (46). It has been reported that inactivation of APC
seems to be critical for the development of adenomas (47).
One possibility for explaining this phenomenon is that p53
mutation may occur before APC mutation, or that mutations
in both genes may have occurred within such a short period
that the tumors lacked an apparent adenoma stage. How-
ever, unless the precursor of de novo carcinoma is clarified,
this question may be never resolved. BGP-positive foci
might be such a candidate, potentially being early preneo-
plastic lesions with no apparent morphological change. Fur-
ther investigations are necessary to determine the relation-
ship between BGP-positive foci and the genetic alteration of
p53 and APC, which may be a major genetic change for
precancerous lesions of de novo carcinoma of the colon and
rectum.
ACKNOWLEDGMENT
This work was supported in part by Grant-in-Aid (No.
11671254) for Scientific Research from the Ministry of
Education, Japan.
Reprint requests and correspondence: Michio Ogawa, M.D.,
Department of Surgery II, Kumamoto University School of Med-
icine, 1–1-1 Honjo, Kumamoto 860 – 8556, Japan.
Received Mar. 15, 1999; accepted Sep. 7, 1999.
REFERENCES
1. Morson BC. The polyp-cancer sequence in the large bowel.
Proc R Soc Med 1974;67:451–7.
2. Fearon ER, Jones PA. Progressing toward a molecular descrip-
tion of colorectal cancer development. FASEB J 1992;6:2783–
90.
3. Vogelstein B, Fearon ER, Hamilton SR, et al. Genetic alter-
ations during colorectal-tumor development. N Eng J Med
1988;319:525–32.
4. Shimoda T, Ikegami M, Fujisaki J, et al. Early colorectal
carcinoma with special reference to its development de novo.
Cancer 1989;64:1138 – 46.
5. Minamoto T, Sawaguchi K, Mai M, et al. Infrequent K-ras
activation in superficial-type (flat) colorectal adenomas and
adenocarcinomas. Cancer Res 1994;54:2841– 4.
6. Yamagata S, Muto T, Uchida Y, et al. Lower incidence of
K-ras codon 12 mutation in flat colorectal adenomas than in
polypoid adenomas. Jpn J Cancer Res 1994;85:147–51.
7. Mueller J, Mueller E, Hoepner I, et al. Expression of bcl-2 and
p53 in de novo and ex-adenoma colon carcinoma: A compar-
ative immunohistochemical study. J Pathol 1996;180:259 – 65.
AJG – Vol. 95, No. 1, 2000
8. Cori CF, Cori GT. Mechanism and formation of hexosemono-
phosphate in muscle and isolation of a new phosphate ester.
Proc Soc Exp Biol Med 1936;34:702–12.
9. Davis CH, Schliselfeld LH, Wolf DP, et al. Interrelationships
among glycogen phosphorylase isozymes. J Biol Chem 1967;
242:4824 –33.
10. Newgard CB, Hwang PK, Fletterick RJ. The family of glyco-
gen phosphorylase: Structure and function. Crit Rev Biochem
Mol Biol 1989;24:69 –99.
11. Shimada S, Maeno M, Akagi M, et al. Immunohistochemical
detection of glycogen phosphorylase isoenzymes in rat and
human tissues. Histochem J 1986;18:334 – 8.
12. Nakano K, Hwang P, Fletterick RJ. Complete cDNA sequence
for rabbit muscle glycogen phosphorylase. FEBS Lett 1986;
204:283–7.
13. Newgard CB, Littman DR, Genderen CV, et al. Human brain
glycogen phosphorylase. J Biol Chem 1988;263:3850 –7.
14. Sato K, Morris HP, Weinhouse S. Phosphorylase: A new
isozyme in rat hepatic tumors and fetal liver. Science 1972;
178:879 – 81.
15. Shimada S, Maeno M, Misumi A, et al. Antigen reversion of
glycogen phosphorylase isoenzyme in carcinoma and prolif-
erative zone of intestinal metaplasia of the human stomach.
Gastroenterology 1987;93:35– 40.
16. Gelinas RP, Froman BE, McElroy F, et al. Human brain
glycogen phosphorylase: Characterization of fetal cDNA and
genomic sequence. Mol Brain Res 1989;6:177– 85.
17. Takashi M, Koshikawa T, Kurobe N, et al. Elevated concen-
trations of brain type glycogen phosphorylase in renal cell
carcinoma. Jpn J Cancer Res 1989;80:975– 80.
18. Hashimoto K, Tamura K, Otani H, et al. Histochemical and
immunohistochemical studies related to the role of glycogen in
human developing digestive organs. Anat Embryol 1995;192:
497–505.
19. Shimada S, Maeno M, Misumi A, et al. Histochemical study
of phosphorylase in proliferating cells of intestinal metaplasia
and carcinoma of the human stomach. Scand J Gastroenterol
1984;19:965–70.
20. Shimada S, Honmyo U, Yagi Y, et al. Expression of glycogen
phosphorylase activity in minute gastric carcinoma. Am J
Gastroenterol 1992;87:1230 –1.
21. Matsuzaki H, Shimada S, Uno K, et al. Novel subtyping of
intestinal metaplasia in the human stomach: Brain-type gly-
cogen phosphorylase expression in the proliferative zone and
its relationship with carcinogenesis. Am J Clin Pathol 1998;
109:181–9.
22. Brito MJ, Filipe MI, Morris RW. Cell proliferation study on
gastric carcinoma and non-involved gastric mucosa using a
bromodeoxyuridine (BrdU) labeling technique. Eur J Cancer
Prev 1992;1:429 –35.
23. Hollstein M, Sidransky D, Vogelstein B, et al. p53 mutations
in human cancers. Science 1991;253:49 –53.
24. Greenblatt MS, Bennett WP, Hollstein M, et al. Mutations in
the p53 tumor suppressor gene: Clues to cancer etiology and
molecular pathogenesis. Cancer Res 1994;54:4855–78.
25. Levine AJ, Perry ME, Chang A, et al. The role of the p53
suppressor gene in tumorigenesis. The 1993 Walter Hubert
Lecture. Br J Cancer 1994:69:409 –16.
26. Baker SJ, Preisinger AC, Jessup M. p53 gene mutations occur
in combination with 17p allelic deletions as late events in
colorectal tumorigenesis. Cancer Res 1990;50:7717–22.
27. Fearon ER, Vogelstein B. A genetic model for colorectal
tumorigenesis. Cell 1990;61:759 – 67.
28. Wada R, Matsukuma S, Abe H, et al. Histopathological studies
of superficial-type early colorectal carcinoma. Cancer 1996;
77:44 –50.
29. Ohue M, Tomita N, Monden T, et al. A frequent alteration of
AJG – January, 2000
p53 gene in carcinoma in adenoma of colon. Cancer Res
1994;54:4798 – 804.
30. Hasegawa T, Ueda M, Furukawa K, et al. p53 gene mutations
in early colorectal carcinoma, de novo vs. adenoma-carcinoma
sequence. Int J Cancer 1995;64:47–51.
31. Crawford BE, Stromeyer FW. Small nonpolypoid carcinomas
of the large intestine. Cancer 1983;51:1760 –3.
32. Herrera L, Hanna S, Castillio N, et al. Primary de novo
adenocarcinoma of the colon measuring 8 mm in diameter
with lymph node metastasis. Dis Colon Rectum 1991;34:
275–9.
33. Kuramoto S, Oohara T. Flat early cancers of the large intes-
tine. Cancer 1989;64:950 –5.
34. Ignacio PC, Baldwin BA, Vijayan VK, et al. Brain isozyme of
glycogen phosphorylase: Immunohistochemical localization
within the central nervous system. Brain Res 1990;529:42–9.
35. Shi SR, Key ME, Kalra KL. Antigen retrieval in formalin-
fixed, paraffin-embedded tissues. An enhancement method for
immunohistochemical staining based on microwave oven
heating of tissue sections. J Histochem Cytochem 1991;39:
741– 8.
36. UICC. TNM Atlas: An illustrated guide to the TNM/pTNM
classification of malignant tumors, 3rd ed. Heidelberg: Spring-
er-Velag, 1993.
37. Sato K, Satoh K, Sato T, et al. Isozyme patterns of glycogen
phosphorylase in rat tissues and transplantable hepatomas.
Cancer Res 1976;36:487–95.
38. Matsuda M, Misumi A, Shimada S, et al. Quantitative electron
microscopic study on grades of atypia in adenomas of the
human large bowel. Dis Col Rect 1989;32:57– 63.
BGP in Carcinogenesis of Colorectal Cancer 263
39. McKenzie KJ, Purnell DM, Shamsuddin AKM. Expression of
carcinoembryonic antigen, T-antigen and oncogene products
as markers of neoplastic and preneoplastic colonic mucosa.
Hum Pathol 1987;18:1282– 6.
40. Xu HM, Sakamoto K, Shamsuddin AK. Detection of the tumor
marker D-galactose-␤-(1–3)-N-acetyl-D-galactosamine in co-
lonic cancer and precancer. Arch Pathol Lab Med 1992;116:
1234 – 8.
41. Yang GY, Shamsuddin AM. Gal-GalNac: A biomarker of
colon carcinogenesis. Histol Histopathol 1996;11:80 – 6.
42. Bird RP. Observation and quantification of aberrant crypts in
the murine colon treated with a colon carcinogen. Cancer Lett
1987;37:147–51.
43. Pretlow TP, Barrow BJ, Ashton WS, et al. Aberrant crypts:
Putative preneoplastic foci in human colonic mucosa. Cancer
Res 1991;51:564 –7.
44. Shivapurkar N, Huang L, Ruggeri B, et al. K-ras and p53
mutations in aberrant crypt foci and colonic tumors from colon
cancer patients. Cancer Lett 1997;115:39 – 46.
45. Takayama T, Katsuki S, Takahashi Y, et al. Abberant crypt
foci of the colon as precursors of adenoma and cancer. N Engl
J Med 1998;339:1277– 84.
46. Aoki T, Takeda S, Yanagisawa A, et al. APC and p53 muta-
tions in de novo colorectal adenocarcinomas. Human Mutation
1994;3:342– 6.
47. Ichii S, Takeda S, Horii A, et al. Detailed analysis of genetic
alterations in colorectal tumors from patients with and without
familial adenomatous polyposis (FAP). Oncogene 1993;8:
3801–3.