- Email: [email protected]
A putative CEA moiety is shared by the cotton-top tamarin (Saguinus oedipus) and humans
CANCER LETTERS Cancer Letters 77 (19Y4) 7-13
A putative CEA moiety is shared by the cotton-top tamarin (Saguinus oedipus) and humans M. Tobi*“, “Departments hMarmoset
of Medicine Research
(Received
M. Memona,
and Parholog!.
Cenrer
(MA RCOR).
29 July 19Y3: revision
Wame
K. Kithier”,
State
University
received
Universiry School
of Tmmrssru
38 October
N. Clappb of Medicine.
Medical
Cmrcr.
1993: accepted
Detroit.
OaX Ridge,
29 October
MI. USA TN,
LISA
1993)
Abstract have been described in non-human primates. development of colorectal cancer. CEA expression has not been fully evaluated in a lower-order primate, the cotton-top tamarin (Scrgtrinus o&pus), an animal model for colitis and colorectal cancer. We found increased levels of CEA in both colonic washings and tissues of these animals using a commercially available kit, CEA AIA-PACK (Tosoh Medics, Foster City, CA). In contrast, we observed that other CEA kits failed to detect CEA in tamarins. To elucidate the nature of the CEA-like protein detected, we used the two component monoclonal antibodies used in the CEA AIA-PACK kit, and identified the reactive molecules by Western blotting. A band of approximately M, 50 000 was found to be common to samples from both humans and the tamarins. Minimal binding was observed with NCA antibody. We conclude that a CEA-like molecule shared by humans and tamarins may play a role in the pathogenesis of colitis and cancer in both species. CEA-like
molecules
immunologically
distinct
from those in humans
These primates do not share the human predilection for colitis and subsequent
Key ~cords; Colorectal
cancer;
Colitis;
NCA; Monoclonal
antibodies
1. Introduction Despite initial expectations for its use as a marker of neoplasia, carcinoembryonic antigen (CEA) has not proved to have widespread applicability in the effort to reduce the mortality of colorectal cancer [CRC] [I]. Recently, technological advances have made it possible to define CEA functionally as an adhesion molecule [2] and demonstrate * Corresponding 48201. USA.
author,
Division
of Gastroenterology.
0304-38351941.SO6.00 0 1994 Elsevier Scienttfic SSDI 0304-3835(93)03221-P
Department
Publishers
Ireland
an important interface with the inositol pathway of membrane signal transduction [3]. Breakdown products in colonic feces have also been demonstrated that may have important implications in cancer detection [4]. The cotton-top tamarin (CTT) is a spontaneously occurring model for colitis and CRC [5,6]. CEA expression in the CTT has not been described and would be an important antigen system to explore in this naturally occurrof Medicine.
Harper
Hospital
Ltd. All rights reserved.
2WS. 3990 John R. Detroit.
MI
8 ing model. Although CEA-like molecules have been described for other mammals [7,8.9], they are not thought to have identical human homologues. With the description of different groups of epitopes on the CEA molecules defined by monoclonal antibodies (MAbs) [lo], we used anti-CEA MAbs defining different CEA epitopes to determine whether CEA homologues exist in the CTT and if these may be shared by humans. 2. Materials and methods Tissues from CTTs were obtained by colonoscopic biopsy or at post-mortem from normalappearing mucosa. For comparison, non-cancerous tissues were obtained from a patient with long-standing inflammatory bowel disease (IBD) complicated by cancerous changes in the rectum and from other patients with IBD without dysplasia. Effluent samples (colonic washings) were obtained from CTTs and IBD patients without dysplasia and processed as described previously [l I]. Meconium was obtained from a neonate, solubilized in phosphate-buffered saline (Sigma, St. Louis, MO) and clarified by microfuging for 1 min. A membrane-enriched tissue extract (MEE) was prepared from the biopsies by homogenization in extraction buffer (0.01 M Tris-HCl, 0.2 mM CaCl,, pH 7.2) and clarified by centrifugation (1000 x g) and the supernatant sonicated and recentrifuged ( 10 000 x g). This supernatant constituted the MEE, and a Lowry protein estimation [ 121 established the protein content for MEE, effluent and meconium supernatant to serve as a standard in the Western blotting experiments. CEA determinations were performed using commercially available MAb kits (Roche, Abbott and Tosoh Medics) (CEA-AIA) in the CTT samples. In the human samples, the Pharmatope kit was used. The Western blot experiments were designed to identify the molecular moieties binding the antibodies in the CEA-AIA kit and to establish that no cross-reactivity with NCA existed. For the Western blotting experiments, the two MAbs used in the kit, the solid phase and tracer MAbs [13], were kindly supplied by Dr. H. Rittenhouse, Hybritech. San Diego. CA. Purified
hl.
7hhi (‘1 trl. / (‘tmcrr
Lrrr.
77 f IYY4)
7-1.3
CEA antigen and an anti-NCA MAb, T84.1 EC, which also binds CEA, were kind gifts of Dr. J. Shively and Dr. S. Hefta (City of Hope. Duarte, CA). These were used to determine the presence of colonic non-specific cross-reacting antigen (NCA). Western blotting was performed as described previously [ 1 l] in 12% SPS polyacrylamide gels under reducing conditions with 10 pg of protein loaded per lane. Pre-colored molecular weight markers (Amersham, UK) were used to determine the relative mobility (M,) of the samples. Specimens were collected subject to protocols approved by the Ethics Committees of the institutions involved, and informed consent was obtained from the patients. 3. Results Table 1 shows the quantitative levels of CEA determined in the CTTs and humans. No detectable CEA was evident using the Abbott and Roche kits. The CEA levels detected in IBD patient effluent and those of CTTs are comparable using the CEA-AIA PACK kit (432 f 407 ng/ml [n = lo] and 739 + 401 [n = 51, respectively) and similar to those previously reported for humans [ 111. These estimations were obtained in each species using different kits. The findings in selected samples using the Tosoh MedicsTM CEA-AIA PACK kit were confirmed by two independent laboratories (J. Loebel, Tosoh Medics Inc., personal communication). The CEA levels by the CEA-AIA PACK kit (932 f 690) were higher
Table I CEA in cotton-top
tamarins
and
IBD
CEA-AIA
Sample
patienta
PACK
Tamarin
extracts
73.3 f
Tamarin
effluent
739 * 401 (5)
IBD
patient
IBD,
inflammatory
column
937 f
effluent
effluent
AIA
kit. The difference of sample.
Ahhott
690
(IO)
<1
(8) (II)
432 * 407 (10)
in
levels compared
with the levels as determined
0.05 by the paired Student’? “Roche.
325 (6)
the Pharmatope
IBD patlent
ng CEAiml
kits
bowel disease. The last result in the second
represents
PACK
Other,’ (n)
(11)
tamarlns.
is statistically
in human
by the CEA-
ugnificant
(f’ <
t-test). All results arc expressed in
M. Tobi Ed al. /Cuncer
Lett 77 119941 7-13
9
Cl-l I
Cl-T
EXTRACTS
+
n
2
5
6
34
EFF
hi
HU
9
, ::
IO II
, MEC
I2
200-
Fig. I. Western blot using solid phase anti-CEA monoclonal antibody. lmmunoblotting was performed using 5.6pg protein/ml antibody after electrophoresis of a 12% SDS-polyacrylamide gel run under reducing conditions. Relative mobility (M,) is shown on the left, and the type of samples loaded at IOpg protein/lane are shown above the numbered lanes. Cotton-top tamarin extracts are in lanes 1 to 4 and effluent samples in 7 and 8. An extract from a patient with histologically proven rectal cancer and IBD is in lane 9, with human effluent samples in lanes IO and I I and human meconium in lane 12. The lane designated (+) contains the positive CEA control (lane 5), and the M, markers (M) are in lane 6. An arrow indicates the M, - 50000 band.
than those above as determined by the Pharmatope kit in 10 paired human IBD effluent samples, and the difference is significant (P < 0.05 by the paired Student’s t-test). Fig. 1 is a Western blot using the less specific solid-phase Hybritech anti-CEA MAb and shows an M, - 50 000 band common to extracts from CTT (n = 4) and a patient with IBD and CRC. Although no clear bands are seen in CTT effluent. proteolytic breakdown products (M, < 50 000) are evident. By contrast, minimal binding is seen in the lanes with human IBD effluent. An intense M, - 50 000 band is, however, seen in the lane containing human meconium. Fig. 2 is a blot using the
highly specific anti-CEA tracer MAb and shows binding in essentially the same set of samples (1 CTT and 1 human effluent sample excepted), as in Fig. 1. A similar M, - 50 000 band is seen in the CTT effluent and extracts as well as in the extract of the patient with IBD and CRC. As in Fig. 1, no bands are seen in human effluent samples from IBD patients without dysplasia. In contrast to Fig. I, human meconium shows no reactivity with this more specific MAb. Species cross-reactive with anti-NCA MAb are shown in Fig. 3. An A4, - 90 000 NCA moiety is seen in human meconium, which corresponds to previously reported M, ranges for NCA [ 141. No
IO
CTT EFF HU
M
CTT EXT.
+
1234567891011
“U
EFF
MEC
zoo97.4-
.
46-
30-
21.5-
I4.3-
Fig. 2. Western blot using tracer antt-CEA
monoclonal
samples that were not included (one tamarin
antibody. The samples are the same as seen in Fig. I, except for two effluent
and one human).
The M,. ts indicated
on the left. and the type of sample above the
numbered lanes. The CTT effluent sample is seen in lane I. and human extract of normal colonic mucosa from an IBD patient with rectal cancer is in lane 2. M, markers (M) are in lane 3. and CEA and 9-10
human effluent samples, Lane CEA
1I
(+) (obscured by artifact)
is in lane 4. Lanes 5-X contain tamarin
contains human meconium. The arrow depicts the locdtion of the common ,%I, -
moiety. All lanes arc loaded wtth samples standardized
dominant bands equal in intensity to bands seen in Figs, 1 and 2 are evident. In most samples, multiple bands of both high and lower molecular weight species are visible. CEA is also recognized by this MAb which binds both CEA and NCA. 4. Discussion A CEA-like molecule of M, - 50 000 can be demonstrated in CTTs and may also be shared by patients with IBD complicated by CRC. To our knowledge, this is the first description of a putative common CEA antigen. Studies with specific
50 000
at IOpg protetn/lanc.
molecular probes at the mRNA level will have to be conducted to elucidate whether this moiety is distinct. an underglycosylated form of CEA or part of a larger protein species that has undergone proteolysis. The antigen appears to be distinct from NCA; however, further study using a specific anti-50/90 NCA antibody (Dr. W.J. Allard, Miles Diagnostics. personal communication) may help to elucidate this question. A similar-sized CEA moiety has been described in human feces [4]. but its importance is yet to be established. The ease of testing by the Tosoh MedicsTM-CEA-AIA PACK kit makes for a convenient method of assessing the
M. Tobi et al. /Cuncer
Lett. 77 (1994)
MEC I
11
7-13
, CTTEXT.
2
3
4
,
+
M
z
5
6
7
8
HU
9
1011
&
I2
21.5-
l4.3-
Fig. 3. Western blot using anti-NCA monoclonal antibody. An immunoblot containing the same samples as described in Fig. 1 are reacted with T84.1 EC anti-NCA monoclonal antibody and also reacts with CEA. The M, values based on the pre-colored molecular weight markers (lane 7) are shown on the left. The samples are indicated above the numbered lanes. Human meconium is in lane 1. tamarin tissue extracts in lanes 2-5 and CEA in lane 6. Tamarin effluent samples are in lanes 8-9, with a human extract of normalappearing colonic mucosa from a patient with rectal CRC and IBD in lane 10. Human effluent samples are in lanes 1 I and 12. The band at M, 68 000 is common to all lanes (including the marker lane) and is thought to be albumin, a contaminating artifact.
clinical significance of this CEA molecule in CTTs and humans. In the last two decades, attention has been directed to finding animal models for human CRC to evaluate the role of immunologically defined markers such as CEA. The lower-order primates (macaques, rhesus) were shown to express a CEAlike molecule in pulmonary and splenic tissues, but the molecule was shown to be similar to NCA [8]. Efforts were directed at examining blood levels of higher-order primates, such as chimpanzees and gorillas [9]. Higher levels were detected in these
and other primates than seen in humans using a Hoffman-La Roche radioimmunoassay. The CEA molecules, although of similar size to human ( - 200 000 Da), were not shown to be immunologically identical to human CEA. The mean levels detected were in the order of 25 and 32 ng/ml for chimpanzees and gorillas, respectively. Those are far lower than the levels in CTTs as determined using the Tosoh Medics kit. The CEA-like protein from mouse colon is approximately 120 000 Da in size, showing highly conserved N-terminal homology with the human but felt to be part of a sepa-
I2
rate murine CEA system [7]. The higher levels in the same effluent samples of IBD patients using the CEA-AIA PACK kit suggest that additional CEA-like substances exist in such samples which might remain undetected by other anti-CEA antibodies in routine use. The CEA moiety we describe does not seem to be recognized by anti-NCA antibody, despite a molecular weight similar to that of the lower molecular weight NCA [14]. Further study of the amino acid sequences or cDNA would be necessary to completely resolve this issue. The epitope recognized by the solid phase MAb belongs to the Gold group IV and the tracer to the more specific Gold group I [ 131. This may explain the observation that no specific antigens are detected in meconium using the tracer because the meconium moiety appears to lack the epitope recognized by the more specific MAb. The antiNCA MAb also does not appear to recognize the Mr - 50 000 species in meconium (Fig. 3). One of the antibodies in the Roche kit reacts with an epitope near the C-terminal domain of the CEA molecule, and CEA reactivity is absent when CTT samples are assayed with the kit. This suggests that the CEA-AIA PACK kit antibodies react with epitopes in the immunoglobin-like loop domains on a CEA molecule which excludes the A3B3 domain. This. however, is speculative, because of the observation that different Gold class antibodies may react with epitopes throughout the CEA molecule [15] and because the structure of the 50 kDa species is yet to be elucidated. The M, 50000putative CEA molecule in CTTs, however, does seem to possess both epitopes but apparently does not react intensely with an anti-NCA MAb. This suggests that it might be a specific marker for CRC, which is supported by its presence in the normal-appearing mucosa of a patient with longstanding IBD and CRC and its absence in the effluent samples of IBD patients without dysplasia. Other antigens have been described as being shared by humans and CTTs. These include the Leb [16] and those recognized by MAbs FBB 3161, CaCo 3161 and Span-l 161. The question of CEA and CEA-like antigens acting as autoantigens in the CTT should be addressed. Other similarities and differences between the CTT and humans have recently been summarized [6]. One
of the important differences between CRC in CTT and humans yet to be explained is the paucity of liver metastasis in the CTT [ 171. Since hepatic uptake of CEA may be important in the development of liver metastasis, the smaller CTT CEA molecule might lack the critical peptides necessary for uptake (P. Thomas, personal communication) and hence explain the paucity of liver involvement in CTTs with CRC. 5. Acknowledgements The authors wish to express their gratitude to Drs. J. Bodzin, G. Herman, T. Heller, V. Kaila, S. Chintalapani, S. Prabhu. N. Hassan and R. Ferguson for their able assistance. We also thank M. Henke. E. Darmon, L. Nakayama and P. Spence for technical assistance and D. Farrah for expert typographical assistance, and J. Shively, S. Hefta, P. Thomas, B. Wolfert and H. Rittenhouse for comments and suggestions. This article is dedicated to the memory of Joseph M. Katz. 6. References Zamcheck, N., Liu, P., Thomas, P. and Steele. G. (1988) Search for useful biomarkers of early malignant tumors. In: Basic and Clinical Perspectives of Colorectal Polyps and Cancer (Progress in Clinical and Biological Research. vol. 279). pp. 251-275. Editors: G. Steele, R.W. Burt. S.J. Winawer and J.P. Karr. Liss, New York. Benchimol, S., Fuks. A.. Jothy. S., Beauchemin. N.. Shirota, K. and Stanners. C.P. ( 1989) Carcinoembryonic antigen, a human marker. functions as an intercellular adhesion molecule. Cell, 57. 327-34. Hefta. S.. Hefta. L.J.F.. Lee. T.P.. Paxton. R.J. and Shively. J.E. (1988) Carcinoembryonic antigen is anchored to membranes by covalent attachment to a glycosylphophatidylinositol moiety: identification of the ethanolamine linkage site. Proc. Natl. Acad. Sci. USA. 85. 4648-52. Henslee, J.G.. Vachula, E.J., Paxton. R.J., Matias, MS.. Shively, J.E., Rittenhouse. H.G. and Tomita. J.T. (1993) Indentitication and characterization of new antigenic fragments related to carcinoembryonic antigen in adult feces. Cancer Res.. 52. 4175-82. Kim, H.-S. and Berstad. A. (1992) Experimental colitis in animal models. Scan. J. Gastroenterol.. 27, 529-37. Tobi, M.. Kaila. V.. Chintalapani, S., Kithier. K.. Henke. M. and Clapp. N. (1993) An antigenic profile in cottontop tamarins (Saguinus oedipus):a model for human inflammatory bowel disease and colorectal cancer. In: The Cotton-Top Tamarin: a Primate Model for the Study of
M. Tohi et al. / Cuncer Lerr. 77
7
8
9
10
I1
( 1994) 7-
I3
13
Colon Carcinogenesis. Editor: N. Clapp. CRC Press. Boca Raton (in press). Beauchemin. N.. Turbide. C.. Afar. D.. Bell. J.. Raymond, M.. Stanners. C.P. and Fuks. A. (1989) A mouse analogue of human carcinoembryonic antigen. Cancer Res.. 49. 2017-21. Engvall, E.. Vuento. M. and Ruoslahti, E. (1976) A monkey antigen cross reactant with carcinoembryonic antigen, CEA. Br. J. Cancer, 34. 341-45. Haagensen, D.E. Jr., Metzgar. R.S., Swenson. B.. Dilley. W.G.. Cox, C.E.. Davis. S.. Murdoch, J.. Zamcheck. N. and Wells, S.A. (1982) Carcinoembryonic antigen in nonhuman primates. J. Nat]. Cancer Inst.. 69. 1073-6. Hammarstrom, S.. Shively, J.E., Paxton, R.J. et al. (1989) Antigenic sites in carcinoembryonic antigen. Cancer Res.. 49, 4852-S. Tobi, M., O’Kieffe, D.. Trujillo. N.. Nochomovitz. L.E. and Steinberg, W.M. (1992) Detection of carcinoembryonic antigen in colonic effluent by specific anti-CEA monoclonal antibodies. Cancer Lett., 67, 47-54.
I2
I3
I4
I5
16
I7
Lowry. O.H.. Rosenbrough. N.J.. Farr. A.L. and Randall. R.J. (1951) Protein measurements with folin phenol reagent. J. Biol. Chem., 193. 265-275. Bormer. O.P.. Thrane-Steen, K. (1991) Epitope group specificity of six immunoassays for carcinoembryonic antigen. Tumor Biol.. 12, 9-15. Miyazaki. T.. Hachiya. T.. Kajita. Y.. Ochi. I’.. Fujiyama. Y.. Hosoda S. (1982) Immunologtcal similarity of NCA (non-specific cross-reacting antigen) in feces with @,-acid glycoprotein. Clin. Chim. Acta, 122, 161-8. Kuroki. M.A., et al.. A reference of the Gold classification of the anti-CEA MAbs to the domain structure of the CEA molecule, Fourth International CEAiPSG Workshop, Boston. 1993. Steplewski. Z. (1985) Monoclonal antibody-defined antigen detected in colonic tissues of cotton-top tamarins. .Squ~nus oedipus ordipus. Dig. Dis. Sci., 30, 14lS- 146s. Lushbaugh. C.. Humason. G. and Clapp, N. (1985) Histology of colon cancer in Squinus ovdipus oediptrc. Dig. Dis. Sci.. 30. I19S-2%.
A putative CEA moiety is shared by the cotton-top tamarin (Saguinus oedipus) and humans M. Tobi*“, “Departments hMarmoset
of Medicine Research
(Received
M. Memona,
and Parholog!.
Cenrer
(MA RCOR).
29 July 19Y3: revision
Wame
K. Kithier”,
State
University
received
Universiry School
of Tmmrssru
38 October
N. Clappb of Medicine.
Medical
Cmrcr.
1993: accepted
Detroit.
OaX Ridge,
29 October
MI. USA TN,
LISA
1993)
Abstract have been described in non-human primates. development of colorectal cancer. CEA expression has not been fully evaluated in a lower-order primate, the cotton-top tamarin (Scrgtrinus o&pus), an animal model for colitis and colorectal cancer. We found increased levels of CEA in both colonic washings and tissues of these animals using a commercially available kit, CEA AIA-PACK (Tosoh Medics, Foster City, CA). In contrast, we observed that other CEA kits failed to detect CEA in tamarins. To elucidate the nature of the CEA-like protein detected, we used the two component monoclonal antibodies used in the CEA AIA-PACK kit, and identified the reactive molecules by Western blotting. A band of approximately M, 50 000 was found to be common to samples from both humans and the tamarins. Minimal binding was observed with NCA antibody. We conclude that a CEA-like molecule shared by humans and tamarins may play a role in the pathogenesis of colitis and cancer in both species. CEA-like
molecules
immunologically
distinct
from those in humans
These primates do not share the human predilection for colitis and subsequent
Key ~cords; Colorectal
cancer;
Colitis;
NCA; Monoclonal
antibodies
1. Introduction Despite initial expectations for its use as a marker of neoplasia, carcinoembryonic antigen (CEA) has not proved to have widespread applicability in the effort to reduce the mortality of colorectal cancer [CRC] [I]. Recently, technological advances have made it possible to define CEA functionally as an adhesion molecule [2] and demonstrate * Corresponding 48201. USA.
author,
Division
of Gastroenterology.
0304-38351941.SO6.00 0 1994 Elsevier Scienttfic SSDI 0304-3835(93)03221-P
Department
Publishers
Ireland
an important interface with the inositol pathway of membrane signal transduction [3]. Breakdown products in colonic feces have also been demonstrated that may have important implications in cancer detection [4]. The cotton-top tamarin (CTT) is a spontaneously occurring model for colitis and CRC [5,6]. CEA expression in the CTT has not been described and would be an important antigen system to explore in this naturally occurrof Medicine.
Harper
Hospital
Ltd. All rights reserved.
2WS. 3990 John R. Detroit.
MI
8 ing model. Although CEA-like molecules have been described for other mammals [7,8.9], they are not thought to have identical human homologues. With the description of different groups of epitopes on the CEA molecules defined by monoclonal antibodies (MAbs) [lo], we used anti-CEA MAbs defining different CEA epitopes to determine whether CEA homologues exist in the CTT and if these may be shared by humans. 2. Materials and methods Tissues from CTTs were obtained by colonoscopic biopsy or at post-mortem from normalappearing mucosa. For comparison, non-cancerous tissues were obtained from a patient with long-standing inflammatory bowel disease (IBD) complicated by cancerous changes in the rectum and from other patients with IBD without dysplasia. Effluent samples (colonic washings) were obtained from CTTs and IBD patients without dysplasia and processed as described previously [l I]. Meconium was obtained from a neonate, solubilized in phosphate-buffered saline (Sigma, St. Louis, MO) and clarified by microfuging for 1 min. A membrane-enriched tissue extract (MEE) was prepared from the biopsies by homogenization in extraction buffer (0.01 M Tris-HCl, 0.2 mM CaCl,, pH 7.2) and clarified by centrifugation (1000 x g) and the supernatant sonicated and recentrifuged ( 10 000 x g). This supernatant constituted the MEE, and a Lowry protein estimation [ 121 established the protein content for MEE, effluent and meconium supernatant to serve as a standard in the Western blotting experiments. CEA determinations were performed using commercially available MAb kits (Roche, Abbott and Tosoh Medics) (CEA-AIA) in the CTT samples. In the human samples, the Pharmatope kit was used. The Western blot experiments were designed to identify the molecular moieties binding the antibodies in the CEA-AIA kit and to establish that no cross-reactivity with NCA existed. For the Western blotting experiments, the two MAbs used in the kit, the solid phase and tracer MAbs [13], were kindly supplied by Dr. H. Rittenhouse, Hybritech. San Diego. CA. Purified
hl.
7hhi (‘1 trl. / (‘tmcrr
Lrrr.
77 f IYY4)
7-1.3
CEA antigen and an anti-NCA MAb, T84.1 EC, which also binds CEA, were kind gifts of Dr. J. Shively and Dr. S. Hefta (City of Hope. Duarte, CA). These were used to determine the presence of colonic non-specific cross-reacting antigen (NCA). Western blotting was performed as described previously [ 1 l] in 12% SPS polyacrylamide gels under reducing conditions with 10 pg of protein loaded per lane. Pre-colored molecular weight markers (Amersham, UK) were used to determine the relative mobility (M,) of the samples. Specimens were collected subject to protocols approved by the Ethics Committees of the institutions involved, and informed consent was obtained from the patients. 3. Results Table 1 shows the quantitative levels of CEA determined in the CTTs and humans. No detectable CEA was evident using the Abbott and Roche kits. The CEA levels detected in IBD patient effluent and those of CTTs are comparable using the CEA-AIA PACK kit (432 f 407 ng/ml [n = lo] and 739 + 401 [n = 51, respectively) and similar to those previously reported for humans [ 111. These estimations were obtained in each species using different kits. The findings in selected samples using the Tosoh MedicsTM CEA-AIA PACK kit were confirmed by two independent laboratories (J. Loebel, Tosoh Medics Inc., personal communication). The CEA levels by the CEA-AIA PACK kit (932 f 690) were higher
Table I CEA in cotton-top
tamarins
and
IBD
CEA-AIA
Sample
patienta
PACK
Tamarin
extracts
73.3 f
Tamarin
effluent
739 * 401 (5)
IBD
patient
IBD,
inflammatory
column
937 f
effluent
effluent
AIA
kit. The difference of sample.
Ahhott
690
(IO)
<1
(8) (II)
432 * 407 (10)
in
levels compared
with the levels as determined
0.05 by the paired Student’? “Roche.
325 (6)
the Pharmatope
IBD patlent
ng CEAiml
kits
bowel disease. The last result in the second
represents
PACK
Other,’ (n)
(11)
tamarlns.
is statistically
in human
by the CEA-
ugnificant
(f’ <
t-test). All results arc expressed in
M. Tobi Ed al. /Cuncer
Lett 77 119941 7-13
9
Cl-l I
Cl-T
EXTRACTS
+
n
2
5
6
34
EFF
hi
HU
9
, ::
IO II
, MEC
I2
200-
Fig. I. Western blot using solid phase anti-CEA monoclonal antibody. lmmunoblotting was performed using 5.6pg protein/ml antibody after electrophoresis of a 12% SDS-polyacrylamide gel run under reducing conditions. Relative mobility (M,) is shown on the left, and the type of samples loaded at IOpg protein/lane are shown above the numbered lanes. Cotton-top tamarin extracts are in lanes 1 to 4 and effluent samples in 7 and 8. An extract from a patient with histologically proven rectal cancer and IBD is in lane 9, with human effluent samples in lanes IO and I I and human meconium in lane 12. The lane designated (+) contains the positive CEA control (lane 5), and the M, markers (M) are in lane 6. An arrow indicates the M, - 50000 band.
than those above as determined by the Pharmatope kit in 10 paired human IBD effluent samples, and the difference is significant (P < 0.05 by the paired Student’s t-test). Fig. 1 is a Western blot using the less specific solid-phase Hybritech anti-CEA MAb and shows an M, - 50 000 band common to extracts from CTT (n = 4) and a patient with IBD and CRC. Although no clear bands are seen in CTT effluent. proteolytic breakdown products (M, < 50 000) are evident. By contrast, minimal binding is seen in the lanes with human IBD effluent. An intense M, - 50 000 band is, however, seen in the lane containing human meconium. Fig. 2 is a blot using the
highly specific anti-CEA tracer MAb and shows binding in essentially the same set of samples (1 CTT and 1 human effluent sample excepted), as in Fig. 1. A similar M, - 50 000 band is seen in the CTT effluent and extracts as well as in the extract of the patient with IBD and CRC. As in Fig. 1, no bands are seen in human effluent samples from IBD patients without dysplasia. In contrast to Fig. I, human meconium shows no reactivity with this more specific MAb. Species cross-reactive with anti-NCA MAb are shown in Fig. 3. An A4, - 90 000 NCA moiety is seen in human meconium, which corresponds to previously reported M, ranges for NCA [ 141. No
IO
CTT EFF HU
M
CTT EXT.
+
1234567891011
“U
EFF
MEC
zoo97.4-
.
46-
30-
21.5-
I4.3-
Fig. 2. Western blot using tracer antt-CEA
monoclonal
samples that were not included (one tamarin
antibody. The samples are the same as seen in Fig. I, except for two effluent
and one human).
The M,. ts indicated
on the left. and the type of sample above the
numbered lanes. The CTT effluent sample is seen in lane I. and human extract of normal colonic mucosa from an IBD patient with rectal cancer is in lane 2. M, markers (M) are in lane 3. and CEA and 9-10
human effluent samples, Lane CEA
1I
(+) (obscured by artifact)
is in lane 4. Lanes 5-X contain tamarin
contains human meconium. The arrow depicts the locdtion of the common ,%I, -
moiety. All lanes arc loaded wtth samples standardized
dominant bands equal in intensity to bands seen in Figs, 1 and 2 are evident. In most samples, multiple bands of both high and lower molecular weight species are visible. CEA is also recognized by this MAb which binds both CEA and NCA. 4. Discussion A CEA-like molecule of M, - 50 000 can be demonstrated in CTTs and may also be shared by patients with IBD complicated by CRC. To our knowledge, this is the first description of a putative common CEA antigen. Studies with specific
50 000
at IOpg protetn/lanc.
molecular probes at the mRNA level will have to be conducted to elucidate whether this moiety is distinct. an underglycosylated form of CEA or part of a larger protein species that has undergone proteolysis. The antigen appears to be distinct from NCA; however, further study using a specific anti-50/90 NCA antibody (Dr. W.J. Allard, Miles Diagnostics. personal communication) may help to elucidate this question. A similar-sized CEA moiety has been described in human feces [4]. but its importance is yet to be established. The ease of testing by the Tosoh MedicsTM-CEA-AIA PACK kit makes for a convenient method of assessing the
M. Tobi et al. /Cuncer
Lett. 77 (1994)
MEC I
11
7-13
, CTTEXT.
2
3
4
,
+
M
z
5
6
7
8
HU
9
1011
&
I2
21.5-
l4.3-
Fig. 3. Western blot using anti-NCA monoclonal antibody. An immunoblot containing the same samples as described in Fig. 1 are reacted with T84.1 EC anti-NCA monoclonal antibody and also reacts with CEA. The M, values based on the pre-colored molecular weight markers (lane 7) are shown on the left. The samples are indicated above the numbered lanes. Human meconium is in lane 1. tamarin tissue extracts in lanes 2-5 and CEA in lane 6. Tamarin effluent samples are in lanes 8-9, with a human extract of normalappearing colonic mucosa from a patient with rectal CRC and IBD in lane 10. Human effluent samples are in lanes 1 I and 12. The band at M, 68 000 is common to all lanes (including the marker lane) and is thought to be albumin, a contaminating artifact.
clinical significance of this CEA molecule in CTTs and humans. In the last two decades, attention has been directed to finding animal models for human CRC to evaluate the role of immunologically defined markers such as CEA. The lower-order primates (macaques, rhesus) were shown to express a CEAlike molecule in pulmonary and splenic tissues, but the molecule was shown to be similar to NCA [8]. Efforts were directed at examining blood levels of higher-order primates, such as chimpanzees and gorillas [9]. Higher levels were detected in these
and other primates than seen in humans using a Hoffman-La Roche radioimmunoassay. The CEA molecules, although of similar size to human ( - 200 000 Da), were not shown to be immunologically identical to human CEA. The mean levels detected were in the order of 25 and 32 ng/ml for chimpanzees and gorillas, respectively. Those are far lower than the levels in CTTs as determined using the Tosoh Medics kit. The CEA-like protein from mouse colon is approximately 120 000 Da in size, showing highly conserved N-terminal homology with the human but felt to be part of a sepa-
I2
rate murine CEA system [7]. The higher levels in the same effluent samples of IBD patients using the CEA-AIA PACK kit suggest that additional CEA-like substances exist in such samples which might remain undetected by other anti-CEA antibodies in routine use. The CEA moiety we describe does not seem to be recognized by anti-NCA antibody, despite a molecular weight similar to that of the lower molecular weight NCA [14]. Further study of the amino acid sequences or cDNA would be necessary to completely resolve this issue. The epitope recognized by the solid phase MAb belongs to the Gold group IV and the tracer to the more specific Gold group I [ 131. This may explain the observation that no specific antigens are detected in meconium using the tracer because the meconium moiety appears to lack the epitope recognized by the more specific MAb. The antiNCA MAb also does not appear to recognize the Mr - 50 000 species in meconium (Fig. 3). One of the antibodies in the Roche kit reacts with an epitope near the C-terminal domain of the CEA molecule, and CEA reactivity is absent when CTT samples are assayed with the kit. This suggests that the CEA-AIA PACK kit antibodies react with epitopes in the immunoglobin-like loop domains on a CEA molecule which excludes the A3B3 domain. This. however, is speculative, because of the observation that different Gold class antibodies may react with epitopes throughout the CEA molecule [15] and because the structure of the 50 kDa species is yet to be elucidated. The M, 50000putative CEA molecule in CTTs, however, does seem to possess both epitopes but apparently does not react intensely with an anti-NCA MAb. This suggests that it might be a specific marker for CRC, which is supported by its presence in the normal-appearing mucosa of a patient with longstanding IBD and CRC and its absence in the effluent samples of IBD patients without dysplasia. Other antigens have been described as being shared by humans and CTTs. These include the Leb [16] and those recognized by MAbs FBB 3161, CaCo 3161 and Span-l 161. The question of CEA and CEA-like antigens acting as autoantigens in the CTT should be addressed. Other similarities and differences between the CTT and humans have recently been summarized [6]. One
of the important differences between CRC in CTT and humans yet to be explained is the paucity of liver metastasis in the CTT [ 171. Since hepatic uptake of CEA may be important in the development of liver metastasis, the smaller CTT CEA molecule might lack the critical peptides necessary for uptake (P. Thomas, personal communication) and hence explain the paucity of liver involvement in CTTs with CRC. 5. Acknowledgements The authors wish to express their gratitude to Drs. J. Bodzin, G. Herman, T. Heller, V. Kaila, S. Chintalapani, S. Prabhu. N. Hassan and R. Ferguson for their able assistance. We also thank M. Henke. E. Darmon, L. Nakayama and P. Spence for technical assistance and D. Farrah for expert typographical assistance, and J. Shively, S. Hefta, P. Thomas, B. Wolfert and H. Rittenhouse for comments and suggestions. This article is dedicated to the memory of Joseph M. Katz. 6. References Zamcheck, N., Liu, P., Thomas, P. and Steele. G. (1988) Search for useful biomarkers of early malignant tumors. In: Basic and Clinical Perspectives of Colorectal Polyps and Cancer (Progress in Clinical and Biological Research. vol. 279). pp. 251-275. Editors: G. Steele, R.W. Burt. S.J. Winawer and J.P. Karr. Liss, New York. Benchimol, S., Fuks. A.. Jothy. S., Beauchemin. N.. Shirota, K. and Stanners. C.P. ( 1989) Carcinoembryonic antigen, a human marker. functions as an intercellular adhesion molecule. Cell, 57. 327-34. Hefta. S.. Hefta. L.J.F.. Lee. T.P.. Paxton. R.J. and Shively. J.E. (1988) Carcinoembryonic antigen is anchored to membranes by covalent attachment to a glycosylphophatidylinositol moiety: identification of the ethanolamine linkage site. Proc. Natl. Acad. Sci. USA. 85. 4648-52. Henslee, J.G.. Vachula, E.J., Paxton. R.J., Matias, MS.. Shively, J.E., Rittenhouse. H.G. and Tomita. J.T. (1993) Indentitication and characterization of new antigenic fragments related to carcinoembryonic antigen in adult feces. Cancer Res.. 52. 4175-82. Kim, H.-S. and Berstad. A. (1992) Experimental colitis in animal models. Scan. J. Gastroenterol.. 27, 529-37. Tobi, M.. Kaila. V.. Chintalapani, S., Kithier. K.. Henke. M. and Clapp. N. (1993) An antigenic profile in cottontop tamarins (Saguinus oedipus):a model for human inflammatory bowel disease and colorectal cancer. In: The Cotton-Top Tamarin: a Primate Model for the Study of
M. Tohi et al. / Cuncer Lerr. 77
7
8
9
10
I1
( 1994) 7-
I3
13
Colon Carcinogenesis. Editor: N. Clapp. CRC Press. Boca Raton (in press). Beauchemin. N.. Turbide. C.. Afar. D.. Bell. J.. Raymond, M.. Stanners. C.P. and Fuks. A. (1989) A mouse analogue of human carcinoembryonic antigen. Cancer Res.. 49. 2017-21. Engvall, E.. Vuento. M. and Ruoslahti, E. (1976) A monkey antigen cross reactant with carcinoembryonic antigen, CEA. Br. J. Cancer, 34. 341-45. Haagensen, D.E. Jr., Metzgar. R.S., Swenson. B.. Dilley. W.G.. Cox, C.E.. Davis. S.. Murdoch, J.. Zamcheck. N. and Wells, S.A. (1982) Carcinoembryonic antigen in nonhuman primates. J. Nat]. Cancer Inst.. 69. 1073-6. Hammarstrom, S.. Shively, J.E., Paxton, R.J. et al. (1989) Antigenic sites in carcinoembryonic antigen. Cancer Res.. 49, 4852-S. Tobi, M., O’Kieffe, D.. Trujillo. N.. Nochomovitz. L.E. and Steinberg, W.M. (1992) Detection of carcinoembryonic antigen in colonic effluent by specific anti-CEA monoclonal antibodies. Cancer Lett., 67, 47-54.
I2
I3
I4
I5
16
I7
Lowry. O.H.. Rosenbrough. N.J.. Farr. A.L. and Randall. R.J. (1951) Protein measurements with folin phenol reagent. J. Biol. Chem., 193. 265-275. Bormer. O.P.. Thrane-Steen, K. (1991) Epitope group specificity of six immunoassays for carcinoembryonic antigen. Tumor Biol.. 12, 9-15. Miyazaki. T.. Hachiya. T.. Kajita. Y.. Ochi. I’.. Fujiyama. Y.. Hosoda S. (1982) Immunologtcal similarity of NCA (non-specific cross-reacting antigen) in feces with @,-acid glycoprotein. Clin. Chim. Acta, 122, 161-8. Kuroki. M.A., et al.. A reference of the Gold classification of the anti-CEA MAbs to the domain structure of the CEA molecule, Fourth International CEAiPSG Workshop, Boston. 1993. Steplewski. Z. (1985) Monoclonal antibody-defined antigen detected in colonic tissues of cotton-top tamarins. .Squ~nus oedipus ordipus. Dig. Dis. Sci., 30, 14lS- 146s. Lushbaugh. C.. Humason. G. and Clapp, N. (1985) Histology of colon cancer in Squinus ovdipus oediptrc. Dig. Dis. Sci.. 30. I19S-2%.