Proc. Natl. Acad. Sci. USAVol. 90, pp. 3466-3470, April 1993Medical Sciences

Decreased expression of Mac-2 (carbohydrate binding protein 35)and loss of its nuclear localization are associated with theneoplastic progression of colon carcinomaMARGARET M. LOTZ, CHARLES W. ANDREWS, JR., CYNTHIA A. KORZELIUS, EDWARD C. LEE,GLENN D. STEELE, JR., ASTRID CLARKE, AND ARTHUR M. MERCURIO*Laboratory of Cancer Biology, Deaconess Hospital, Harvard Medical School, Boston, MA 02115

Communicated by Phillips W. Robbins, January 7, 1993

ABSTRACT The Mac-2 lectin (carbohydrate binding pro-tein 35) is a soluble, 32- to 35-kDa phosphoprotein that bindsgalactose-containing glycoconjugates. We report here that thecolonic epithelium is a major site of Mac-2 expression in vivobased on immunohistochemistry ofhuman tissue specimens. Inthis epithelium, proliferating cells at the base of the crypts donot express Mac-2 but its expression increases with differen-tiation along the crypt-to-surface axis. Mac-2 expression isconcentrated in the nuclei of these differentiated epithelial cells.The progression from normal mucosa to adenoma to carcinomais associated with significant changes in Mac-2 nuclear local-ization and expression. In all adenomas (9/9) and carcinomas(13/13) examined, Mac-2 was not present in the nucleus butwas localized in the cytoplasm. Sequencing of Mac-2 cDNAsfrom normal mucosa and carcinoma revealed no specificmutations that could account for this loss of nuclear localiza-tion. We also observed a 5- to 10-fold decrease in Mac-2 mRNAlevels in cancer compared to normal mucosa as well as asignificant reduction in the amount ofMac-2 protein expressed.These observations suggest that Mac-2 exclusion from thenucleus and its decreased expression may be related to theneoplastic progression of colon cancer.

The Mac-2 lectin is a soluble, 32- to 35-kDa phosphoproteinthat binds galactose-containing glycoconjugates (1-5). Mo-lecular cloning studies revealed that Mac-2 is identical tocarbohydrate binding protein 35 (CBP 35), a lectin charac-terized initially in 3T3 fibroblasts (3, 4, 6). It is also identicalto IgE binding protein (7) and a 32-kDa tumor-associatedlectin (8, 9). The COOH-terminal domain of Mac-2, whichcontains a conserved carbohydrate binding motif, is homol-ogous to a related family of 14-kDa galactose-specific lectins(reviewed in ref. 5). Collectively, Mac-2 and the 14-kDalectins are referred to as S-lectins because their carbohydratebinding function was thought to be thiol-dependent (10), anassumption that has recently been questioned (2).The biological functions of Mac-2 remain elusive. Its

putative role in cell adhesion has not been substantiated (11).Mac-2 may be associated with cell growth and differentiationbecause it is found in the nucleus of 3T3 fibroblasts and itsnuclear localization may be related to the proliferative stateof these cells (5). Also, the expression of Mac-2 occurs as afunction of macrophage differentiation (12). These reportsare consistent with the recent finding that a homologue of the14-kDa S-lectin is cytostatic for embryonic fibroblasts (13).Other in vitro studies have suggested that the expression ofMac-2 increases in oncogenically transformed and metastaticcells (14-16).

Progress in elucidating the function of Mac-2 has beenhampered, in part, by a lack of data on its expression in vivo.

In the present study, we found that Mac-2 is largely anepithelial-specific lectin with high expression seen in thecolonic epithelium. The progression from normal colonicmucosa to adenoma to carcinoma is characterized by amarked reduction in Mac-2 mRNA and protein levels and aloss of its nuclear localization. These observations suggestthat Mac-2 exclusion from the nucleus and its decreasedexpression may be related to the neoplastic progression ofcolon cancer.

MATERIALS AND METHODSTissue Specimens. Specimens of normal colonic mucosa,

colonic polyps, and colonic adenocarcinomas were obtainedfrom patients at the time of surgery. Pathological examinationwas performed on all specimens to assess the presence ofpreneoplastic and neoplastic disease. Tissue procurementwas in accordance with institutional review board standardsand procedure.

Immunohistochemistry. Specimens of colonic tissue wereformalin-fixed and paraffin embedded. Sections (3 ,um) wereincubated with the Mac-2-specific monoclonal antibody(mAb) M3/38 (12) at a concentration of 7 pg/ml. Boundantibody was detected by the peroxidase-antiperoxidasetechnique. The sections were counterstained with hematox-ylin.RNA Blots. RNA was extracted from normal mucosa and

tumors using guanidium isothiocyanate. The RNA was trans-ferred onto nylon membranes (Biotrans; ICN) and hybridizedwith a cDNA probe for human Mac-2 that was obtained byRNA PCR as described below. A cDNA probe for 18S rRNAprovided by Dylan Edwards (University of Calgary) was alsoused.

Protein Blots. Frozen specimens of normal mucosa andtumor were resuspended by probe sonication in Laemmlisample buffer. Samples (25 ,ug of protein) were resolved bySDS/PAGE (12%) and then transferred to nitrocellulose. Theimmunoreaction was performed by incubating the nitrocel-lulose with M3/38 (5 ,ug/ml) in Tris-buffered saline with0.05% Tween 20. Bound antibody was detected using aperoxidase-conjugated goat anti-rat IgG.RNA PCR Cloning. RNA PCR (Perkin-Elmer/Cetus) was

performed according to the manufacturer's instructions. Oli-gonucleotide primers were chosen that flanked the 5' and 3'ends of the Mac-2 coding sequence (3, 17). The primers usedwere GGAGCCAGCCAACGAGCGGAAAATG (upstreamprimer) and GCCCCTTTCAGATTATATCATGGT (down-stream primer). First-strand cDNA synthesis was generatedby incubating total RNA (2 ,tg) with the downstream primerand reverse transcriptase at 42°C for 15 min and 99°C for 5

Abbreviations: mAb, monoclonal antibody; CBP 35, carbohydratebinding protein 35.*To whom reprint requests should be addressed.

3466

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

July

25,

202

1

Proc. Natl. Acad. Sci. USA 90 (1993) 3467

min and then cooling to 5°C. PCR was performed as follows:The mixture was heated to 95°C for 3 min and then cooled to80°C. At this point, Vent polymerase (New England Biolabs)was added and 30 cycles were completed under the followingconditions: 1 min at 60°C, 1 min at 75°C, and 1 min at 95°C.Finally, the mixture was incubated at 60°C for 7 min andcooled to 4°C.One microliter of the PCR product was then used for

ligation in the TA cloning system (Invitrogen, San Diego),and subcloning was performed as recommended by themanufacturer. Positive colonies were analyzed by dideoxyDNA sequencing (Sequenase; United States Biochemical) oftheir inserts.

RESULTSImmunohistochemical Analysis of Mac-2 Expression. The

expression of Mac-2 in specimens of human colonic tissuewas examined by immunohistochemistry using the M3/38mAb. The degree ofimmunostaining was graded according tothe percentage of reactive cells as follows: <1% = 0, 1-25%= 1+, 26-75% = 2+, >75% = 3+. In the 15 sections ofnormal colonic mucosa that were examined, Mac-2 stainingwas confined to epithelial cells (Fig. 1A). There was noevidence of stromal cell staining with the exception ofscattered macrophages in the lamina propria. The distribu-tion of reactivity was distinct because a gradient of stainingwas noted along the crypt-to-surface axis. All specimensdemonstrated 3+ reactivity in the upper third ofthe crypt and

A

surface cells. In most specimens, cells at the base ofthe cryptwere nonreactive. This pattern of Mac-2 staining indicatedthat the proliferating cells at the base of the crypts expressedrelatively little Mac-2 in comparison to the more differenti-ated cells along the crypt-to-surface axis.A striking feature of the Mac-2 immunostaining of normal

colon was that it was largely nuclear (Fig. 1 A-C). Intensenuclear staining was evident in all specimens of normalmucosa examined (Fig. 1 B and C). This nuclear staining wasparticularly apparent in the middle third of the crypts (Fig. 1A and B). In surface epithelial cells in the upper third of thecrypts, basal and focally basolateral cytoplasmic staining wasseen along with intense nuclear staining (Fig. 1 B and C).The distinct localization of Mac-2 staining in normal co-

lonic mucosa prompted us to examine its localization in coloncarcinoma. The Mac-2 immunostaining pattern of 13 primarycolonic adenocarcinomas was markedly distinct from that ofnormal mucosa. The most salient feature of this stainingpattern was a loss of Mac-2 nuclear localization (Fig. 1D).Reactivity of Mac-2 in carcinomas was predominantly cyto-plasmic with only rare isolated staining of nuclei seen. Inaddition to this shift in Mac-2 localization, we also noted asignificant decrease in the intensity of Mac-2 staining com-pared to normal mucosa (compare Fig. 1 A-C with D). TheMac-2 staining of adenocarcinomas was heterogeneous be-cause not all areas of tumor exhibited staining (Fig. 1D).Specifically, only 54% of the carcinoma specimens exhibited3+ staining.

*J i.i,1 }' iX j i ;. Dj 1

e~~~tt J ' S

[R!; ..1 if 7X'sb

,

.A

qA'.-

"44 ....

FIG. 1. Immunohistochemical analysis of Mac-2 expression in normal colonic mucosa and colon carcinoma. Paraffin-embedded sections (3,um) of colonic tissue were immunostained with M3/38, a Mac-2 specific mAb (12). Bound antibody was detected by the peroxidase-antiperoxidase technique. Sections were counterstained with hematoxylin. (A) Normal colonic mucosa. Arrows indicate the base of coloniccrypts. (x25.) (B) Normal colonic mucosa. Arrow indicates the direction of the crypt-to-surface axis. (x65.) (C) Normal colonic mucosa. (x260.)(D) Colon adenocarcinoma. (x65.)

Medical Sciences: Lotz et al.

"W- 4.-A " elfflt", AP. .",.- '404Li-,, '. -AF

1.4 04.W Aw A.

a

i.., '. 4-I-..efRe..

qpjpp.:".i.

,OR .., A4w!,....W T

Dow

nloa

ded

by g

uest

on

July

25,

202

1

3468 Medical Sciences: Lotz et al.

The above results prompted us to study possible changesin Mac-2 staining that may occur during the progression fromnormal mucosa to carcinoma. It is widely accepted that mostcolon carcinomas arise from adenomatous polyps (18). Innine adenomas, Mac-2 staining was considerably less intensethan that seen in normal mucosa. However, the percentage ofreactive cells was high, with 78% of the adenomas exhibiting3+ staining. These differences are evident in Fig. 2A in whicha region of normal mucosa is adjacent to adenomatous tissue.The immunostaining of adenomas was much more uniformthan that of carcinomas because most of the adenomatous

p tA-3' y 4I:~~~~~~~~~~~~~~~~~~f./~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~4

{~~- q

colonic poys.Cloi

FIG. 2. Immunohistochemical analysis of Mac-2 expression incolnicpolps.Colnicpoiyps were processed for Mac-2 immuno-staining as described in the legend to Fig. 1. (A) Adenoma (openaqrrowhea,d) and adjacePnt normal mucosaq (solid arrowhe-ad). (x25.)(B) Adenoma. Arrow points to a nucleus. (x65.) (C) Hyperplasticpolyp. In this section, the surface of a hyperplastic polyp is shown.(x65.)

cells stained with Mac-2. However, the absence of nuclearstaining in carcinomas was also apparent in adenomas. In thehigh-power photomicrograph shown in Fig. 2B, none of thenuclei exhibited Mac-2 staining. The Mac-2 staining seen isexclusively cytoplasmic. Thus, the neoplastic progressionfrom normal mucosa to adenoma to carcinoma is associatedwith a loss of Mac-2 nuclear localization. One test of thiscorrelation would be to examine Mac-2 staining in colonicpolyps that are not preneoplastic. Hyperplastic polyps aresmall colonic polyps that do not normally progress to cancer(18). The Mac-2 staining pattern of these polyps was strik-ingly different than that of adenomatous polyps (Fig. 2C).Examination of eight hyperplastic polyps revealed intensenuclear staining in all of the specimens with focal areas ofbasolateral cytoplasmic staining as seen in normal mucosa.

Expression of Mac-2 Protein and mRNA. The immunostain-ing results indicated that the progression from normal colonicmucosa to adenocarcinoma was associated with a decrease inMac-2 expression. Additional evidence to support this pos-sibility was obtained by examining the relative amounts ofMac-2 protein in normal mucosa and tumor. Mac-2 expres-sion was detected by immunoblotting tissue extracts of colonadenocarcinoma and distal normal colonic mucosa obtainedfrom the same patient. As shown in Fig. 3, only a 35-kDaprotein, which corresponds to the molecular mass of Mac-2,was detected on the immunoblots using the M3/38 mAb. Thisobservation confirms that this mAb reacted only with Mac-2in the immunostaining studies described above. Analysis ofthe immunoblot shown in Fig. 3 indicates a significantdecrease in the intensity of the 32-kDa band in cancer relativeto normal mucosa in tissue pairs obtained from six differentpatients. These data corroborate the decrease in the intensityof Mac-2 immunostaining of cancer specimens compared tonormal mucosa that was apparent in Fig. 1.The possibility that the observed decrease in Mac-2 ex-

pression resulted from a decrease in the expression of itsmRNA was examined by Northern blotting. RNA was ex-tracted from pairs of normal mucosa and cancer obtainedfrom five different patients and hybridized with a Mac-2cDNA probe obtained from normal colonic mucosa (Fig. 4).A single 1.3-kb transcript was observed in each sample. Theintensity of this 1.3-kb band was significantly reduced in thecancer specimens compared to normal mucosa. Densitomet-ric scanning of these blots indicated a 5- to 10-fold decreasein the level of Mac-2 in cancer relative to the amount of 18SRNA in each sample.cDNA Cloning and Sequencing of Mac-2. We thought it

important to clone and sequence Mac-2 from normal colonicmucosa and tumor for several reasons. Foremost, the differ-ence in nuclear and cytoplasmic localization could resultfrom mutations in the Mac-2 coding sequence. Such muta-

N TNT N T N T N T N T97.4 -66.2 -

42.7 -

31-

21.5 -

FIG. 3. Immunoblot analysis of Mac-2 expression in normalcolonic mucosa (N) and colon tumors (T). Detergent extracts (25 pgof protein) of tumor and distal normal colon (at least 10 cm fromtumor) obtained from six different patients were resolved by SDS/PAGE (12%) and immunoblotted with the Mac-2-specific mAbM3/38 (12). The migration positions of molecular mass standards areindicated. The Mac-2 immunoreactive band exhibits a molecularmass of 32 kDa.

Proc. Natl. Acad. Sci. USA 90 (1993)

Dow

nloa

ded

by g

uest

on

July

25,

202

1

Proc. Natl. Acad. Sci. USA 90 (1993) 3469

AN TNT N T N T N T

4 . -B

N T N T N T N T N T

FIG. 4. Northern blot analysis of Mac-2 mRNA expression innormal colonic mucosa (N) and colon tumors (T). (A) RNA (20 ,ug)extracted from normal mucosa and tumor obtained from five differ-ent patients was resolved by agarose electrophoresis and hybridizedwith a 32P-labeled Mac-2 cDNA probe. A 1.3-kb RNA transcript wasdetected using this probe. (B) The relative amount of RNA in eachsample was assessed by hybridization with a cDNA probe specificfor 18S rRNA (1.8 kb).

tions have been shown to affect the nuclear localization ofother proteins (19). In addition, the four published sequencesof human Mac-2 are all derived from tumor cell cDNAlibraries (3, 9, 17, 20). No sequence has been reported froma nontumorigenic cell line or normal tissue.We generated Mac-2 cDNA by RNA PCR of mRNA

obtained from normal colonic mucosa, adenocarcinoma, andclone A, a colon carcinoma cell line (21). The nucleotidesequences of all three cDNAs were identical (Fig. 5). How-ever, we did observe allelic variation at codons 191 and 292upon sequencing several clones ofeach cDNA (Fig. 5). Thesesequence variations cannot be considered tumor specificbecause they were evident in the normal and tumor se-quences. Our sequence of human Mac-2 is identical to thosepublished by Robertson et al. (20) and Oda et al. (17) with theexception of the allelic variations that we noted.

DISCUSSIONThe data presented here provide new insight into the behav-ior and possible biological function of Mac-2. We found thatthe colonic epithelium is a major site of Mac-2 expression invivo. Expression in the colonic epithelium, as well as otherepithelia (unpublished data), is associated with terminallydifferentiated cells and not with proliferating crypt cells, anobservation that had been foreshadowed in an earlier study(22). Perhaps the most interesting observation we made isthat Mac-2 expression is concentrated in the nuclei of differ-entiated, colonic epithelial cells. Nuclear localization ofMac-2 (CBP 35) in cultured 3T3 fibroblasts has been estab-lished (23, 24). This report extends these observations bydocumenting that Mac-2 exists as a nuclear protein in vivo. Akey difference, however, between our data and previous invitro studies is that the nuclear localization ofMac-2 (CBP 35)in 3T3 cells is associated with proliferating cells (25, 26). Thein vivo data reported here indicate that the nuclear localiza-tion of Mac-2 is associated with cells in the upper two-thirdsof colonic crypts. These cells, for the most part, are notdividing, as evidenced by the fact that they do not incorporate[3H]thymidine, and they are assumed to be differentiated(27). These observations imply that Mac-2 nuclear localiza-tion correlates with either cessation of cell division or dif-ferentiation in the intestinal epithelium.

ATG GCA GAC AAT TTT TCG CTC CAT GAT GCG TTA TCT GGG TCT GGA AAC CCAM A D N F S L H D A L S G S G N P

AAC CCT CAA GGA TGG CCT GGC GCA TGG GGG AAC CAG CCT GCT GGG GCA GGGN P Q G W P G A W G N Q P A G A G

GGC TAC CCA GGG GCT TCC TAT CCT GGG GCC TAC CCC GGG CAG GCA CCC CCAG Y P G A S Y P G A Y P G 0 A P P

GGG GCT TAT CCT GGA CAG GCA CCT CCA GGC GCC TAC GGA GCA CCT GGAG A Y P G 0 A P P G A Y G A P G

GCT TAT CCC GGA GCA CCT GCA CCT GGA GTC TAC CCA GGG CCA CCC AGC GGCA Y P G A P A P G V Y P G P P S G

CCT GGG GCC TAC CCA TCT TCT GGA CAG CCA AGT GCC GGA GCC TAC CCTP G A Y P S S G 0 P S A G A Y P

GCC ACT GGC CCC TAT GGC GCC CCT GCT GGG CCA CTG ATT GTG CCT TAT AACA T G P Y G A P A G P L I V P Y N

CTG CCT TTG CCT GGG GGA GTG GTG CCT CGC ATG CTG ATA ACA ATT CTG GGCL P L P G G V V P R M L I T I L G

ACG GTG AAG CCC AAT GCA AAC AGA ATT GCT TTA GAT TTC CAA AGA GGG AATT V K P N A N R I A L D F 0 R G N

5117

10234

15351

20468

25585

306102

357119

408136

459153

GAT GTT GCC TTC CAC TTT AAC CCA CGC TTC AAT GAG AAC AAC AGG AGA GTC 510D V A F H F N P R F N E N N R R V 170

ATT GTT TGC AAT ACA AAG CTG GAT MT AAC TGG GGA AGG GAA GAA AGA CAG 561I V C N T K L D N N W G R E E R 0 187

TCG GTT TTC CCA TTT GAA AGT GGG AAA CCA TTC AAA ATA CAA GTA CTG GTT 612S V F P F E S G K P F K I 0 V L V 204

GAA CCT GAC CAC TTC AAG GTT GCA GTG AAT GAT GCT CAC TTG TTG CAG TAC 663E P D H F K V A V N D A H L L 0 Y 221

AAT CAT CGG GTT AAA AAA CTC AAT GAA ATC AGC AAA CTG GGA ATT TCT GGT 714N H R V K K L N E I S K L G I S G 238

GAC ATA GAC CTC ACC AGT GCT TCA TAT ACC ATG ATA TAA 753D I D L T S A S Y T M I Z 251

FIG. 5. Complete nucleotide sequence of human Mac-2 and itstranslated amino acid sequence (single-letter code). cDNA wasobtained by RNA PCR from normal colonic mucosa, colon carci-noma, and clone A, a colon carcinoma cell line, and analyzed bydideoxy DNA sequencing. The nucleotide sequences of all threecDNAs were identical. Boxed codons (191 and 292) indicate sites ofallelic variation that were seen upon sequencing multiple clones ofeach cDNA. The allelic variant for codon 191 was CAT (H) and forcodon 292 it was CCC (P).

The morphological transition of normal colonic mucosa toadenoma to carcinoma has provided one of the best modelsfor studying the molecular basis of neoplastic progression(28). In the present study, we were able to associate thisprogression with changes in the expression and nuclearlocalization of Mac-2. Specifically, the progression fromnormal mucosa to carcinoma involves a significant decreasein Mac-2 protein and mRNA levels as well as a loss of itsnuclear localization. These findings are of interest becauseMac-2 had been identified originally as a protein whoseexpression increases in oncogenically transformed and met-astatic cells (14, 15). There are several reasons that mayexplain the discrepancy between these data and our results.Much of the work on Mac-2 expression in cancer is based onthe use of cell lines, especially oncogenically transformedfibroblasts (14, 15). It is possible that in vitro culture condi-tions affect Mac-2 expression or, more importantly, thattransformed embryonic fibroblasts behave quite differentlyfrom colon carcinoma cells. In this context, we were not ableto detect Mac-2 expression in stromal fibroblasts by immu-noperoxidase staining. Thus, the data presented here consti-tute a study of Mac-2 in the neoplastic progression ofintestinal epithelial cells, a cell type that exhibits high levelsof Mac-2 expression in vivo.The difference in Mac-2 nuclear localization between nor-

mal mucosa and adenomas/adenocarcinomas was readilyapparent from the immunostaining results. Such differencesin nuclear localization are observed when mutations occur inthe nuclear localization sequences ofspecific proteins such assimian virus 40 large tumor (T) antigen (19, 29, 30). For thisreason, we compared the cDNA sequence of Mac-2 from

Medical Sciences: Lotz et al.

I

Dow

nloa

ded

by g

uest

on

July

25,

202

1

3470 Medical Sciences: Lotz et al.

normal colon, from colon carcinoma, and from a coloncarcinoma cell line. No mutations were found that woulddistinguish normal mucosa from neoplastic tissue. One con-clusion from these results is that Mac-2 may belong to acategory of nuclear proteins that depend on other proteins fortheir transport into the nucleus (19). This possibility is alsoevidenced by the fact that Mac-2 does not contain a well-defined nuclear localization sequence similar to T antigen(Fig. 5), though it is possible that its nuclear localizationinvolves a much larger protein sequence (31). These obser-vations on Mac-2 are similar to recent findings on the nuclearlocalization of the p53 tumor suppressor protein in breastcarcinomas (32). A significant fraction of breast cancersexamined (37%) excluded p53 from the nucleus but most ofthese cancers contained wild-type p53 alleles. These dataprovide a compelling reason for identifying proteins withwhich Mac-2 associates and for elucidating its mode ofnuclear transport.The decreased expression of Mac-2 in colon carcinomas

was indicated initially by the immunostaining data and cor-roborated by analysis of its RNA and protein levels. Thedown-regulation of gene expression in cancer can result fromdeletion, mutation, or rearrangement of genomic DNA orfrom aberrations in either transcriptional or translationalregulation (33). In a Southern analysis of three paired spec-imens of normal colonic mucosa and carcinoma, we found noevidence for gross deletion or rearrangement of the Mac-2gene (unpublished data). This preliminary result, along withthe observation that Mac-2 mRNA levels are reduced intumors, suggests that decreased Mac-2 expression is causedby altered expression of an intact Mac-2 gene. Additionalstudies are required to verify this possibility and to define thenature of such control mechanisms.The results presented in this paper justify further work to

investigate the possibility that Mac-2 is a tumor suppressorprotein. At present, this justification is based on its decreasedexpression and loss ofnuclear localization in tumors. The keytest of this possibility will be to determine if increasing thelevels of nuclear Mac-2 in colon carcinoma cells alters theirtransformed phenotype. Experiments of this type have beendone with known tumor suppressor proteins including p53and RB (34-38). In both cases, expression of the wild-typeprotein inhibited tumor cell growth in vitro. Several lines ofevidence suggest that Mac-2 expression also may affect cellgrowth or differentiation. Its expression in vivo is associatedwith terminally differentiated epithelial cells. In addition, itsnuclear localization and state of phosphorylation have beencorrelated with the proliferative state of 3T3 fibroblasts(24-26). Lastly, a 14-kDa galactose-specific lectin that ishomologous to the COOH-terminal domain of Mac-2 inhibitsthe growth of embryonic fibroblasts (13). Collectively, theavailable data provide reason for investigating the role ofMac-2 in the growth and differentiation of normal and ma-

lignant colonic epithelial cells.

This work was supported by National Institutes of Health GrantsCA 42276 and 44704. M.M.L. was funded by a National ResearchService Award. A.M.M. is the recipient of an American CancerSociety Faculty Research Award.

1. Woo, H. J., Shaw, L. M., Messier, J. M. & Mercurio, A. M.(1990) J. Biol. Chem. 265, 7097-7099.

2. Woo, H. J., Lotz, M., Jung, J. U. & Mercurio, A. M. (1991) J.Biol. Chem. 266, 18419-18422.

3. Cherayil, B. J., Weiner, S. J. & Pillai, S. (1989) J. Exp. Med.170, 1959-1972.

4. Cherayil, B. J., Chaitovitz, S., Wong, C. & Pillai, S. (1990)Proc. Natl. Acad. Sci. USA 87, 7324-7328.

5. Wang, J. L., Laing, J. G. & Anderson, R. L. (1991) Glycobi-ology 1, 243-252.

6. Jia, S. & Wang, J. L. (1988) J. Biol. Chem. 263, 6009-6011.7. Laing, J. G., Robertson, M. W., Gritzmacher, C. A., Wang,

J. L. & Liu, F.-T. (1989) J. Biol. Chem. 264, 1907-1910.8. Raz, A., Pazerini, G. & Carmi, P. (1989) Cancer Res. 49,

3489-3493.9. Raz, A., Carmi, P., Raz, T., Hogan, V., Mohamed, A. &

Wolman, S. R. (1991) Cancer Res. 51, 2173-2178.10. Drickamer, K. (1988) J. Biol. Chem. 263, 9557-9560.11. Mercurio, A. M. (1990) Curr. Opin. Cell Biol. 2, 845-849.12. Ho, M.-K. & Springer, T. A. (1982) J. Immunol. 128, 1221-

1228.13. Wells, V. & Mellucci, L. (1991) Cell 64, 91-97.14. Crittenden, S. L., Roff, C. F. & Wang, J. L. (1984) Mol. Cell.

Biol. 4, 1252-1259.15. Raz, A., Zhu, D., Hogan, V., Shah, N., Raz, T., Karkash, R.,

Pazerini, G. & Carmi, P. (1990) J. Cancer 46, 871-877.16. Irimura, T., Matsushita, Y., Sutton, R. C., Carralero, D.,

Ohannesian, D. W., Cleary, K. R., Ota, D. M., Nicolson,G. L. & Lotan, R. (1991) Cancer Res. 51, 387-393.

17. Oda, Y., Leffler, H., Sakakura, Y., Kasai, K. & Barondes,S. H. (1991) Gene 99, 279-283.

18. Ming, S. I. & Goldstein, H. (1992) Gastrointestinal Pathology(Saunders, Philadelphia).

19. Silver, P. A. (1991) Cell 64, 489-497.20. Robertson, M. W., Albrandt, K., Keller, D. & Liu, F.-T. (1990)

Biochemistry 29, 8093-8100.21. Daneker, G. W., Piazza, A. J., Steele, G. D. & Mercurio,

A. M. (1988) Cancer Res. 49, 681-686.22. Flotte, T. J., Springer, T. A. & Thorbeck, G. J. (1983) Am. J.

Pathol. 111, 112-124.23. Moutsatsos, I., Davis, J. M. & Wang, J. L. (1986) J. Cell Biol.

102, 477-483.24. Moutsatsos, I. K., Wade, M., Schindler, M. & Wang, J. L.

(1987) Proc. Natl. Acad. Sci. USA 84, 6452-6456.25. Cowles, E., Argwal, N., Anderson, R. L. & Wang, J. L. (1990)

J. Biol. Chem. 265, 17706-17712.26. Haman, K., Cowles, E. A., Wang, J. L. & Anderson, R. L.

(1991) Exp. Cell Res. 196, 82-91.27. Gordon, J. I. (1989) J. Cell Biol. 108, 1187-1194.28. Fearon, E. R. & Vogelstein, B. (1990) Cell 61, 759-767.29. Lanford, R. E. & Butel, J. S. (1984) Cell 37, 801-813.30. Kalderon, D., Roberts, B. L., Richardson, W. D. & Smith,

A. E. (1984) Cell 39, 499-509.31. Standiford, D. M. & Richter, J. D. (1992) J. Cell Biol. 118,

991-1002.32. Moll, U. M., Riou, G. & Levine, A. J. (1992) Proc. Natl. Acad.

Sci. USA 89, 7262-7266.33. Lee, S. W., Tomasetti, C. & Sager, R. (1991) Proc. Natl. Acad.

Sci. USA 88, 2825-2829.34. Shohat, O., Greenberg, M., Reisman, D., Oren, M. & Rotter,

V. (1987) Oncogene 1, 277-283.35. Baker, S. J., Markowitz, S., Fearon, E. R., Willson, J. K. V.

& Vogelstein, B. (1990) Science 249, 912-915.36. Mercer, W. E., Sheilds, M. T., Amin, M., Sauve, G. J., Ap-

pella, E., Romano, J. W. & Ullrich, S. J. (1990) Proc. Natl.Acad. Sci. USA 87, 6166-6170.

37. Huang, H.-J., Yee, J.-K., Shew, J.-Y., Chen, P. L., Bookstein,R., Freidman, T., Lee, E. Y.-H. P. & Led, W. H. (1988)Science 242, 1563-1566.

38. Takahashi, R., Hashimoto, T., Xu, H.-J., Hu, S.-X., Matsui,T., Miki, T., Bigo-Marshall, H., Aaronson, S. A. & Benedict,W. F. (1991) Proc. Natl. Acad. Sci. USA 88, 5257-5261.

Proc. Natl. Acad. Sci. USA 90 (1993)

Dow

nloa

ded

by g

uest

on

July

25,

202

1