ICANCERRESEARCH56, 4841-4845, November 1, 1996]

Advances in Brief

Widespread Loss of Gelsolin in Breast Cancers of Humans, Mice, and Rats1

Harold L. Asch,2 Karen Head, Yan Dong, Farah Natoli, Janet S. Winston, James L. Connolly, and Bonnie B. Asch

Departments of Experimental Pathology [H. L A., K. L H., Y. D., F. N., B. B. A.] and Pathology [J. S. WI, Roswell Park Cancer Institute, Buffalo, New York 14263, and Department of Pathology, Beth Israel Hospital, Boston, Massachusetts 02215 Ii. L C.)

Abstract Materialsand Methods

Down-regulation of gelsolin, an actin-binding protein, is frequently found Mammary Tissues and Cell Lines. Human breast tissues included 2 in several types of transformedcells and tumors.The presentstudy demon reduction mammoplasties, 1 of which had proliferative changes without atypia, strates that gelsolin protein and RNA were absent or markedly reduced in 3 nonmalignant tissues immediately adjacent to cancers, 29 sporadic, primary human breast cancer celllines relative to “normal―mortalhuman mammary invasive breast cancers (1 tubular and 28 poorly differentiated ductal carcino epithelial cells and benign, immortalized cell lines. Moreover, actin filaments mas), and 1 metastasis to chest wall. Normal mouse mammary tissues were were usually attenuated coincident with the reduction in gelsolin. GeISOIIn from pregnant BALB/c mice, and mammary tumors were from BALB/cfC3H was also missingor greatlydecreasedin 70% of3Ohumansporadic,invasive animals infected with mouse mammary tumor virus. Normal rat mammary breast carcinomas examined by immunocytochemistry and in 100% of virally tissues were provided by Dr. M. Ip from virgin Sprague-Dawley females, and induced mouse and chemically induced rat mammary carcinomas evaluated rat mammary carcinomas induced by DMBA treatment were a gift from Dr. C. by Northernanalysis.Southernanalysisrevealedno nutjormutationsin the Ip. HMECs from a reduction mammoplasty (Clonetics, Inc., San Diego, CA) gelsolin gene of human breast cancer cells. Our results show that partial or were grown in short-term culture (9—12passages) and represent human nor total lossofgelsolin expression is common to the majority ofbreast cancers of mal, mortal mammary cells. Benign immortal breast epithelial cell lines were diverse etiologiesin three animal species and point to gelsolin as a candidate HBL100, MCF1OA,and l84AlN4. Tumor cell lines included 184A1N4TH, suppress@w of breast cancer. SKBR3, MCF7, MDA-MB-231, and T47D. Immunocytochemistry and Analysis by Western, Northern, and South Introduction em Blots. Acetone-fixed,frozenbreasttissue sections (4—6p@m)wereincu bated in anti-gelsolin MAb (Sigma Chemical Co., St. Louis, MO), washed, and The AF3 cytoskeleton is directly or indirectly involved in many vital cell incubated with a fluorescein-tagged secondary antibody for IF or with a functions such as cell shape, motility, cytokinesis, endocytosis, mRNA local biotin-tagged secondary antibody for the ABC IP technique (Vector Labora ization, and growth regulation (1—3).The AF network has been implicated as tories, Burlingame, CA) using the diaminobenzidine color reagent. Additional both a target and mediator of signal transduction initiated through receptor breast tissues were fixed in formalin and embedded in paraffin; 4-,im-thick tyrosine kinases and extracellular matrix-integrin systems (4—8).In response sections were microwaved for 10 mm, followed by the ABC IP protocol using to many extracellular signals, cells change their shape, their adhesion to matrix a Ventana 320 Automated Slide staining system (Ventana Medical Systems, components, and their interactions with adjacent cells (5—9).Cells possess a Tucson, AZ). Cells grown on glass coverslips were fixed in 2% paraformal complex army of proteins that bind to actin in its filamentous or monomeric dehyde, stained by IF with the anti-gelsolin MAb, and subsequently stained globular forms to regulate the assembly/disassembly, architecture, and distri with rhodamine phalloidin (30 units/ml) to detect filamentous actin in the same bution ofAFs (1, 2, 9). A key determinant of AF length is gelsolin, a Mr90,000 cells. Negative primary antibody controls for tissue and cultured cell staining actin-binding protein that severs AFs, caps the fast-growing (barbed) ends, and promotes nucleation of polymerization (10). included use of an anti-ras MAb (Chemicon International, Inc., Temecula, CA), which was unreactivewith humanbreastcells (datanot shown). Evidence in support of gelsolin as a tumor suppressor has been steadily accruing.Gelsolin was markedlydiminishedin H-ms-transformedmouse fibro Immunohistochemistry staining was evaluated independently by two or three investigators (H. L. A. and B. B. A., and in some cases, J. S. W.), and blasts (1 1), human fibroblasts, and epitheial cells transformed by SV4O virus (12), gastric caitinoma cell lines (13), bladder cancer cell lines, and the majority of cancers were categorized as having reduced gelsolin if two-thirds or more of bladder cancers (14). Transfection and mutation studies have confirmed the tumor tumor cells in the sample were either negative or weakly stained. The percent suppressor function ofgelsolin for both epitheial and fibroblastic tumor cells (11, age of reactive and unreactive cells was estimated from examination of the 14). Two reports on mammary cancers indicated gelsolin was greatly reduced in entire tissue section. The results for the great majority of cancers were all tumorsexamined.The proteinwas undetectableby immunocytochemistryin12 clear-cut. The rare disagreements that arose about a staining pattern were of 12 human breast cancers (15), and gelsolin RNA was reduced 4—5-foldin resolved by simultaneous examination and discussion of the sample by all mouse mammary tumors (16). The key findings of the present study are: (a) three investigators. gelsolin expression was drastically reduced in the majority of human, mouse, and Cultured mammary cells were harvested in an antiprotease cocktail and rat mammarytumorsexaminedas well as in most humanbreastcancercell lines; solubilized in Laemmli buffer for Western blotting (17). Equal protein loads of (b) down-regulation of gelsolin protein in the human tumor cell lines was due to cell extracts (Dot/Metric protein assay; Geno Technology, St. Louis, MO) were decreased gelsolin RNA and may not involve major mutations of the gene; (c) in run on 7.5% SDS-PAGE minigels, and Western blots probed with anti-gelsolin most breast cancer cells, reduced gelsolin protein expression was associated with MA], were visualized by chemiluminescence (Pierce, Rockford, IL). Immu decreasedAFs. noreactive bands were quantitated by volume densitometry. Equal protein transfer for each lane was monitored by India ink staining of blots after immunostaining. The anti-gelsolin MAb was specific for a Mr 90'000 band. Received 8/8/96; accepted 9/18/96. The costs of publication of this article were defrayed in part by the payment of page Parallel Western blots were probed with an anti-actin MAb (Sigma), which charges. This article must therefore be hereby marked advertisement in accordance with reacts with the carboxy terminus of all known mammalian actins. For Northern 18 U.S.C. Section 1734 solely to indicate this fact. analysis, total RNA was extracted from cultured cells or tissues by ThI reagent I Supported by the Roswell Park Alliance Foundation and Grants CA56609 and CA62014 from the National Cancer Institute, NIH. (Molecular Research Center, Cincinnati, OH). Genomic DNA was isolated 2 To whom requests for reprints should be addressed. from cultured cells and digested with an endonuclease for Southern analysis. 3The abbreviations used are: ÀY,actin filament; DMBA, 7,l2-dimethylbenzanthra Northern and Southern blots were probed with a full-length cDNA clone of cene; HMEC, human mammary epitheial cell; MAb, monoclonal antibody; IF, immuno fluorescence; IP, immunoperoxidase; PIP2, phosphatidylinositol 4,5-trisphosphate; PLD, human cytoplasmic gelsolin (a gift of Dr. D. Kwiatkowski) following standard phospholipase D. procedures. 4841

Downloaded from cancerres.aacrjournals.org on September 29, 2021. © 1996 American Association for Cancer Research. @@@ /@@@ ‘@ @J@:-.S •@ “ @, @‘

LOSS OF GELSOLININ BREASTCANCER

Results (Fig. 1, b and d), positive (Fig. le) or weakly stained (Fig. lg), whereas some tumors were heterogeneous, with tumor cell reactions Gelsolin Expression in Normal and Malignant Human Breast ranging from negative to positive (Fig. 1, f and h). In a few tumors, Tissues. Immunocytochemistry on frozen sections of human reduc distinct regions of positive tumor cells were adjacent to regions of tion mammoplasty epithelium or of normal-appearing epithelium ad negative tumor cells (Fig. lfj; in these cases, the stroma stained jacent to or within breast cancers produced moderate-to-strong stain equally in both regions, thereby demonstrating the evenness of the ing in luminal cells and an intense reaction in myoepithelial cells specific immunoreaction. Gelsolin reactivity was primarily cytoplas (Figs. 1, a, c, and h). The moderate-to-strong reactivity of stromal cells provided a positive control for the staining procedure (Fig. 1). mic, but occasionally, nuclei appeared to stain (Fig. 1). Expression of gelsolin in the epithelium of benign breast tissue with Gelsolinand Actin Expression in Malignant and Nonmalignant proliferative disease was also strong (data not shown). In contrast, Human Breast Cell Cultures. Gelsolin expression in cultured human gelsolin expression in tumor cells was undetectable or greatly reduced breast cells has not been examined previously. By immunocytochemistry, in frozen sections of 11 of 14 infiltrating ductal carcinomas compared the mortal HMECs displayed the most intense reactivity for gelsolin, to nonmalignant epithelium (Fig. lb). In paraffin-embedded tissue which usually appeared as a diffuse cytoplasmic stain (Fig. 2a), but a few sections stained by IP, gelsolin was undetectable or greatly reduced in cells had a punctate pattern (data not shown). Gelsolin expression was 10 of 16 breast tumors (Fig. 1, d-g). Thus, 70% of 30 human breast substantially reduced in essentially all of five cell lines derived from carcinomas examined by IF or IP had little or no detectable gelsolin. breast cancers (Fig. 2, d-g) as compared to cells derived from nonmalig The tumors displayed a spectrum of reactivities. Most tumors had nant human breast tissues (Fig. 2, a-c). Although reduced, the pattern homogeneous staining, with more than 90% of tumor cells negative observed in SKBR3 cells (Fig. 2d) appeared to be intermediate between

@ U@:'@4J. . 5..,‘* @ i• ‘ @4: @ .@- .5 ,@ @ . ‘@ .,,

\_.5 .5..',,. @@@ :-@, S @@@ .,(@...... ,- ‘F. . @: @ . ‘5 @ ) .— -.;@._.(. @‘@c @ . .‘@ ?@ . @@‘5 ,: @, @,•‘-. @,‘\: ‘@@‘ - @ I ‘@‘‘@•@•.-@•;.2 .;‘@‘@, .1@. @@ .@,‘-. .-.-.,, ..@ . .@., @@@‘:t-@ • @ ,5-.._ b @ .@ 9'. -5. I : , --.: . _‘- . @@ ..@•5) a @ .. .‘,. --.‘ S @ - ‘ , . p S ‘ •••@#S, •@ S @ ,., I 1@' . .5 7 : . @Iz..@. ..,@.i @ fl.l @‘th c@ q lQ4•@4I@øs@i@d “ @ 15

@ i1@' (‘@]I A. ‘tfr, ‘4' j ‘It @@@ t ,,,; @.t @-y@7k.‘@c(if -.-‘@k@'@'-2 @@@ . ‘S ‘@ I “@ @:. ‘@‘ t . .- ‘ . . •, /@ S @-L @@@ @,at,>.... V. • .;, ,:.. @‘ ‘.. @@@ t@@' @.: •,,., k @b@1 @c'@ r @@ ,I I @. I @@@ @‘,. 0 • .. ,@) ,, @ @1@ @‘ • @.t

@ ,. •W@;.4J@ . @.. .T

Fig. I. Expression of gelsolin in human breast tissues. In a and b, frozen tissue sections were stained by IF using an anti-gelsolin MAb; in c.h, paraffin-embedded tissue sections were stained by IP with the same MAb and counterstained with hematoxylin, except for c, which was not counterstained. Normal epithelium (arrowheads in a; arrow in c) and stroma (arrow in a) in reduction mammoplasties (a and c) are strongly stained, whereas tumor cells in invasive carcinomas (b and d.h) displayed a range of reactivities, including no or very little staining (b, d, and g), heterogeneous staining (1and h), and moderate to strong staining (e). Note positive stroma (arrowhead) in d. In f, a tumor has all tumor cells positive (arrowhead) in one region and negative (arrow) in an adjacent region. Note even stromal reactivity in both regions. In h. heterogeneous tumor cell reactivity is seen, including weak (arrowheatf), moderate-strong, negative (small arrow), along with strongly reactive, normal-appearing ductal epithelium (large arrow). X 87. 4842

Downloaded from cancerres.aacrjournals.org on September 29, 2021. © 1996 American Association for Cancer Research. @@@@ - S 5

LOSS OF GELSOLININ BREASTCANCER

Gelsolin F..Actin and MCF1OA cells (data not shown), whereas their tumorigenic deriva tive, the 184A1N4TH cells, gave a reaction similar to SKBR3 cells (data not shown). abCdefg In general, Western blot results using the anti-gelsolin MAb paralleled those obtained by immunocytochemistry, with most malignant cell cx tracts having a diminished Mr 90'000 band compared to benign cell extracts (Fig. 3). Densitometric analysis indicated there were 3-fold or fewer differences in gelsolin levels between HMEC and the HBL100, l84AlN4, l84AlN4Th, and SKBR3 cells, whereas MDA-MB-231 had 20-fold less and T47D and MCF7 cells had 50-fold less gelsolin than HMECs. Two anomalies apparent in the anti-gelsolin Western blots were the strong gelsolin band for the malignant l84AlN4Th cells and the lack ofreactivity in extracts ofthe benign MCF1OA cells (Fig. 3). The absence of reactivity was obtained with three different extracts of the MCF1OA cells, despite inclusion of antiproteases during extraction and the presence of a substantial immunoreactive actin band on the blots (see below). Consistent with gelsolin protein expression data, Northern analysis showed that gelsolin transcripts were significantly reduced in four of five tumor cell lines compared to the HMECs and benign cells (Fig. 4, top panel); 184A1N4Th had high levels ofgelsolin RNA, consistent with the protein detected on Western blots. Southern analysis of genomic DNA from the cultured cells after restriction with BamHI or Bc!! produced similar banding patterns, indicating that there are no major mutations in the gelsolin gene of cancer compared to noncancer cells (Fig. 5). To examine AFs, rhodamine phalloidin was used. Double-staining of cultured cells for AFs and gelsolin revealed that the AF cytoskel eton was diminished or disordered in tumor cells coincident with reduced gelsolin expression (Fig. 2). Western blots probed with an anti-actin MAb paralleled the F-actin findings, with some exceptions (Fig. 3). Compared to HMECs and benign cells, three of the cancer cell lines (MCF7, SKBR3, and T47D) had reduced actin band in tensity, whereas MDA-MB-231 and 184A1N4TH cells were not different. Gelsolin RNA in Normal and Malignant Mouse and Rat Mam mary Tissues. Results of Northern analysis for gelsolin RNA from both chemically induced rat and virally induced mouse mammary tumors were consistent with data from the human tissues and cultured cells. Gelsolin transcripts were dramatically reduced in six of six rat and eight of eight mouse tumors (Fig. 4).

I—

Fig. 2. Expression of gelsolin and actin in human breast-derived benign and malignant cultured cells. Cultured cells were stained with an anti-gelsolin MAb in combination with @@@ a fluorescein-tagged secondary antibody (left column) and then by rhodamine phalloidin S 9OkD (right column) to visualize AFs. Each row represents fluorescence in the same cells. Rows a-c, benign cells; rows d-g, tumor cells. HMECs in a exhibit strong gelsolin and AF expression, while the staining is weaker but still prominent in the benign cell lines HBL100 (b) and MCFIOA (c). Little or no gelsolin was detected in most MCF7 (e), @@@@ MDA-MB-231(f),andT47D(g)cells,andgelsolinwassubstantiallyreducedinSKBR3 43kD (‘0cells.AFswereverysparseinMCF7(e)andT47D(g)cellsandweredecreasedin SKBR3 (d) and MDA231 (J) cells. The latter cells had AFs localized mainly in the peripheral cytoplasm, whereas HMECs, the benign cells, and SKBR3 cells had AFs throughout their cytoplasm. The reactions of I84A1N4 cells resembled the other benign cells, and the 184A1N4Th tumor cells were similar to SKBR3 cells (data not shown). X187. Fig. 3. Expression of gelsolin and actin in human breast cells as revealed by Western blot analysis. Equal total protein loads of cell extracts were run on SDS-PAGE, and Western blots were probed with anti-gelsolin MAb (toppanel) or anti-actin (bottompanel) that of the other malignant cells and that of benign immortal cells. For and visualized by chemiluminescence. The resulting bands on film were quantified by densitomery (see “Results―).Equaltotal protein transfer in each lane was verified by HBL100 cells (but not for others), gelsolin expression was density de post-Western India ink staining (data not shown). Cell extracts are indicated above pendent, with maximum intensity displayed in confluent cultures (Fig. corresponding lanes (cancer cells in first five lanes on the left). Note reduction of gelsolin (Mr 90,000 band; 90W) in three of five cancer cells. For MCFIOA, a gelsolin band was 2!@)andbarely detectable levels in sparse cultures (data not shown). The not detected, but actin was abundant. Total actin (Mr 43,000 band; 43W) was similar in immortal benign 184A1N4 cells had a staining pattern similar to HBL100 benign and two of five tumor cells. 4843

Downloaded from cancerres.aacrjournals.org on September 29, 2021. © 1996 American Association for Cancer Research. @@@@@@@@ 18S 8 g@ 2

LOSS OF GELSOLININ BREASTCANCER

S carcinomas were induced by mouse mammary tumor virus, and the rat neoplasms were induced by DMBA, whereas the human tumors were m =@l mI-c#@ sporadic and of unknown origin. These results expand on earlier reports wherein gelsolin protein was undetectable in a small set of @ 2.6 S human breast cancers (15), and gelsolin RNA was markedly decreased in mouse mammary tumors arising from spontaneously transformed hyperplastic cells ( 16). Partial or complete loss of gelsolin expression, therefore, appears to be one of the most frequently occurring defects in breast cancer, regardless of etiology and across three animal spe z cies. Results with the cultured human breast cells were consistent with the findings in tissues. Gelsolin expression was not substantially different from normal and benign cells in the 184A1N4TH and @@@ 2.6 @@::1 SKBR3 tumor lines but was drastically decreased in T47D, MCF7, and MDA-MB-23l cells. The reduced gelsolin protein was associated with decreased gelsolin RNA in the latter cell types, but the size of laS both the protein and the transcript was the same as in benign cells, suggesting there were no major mutations in the coding sequence of the gelsolin gene. Likewise, large deletions or other gross rearrange I., @ V) @O s@- ments were not detected in the gene; therefore, the down-regulation is @ ..@ — ri !‘@ a. @. @- I- I- I- I- I- I-' probably due to some other mechanism. These results are consistent r' with the fact that the chromosome location of the gelsolin gene, 9q33, @ is not a locus which has been associated with loss of heterozygosity in human breast cancers. Although loss of heterozygosity often charac @ terizes tumor suppressor genes in cancer, this is not always the case, as found with maspin and BRCA1 in sporadic breast cancers (18, 19). The lack of an immunoreactive gelsolin band on Western blots Fig. 4. Northern analysis of gelsolin expression in cultured human breast cells and in using several different extracts of MCF1OA cells is difficult to explain normal and malignant mouse and rat mammary tissues. Total RNA from cells or tissues was subjected to standard Northern blotting and probed with a full-length cDNA of the since the same MAb detected abundant gelsolin in them by IF (Fig. 2). human gelsolin gene (upper strip of each panel), which also reacts with rodent gelsolin The gelsolin band was present in extracts of other benign cells, and transcripts.Toppanel,benign(HBLIOO,HMEC,andMCF1OA)orcancer(I84AIN4Th, India ink staining showed excellent transfer of equal total protein in MDA-MB-23l, MCF7, T47D, and SKBR3) human breast cells in culture. Middle panel, virgin rat mammary gland and DMBA-induced rat mammary tumors (TI-Tb). Bottom all lanes run on the gels as well as confirming general integrity of panel, normal mammary gland and placenta from pregnant mouse and primary mouse protein bands (data not shown). Probing of replicate blots with an mammary tumors (T1—T8).The single gelsolin band at 2.6 kb is dramatically reduced in anti-actin MAb demonstrated a substantial, undegraded actin band in four of five human cancer cell lines, six of six rat mammary tumors, and eight of eight mouse mammary carcinomas. Blots were subsequently probed with a eDNA for 18S the same MCF1OA extract. A possible explanation for these data is rRNA for lane load verification (lower strip of each panel). that the gelsolin protein epitope in MCF1OA cells is either mutated or modified. However, ample gelsolin RNA of correct size was present in these cells. Discussion The abundant gelsolin expression in mortal and immortal benign Our data show that gelsolin was absent or greatly reduced in 70% human breast cells as well as in epithelium of breast tissue containing of human invasive breast cancers examined and in 100% of primary proliferative changes without atypia (our study and Ref. 15) implies mammary tumors in two rodent models of the disease. The mouse that its down-regulation most likely occurs at a later stage in neoplas

@ n I-

@ Fig. 5. Southern analysis of the gelsolin gene in @@@ human breast cells. Genomic DNA extracted from ,, . ‘, , 4 5 @@@@ human breast cell lines was restricted with BamHI S - or Bctl and analyzed by Southern blotting using a . full-length gelsolin cDNA. Differences in banding patterns of DNA fragments were not apparent be tween malignant and benign cells. The different @@@@ intensities observed among lanes were due to van- .. —@ I•@ I@ ations in loads.

@ ‘ . k*S.l b a BamHI b ilelI 4844

Downloaded from cancerres.aacrjournals.org on September 29, 2021. © 1996 American Association for Cancer Research. LOSS OF GELSOUN IN BREASTCANCER tic progression. Indeed, Medina et al. (16) found that gelsolin RNA defects in breast cancer cells by changing their response to external levels were normal in mouse mammary hyperplastic preneoplasias but signals. Experiments to determine if gelsolin functions as a tumor decreased in the cancers. The exact role of the loss of gelsolin in suppressor of breast cancer are in progress. mammary tumongenesis remains to be determined. The absence of gelsolin per se apparently does not induce mammary tumorigenesis Acknowledgments because an increased incidence of mammary carcinomas has not been noted in gelsolin-null mice (20). However, such mice might be more Reduction mammoplasties were obtained from the Department of Pathol susceptible to carcinogen-induced mammary tumorigenesis. Because ogy, Beth Israel Hospital, and all other human breast tissues were from the Tissue Procurement Facility, Dr. Harry Slocum, Director, Department of gelsolin is a component of the AF cytoskeleton, its loss might affect Pathology, Roswell Park Cancer Institute. We thank Drs. Margot and Clement the invasive ability of cells. However, MDA-MB-23l cells are inva Ip for the rat mammary tissues, Dr. David Kwiatkowski for the cDNA clone of sive in in vitro assays, whereas T47D and MCF7 cells are not (21); human gelsolin,Dr. MarthaStampferfor the breast cell lines, A1N4 and therefore, invasiveness is either independent of gelsolin, or loss of A1N4Th, and Dr. Tom Fanning for MCFIOA cells. We gratefully acknowl gelsolin is not sufficient for this ability. Loss of gelsolin may be edge the excellent technical assistance of Marlene Andrews and Nancy Kaul. related to deregulation of growth control and tumorigenicity. The three tumor lines (MCF7, MDA-MB-23l, and T47D) with the greatest References reduction in gelsolin are all tumorigenic in nude mice, whereas I. Janmey, P. A., and Chaponnier, C. Medical aspects of the actin cytoskeleton. Curr. SKBR3 cells are not (21). 184A1N4TH cells, which were transformed Opin. Cell Biol., 7: 111—117,1995. by the introduction of two oncogenes, an activated H-ras and SV4O 2. Theriot,J. A. Regulationof the actincytoskeletonin living cells.Semin.Cell Biol., 5: 193—199,1994. T-antigen, had no major change in gelsolin expression yet are also 3. Iwig, M., Czeslick,E.,Muller, A., Gruner,M., Spindler,M., andGlaesser,D.Growth tumorigenic in nude mice (22). However, these oncogenes are rarely regulation by cell shape alteration and organization of cytoskeleton. Eur. J. Cell Biol., or never, respectively, involved in human breast cancers (23); there 67: 145—157,1995. 4. Boonstra, J., Rijken, P., Humbel, B., Cremers, F., Verkleij, A., and van Bergen en fore, their pathway to mammary tumorigenesis may differ substan Henegouwen, P. The epidermal growth factor. Cell Biol. tnt., 19: 413—430,1995. tially from that of “naturallyoccurring―humanbreast cancers. 5. Ridley, A. J. Membrane ruffling and signal transduction. BioEssays, 16: 321—327, At least two different mechanisms may be responsible for the AF 1994. 6. Prendergast, 0. C., and Gibbs, J. B. Pathways of ras function: connections to the actin changes found in the breast cancer cell lines. For three of the lines cytoskeleton. Adv. Cancer Res., 62: 19—64,1993. (MCF7, T47D, and SKBR3), Western blotting for total actin (monomeric 7. Gips, S. J., Kandzari, D. E., and Goldschmidt-Clermont, P. J. Growth factor receptors, phospholipases, phospholipid kinases and actin reorganization. Semin. Cell Biol., 5: plus filamentous) yielded lower acm band intensity, suggesting that the 201—208,1994. total pool of actin available for polymerization may be diminished. In 8. Plopper, G. E., McNamee, H. P., Dike, L. E., Bojanowski, K., and Ingber, D. E. contrast, the abundant total actin bands found in MDA-MB-23l and Convergence of integrin and growth factor receptor signaling pathways within the focal adhesion complex. Mol. Biol. Cell, 6: 1349—1365, 1995. 184A1N4Th cells suggest that these tumor cells may have defects that 9. Zigmond, S. Signal transduction and actin filament organization. Cuff. Opin. Cell interfere with polymerization or promote depolymerization of AFs. Sev Biol., 8: 66—73,1996. cml actin-binding proteins are subject to alteration in malignant transfor 10. Yin, H. L. Gelsolin: calcium- and polyphosphoinositide-regulated modulating pro tein. Bioessays, 7: 176—179,1988. mation and might contribute to aberrant regulation of AFs (1); therefore, I 1. Mtlllauer, L., Fujita, H., Ishizaki, A., and Kuzumaki, N. Tumor-suppressive function the status of AFs in the cancer cells may or may not be related to of mutatedgelsolinin ras-transformedcells.Oncogene,8:2531—2536,1993. decreased gelsolin. On the one hand, less gelsolin, which severs AFs, 12. Vandekerckhove,J.,Bauw,G., Vancompernolle,K.,Honoré,B.,andCelis,J. Com parative two-dimensional gel analysis and microsequencing identifies gelsolin as one might result in increased numbers and/orlength ofAFs (20). On the other of the mostprominentdownregulatedmarkersof transformedhumanfibroblastand hand, gelsolin also acts in nucleation of AF polymerization, so its loss epithelial cells. J. Cell Biol., 111: 95—102,1990. 13. Moriya, S., Yanagihara,K.,Fujita, H., andKuzumaki,N. Differential expressionof might cause a reduction of AFs. HSP9O,gelsolinandGST-srinhumangastriccarcinomacelllines.mt.J. Oncol.,5: Gelsolin interacts with key components of several signal transduc 1347—1351,1994. tion pathways, and its loss might adversely affect their availability or 14. Tanaka, M., MUIlauer, L., Ogiso, Y., Fujita, H., Moriya, S., Furuuchi, K., Harabayashi, T., Shinohara, N., Koyanagi, T., and Kuzumaki, N. Gelsolin: a candi activity (1, 24). Calcium ions promote the binding of gelsolin to AFs, date for suppressor of human bladder cancer. Cancer Res., 55: 3228—3232, 1995. whereas phosphoinositides, in particular PIP2, are the only known 15. Chaponnier, C., and Gabbiani, G. Gelsolin modulation in epitheial and stromal cells physiological molecules that can dissociate gelsolin from actin (1). of mammarycarcinoma.Am. J. Pathol.,134: 597—603,1989. 16. Medina, D., Kittrell, F. S., Obom, C. J., and Schwartz, M. Growth factor dependency When gelsolin is bound to PIP2, the latter cannot be hydrolyzed by and gene expression in preneoplastic mouse mammary epithelial cells. Cancer Res., phospholipase Cy, thereby preventing generation of the second mes 53: 668—674,1993. 17. Asch, H. L., Mayhew,E., Lazo,R. 0., andAsch, B. B. Lipids covalentlyboundto sengers, diacylglycerol and inositol 1,4,5-trisphosphate. Moreover, keratins of mouse mammary epithelial cells. Biochem. Mol. Biol. Int., 29: 1161— recent data have shown that a physical association between gelsolin 1169, 1993. and PLD coordinately regulates AF reorganization and PLD signaling 18. Thompson, M. E., Jensen, R. A., Obermiller, P. 5., Page, D. L., and Holt, 1. T. Decreased expression of BRCA1 accelerates growth and is often present during (24). Gelsolin stimulates activity of PLD, but binding of nucleoside sporadic breast cancer progression. Nat. Genet., 9: 444—450, 1995. triphosphates to gelsolin prevents its interaction with the enzyme. 19. Zou, Z., Anisowicz, A., Hendrix, M. J. C., Thor, A., Neveu, M., Sheng, S., Rafidi, K., Decreased gelsolin might, therefore, result in less PLD activity and Seftor, E., and Sager, R. Maspin, a serpin with tumor-suppressing activity in human mammary epitheial cells. Science (Washington DC), 263: 526—529, 1994. consequently fewer AFs. The fact that many signaling proteins bind to 20. Witke,W., Sharpe,A.H., Hartwig,J.H., Azuma,T., Stossel,T.P.,andKwiatkowski, the actin cytoskeleton upon activation has led to the hypothesis that D. J. Hemostatic, inflammatory, and fibroblast responses are blunted in mice lacking gelsolin. Cell, 81: 41—51,1995. AFs form a matrix for aligning signal transduction components in 21. Thompson,E.W., Paik,S., Brunner,N., Sommers,C.L., Zugmaier,G., Clarke,R., proper sequence to optimize speed and targeting of signaling (4). Shima, T. B., Torn, J., Donahue, S., Lippman, M. E., Martin, G. R., and Dickson, Thus, perturbation of AFs due to alterations in gelsolin and/or other R. B. Association of increased basement membrane invasiveness with absence of estrogen receptor and expression of vimentin in human breast cancer cell lines. J. actin-binding proteins could produce aberrant signal transduction. Cell. Physiol., 150: 534—544,1992. In summary, our results demonstrate the widespread loss of gelsolin 22. Clark, R., Stampfer,M. R., Milley, R., O'Rourke, E., Walen, K., Kriegler, M., in breast cancers of humans and rodents. Insufficient gelsolin may Kopplin, J., and McCormick, F. Transformation of human mammary epithelial cells by oncogenic retroviruses. Cancer Res., 48: 4689—4694,1988. impair regulation of the AF cytoskeleton as well as derange various 23. Clark, G. J., and Der, C. J. Aberrantfunctionof the Ras signaltransductionpathway signal transduction pathways, such as those involving PIP2 or PLD in human breast cancer. Breast Cancer Res. Treat., 35: 133—144,1995. and those whose activated components bind to the AF cytoskeleton. 24. Steed,P. M., Nagar, S., and Wennogle,L. P. PhospholipaseD regulationby a physical interaction with the actin-binding protein gelsolin. Biochemistry, 35: 5229— Such aberrations could contribute to many growth and behavioral 5237, 1996. 4845

Downloaded from cancerres.aacrjournals.org on September 29, 2021. © 1996 American Association for Cancer Research. Widespread Loss of Gelsolin in Breast Cancers of Humans, Mice, and Rats

Harold L. Asch, Karen Head, Yan Dong, et al.

Cancer Res 1996;56:4841-4845.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/56/21/4841

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/56/21/4841. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.

Downloaded from cancerres.aacrjournals.org on September 29, 2021. © 1996 American Association for Cancer Research.