perspectives in cancer research · [cancer research 55, 1856-1862, may 1, 1995] perspectives in...

[CANCER RESEARCH 55, 1856-1862, May 1, 1995]

Perspectives in Cancer Research

Molecular Insights into Cancer Invasion: Strategies for Prevention and Intervention

Elise C. Kohn and Lance A. LiottaLaboratory of Pathology, Division of Cancer Biology, Diagnosis and Centers, National Cancer Institute, NÃŒH,Bethesda, Maryland 28092

Abstract

The diagnosis and treatment of solid tumors usually begins at a latestage when most patients already have occult or overt metastasis. Manyyears of cancer progression precede diagnosis of most solid tumors.Novel noncytotoxic therapeutics may be specially suited for administration during this interval. An important window of intervention canbe defined as the period during which transition from a hyperprolif-erative state to acquisition of the capacity for invasion and metastasisoccurs. Investigation of the molecular basis of invasion is uncoveringstrategies for delaying progression of preinvasive carcinoma and treatment of primary tumors and established metastasis. Although tumorcell invasion might not be rate limiting for the growth of metastasis,anti-invasive agents can block tumor angiogenesis and thereby indirectly block metastasis growth. Two classes of molecular anti-invasiontargets exist: (a) cell surface and extracellular proteins, which mediatesensing, adhesion, and proteolysis; and (hi signal transduction pathways, which regulate invasion, angiogenesis, and proliferation. Bothcategories of targets yield treatment approaches that are now beingtested in the clinic. Metalloproteinase inhibitors, such as BB94, arebased on the recognition that metalloproteinases play a necessary rolein invasion and angiogenesis. The orally active signal transductioninhibitor carboxyamidotriazole modulates non-voltage-gated calciuminflux-regulated signal pathways and reversibly inhibits tumor invasion, growth, and angiogenesis. Blockade of invasion, angiogenesis, orcellular signal pathways is likely to generate a cytostatic, rather than acytotoxic effect. Cytostatic therapy constitutes an alternative paradigmfor clinical translation that may complement conventional cytotoxictherapy. For patients with newly diagnosed solid tumors, long-termcytostatic therapy could potentially create a state of metastasis dormancy or delay the time to overt relapse following cytotoxic agent-induced remission. Clinical toxicity and pharmacology using oral cytostatic agents in phase I trials and in adjuvant settings will provide animportant foundation for the translation of this approach to the preinvasive carcinoma period.

Introduction

The overt manifestation and initial presentation of cancer usuallyoccur at a late stage in the disease process when the capacity forinvasion has already been unleashed. By the time of diagnosis, a highproportion of patients have occult or clinically detectable metastasis.The capacity of conventional cytotoxic approaches to succeed in theface of this advanced, accelerating disease has unfortunately beenlimited (1, 2). In contrast to the short time between disease presentation and established metastasis, the chemoprevention period (Fig. 1)may extend back 10 years or more (3-6). For breast cancer, the period

of transition from hyperproliferative but noninvasive disease (3, 5, 7,8) to invasive cancer is estimated to average 6 years (3, 5).

We have recently obtained direct genetic evidence for the previously unproved assumption that in situ breast cancer is a clonalprecursor of invasive carcinoma (7). Carcinoma in situ of the breastcan now be viewed as a field disease spawning the clonal expansionof hyperproliferating cells that permeate the ductal system (3, 5, 7, 8).The neoplasm may remain confined to the intraductal space untiladditional genetic progression events occur that instill the capacity forinvasion. It is assumed that dissemination is impossible until this step

has occurred. The time period after the invasive carcinoma grows toreach the minimum threshold size of detection (0.25 cm in diameter),but prior to establishment of the first metastasis, can be less than 1year (5, 8). Thus, incorporating the transition of preinvasive carcinoma to invasion provides a much larger window for screening andintervention compared to the conventional goal which focuses onestablished invasive carcinomas.

The critical pathological turning point that defines this new prevention target is the initiation of local invasion leading to the dissemination of tumor cells. Tumor-induced neovascularization occurs in

parallel with the transition to invasion and provides a vascular entryportal for dissemination which may precede evident primary tumoroutgrowth by many years (9-11) (Fig. 1). Circulating tumor cells can

be detected in patients who never succumb to metastasis (9). Thisclinical observation is in keeping with experimental studies indicatingthat metastasis initiation from circulating tumor cells is a low probability event. Less than 0.05% of circulating tumor cells are successful(10, 12). The first metastasis is established only after a sufficientlylarge number of tumor cells containing the necessary genotypic andphenotypic changes have entered the tumor venous and lymphaticdrainage (12, 13).

Agents that either block angiogenesis or retard invasion can offeran alternative approach for arresting neoplastic progression at thestage of hyperproliferative noninvasive cancer. Using the example ofbreast cancer, in situ carcinoma remains unvascularized, is strictlylimited in size, and is separated from the stromal lymphatics by abasement membrane barrier (14). Thus, this stage of cancer may beeffectively benign. Furthermore, since the incidence of in situ carcinoma decreases in later ages (5, 8), we can hypothesize that thismultifocal ductal lesion may ultimately undergo regression if progression to invasive cancer does not occur. Preventing the malignantevolution of incipient in situ carcinoma may be just as clinicallyeffective as blocking neoplastic progression at an earlier stage.

Molecular Checkpoints That Regulate Invasion

The molecular characterization of invasion and angiogenesis hasled to the identification of two categories of checkpoints that constitute intervention targets. The first category encompasses cell surfaceand secreted proteins. The second category is defined as regulatoryproteins and pathways inside the cell. Positive and negative elementsplay a role for both categories of checkpoints (12, 15, 16). Moleculesfalling under the first category include cell surface and secretedproteins such as adhesion receptors, degradative enzymes and theirinhibitors, and motility-stimulating cytokines (17-22). The second

checkpoint category includes regulatory proteins and pathways suchas calcium-mediated signaling, G-protein activation, and tyrosinephosphorylation events (23-26). Examination of the structure and

function of checkpoint molecules provides the basis for developmentof agents that can potentially block tumor invasion or growth.

Cancer Invasion: Spatial and Temporal Regulation at theCellular Level

Received 12/30/94; accepted 2/23/95.

Invasion is the active translocation of neoplastic cells across tissueboundaries and through host cellular and extracellular matrix barriers.

1856

Association for Cancer Research. by guest on August 31, 2020. Copyright 1995 Americanhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from

https://bloodcancerdiscov.aacrjournals.org

MOLECULAR INSIGHTS INTO CANCER INVASION

DIAGNOSIS PERIOD

INHERITEDRISK

CARCINOMA IN SITU

ANGIOGENESIS

Fig. 1. Time windows for intervention versus treatment. Diagnosis and treatment ofcancer usually are performed late in the disease course, when a high proportion of patientsalready have established metastasis. In contrast, progression through dysplasia to carcinoma in situ and the acquisition of the invasive and angiogcnic phenotype may extendback over a period of 5-10 years and offer a window for intervention prior to metastasis

onset.

Invasion is not simply due to growth pressure but involves additionalgenetic deregulation over and above those molecular events that causeuncontrolled proliferation. Expression of the malignant phenotype isprobably not caused by one single gene or protein. Instead, invasionresults from an imbalance and deregulation of what are normallystimulatory and inhibitory events. This is especially important in thatall of the components of invasion play important roles in tightlycontrolled normal physiological events. At the biochemical level, themechanism of invasion used by tumor cells, may parallel, or besimilar to, that used by nonmalignant cells, which traverse tissueboundaries under normal physiological conditions (16).

Examples of physiological invasion are smooth muscle cell migration from the media to the intima (27), angiogenesis (11, 28), embry-

ogenesis and morphogenesis (29), nerve growth cone extension andhoming (30), and trophoblast implantation (31). In contrast to malignant invasion, physiological invasion is tightly regulated and ceaseswhen the stimulus is removed. Angiogenesis will cease when thesource of the angiogenic factor is removed (11). Neurite migration andinvasion is arrested when the neuntes sense destination arrival cues(30). Differentiation of trophoblasts terminates their physiologicalinvasion (31). In counterdistinction, invading tumor cells appear tohave lost the control mechanisms which prevent normal cells frominvading neighboring tissue at inappropriate times and places (16, 32).

If one assumes that the malignant tumor cell is inappropriatelyexpressing a pre-existing normal cell program of physiological inva

sion, then the fundamental difference between normal and malignantcells is regulation. The difference must lie in the proteins that start,stop, or maintain the invasion program at times and places that areinappropriate for nonmalignant cells (12, 16, 32). A major goal,therefore, is to understand what signals and signal transduction pathways are perpetually activated or unrestrained in malignant invasioncompared to physiological invasion.

Cellular Invasion: Action and Reaction at the Pseudopod

Regulation of molecular events necessary for invasion, malignantor physiological, requires spatial and temporal coordination and cyclicon-off processes at the level of the individual cell. Cellular invasion is

dependent on the coordinated activity of a series of interacting proteins extending from inside the cell, through the plasma membrane, tothe cell surface and pericellular microenvironment (Fig. 2). For cellsmigrating within a three dimensional matrix, such as penetrating abasement membrane, protrusion of a cylindrical pseudopod is the firststep, prior to translocation of the whole cell body (33, 34). Pseudopo-

dial cell fragments, devoid of nuclei, retain pertinent sensing anddirectional locomotion capacity (33, 34), proving that these organs oftranslocation have all the sensory and motor equipment necessary forcrawling.

A number of observations support the central role of cytoskeleton-

driven pseudopodia as organs of motility and invasion. Pseudopodiaprotrude at the point of surface stimulation by appropriate ligandswhich cause migration (33-35), and they appear to aggregate or

concentrate cell surface degradative enzymes and adhesion receptors(35). Pseudopods have an internal supporting actin structure thatcauses the pseudopod to extend into free space at the confined area ofligand binding (34, 36).

A series of proteins on the surface of the pseudopod coordinatesensing, protrusion, burrowing, and traction. Proteinases at the pseudopod tip may locally disrupt the extracellular matrix and permitforward extention. However, the balance must switch from proteolysisto adhesion in order for the advancing pseudopod to grip the matrixand pull the cell forward. General unregulated proteolysis alonecannot be responsible for the entire invasion cascade because it wouldremove the substratum necessary for proper traction. The cell would,in effect, dig itself into a hole and become immobilized. Consequently, to achieve forward locomotion, the invading cell coupleslocal proteolysis (16) with coordinated and temporally limited attachment and deattachment (37). At the leading edge of the cell, protein-

ases bound to activator proteins and receptors may be colocalized onthe advancing tip of the pseudopodia. When the cell moves into thezone of lysis, adhesion is required and proteolysis must be shut down(37, 38). At the rear of the cell, dissociation from adjacent cells, anddetachment (18, 19) from previous attachment sites are necessary torelease the cell (Fig. 2).

Regulation of Invasion Extends to the Cell-Matrix Interface

The basement membrane and interstitial stroma play an importantregulatory role in malignant as well as physiological invasion. Basement membranes constitute barriers that confine the movement ofcells to specific tissue compartments. The basement membrane may

Fig. 2. Spatial and temporal regulation of cell invasion of the extracellular matrix.Physiological or malignant invasion can be viewed as motility coupled to regulatedadhesion and proteolysis. The advancing pseudopod, extended through calcium-regulated

actin polymerization, may focus the action of cell surface proteinases, enzymes, receptors,and activators. Extracellular matrix degradation must be balanced by antiproteolysis toallow for adhesive traction. Signal transduction pathways cycle the individual invadingcell through pseudopod protrusion, proteolysis, antiproteolysis, adhesion, and detachment.

1857




also be a storage depot for latent proteinases and cytokines includingangiogenesis factors, which can be activated or released by the invading cell pseudopodia. Abnormal, dysfunctional, or fragmentedbasement membranes accompany physiological as well as malignantinvasion. Basement membranes may be defective due to altered assembly or disruption by augmented local activity of matrix-degrading

proteinases. It is now well established that one way the integrity of thebasement membrane can be regulated is by the balance betweenmetalloproteinases and metalloproteinase inhibitors (14, 20, 29). Under this hypothesis, agents or treatments that maintain the integrity ofthe epithelial basement membrane may be an alternative approach toretard invasion or angiogenesis. For example, carcinoma in situ lesions of the breast, if induced to remain confined within the ductalepithelial basement membrane, may remain in a dormant state, unableto become vascularized, and unable to disseminate.

Anti-invasion Defines a New Category of Cancer

Chemoprevention

Arresting or retarding molecular events involved in tumor invasionand angiogenesis defines a new category of Chemoprevention agents.Overlapping regulation of growth and invasion occurs at the cellularand extracellular level. Consequently, anti-invasion agents may also

be cytostatic (23, 26). Chemoprevention has traditionally been definedas the use of one or several chemical agents to prevent the occurrenceof cancer (4, 6, 23). Chemoprevention agents have been placed intotwo major categories, blocking agents and suppressing agents (4, 6).Blocking agents have a barrier function to prevent cancer-producing

substances from reaching or reacting with cellular target sites. Giventhe time at which initiation and promotion events occur in stem cells,blocking agents, in order to be effective, may need to be administeredat an early age much before adulthood. According to the report ofWattenberg (6), suppressing agents act after exposure to cancer-

producing compounds and prevent the evolution of the neoplasticprocess in cells that otherwise would become malignant.

In theory, some suppressing agents stimulate differentiation, whileothers reverse the consequences of activated oncogenes and suppressor genes. For example, retinoids and vitamin D compounds have thecapacity to induce differentiation, and H-ras oncogene action can be

blocked by monoterpenes or isoprenylation inhibitors (25). Anothergroup of suppressing agents selectively inhibits proliferation of potentially malignant cells (6). Under this group, hormone antagonistsand hormone metabolism inhibitors are suppressing agents that retardcell proliferation.

The mission previously posed for classical Chemoprevention agentsdid not encompass the full spectrum of malignant progression. Thetransition period from the hyperproliferative state to the emergence oftumor cells with lethal capacity can occupy a major fraction of theChemoprevention window. This affords a potentially safer point oftherapeutic entry in late adulthood compared to high risk treatment atearlier ages, which theoretically would be required to block initiationof transformation.

Matrix Metalloproteinases: From Correlation to ViableTherapeutic Target

A positive correlation between tumor aggressiveness and proteaselevels has been documented for all four classes of proteases includingserine, aspartyl, cysteinyl, and metal atom dependent (20, 38-47). In

addition to proteolytic activity, augmented heparanase activity hasalso been associated with malignancy (45). All of these classes ofenzymes may be equally important at one or more stages of invasion.Nevertheless, progress made in the field of metalloproteinases servesas an illustrative example of how specific inhibitors can be used to

simultaneously test the proteinase hypothesis and explore therapeuticstrategies.

Significant evidence has accumulated to directly implicate members of the gene family of matrix metalloproteinases in tumor invasionand metastasis formation. These enzymes, each of which is secreted asa proenzyme that requires activation, are divided into three generalsubclasses (48-50): (a) interstitial collagenase; (b) stromelysins; and

(c) gelatinases (type IV collagenases). Interstitial collagenase degrades type I collagen, as well as the fibrillar collagens II, III, and X.The stromelysins degrade proteoglycan core protein, laminin, fi-

bronectin, gelatin, and the nonhelical portions of basement membranecollagens. Stromelysin 3 is associated with human breast cancerstremai cells, but its substrate specificity has yet to be defined.Matrilysin, but not Stromelysin 1 or Stromelysin 2, is prominent inhuman gastric and colonie carcinomas. In epidermoid carcinomas ofthe head and neck, Stromelysin 2 as well as interstitial collagenasemRNAs were detected in invasive cancer cells and associated stromalcells, whereas adjacent normal tissues were negative. Furthermore,Stromelysin 2 mRNA was localized in tumor cells arranged alongdisrupted basement membranes. These studies suggest that there maybe organ or cell type specificity associated with the up-regulation of

proteolytic activity during malignant conversion.The third group of enzymes of the MMP1 gene family are the type

IV collagenases. The M, 72,000 and MT 92,000 enzymes (MMP-2,MMP-9) were orginally named for their ability to degrade pepsinized,triple-helical type IV collagen. In addition, they possess potent ge-

latinolytic activity and can degrade collagen types V, VII, IX, and X;fibronectin; and elastin. The two gelatinase enzymes arise from separate mRNA transcripts on separate genes. They are distinct fromother members of the matrix metalloproteinases in that they possessa unique region immediately adjacent to the metal-binding domainthat is homologous to the gelatin-binding domain of fibronectin.

Gelatinases A and B differ from other metalloproteinases by theirability to interact, as latent proenzymes, with the endogenous TIMPs(TIMP-1 and TIMP-2). The gelatinase A proenzyme forms a complexwith TIMP-2, and the Mr 92,000 gelatinase B proenzyme binds withTIMP-1 (49-55). TIMPs are not known to bind other latent matrix

metalloproteinases but will inhibit all matrix metalloproteinases oncethese enzymes are activated. Thus, TIMP-2 preferentially recognizes

the latent form of gelatinase A.Correlative evidence for the involvement of gelatinase A and gel

atinase B in the invasive phenotype is abundant (56-59). Induction ofthe malignant phenotype using the H-ras oncogene has been shown toenhance expression of gelatinases A and B (56). Most invasive co-Ionic, gastric, ovarian, and thyroid adenocarcinomas or the desmo-plastic stroma of breast lesions have been shown to be immunoreac-

tive for gelatinase A, whereas benign proliferative disorders of thebreast and colon and normal colorectal and gastric mucosa and benignovarian cysts show decreased or negative staining for this enzyme (57,58). However, gelatinase A was also detected in benign disorders inwhich the tissue was undergoing remodeling and repair such as ininflammation, fibrosis, and distortion of normal follicles. In contrast,normal thyroid, goiter, and Graves disease tissue showed little or noimmunoreactivity.

Connective tissue destruction by metalloproteinases is largely irreversible. Consequently, selective inhibitors of these enzymes offerimportant tools with which to study the role of MMPs in physiologicaland malignant invasion (59-64). MMP inhibitors also constituteviable strategies for disease intervention. Structure-function informa

tion about MMP enzyme domains, activation mechanisms, and the

1The abbreviations used are: MMP, matrix metaltoproteinase; TIMP, tissue inhibitor

of metalloproteinases; CAI, carboxyamidotriazole.

1858




interaction with TIMPs has led to a number of successful inhibitorstrategies. One approach is to use TIMPs or TIMP fragments directlyas selective inhibitors of MMP activation or activity (46-53, 59-63).A second strategy is peptide inhibitors, which mimic the amino-terminal MMP motif that maintains the latent enzyme state (64-67).A third example is synthetic compounds, which compete for thesubstrate or bind to the active site (64, 68-70).

In keeping with the concept of parallel mechanisms for physiological and malignant invasion, all three types of metalloproteinaseinhibitors have been shown to inhibit physiological invasion models.TIMP-1 and TIMP-2 block trophoblast invasion. TIMP-1, TIMP-2,and synthetic substrate inhibitors inhibit angiogenesis in vivo (63, 64,70). Gelatinase A peptide inhibitors and synthetic substrate inhibitorsblock neurite outgrowth (69). Synthetic substrate inhibitors also havebeen shown to prevent smooth muscle cell migration from the vascular intima to the media following arterial balloon injury (27).Collectively, these data support a role for collagenolytic activity in atleast two functional processes contributing to metastasis. That is,collagenases are involved in tumor cell invasion as well as neovas-cularization, upon which solid tumor growth is dependent. These areimportant considerations in the design of proteolytic inhibitors forpotential use as therapeutic agents.

Native or recombinant TIMP-1 has been shown to inhibit in vitroinvasion of human amniotic membranes and in vivo metastasis inanimal models (47). Furthermore, transfection of antisense TIMP-1RNA into mouse 3T3 cells, which down-regulates TIMP-1 expression, enhances their ability to invade human amnionic membranes andto form metastatic tumors in athymic mice (60). TIMP-2 is a M,21,000, nonglycosylated protein that shows 37% identity and 65.6%overall homology to TIMP-1, yet the two proteins are immunologi-cally distinct and are encoded by genes on separate chromosomes(51-53). Exogenous addition of TIMP-2 has also been shown tosuccessfully inhibit in vitro tumor cell invasion of extracellular matrices (61, 62). Overexpression of TIMP-2 in invasive and metastaticras-transformed rat embryo fibroblasts suppressed the formation oflung mÃ©tastasesfollowing i.v. injection in nude mice (62).

In vitro studies of the mechanism of activation of MMPs enzymeshas yielded new insights into the molecular basis of proenzymelatency. Activation is produced by the loss of an 80-residue NH2-terminal domain containing the sequence PRCGXPDV, which ishighly conserved among family members (65-67). The latency of themetalloproteinases appears to be maintained by a specific metalatom-sulfhydryl side chain interaction with the cystinyl residue. Thisautoinhibitory consensus peptide has been used to develop a family ofinhibitor peptides, some of which have a 50% inhibitory concentrationof 3 /UM.These inhibitors require the single cysteinyl residue to beeffective in blocking the enzyme or halting invasion (65-67). Thecysteinyl-containing peptide prevented more than 80% of the invasionof reconstituted extracellular matrix by human fibrosarcoma, breastcarcinoma, or melanoma cells. Although effective and specific, theclinical utility of these cysteinyl peptide inhibitors is currently limiteduntil orally active derivatives are available.

The largest current group of synthetic metalloproteinase inhibitorsconsists of the collagen substrate analogues. All are less than sixamino acids long (64). Up to 3 of the amino acids are placed on 1 sideof the collagen scissle bond, and the peptide is linked to a Zn2+binding moiety. This design targets the Zn2+ atom coordinated withinthe active site of all MMP molecules. A variety of Zn2+-dictating

moieties have been used, among them hydroxamic acid, thiols, andcarboxyalkyl groups. It is theorized that these compounds act bybinding to the active site of the metalloproteinase. An example compound is BB94 (Batimastat) (64, 68, 71). BB94 is active in vitro witha 50% inhibitory concentration of 20 nMfor stromelysin and 4 nMfor

gelatinase A. Administration of BB94 i.p. to athymic nu/nu micebearing fragments of human colon carcinoma inhibited the growth ofthe primary tumor, caused a marked reduction in the incidence oftumor invasion of local tissue, and reduced the incidence of spontaneous metastasis (68, 71). In a separate study, BB94 significantlydelayed the growth of transplanted B16-BL6 primary melanoma andreduced the size of established metastasis. As a test of the theory thatmetalloproteinases may play a role in peritoneal invasion of ovariancancer (71), BB94 was shown to decrease tumor burden and prolongsurvival of mice bearing human ovarian carcinoma xenografts. Theefficacy seen in these preclinical studies supports the basic hypothesisthat metalloproteinases are important effectors of invasion and metastasis. The suppression of tumor growth may be indirectly linked toan effect on angiogenesis (16, 63, 70). Furthermore, a component oftumor expansion may be dependent on the invasion of adjancenttissue. This strong rationale, combined with promising preclinicaldata, has been the basis for ongoing clinical trials exploring the use ofBB94 in patients with advanced cancer.

Calcium Homeostasis Modulators Block Invasion, Growth,and Angiogenesis

The second molecular checkpoint category contains the regulatoryproteins and signaling pathways of the cell. Cellular communicationcircuits can be modulated to produce a positive or negative downstream action depending on the location and type of intervention.Examples are differentiating agents, enzyme modulators, and directsignal transduction effectors (72-77). Targeting signal transductionpathways provides a novel approach to down-regulate the invasiveprocess, and one that may potentially allow earlier intervention duringthe invasion prevention window described above. The accumulationof molecular events during cancer progression changes the signalinghomeostasis of cancer cells through mutation, altered expression, andthe development of autocrine loops upon which the cells depend. Thegoal of signal transduction therapy is to suppress hyperactive oraberrantly active signaling pathways of malignant proliferation andinvasion. This class of agents may therefore be expected to be cyto-static rather than cytotoxic. The redundancy in signaling pathways innonmalignant cells can theoretically create checks and balances thatwould protect against normal cell toxicity during continuous exposureto signal modulatory drugs (23, 26, 78-84). In contrast, classicalchemotherapy agents, which can inhibit invasion and metastasis bytheir cytotoxic activity, target DNA synthesis, repair, and transcrip-tional events. The frequent global toxicity patterns of cytotoxic agentsis a consequence, in part, of the central action at the DNA level.

There are several signaling pathways to which drug developmentmay be directed. These include nuclear events, such as transcription;however, drug delivery may be a problem. It may be difficult toselectively target nuclear events either in directing the drug to its siteof action or to direct it to the specific event in question. For example,the AP-1 transcription factor is important in the process of invasionand metastasis by its activation of transcription of stromelysin andgelatinase B (39, 40, 77), as well as being important in signalingevents of proliferation. Targeting of AP-1 as a mechanism by whichto intervene therapeutically will require not only delivering the agentto AP-1 but also creating a level of selectivity. Since signals originating peripherally in the cell may route through several subsequentpathways, peripheral signal targets may allow for more selectivity byallowing normal checks and balances within the signaling homeostasis of the cell to intervene (23, 24, Â¿6,78-80, 83-87). Signalingevents in the cytoplasm or at the nucleus, such as post-translationalmodifications and transmembrane signaling, may provide more accessible checkpoints. Transmembrane or cytoplasmic signaling events

1859




proposed for drug intervention have included inhibition of post-

translational modifications such as isoprenylation (25, 88), modulation of protein kinase activity (76), and alteration in ion homeostasis,such as calcium influx (23, 26, 75, 81).

Altered signaling pathways have been tied to the process of invasion and metastasis. A prime example is the introduction of activatedH-ra.i into primary rat embryo fibroblast cells. This transfection (89)

resulted in development of a malignant and invasive phenotype,characterized by production of gelatinase B and marked pulmonarycolonization after tail vein inoculation of transfected cells. The invasive phenotype could be differentiated from the tumorigenic phenotype by further transfection with adenovirus Eia. Similar effects havebeen observed with expression of other oncogenes, such as Her-2/neu(90). When Eia was transfected into new-transformed NIH-3T3 cells,it abrogated the metastasis-associated properties imparted by neu andalso its transforming effects (90). Re-expression of neu resulted in

return of transforming potential without metastatic capabilities. Thesestudies reinforce the overlap between oncogene-induced tumor devel

opment and dissemination but underscore that distinct signals must bein place for the functional separation of transforming activity andmetastasis to occur. The invasive phenotype may be regulated bysignal transduction pathways downstream from activated oncogenes.

Intracellular calcium is an important regulator of subsets of thethree transmembrane signaling categories: (a) ion fluxes; (b) phos-phorylation events; and (c) guanine nucleotide (G)-binding protein-mediated second messenger production (26, 72, 85-87, 91). Intracel

lular calcium may be mobilized from within the cell fromendoplasmic reticulum stores released in response to inositol trisphos-phate, or it may enter the cell through voltage-gated or non-voltage-gated calcium channels (24, 78, 79, 81, 83-87, 91). Most epithelial

cells, parental cells of carcinomas, are electrically neutral and do nothave voltage-gated calcium channels as their primary method ofcalcium entry. This defines non-voltage-gated calcium influx as the

primary regulator of cellular signaling pathways. We have identifiedan inhibitor of non-voltage-gated calcium pathways, CAI, which wehave shown to inhibit tumor cell proliferation, invasion, and angio-genesis both in vitro and in vivo (26, 78-84). The experimental

efficacy of CAI suggests that this mode of intervention of the signaling ofinvasion may be a viable therapeutic option and will be discussed as aexample of this therapeutic checkpoint target of invasion.

Early studies of CAI demonstrated that it inhibited non-voltage-

gated calcium influx in response to several agonists (83, 84) in theconcentration range of 1-10 /LIM.The CAI-sensitive calcium influx

events have been linked to downstream signaling pathways includingphospholipase A2, calcium-sensitive phospholipase Cy, and proteintyrosine kinase activity (81, 83, 84).2 Modulation of these signaling

pathways by cellular preincubation with this concentration range ofCAI resulted in inhibition of the component parts of invasion. CAIexposure inhibited tumor cell migration (78) in response to autocrinemotility factor (22), a G protein-mediated event, and type IV collagen,

shown to involve calcium mobilization (92); adhesion to tissue cultureplastic and type IV collagen has also been demonstrated (78). Cellexposure to CAI inhibits production of gelatinase A by reduction ingene expression (80). Coupled with inhibition of proliferation of awide variety of malignant human cell types in vitro (79, 81), CAIcould thus inhibit both tumor growth and invasive potential ofmalignant cells.

Angiogenesis, which may be permissive for the progression fromcarcinoma in situ to invasive carcinoma, is a form of regulated orphysiological invasion requiring the same components of adhesion,

: Unpublished data

proteolysis, and migration as do malignant cells. Human endothelialcell incubation with CAI resulted in the same inhibition of invasion aswas seen for malignant cells, i.e., reduction in adhesion and inmigration in response to an array of extracellular matrix componentproteins and production of gelatinase A (82). This inhibitory activitywas confirmed in vivo and all effects, in vitro and in vivo, occurred inthe same concentration range as the inhibition of malignant invasionand the inhibition of signaling pathways.

Preclinical studies of CAI in mice demonstrated oral bioavailabilityand the ability to attain targeted signal inhibitory plasma concentrations without marked toxicity. Daily administration of CAI to humanxenograft-bearing mice resulted in reduction in total tumor burden,

tumor incidence, and metastatic dissemination (79). A similar markedreduction in experimental mÃ©tastaseswas observed with cellular pretreatment in vitro prior to tail vein inoculation or p.o. administrationof CAI prior to or after tail vein inoculation with CAI-naive cells (79).

Only minor gross or histological toxicity has been observed with p.o.dosing of CAI in preclinical studies. A phase I clinical trial of CAI inpatients with refractory tumors has demonstrated the ability to attainplasma concentrations in the range that inhibits signaling and invasionin vitro, 1-10 /AM, and shows an acceptable pattern of toxicity and

promising evidence of cytostatic clinical activity. Phase III trials arein the planning stage.

Cytostatic Therapy: A New Paradigm

Identification of the molecular basis of tumor dissemination isdriving drug development to discover or synthesize inhibitors thatblock key pathways in this process. Many of the checkpoints involvedin cellular invasion have been shown to regulate angiogenesis (16),suggesting a dual therapeutic role for these inhibitors. Drugs identifiedto date, including metalloproteinase inhibitors (64) and signalinginhibitors (23, 26), have been shown to be reversible inhibitors of bothinvasion and proliferation. In fact, the general class of signal transduction inhibitors is expected to produce a cytostatic rather than acytotoxic action.

The current cytotoxic treatment paradigm is aimed at identifyingand applying drugs that cause a measurable reduction in tumor mass.This class of chemotherapeutic agents also has broad toxicity, particularly for rapidly growing normal tissues. Moreover, the reduction intumor bulk produced by cytotoxic chemotherapy is all too oftenassociated with tumor regrowth (1, 2), followed by repeat cycles ofchemotherapy and regrowth, ultimately leading to a lethal outcome.While effective in reducing tumor burden, classical chemotherapy hasproduced true cures in only a small percentage of patients withmetastatic cancer.

The goal of the cytostatic paradigm of intervention is to holdmetastastic dissemination at bay, slow or halt proliferation, and generate a period of disease stabilization (Fig. 3). For patients withestablished systemic metastasis, the goal of chronic treatment wouldbe to place, and keep, the metastatic colonies in a state of dormancy.Over time it is possible that the high death rate of tumor cells,especially nonhematological malignancies, may overbalance the arrested growth yielding slow regression. Disease stabilization or slowregression as an end point is contrary to current standards of clinicalinvestigation, which seek immediate tumor mass reductions. Consequently, under the paradigm of cytostasis, new end points of clinicalefficacy will have to be developed, such as time to progression,quality of life, and survival. In addition, it will be important to developnew methods to monitor tumor cytostasis in the patient.

Because chronic treatment over long time periods is required forcytostatic agents, they must be preferably of a formulation that isreadily administered and well tolerated, and therefore not compliance

1860




RELAPSE

CYTOTOXIC

CHEMOTHERAPY REGRESSION

+

VERSUSRx

REGRESSIONCOMBINATION

CYTOTOXIC THEN

CYTOSTATIC

HYPOTHESIS: CYTOSTATIC THERAPY:-DISEASE STABILIZATION?

-IMPROVED QUALITY OF LIFE'

-PROLONGATION OF LIFE'

-SUPPRESSION OF DRUG RESISTANCE?

-CHEMOPREVENTION?

Fig. 3. Cancer treatment paradigms. Anti-invasion therapy, anti-angiogenesis, signal

transduction therapy, abrogation of the action of ocogenes, and normalization of unrestrained growth pathways are all likely to be cytostatic rather than directly cytotoxic. Thisrequires a redefinition of the basic premises underlying clinical trials (1, 2, 78, 79. 81).Cytotoxic therapy induced regression is often followed by relapse (top). In contrast, cytostatictherapy must be administered over a long period of time (middle}. Combination of these twomodalities can potentially extend the time to recurrence and prolong survival (bottom).

limiting. If this criterion is met, the potential benefit for advanceddisease is a reduction in the requirement for cytotoxic therapy, maintenance or reduction in the use of pain medication, stabilization orimprovement in performance status, and ultimately delayed time toprogression and extended survival. Indeed, improved quality of lifealone may be an important criteria for efficacy of novel cytostaticagents (1). The trial agents discussed here and in the literature willconstitute the first generation of invasion inhibitors and cytostaticagents. Toxicity and pharmacology in phase I trials and in adjuvantsettings will provide an important foundation to move from thetreatment arena to the invasion prevention period (Fig. 1).

Clinical cohorts for whom invasion prevention approaches might beapplicable include patients with genetic risks of invasive cancers (4,15). These cancers frequently begin the hyperproliferation period atearly ages, and the clinical task of preventing tumor initiation andpromotion in the setting of a defective germ line may prove difficult.Instead this population could be targeted for intervention during theprogression period. The goal would be detection and managementprior to invasion and metastasis. If successful, cytostatic therapy forpatients genetically at high risk could provide an adjunct to surgicalextirpation, perhaps minimizing the extent of surgery required forthese patients with broad tissue/organ fields at risk. Information gleanedfrom these obligate cancer cohorts can then be used to design subsequentclinical investigations of cytostatic invasion inhibitor therapies.

Conclusion

The recognition that the process of malignant invasion is one ofderegulated physiological invasion and is a dynamic process of temporal and spatially defined molecular events opens a new arena fortherapeutic gain. Since physiological invasive processes are strictlydelimited and under tight control, intervention directed against ma

lignant invasion may be feasible, specific, and easily targeted. Cytostatic agents targeted to communication pathways that regulate proliferation or key proteins involved in invasion may be a way tointroduce regulation into the molecularly rigid malignant cell and putbrakes against its otherwise inexorable progression.

References1. Astrow, A. B. Commentary: rethinking cancer. Lancet, 343: 494-495, 1994.2. Bailar, J. C, and Smith. E. Progress against cancer? N. Engl. J. Med., 314: 1226-

1232, 1986.3. Gallager, H. S., and Martin, J. E. The study of mammary carcinoma by mammography

and whole organ sectioning, early observations. Cancer (Phila.), 23: 855-866, 1969.4. Kelloff, G. J., Boone, C. W., Steele, V., et al. Progress in cancer chemoprevention:

perspectives on agent selection and short-term clinical intervention trials. CancerRes., 54: 2015-2024, 1994.

5. Spratt, J. S., Greenberg, R. A., and Heuser, L. S. Geometry, growth rates, and durationof cancer and carcinoma in .Â¥//Â«of the breast before detection by screening. CancerRes.. 46: 970-974, 1986.

6. Wallenberg, L. W. Prevention therapy: basic science and the resolution of the cancerproblemâ€”presidential address. Cancer Res., 5.ÃŽ:5890-5896, 1993.

7. Zhuang, Z., Merino, M., Chuaqui, R., Liotta, L. A., and Emmert-Buck, M. Identical

allelic loss on chromosome 1Iql3 in microdissected in situ and invasive human breastcancer. Cancer Res., 55: 467-471, 1995.

8. Bedwani, R., Vana, J., Rosner, D., et al. Management and survival of female patientswith minimal breast cancer. Cancer (Phila.), 47: 2769-2778. 1981.

9. Melamed, M. Circulating cancer cells. In: L. G. Koss (ed.), Diagnostic Cytology andIts Histopathologic Bases, Ed. 4, Vol. 2, pp. 1403-1419. Philadelphia: J. B. LippincottCompany, 1992.

10. Liotta, L. A., Kleinerman, J., and Saidel, G. Quantitative relationships of intravasculartumor cells: tumor vessels and pulmonary mÃ©tastasesfollowing tumor implantation.Cancer Res., 34: 997-1003, 1974.

11. Folkman, J. Tumor angiogenesis: therapeutic implications. N. Engl. J. Med., 285:1182-1186, 1971.

12. Nicolson, G. L. Gene expression, cellular diversification and tumor progression to themetastatic phenotype. Bioessays, 13: 337-342, 1991.

13. Liolta, L. A., Tryggvason. K.. Garbisa, S., Gehron-Robey, P., and Abc, S. Partialpurification and characterization of a neutral proteinase which cleaves type IVcollagen. Biochemistry, 20: 100-104, 1981.

14. Barsky, S. H., Siegal. G. P., Jannotta, F., and Liotta L. A. Loss of basement membranecomponents by invasive tumors but not by their benign counterparts. Lab. Invest., 49:140-147, 1983.

15. Fearon, E. R., and Vogelstein, B. A genetic model for colorectal tumorigenesis. Cell,61: 759-767, 1990.

16. Liotta, L. A., Steeg, P. S., and Stetler-Stevenson. W. G. Cancer metastasis and angiogenesis; an imbalance of positive and negative regulation. Cell, 64: 327-336, 1991.

17. Hynes, R. O., and Lander, A. D. Contact and adhesive specificities in the associations,migrations, and targeting of cells and axons. Cell, 68: 303-322, 1992.

18. Seftor, R. E. B., Seftor. E. A., Gehlsen. K. R., Steller-Stevenson. W. G., Ruoslahti, E.,

and Hendrix, M. J. C. Role of the av03 integrin in human melanoma cell invasion.Proc. Nati. Acad. Sci. USA, 89: 1557-1561, 1992.

19. Frixen, U. H., Behrens, J., Sachs, M., el al. E-cadherin-mediated cell-cell adhesion prevents invasivenss of human colorectal carcinoma cells. J. Cell Biol., 113:173-185, 1991.

20. Liotta, L. A., Tryggvason, K., Garbisa, S., Hart, I., Foltz, C. M., and Shafie, S.Metastatic potential correlates with enzymatic degradation of basement membranecollagen. Nature (Lond.), 284: 67-68, 1980.

21. Rong, S., Sega], S., Anver, M., Resau, J. H., and Vande Woude, G. F. Invasivenessand metastasis of NIH 3T3 cells induced by Met-hepatocyte growth factor/scalierfactor autocrine stimulation. Proc. Nati. Acad. Sci. USA, 91: 4731-4735, 1994.

22. Stracke, M. L., Krutzsch, H. C, Unsworth, E. J., Ã¢restad.A., Cioce, V., Schiffmann, E.,and Liotta. L. A. Identification, purification, and partial sequence analysis of autotaxin, anovel motility-stimulating protein. J. Biol. Chem., 267: 2524-2529, 1992.

23. Powis, G. Signaling targets for anticancer drug development. Trends Pharmacol. Sci.,12: 188-194, 1991.

24. Irvine. R. F. Inositol phosphates and Ca2* entry: toward a proliferation or a simpli

fication, FASEB J., 6: 3085-3091, 1992.

25. Gibbs, J. B., Oliff, A., and Kohl, N. E. Farnesyltransferase inhibitors: ras researchyields a potential cancer therapeutic. Cell, 77: 175-178, 1994.

26. Cole, K. A., and Kohn, E. C. Calcium-mediated signal transduction: biology, biochemistry, and therapy. Cancer Metastasis Rev., 13: 33â€”41,1994.

27. Bendeck, M. P., Zempo, N., Clowes, A. W., and Reidy, M. A. Smooth muscle cellmigration and matrix metalloproteina.se expression after arterial injury in the rat. Circ.Res., 75: 539-545, 1994.

28. Weidner, N. Tumor angiogenesis: reveiw of current applications in tumor prognostication. Semin. Diagn. Pathol., 10: 302-313, 1993.

29. Talhouk, R. S., Bisell, M. J., and Werb, Z. Coordinated expression of extracellularmatrix degrading proteinascs and their inhibitors regulates mammary epithelial function during involution. J. Cell Biol., 118: 1271-1282, 1992.

30. Wu, D-Y.. and Goldberg, D. J. Regulated tyrosine phosphorylation at the tips ofgrowth cone filopodia. J. Cell Biol., 123: 653-664, 1993.

31. Lola, P. K., and Graham, C. H. Mechanisms of trophoblast Â¡nvasiveness and theircontrol: the role of proteases and protease inhibitors. Cancer Metastasis Rev., 9:369-380, 1990.

32. Schirrmacher, V. Cancer Metastasis: experimental approaches, theoretical concepts,

1861




and impacts for treatment strategies. Adv. Cancer Res., 43: 1-73, 1985.33. Condeelis, J. Life at the leading edge: the formation of cell protrusions. Annu. Rev.

Physiol., 9: 411-444, 1993.34. Stossel, T. P. On the crawling of animal cells. Science (Washington DC), 260:

1086-1094, 1993.35. Guirguis, R.. Margulies, I., Taraboletti, G., and Liotta, L. Cytokine-induced

pseudopodial protrusion is coupled to tumour cell migration. Nature (Lond.), 329:261-263, 1987.

36. Dong, C, Aznavoorian, S., and Liotta, L. A. Two phases of pseudopod protrusion intumor cells revealed by a micropipette. Microvasc. Res., 47: 55-67, 1994.

37. Damsky, C. H., and Werb, Z. Signal transduction by inlegrin receptors for extracellular matrix: cooperation and processing of extracellular information. Curr. Opin. CellBiol., 4: 772-781, 1992.

38. Blasi, F. Urokinase and urokinase receptor: a paracrine/autocrine system regulalingcell migration and invasiveness. Bioessays, 15: 105-111, 1993.

39. Sato. H., Takino. T., Yasunori, O., Cao, J., Shinagawa, E. Y.. and Seiki, M. A matrixmetalloproteinasc expressed on the surface of invasive tumor cells. Nature (Lond.),370: 61-65, 1994.

40. Sato, 11., Kita, M., and Seiki, M. v-Src activates the expression of 92-kDa type IVcollagenase gene through the AP-1 site and the GT box homologous to retinoblastomacontrol elements. J. Biol. Chem., 268: 23460-23468, 1993.

41. Axelrod, J. H., Reich, R., and Mishkin, R. Expression of human recombinantplasminogen activators enhances invasion and experimental metastasis of \\-ras-transformed NIH 3T3 cells. Mol. Cell. Biol., 9: 2133-2141, 1989.

42. Rochefort, H., Capony, F.. and Garcia, M. Cathepsin D: a protease involved in breastcancer metastasis. Cancer Metastasis Rev., 9: 321-331, 1990.

43. Witty, J. P., Mcdonnell, S., Newell, K. J., Cannon, P., Navre, M., Tressler, R. J., andMatrisian, L. M. Modulation of matrilysin levels in colon carcinoma cell lines affectstumorigenicity in vivo. Cancer Res., 54: 4805-4812. 1994.

44. Templeton, N. S., Brown, P. D., Levy, A. T.. Margulies, I. M. K., Liotta. L. A., andStetler-Stevenson, W. G. Cloning and characterization of human tumor cell interstitialcollagenase. Cancer Res., 50: 5431-5437, 1990.

45. Nakajima, M.. Morikawa, K., Fabra. A., el al. Influence of organ environment onextracellular matrix degradative activity and metastasis of human colon carcinomacells. J. Nati. Cancer Inst., 82: 1890-1898, 1990.

46. Mignatti, P.. Robbins, E., and Rifkin, D. B. Tumor invasion through the human.miniimi membrane: requirement for a proteinasc cascade. Cell, 47: 487-498,

1986.47. Thorgeirsson, U. P.. Liotta. L. A., Kalebic, T., er al. Effect of the natural protease

inhibitors and a chemoattractant on tumor invasion in vitro. J. Nati. Cancer Inst., 69:1049-1054, 1982.

48. Matrisian. L. M. Metalloproteinases and their inhibitors in matrix remodeling. TrendsGenet., 6: 121-125, 1990.

49. Wilhelm, S. M., Collier, I. E.. Marmer, B. L., el al. SV40-transformed human lungfibroblasts secrete a 92-kDa type IV collagenase which is identical to that secreted bynormal human macrophages. J. Biol. Chem.. 264: 17213-17221, 1989.

50. Fessier, L., Duncan, K., and Tryggvason, K. Identification of the procollagen IVcleavage products produced by a specific tumor collagenase. J. Biol. Chem., 259:9783-9789, 1984.

51. Stetler-Stevenson, W. G., Krutzsch, H. C., and Liolta. L. A. Tissue inhibitor ofmetalloproteinase (TIMP-2): a new member of the metalloproteinase inhibitor family.J. Biol. Chem., 264: 17374-17378, 1989.

52. Goldberg, G. I., Marmer, B. L., Grant, G. A., el al. Human 72-kilodalton type IV

collagenase forms a complex with a tissue inhibitor of metalloproleinases designatedTIMP-2. Proc. Nati. Acad. Sci. USA. S6: 8207-8211. 1989.

53. DeClerck, Y. A., Yean, T. D.. Ratzkin, B. J., el al. Purification and characterizationof two related but distinct metalloproteinasc inhibitors secreted by bovine aorticendothelial cells. J. Biol. Chem., 264: 17445-17453, 1989.

54. Murphy, G., Cawston, T., and Reynolds, J. An inhibitor of collagenase from humanamniolic fluid: purification, characterization, and action on metalloproteinases.Biochem. J., 795: 167-173, 1981.

55. Howard, E. W., Bullen, E. C., and Banda, M. J. Preferential inhibition of 72 kDa and92 kDa gelatinases by TIMP-2. J. Biol. Chem., 266: 13070-13075, 1991.

56. Ura, H., Houli!. R. D., Reich. R., el al. Expression of type IV collagenase andprocollagen genes and its correlation with the tumorigenic, invasive and metastaticabilities of oncogene-transformed human bronchial epithelial cells. Cancer Res., 49:4615-4622, 1989.

57. Monteagudo, C., Merino. M., San Juan, J., et al. Immunohistologic distribution oftype IV collagenase in normal, benign, and malignant breast tissue. Am. J. Pathol..136: 585-592, 1990.

58. Levy, A., Cioce, V., Sobel, M. E., el al. Increased expression of the 72 kDa type IVcollagenase in human colonie adenocarcinoma. Cancer Res., 5/: 439-444, 1991.

59. Schultz, R. M., Silberman, S.. Persky, B., el al. Inhibition by recombinant tissueinhibitor of metalloproteinases of human amnion invasion and lung colonization bymurine B16 FIO melanoma cells. Cancer Res., 4K: 5539-5545, 1988.

60. Khokha. R., Waterhouse, P., Yagel, S., el al. Amisense RNA-induccd reduction in

murine TIMP level confers oncogenicity on Swiss 3T3. Science (Washington DC),243: 947-950, 1989.

61. Albini, A.. Melchiori, A., Santi, L., el al. Tumor cell invasion inhibited by TIMP-2.J. Nati. Cancer Inst., 83: 775-779, 1991.

62. DeClerck, Y. A., Perez, N., Shimada, H., el at. Inhibition of invasion andmetastasis in cells transfecled with an inhibitor of metalloproteinases. CancerRes., 52: 701-708. 1992.

63. Mignatti, P., Tsuboi, R., Robbins, E., and Rifkin, D. In vitro angiogenesis on thehuman amniotic membrane: requirement for basic fibroblast growth factor-inducedproteinases. J. Cell Biol., 108: 671-682, 1989.

64. Schwartz, M. A., and Van Wart, H. E. Eight synthetic inhibitors of bacterial andmammalian interstitial collagenases. Prog. Med. Chem., 29: 271-334, 1992.

65. Stetler-Stevenson, W. G., Krutzsch, H. C., Wacher, M. P.. Margulies, 1. M. K., and

Liotta, L. A. The activation of human type IV collagenase proenzyme. Sequenceidentification of the major converstion product following orangomercurial activation.J. Biol. Chem., 264: 1353-1356, 1989.

66. Melchiori, A., Albini, A., Ray, J. M., et al. Inhibition of tumor cell invasion by ahighly conserved peptide sequence form the matrix metalloproteinase enzyme pro-segment. Cancer Res., 52: 2353-2356, 1992.

67. Fotouhi, N., Lugo, A., Visnick, M., Lusch, L., Walsky, R., Coffey, J. W., andHanglow, A. C. Potent peptide inhibitors of slromelysin based on the prodomainregion of matrix metalloproteinases. J. Biol. Chem., 269: 30227-30231, 1994.

68. Wang, X., Fu, X.. Brown, P. D., Crimmin, M. J., and Hoffman, R. M. Matrixmetalloproteinase inhibitor BB-94 (Batimastat) inhibits human colon tumor growthand spread in a patient-like orthotopic model in nude mice. Cancer Res., 54:4726-4728, 1994.

69. Sheffield, J. B., Krasnopolsky. V.. and Dehlinger. E. Inhibition of retinal growth coneactivity by specific metalloproeinase inhibitors in vitro. Dev. Dynamics. 200: 79-88.1994.

70. Galardy, R., Grobelney, D., Foellmer, H. G., and Fernandez, L. A. Inhibition ofangiogenesis by the matrix metalloprotease inhibitor Af-[2/?-2-(hydroxamidocar-bonymethyl)-4-methylpentanoyl)]-[.-tryptophan methylamide. Cancer Res., 54:4715-4718. 1994.

71. Davies. B., Brown, P. D., East, N., Crimmin, M. J., and Balkwill, F. R. A synthetic matrixmetalloproteinase inhibitor decreases tumor burden and prolongs survival of mice bearinghuman ovarian carcinoma xenografts. Cancer Res., 53: 2087-2091, 1993.

72. Yarden, Y., and Urlich, A. Growth factor receptor kinases. Annu. Rev. Biochem., 57:443-478, 1988.

73. Sato, H., and Seiki, M. Regulatory mechanism of 92 kDa type IV collagenasegene expression which is associated with invasiveness of tumor cell. Oncogene,K: 385-405, 1993.

74. Prendergast, G. C., Davide, J. P., Krai, A., Diehl, R.. Gibbs, J. B., Omer. C. A., andKohl. N. E. Negative growth selection against rodent fibroblasts targeted for geneticinhibition of farnesyl transferase. Cell Growth & Differ., 4: 707-713, 1993.

75. Powis, G. Drugs active against growth factor and oncogene phosphatidylinositolsignaling pathways. Semin. Cancer Biol., 3: 343-350. 1992.

76. Schwartz, G. K., Jiang, J., Kelsen, D., and Albino, A. P. Protein kinase C: a noveltarget for inhibiting gastric cancer invasion. J. Nati. Cancer Inst., 85: 402-407, 1993.

77. Kerr, L. D., Miller, D. B., and Matrisian, L. M. TGF-/31 inhibition of transin/

stromelysin gene expression is mediated through nfos binding sequence. Cell, 61:267-278, 1990.

78. Kohn. E. C., and Liotta. L. A. L651582, a novel antiproliferative and antimetastasisagent. J. Nail. Cancer Inst., 82: 54-60, 1990.

79. Kohn, E. C., Sandeen, M. A., and Liotta, L. A. In vivo efficacy of a novel inhibitorof selected signal transduction pathways including calcium, arachidonate, and inositolphosphates. Cancer Res., 52: 3208-3212, 1992.

80. Kohn, E. C., Jacobs, W., Kim, Y. S., Alessandro, R., Stetler-Stevenson, W. G., andLiotta, L. A. Calcium influx regulates matrix metalloproteinase-2 (72 kDa type IVcollagenase, gelatinase A) expression. J. Biol. Chem., 269: 21505-21511, 1994.

81. Kohn, E. C., Felder. C. C., Jacobs, W., Holmes, K. A., Day, A., Freer, R.. and Liotta,L. A. Stucture-function analysis of signal and growth inhibition by carboxyamido-triazole, CAI. Cancer Res., 54: 935-942, 1994.

82. Kohn, E. C., Alessandro, R., Spoonster, J., Wersto, R. W., and Liotta, L. A.Angiogenesis: role of calcium-mediated signal transduction. Proc. Nati. Acad. Sci.USA, 92: 1307-1311, 1995.

83. Gusovsky, F., Lueders, J. E., Kohn, E. C., and Felder, C. C. Muscarinic receptor-mediated tyrosine phosphorylation of phospholipase C-y: an alternative mechanism for cholinergic-induced phosphoinositide breakdown. J. Biol. Chem., 268:7768-7772, 1993.

84. Felder, C. C., Ma, A. L., Liotta, L. A., and Kohn, E. C. The antiproliferative andantimetastatic compound L651582 inhibits muscarinic acetylcholine receptor-stimu

lated calcium influx and arachidonic acid release. J. Pharmacol. Exp. Ther., 257:967-971, 1991.

85. Chapron, Y., Cochet, C., Crouzy, S., Jullien, T., Keramidas, M., and Verdetti, J.Tyrosine protein kinase activity of the EGF receptor is required to induce activationof receptor-operated calcium channels. Biochem. Biophys. Res. Commun., 758:527-533, 1989.

86. Janney, P. A. Phosphoinositides and calcium as regulators of cellular actin assemblyand disassembly. Annu. Rev. Physiol., 56: 169-191, 1994.

87. Meldolesi, J., Clementi, E., Fusolato, C., Zacchetti, D., and Pozzan, T. Ca2* influx

following receptor activation. Trends Pharmacol. Sci., 12: 289-292, 1991.88. Prendergast, G. C., and Gibbs, J. B. Pathways of ras function: connections to the actin

cytoskeleton. Adv. Cancer Res., 62: 19-64, 1993.89. Pozzatti, R., Muschel, R., Williams, J., Padmanabhan, R., Howard, B., Liotta, L., and

Khoury, G. Primary rat embryo cells transformed by one or two oncogenes showdifferent metastatic potentials. Science (Washington DC), 232: 223-227. 1986.

90. Yu, D.. Shi, D., Scanlon, M.. and Hung, M. C. Reexpression of Mew-encodedoncoprotein counteracts the tumor-suppressing but not the metastasis-suppressingfunction of EIA. Cancer Res., 53: 5784-5790, 1993.

91. Tanaguchi, T., Kitagawea, H., Yasue, S., el al. Protein-tyrosine kinase p72syk is

activated by thrombin and is negatively regulated through Cpa2+ regulation inplatelets. J. Biol. Chem., 268: 2277-2279, 1993.

92. Savarese, D. M. F., Russell, J. T., Fatalis, A., and Liotta, L. A. Type IV collagenstimulates an increase in intracellular calcium: potential role in tumor cell motility.J. Biol. Chem., 267: 21928-21935, 1992.

1862



perspectives in cancer research · [cancer research 55, 1856-1862, may 1, 1995] perspectives in...

Documents