(CANCER RESEARCH 36, 2518-2524, July 1976]

Summary

Integral membrane proteins are visualized as intramembrane particles (IMP; also called membrane-associated pantides) at the cleaved surfaces of freeze-fractured plasmamembranes. Topographical distributions of the IMP ofurothelial cell membranes in normal human bladder and fora small series of low-grade noninvasive transitional cellcarcinomas and invasive transitional cell carcinomas areshown to be significantly different. Using several statisticalmethods that test IMP topography vis a vis the random(Poisson) hypothesis, it is demonstrated that IMP are mildlyaggregated in plasma membranes of normal human urothehal cells and that, in noninvasive carcinomas, IMP aggregation is increased. In invasive transitional cell carcinomas,IMP are statistically nonaggregated and are in a randomdistribution in the plane of the membrane. IMP numericaldensities are also altered in the course of neoplastic transformation. IMP are significantly increased in number inplasma membranes in human noninvasive transitional cellcarcinomas but are similar to control values in invasivetumors. Loss of IMP and changes in IMP topography maybe related to tumor invasiveness or they may representan epiphenomenon.

Introduction

Human tumors arising in epithelia may pass through noninvasive stages of development in the course of neoplastictransformation (2, 8). Recent evidence suggests that intninsic tumor cell properties may partly determine the biologicalbehavior of tumors (9, 15). If intrinsic cell changes areessential for the acquisition of the property of invasiveness,then it is reasonable to anticipate that such changes mightfirst appear when tumors pass from a noninvasive to aninvasive stage of tumor development. Previous efforts todemonstrate specific structural differences between cellsfrom invasive carcinomas and cells from tumors at earlierstages of tumor development have been unsuccessful (18,19).

@ Presented at the Conference “EarlyLesions and the Development ofEpithelial Cancer.―October 21 to 23, 1975, Bethesda, Md. This work wassupported by Grant CA-14447 from the National Cancer Institute, NIH. and byUnited States Air Force Contract F33615-75-R-5001 from the AerospaceMedical Research Laboratory, Wright-Patterson Air Force Base, Ohio. Earlyphases of this work were performed in the Department of Pathology, TuftsUniversity School of Medicine, Boston, Mass.

It is generally acknowledged that changes in the plasmamembrane might account for the biological behavior oftumors (7, 13, 24, 29). However, none of the many types ofchanges that have been described in tumor cell membraneshave been correlated successfully with stages of neoplastictransformation. In the current study, normal and malignanturinary bladder epithelia are examined by thin section andfreeze-fracture electron microscopy to determine whethermembrane ultrastructure is modified during neoplastictransformation and if such structural changes can be comelated with tumor invasiveness. Particular attention is dinected to the numerical density and 2-dimensional topognaphy of IMP,2 a distinctive intramembrane component thatcan be visualized by freeze-fracture electron microscopy. Asubstantial body of evidence indicates that IMP representintegral membrane proteins (4, 12, 36) and that their topographical distribution can be modulated by biological control mechanisms (5, 14). In this report, IMP numerical densities and topographical distributions are shown to be sign ificantly different for a small series of noninvasive transitionalcell carcinomas and invasive transitional cell carcinomasarising spontaneously in human urinary bladder. These observations suggest that alterations within the plasma membrane may be related to changes in the biological behaviorof bladder carcinomas.

Materials and Methods

Human Urinary Bladder Specimens. Biopsies of humanurinary bladder were obtained through a cystoscope andfixed immediately (less than 45 sec) for light and electronmicroscopy. Histopathological evaluations of biopsies wereperformed by light microscopy. Biopsy specimens wereclassified into one of 3 categories: Category 1, controls;Category 2, noninvasive transitional cell carcinoma; or Category 3, invasive transitional cell carcinoma. Tumors were

classified as “noninvasive―if histopathological evaluationof many hematoxylin and eosin-stained paraffin sectionsfailed to demonstrate penetration of the unothelial basement membrane by tumor cells, or as “invasive―if tumorclearly extended through the basement membrane. Casesfor which staging was uncertain were eliminated from thisstudy. Accepted for the study were biopsies from 5 patients

2 The abbreviations used are: IMP, intramembrane particles; CD. , coefficient of dispersion; IMPPand IMPE intramembrane particles on fracture faceP and face E. respectively.

2518 CANCER RESEARCH VOL. 36

Changes in Plasma Membrane Structure Associated withMalignant Transformation in Human UrinaryBladder Epithelium1

RonaldS. WeinsteinDepartment of Pathology, Rush Medical College, and Rush-Presbyterian-St. Luke's Medical Center, Chicago, Illinois 60612

Research. on December 9, 2020. © 1976 American Association for Cancercancerres.aacrjournals.org Downloaded from

Integral Membrane Proteins in Cancer

with relatively low-grade noninvasive transitional cell carcinoma, 4 patients with relatively low-grade invasive transitional cell carcinoma, and 4 controls. The controls werefrom 1 patient with incomplete urethral obstruction, 2 patients with benign prostatic hypertrophy, and from an areaof normal-appearing mucosa from 1 patient with transitional cell carcinoma elsewhere in the bladder. It was shownby both light and electron microscopy that the biopsy specimens from all of the controls had normal-appearing bladdermucosa.

Thin-Section Electron Microscopy. Specimens for thinsection and freeze-fracture electron microscopy were divided into 1- to 2-cu mm blocks and were fixed in theoperating room, immediately after surgical removal, eitherin cold 2% glutaraldehyde in 0.1 M cacodylate buffer, pH7.3, or in half-strength Kamnovsky's paraformaldehyde for 1hr (19). For thin-section electron microscopy, tissues werepostfixed in 1% osmic acid in cacodylate buffer, pH 7.3,dehydrated by serial passage through graded ethanol solutions, embedded in Epon 812, sectioned with diamondknives, and stained with unanyl acetate and lead citrate.Thin sections and freeze-fracture replicas were photographed in a Philips EM 300 electron microscope.

Freeze-FractureElectronMicroscopy.Two% glutaraldehyde-fixed tissue blocks were soaked overnight in a cryoprotectant consisting of Millonig's phosphate buffer, pH7.4, containing 20% glycerol (v/v). Blocks were quenched to_1500 in liquid Freon 22, cooled by liquid nitrogen, and

freeze-fractured (22) with a Balzers BAF 301 freeze-etch unitequipped with an electron beam evaporation device (EVM052) and a QSC 201 quartz crystal thin-film-thickness moniton.

For control urothelium, freeze-fracture replicas of plasmamembranes from regions of intermediate cell-intermediatecell apposition, intermediate cell-basal cell apposition, andbasal cell-basal cell apposition were photographed. Luminal membranes and basal cell membranes, at the stromalfront, were excluded. For tumors, sampling was restrictedto plasma membranes in regions of tumor cell-tumor cellapposition. Areas of membrane containing cell-cell junctions were avoided in both controls and tumors.

FrequencyDistributionsof IMP. Electronmicrographsoffreeze-fracture replicas were coded and analyses of IMPtopographical distributions were done by a technician,without knowledge of the clinical history or histopathological diagnosis for the cases. For statistical analysis of IMPtopographical distributions, electron micnognaphs of nelatively large flat areas of the PF fracture face (3) of freezefracture membrane were printed at a final magnification ofx250,000. IMP frequency distributions were determined bysuperimposing a transparency containing a 10-sq cm testgrid (defining a “testfield―),subdivided into 100 1-sq cmsquares, oven individual electron micrographs of freezefracture membranes and recording the number of IMP ineach of the 100 squares on a differential counter. Numericaldensities of IMP, defined as the number of IMP per unit areaof membrane, were calculated from the frequency distnibution data. For individual biopsies, topographical distnibutions of IMP were determined for 8 to 12 membrane fracture faces. EF faces (3) were also evaluated by statisticalmethods.

Statistical Evaluation of IMP Topography Vis-à-VIstheRandom(Poisson)Hypothesis.If the IMP are truely randomly distributed over the membrane fracture face, thenthe freqeuncy distribution of IMP would be expected to bePoisson distributed. Two statistical tests were used to evaluate IMP topography vis-á-visthe Poisson hypothesis. First,the C.D. was used as an index of IMP ordering. This represents the first use of the C.D. statistic for the analysis of

membrane topography. C.D. is defined as variance/meanfor a frequency distribution curve (32). The statistic is anindicator of degree of departure of IMP distributions from atrue random distribution. A characteristic of any Poissondistribution is that C.D. = 1, and any order in the distnibution of IMP would be reflected in the deviation of C.D.values from unity: distribution of IMP in a lattice-like configuration causes the C.D. to go below 1, and aggregation ofIMP causes the C.D. to go above 1 (32). Critical values forthe C.D. statistic have not been published. However, it canbe shown that under suitable conditions, such as those inthis study, the C.D. statistic bears a simple relationship tox2, being distributed approximately as @2/(k—1) for k degrees of freedom where k is the number of domains (i.e.,squares in the test lattice).3 As calculated from values instandard x2 tables, C.D. values over 1.372 are indicative ofstatistically significant IMP aggregation (p < 0.01) for a 100-square lattice, since C.D. = @@,99/k—1 = 135.81/99. C.D.values below 0.708 are indicative of statistically significantordering into a lattice-like arrangement (p < 0.01), sinceC.D. = X@.o/k 1 = 70.06/99. C.D. values within the 0.708to 1.372 range indicate statistical randomness (p < 0.01).Through computer Monte Carlo simulations, we confirmedthat C.D. values are directly related to topography andindependent of IMP numerical densities under conditionsencountered in this study.

Second, the frequency distribution curve for each testfield was also compared with Poisson distribution by thevariance test for goodness of fit of the Poisson distribution(28) and classified as Poisson on non-Poisson at p < 0.01.The percentage of test fields fitting a Poissondistribution isused as an additional index of prevalence of IMP randomness for different test fields within categories.

Results

Fine Structureof TransitionalCell Carcinomas.The ultrastructure of human bladder tumors, as seen by thinsection microscopy, has been described elsewhere (10).Studies on human bladder with the freeze-fracture technique have been limited to a consideration of cell-cell junctions (37). For the purposes of the current study, we exammed the ultnastructureof normal uroepithelium and of bladdentumors in order to determine the level of preservation ofthe freeze-fractured tissue. In general, tissue preservationwas acceptable (Fig. 1). Myelin figures, swollen mitochondna, and dilatedcistemnae,allindicatorsofcellinjury,wererarely encountered.

ElectronMicroscopyof Freeze-FracturedHumanUrothe

3 W. D. Selles, R. S. Weinstein, and I. T. Young. Coefficient of Dispersionas a Statistic for Evaluating Cell Membrane Topography, submitted for publication.

2519JULY 1976

Research. on December 9, 2020. © 1976 American Association for Cancercancerres.aacrjournals.org Downloaded from

IMP counts were made on 10 PF and EF faces for each biopsy.Numbersof patients: controls, 4; noninvasivecarcinoma, 5; invasive carcionma, 4.IMP

numerical densities (IMP/sq@m)PF

fracture face EF fracture faceCategory (lMP@)(IMP@)Controls

690― 105(530-85O)'@(95-115)Noninvasive

1300 125Carcinoma (880-1730)(115-160)Invasive

530 100Carcinoma (400-670) (90-135)

CategoryCoefficientof dis

persion%PoissonControls(4/44)°1 .54 ±0.072'50'Noninvasive

carcinoma (5/48)1.82 ±0.11013.6Invasivecarcinoma (4/48)1 .26 ±0.06181.8

R. S. Weinstein

hal Membranes. The freeze-fracture process splits opencell membranes and generates novel fracture faces whichoriginate from within the interior of the cell membrane (4).Freeze-fractured urothelial cells in normal bladder and intumors are readily recognized by their characteristic sizesand shapes and by their intercellular relationships. Theepithelial-mesenchymal front can be identified in freezefracture replicas for most specimens. For purposes of thisstudy, analysis of membrane topography was restricted tocells above the level of basement membrane in both noninvasive and invasive transitional cell carcinomas.

Two novel fracture faces are generated by freeze-f nactuning cell membranes: “Face PF―has been defined as thefracture face of the inner lamella of the membrane, and“Face EF,― as the fracture face of the outer lamella of the

membrane (3). Both fracture faces bear populations of IMPthat average 7 nm in replica diameter and resemble the IMPwhich are membrane components of many types of cells (4).Henceforth, the Face PF IMP shall be designated IMP1.andthe Face EF IMP shall be designated IMP@:.@

The numerical densities of IMPE:in controls and tumorswere not significantly different (data not included). Analysisof numerical densities of IMP@.in noninvasive transitionalcell carcinomas shows that these values were well abovecontrol values (p < 0.01). In invasive tumors, numericaldensities of IMP,. were not significantly different from control values (Table 1) but were significantly below values fornoninvasive tumors (p < 0.01).

By visual inspection, IMP,. appeared to be similarly distnibuted in the plane of the membrane in both controls (Fig.2) and in invasive transitional cell carcinomas in which theIMP appeared to be either randomly distributed or slightlyaggregated (Fig. 4). IMP1. appeared more aggregated innoninvasive tumors (Fig. 3), although it is noteworthy thatdifferences in IMP, clustering between noninvasive tumors,invasive tumors, and controls were not striking to the eye.IMP@appeared, by visual inspection, to be randomly distnibuted for specimens in all 3 categories.

C.D. Values. Analysis of topographical distributions ofIMP,. by means of the CD. statistic showed significant diffenences among all 3 groups (Table 2). The mean CD. value(±S.D)for controls, 1.54 ±0.072, is indicative of mild IMPaggregation. The mean CD. values for noninvasive transitional cell carcinomas, 1.82 ±0.110, was higher than control values, indicating enhanced 1MP1 aggregation. Themean CD. value for invasive carcinomas, 1.26 ±0.061,indicates that the IMP,. are within the range that is indicativeof a truly random distribution. All 5 noninvasive transitionalcell carcinomas had C.D. values higher than the controlvalues. The percentage of test fields fitting a Poisson hypothesis is highest for invasive carcinoma and lowest fornoninvasive carcinomas, with controls showing an intenmediate level of conformity to the Poisson distribution (Table2).

Comparisons of Populations of Cells. There can be considerable variability of IMP topographical distribution fromcell to cell within a single biopsy specimen, although the“average―distributions are similar from biopsy to biopsy

4 In earlier publications (20, 36), the PF-face was called the A-face and the

EF-face was called the B-face.

Table1Numericaldensitiesof IMPwithin freeze-fracturedcell membranes

in nonneoplastic human urothelium and in transitional cellcarcinomas

(1 Mean for all specimens within the category. One-way analysis

of varianceshowsthat differences in IMPdensities on PFfacesaresignificant (p < 0.01). IMP densities on EF faces are not significantly different.

b Number in parentheses, range of mean values for individualspecimens.

Table 2Statistical analysisof IMP topographical distributions on PF faces

of freeze-fracturedhuman urotheliumAnalysesof EF faces showed no significant differences in the

topographical distributions of IMP. Data are not included for EFfaces.

a Numbers in parentheses: numerator, number of biopsies; de

nominator, total number of test fields for all biopsies within thecategory.

b Mean ±S.D. CD. values were compared by 1-way analysis ofvariance,showing that there is a highly significant difference in IMPclustering among the 3 categories (p < 0.01).

C Fit to a Poisson distribution was determined by the variance test

for goodnessoffit to a Poissondistribution. Acceptanceas Poissonwas atp < 0.01.

within each category. To compare distributions across categonies while accounting for cell-to-cell variability, cumulative distributions were generated from data pooled from allcases within a single category. For each test field, thevariance test for goodness of fit of a Poisson distributionyields a @2variable. This variable can be normalized by astandard formula as follows:

S.D. = (x2

where N = number of degrees of freedom. The cumulativedistribution of normalized x2's is plotted for each category(Chart 1). The curves for pooled data from controls, noninvasive carcinomas, and invasive carcinomas (Chart 1) aresignificantly different from one another (p < 0.01), as determined by the 2-sample Kolmogorov-Smimnov test, confirming that the test fields are drawn from different membranepopulations.

2520 CANCERRESEARCHVOL. 36

Research. on December 9, 2020. © 1976 American Association for Cancercancerres.aacrjournals.org Downloaded from

Integral Membrane Proteins in Cancer

duced in invasive transitional cell carcinomas. In addition,there are significant differences in the topographical distributions of IMP between noninvasive and invasive humancarcinomas. IMP. are mildly aggregated in nonneoplasticepithelia. IMP. aggregation is enhanced in noninvasivetransitional cell carcinomas whereas, in invasive cancinomas, IMP. are randomly distributed in the large majority oftumor cells.

The molecular mechanisms involved in producing thedramatic increases in IMP1.numerical densities in noninvasive transitional cell carcinomas are unknown. High densities of IMP have been associated with high levels of membrane metabolic activity (4). However, there is no reason tobelieve that the level of metabolic activity of plasma membranes of noninvasive tumors is greater than that of invasivetumors of the same histopathological grade. Increased IMPdensities also have been observed in plasma membranes ofcells transformed with oncogenic viruses (35), althoughsuch changes do not invariably accompany transformation(27, 30).

There are a number of plausible explanations for thedifferences in IMP1.topography observed between controlsand tumors. One possibility is that artifacts are introducedat cystoscopy, a technique which requires infusion of unphysiological solutions, such as filtered water and/or hypotonic glycine, into the bladder. Normal unothelium and transitional cell carcinomas differ with respect to gross andmicroscopic morphology, leakiness to bypass diffusion,vasculanity, and stroma. Because of these differences, antifactual alterations in IMP may be preferentially introducedinto certain categories of urothelial cells. The possibility ofcystoscopy artifacts is currently being systematically exammed in this laboratory, and the preliminary data show thatsignificant ultrastructural artifacts can be induced by conventional cystoscopy procedures (B. V. Pauli, R. S. Weinstein, and J. Alnoy, unpublished data).

Another possibility, not necessarily unrelated to the first,is that differences in IMP1.topography can be explained onthe basis of IMP.-cytoskeleton interrelationships. In normalcells, the topographical distributions of some IMP may beregulated by peripheral membrane proteins, elements of thecell cytoskeleton, to which the IMP1.may be bound at thejuxtacytoplasmic surface of the plasma membrane. We suggest that, in noninvasive transitional cell carcinomas, alterations in the cytoskeleton may produce a redistribution ofIMP, . Changes in the cytoskeleton could result either from aprimary structural modification of the cytoskeleton (15) orfrom changes in cell shape produced by other etiologies(16). Although a mechanism involving cytoskeletal pentubations with maintenance of IMP-cytoskeleton binding couldaccount for enhanced IMP1@aggregation in noninvasive tumoms,it is unlikely that this Would account for the randomdistributions of 1MP1in the plasma membranes of invasivetumor cells. Random distributions could be accounted foron the basis of inadequate binding of IMP,. to the cytoskeleton. This would relieve IMP1.of constraints upon their lateralmobility and permit IMP1@to achieve truly random distnibutions within the plane of the membrane. Based upon theseobservations, it is postulated that loss of control of IMP@topography may represent a relatively late step in neoplastictransformation.

STANOARL'0EV/AT/OW(units fromw,ecnl

Chart 1. Cumulative relative frequencies of normalized @2values for thetopographical distributions of IMP in test fields of freeze-fractured humanurinary bladder cell membranes. The @2values were generated by the vanance test for goodness offit of the Poisson distribution and then normalized.The normalization is: S.D. = It(x2 —N)/@/@Np(,where N = the number ofdegrees of freedom. In this instance, N was determined from the number ofclasses of IMP, n, where: n = t(maximum number of IMP/test square) + 11.Numbers of test fields within categories represented by the curves are:invasive carcinomas, 48, controls, 44; and noninvasive carcinomas, 48. IMPtopographical randomness is closely approximated in invasive transitionalcell carcinomas, and IMP aggregation is greatest in noninvasive transitionalcell carcinomas.

Discussion

Integral membrane proteins are defined as amphiphaticproteins that span the lipid bilayer region of the cell membrane (31). They are a heterogenous group of moleculesthat may participate in a number of cellular processes including membrane transport, transmembrane coupling,and cell-cell recognition (16, 21, 26, 30). A substantial bodyof evidence supports the hypothesis that the IMP visualizedby freeze-fracture electron microscopy are proteins or protein aggregates intercalated into the membrane lipid bilayer(4, 12, 34). Changes in the numerical densities (35) andtopographical distributions of IMP (1, 30) have been observed in association with transformation in several tissueculture systems but have not been previously reported insolid tissues undergoing neoplastic transformation.

The topographical distributions of some IMP may be influenced by interactions of the IMP with either other integralmembrane components (1, 33) or elements of the cell cytoskeleton, including peripheral membrane proteins (6, 23),which may place constraints upon the lateral mobility ofIMP. Theoretically, other IMP could be free of these constraints and therefore be capable of achieving truly randomdistributions in diffusion fields within the membrane. Alterations in the topographical distributions of IMP can beinduced by both physiological (5, 14) and unphysiological(11, 17, 25, 33) methods. This multiplicity of factors that mayinfluence topography enormously complicates the task ofascribing changes in IMP topography in pathological statesto specific mechanisms.

In this quantitative electron microscopy study on humanbladder urothelia, we explored the possibility that changesin numbers of IMP and in the topography of IMP may serveas markers for specific stages in neoplastic transformation.We found that numbers of IMP are increased above controllevels in noninvasive transitional cell carcinomas but me

JULY 1976 2521

Research. on December 9, 2020. © 1976 American Association for Cancercancerres.aacrjournals.org Downloaded from

tions on the Occurrence of Nexuses in Benign and Malignant HumanCervical Epithelium. J. Cell Biol., 51: 805-825, 1971.

19. McNutt, N. S., and Weinstein, R. S. Carcinoma of the Cervix: Deficiencyof Nexus Intercellular Junctions. Science, 165: 597-598, 1969.

20. McNutt, N. 5., and Weinstein, R. 5. The Ultrastnucture of the Nexus. ACorrelated Thin-Section and Freeze-Cleave Study. J. Cell Biol., 47: 666-688,1970.

21. McNutt, N. 5. , and Weinstein, R. 5. Membrane Ultrastructure at Mammahan IntercellulanJunctions. Progn. Biophys. Mol. Biol.,26: 45—101,1973.

22. Moor, H., MUhlethalen, K., Waldnen, H., and Fney-Wyssling, A. A NewFreezing-Ultramicrotome. J. Biophys. Biochem. Cytol., 10: 1-10, 1961.

23. Nicolson, G. L., and Painter, R. G. Anionic Sites of Human ErythnocyteMembranes. II. Antispectnin-Induced Transn@embnaneAggregation of theBinding Sites for Positively Charged Colloidal Particles. J. Cell. Biol., 59:395-406. 1973.

24. Pardee, A. B. Cell Division and a Hypothesis of Cancer. NatI. Cancer Inst.Monograph, 14: 7-18, 1964.

25. Pinto da Silva, P. Translational Mobility of the Membrane IntercalatedParticles of Human Erythnocyte Ghosts, pH-Dependent, Reversible Aggregation. J. Cell Biol., 53: 777-787, 1972.

26. Pinto da Silva, P. Membrane Intercalated Particles in Human EnythnocyteGhosts:Sites of Preferred Passageof Water Molecules at Low Temperatune. Proc. NatI. Acad. Sci. U. S., 70: 1339-1343, 1973.

27. Pinto da Silva, P., and Martinez-Palomo, A. Distribution of MembraneParticles and Gap Junctions in Normal and Transformed 3T3 Cells Studied In Situ, in Suspension, and Treated with Concanavalin A. Proc. NatI.Aced. Sci. U. S., 72: 572-576, 1975.

28. Remington, R. D., and Schonk,M. A. In: Statistics with Applications tothe Biologicaland Health Sciences, pp. 246-248. Englewood Cliffs, N. J.:Prentice Hall, Inc., 1970.

29. Rosenblith, J. Z., Ukena, T. E., Yin, H. H.. Beslin, R. D.•and Kamovsky,M. J. A Comparative Evaluation of the Distribution of Concanavalin A-Binding Sites on the Surfaces of Normal, Virally-Transformed, and Protease-Treated Fibroblasts. Proc. NatI. Acad. Sci. U. S.. 70: 1625-1629,1973.

30. Scott, R. E., Funcht, L. T. , and Kensey,J. H. Changes in MembraneStructure Associatedwith Cell Contact. Proc. Natl. Acad. Sci. U. 5. , 70:3631-3635,1973.

31. Singer, S. J., and Nicolson, G. L. The Fluid Mosaic Model of the Structuneof Cell Membranes. Science, 175: 720-725, 1972.

32. Sokal, R. R., and Rohlf, F. J. Introduction to Biostatistics,pp. 72-74. SanFrancisco: W. H. Freeman and Company. 1973.

33. Speth, V., and Wunderlich, F. Membranes of Tetnahymena. II. DirectVisualization of Reversible Transitions in Biomembrane Structure Induced by Temperature. Biochim. Biophys. Acta, 291: 621-628, 1973.

34. Tillack, T. W., Scott, R. E., and Marchesi, V. T. The Structure of Erythnocyte Membranes Studied by Freeze-Etching. II. Localization of Receptorsfor Phytohemagglutinin and Influenza Virus to the IntramembranousParticles. J. Exptl. Med., 135: 1209—1227,1972.

35, Tonpier, 6. , Montagnier, L., Biguand,J-M. . and Vigien, P. A Change ofthe Plasma Membrane Induced by Oncogenic Viruses: QuantitativeStudies with the Freeze-Fracture Technique. Proc. NatI. Aced. Sci. U. S.,72: 1695-1698, 1975.

36. Weinstein, R. S. The Morphology of Adult Red Cells. In: D. MacN.Surgenon(ed), The Red Cell, Ed. 2. pp. 213-268. New York: AcademicPress, Inc., 1974.

37. Weinstein, R. S. , Zel, G. . and Merk, F. B. Quantitation of Occludens,Adherens, and Nexus Cell Junctions in Human Tumors. In: J. Schultzand R. E. Block (ads.), Membrane Transformations in Neoplasia, pp.127-146. New York: Academic Press. Inc., 1974.

Fig. 1. Freeze-fracture replica of a tumor cell in a Grade 1 human noninvasive transitional cell carcinoma. PM, plasma membrane; mit, mitochondnia;mvb, multivesicular body. Arrows, position of nuclear pores within the nuclear membrane. x 36,200.

Fig. 2. PF and EF fracture faces of plasma membranes of 2 unothelial cells in normal human urinary bladder. The PF face bears many more IMP than the EFface. The IMPPappear to be mildly aggregated. ECS, extracellular space. x 75,000.

Fig. 3. PF and EF fracture faces of plasma membranes in a Grade 1 noninvasive transitional cell carcinoma. IMPPare increased in numerical density andshow a tendency to aggregate. x 80,000.

2522

R. S. Weinstein

Acknowledgments

The author gratefully acknowledges the expert technical assistance ofJonathan S. Wallach and William Leonard. I also take pleasure in thankingDr. Ian T. Young, Massachusetts Institute of Technology, Cambridge, Mass..,.and William D. SelIes, New England Medical Center Hospital, Boston, Mass.,for suggestir@gthe statistical tests used in this study and for assisting in theevaluation of the data.

References

1. Barnett, R. E.. Furcht, L. T., and Scott, R. E. Differences in MembraneFluidity and Structure in Contact-Inhibited and Transformed Cells. Proc.NatI. Acad. Sci. U. S., 71: 1992-1994, 1974.

2. Berenblum, I. A Speculative Review: The Probable Nature of PromotingAction and Its Significance in the Understanding of the Mechanism ofCarcinogenesis. Cancer Res., 14: 471-477, 1954.

3. Branton. D., Bullivant, S., Gilula, N. B.. Kannovsky, M. J.. Moon, H.,Mohlethaler, K. , Northcote, D. H.. Packer, L. . Satin. B.. Satir. P., Speth,V. , Staehelin, L. A. , 5teere, R. L., and Weinstein, A. 5. Freeze-EtchingNomenclature. Science, 190: 54-56, 1975.

4, Branton, D., and Deamer, D. Membrane Structure. Protoplasmatologia,Il/E/ 1: 1-70, 1972.

5, Chevalier, J. , Bounguet. J. . and Hugon, J. S. Membnane AssociatedParticles: Distribution in Frog Urinary Bladder Epithelium at Rest andAfter Oxytocin Treatment. Cell Tissue Res., 152: 129-140, 1974.

6. Elgsaeten, A. , and Branton, D. lntnamembnane Particle Aggregation inErythrocyte Ghosts. I. The Effects of Protein Removal. J. Cell Biol., 63:1018-1030, 1974.

7. Emmelot, P. Biochemical Properties of Normal and Neoplastic Cell Sunfaces: A Review. European J. Cancer, 9: 319-333, 1973.

8. Farber, E. Carcinogenesis. Cellular Evolution as a Unifying Thread:Presidential Address. Cancer Res., 33: 2537-2550, 1973.

9. Fidler, I. J. Selection of Successive Tumour Lines for Metastasis. NatureNew Biol., 242: 148-149, 1973.

10. Fulker. M. J.. Cooper, E. H., and Tanaka, T. Proliferation and Ultnastructune of Papillary Transitional Cell Carcinoma of the Human Bladder.Cancer, 27: 71-82, 1971.

11. Hackenbnock, C. R. States of Activity and Structure in MitochondnialMembranes.Ann.N.Y.Acad.Sci.,195:492-504,1972.

12. Hong, K., and Hubbell, W. L. Preparation and Properties of PhospholipidBilayers Containing Rhodopsin. Proc. NatI. Acad. Sci. U.S., 69: 2617-2621, 1972.

13. Inbar, M., and Shinitzky, M. Increase of Cholesterol Level in the SurfaceMembrane of Lymphoma Cells and Its Inhibitory Effect on Ascites TumorDevelopment. Proc. NatI. Acad. Sci. U. S., 71: 2128—2130,1974.

14. Kachadonian, W. A., Wade, J. B., and DiScala, V. A. Vasopressin: Induced Structural Change in Toad Bladder Luminal Membrane. Science,190: 67-69, 1975.

15. Malech, H. L., and Lentz, T. L. Microfilaments in Epidermal Cancer Cells.J. Cell Biol. , 60: 473-482, 1974.

16. Marikovsky, Y. , Brown, C. S., Weinstein, R. S., and Wortis, H. H. Effectsof Lysolecithin on the Surface Properties of Human Erythnocytes. Exptl.Cell Res., 98: 313-324, 1976.

17. McIntyre, J. A., Gilula, N. B., and Karnovsky, M. J. CryoprotectantInduced Redistribution of lntnamembnanous Particles in Mouse Lymphocytes. J. Cell Biol., 60: 192-203, 1974.

18. McNutt, N. S., Henshberg, R. A., and Weinstein, R. S. Further Observa

CANCERRESEARCHVOL. 36

Research. on December 9, 2020. © 1976 American Association for Cancercancerres.aacrjournals.org Downloaded from

. . - .‘ , ,@@ ‘@, - ... 4 @.., .; . .z @‘‘,@‘ ‘ :‘@@ ‘@ â€â€˜@.

. ..: -@ . . . . -,. . . -@ @. a ,,@@ : .@ ..

@ @EF@@

JULY 1976

.. . -@

.@ ‘p

, -

. .,@‘. .‘..

..:44@.@ . : †‘@

@@ a- :@“@‘@:-@‘.. -,.“- :‘@@@ . . @-@@:‘ .@ . •.

. ‘: @. ;3J_:.;I,,, .; . _@@ V @% @,, ,_,@ :- , .- ,:•“ @. ‘@

@ 7_@ - 4!@@ ‘@

.@ :@ ‘.‘@ —•_:@@@ @@:‘@ _•:@@ :@;@‘@:.@.r'..:â€@.@@ ..@-

‘s,___/@@ @!@! •‘ a@@ ::@@

. —. .. . .‘....@S'_. .,@ S@

2523

Integral Membrane Proteins in Cancer

Research. on December 9, 2020. © 1976 American Association for Cancercancerres.aacrjournals.org Downloaded from

“4...@ -

4 @:@‘@ _ ..

..,l

@ .

;@‘.. . f

Fig. 4. PF fracture face of the plasmamembrane of a tumor cell in a Grade 1 to 2invasive transitional cell carcinoma. Thistumor invaded into muscle. IMP@appear tobe randomly distributed. x 80,000.

2524CANCERRESEARCHVOL. 36

R. S. Weinstein

Research. on December 9, 2020. © 1976 American Association for Cancercancerres.aacrjournals.org Downloaded from

1976;36:2518-2524. Cancer Res   Ronald S. Weinstein  Malignant Transformation in Human Urinary Bladder EpitheliumChanges in Plasma Membrane Structure Associated with

  Updated version

  http://cancerres.aacrjournals.org/content/36/7_Part_2/2518

Access the most recent version of this article at:

   

   

   

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

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/36/7_Part_2/2518To request permission to re-use all or part of this article, use this link

Research. on December 9, 2020. © 1976 American Association for Cancercancerres.aacrjournals.org Downloaded from