-
[CANCERRESEARCH38,2651-2660,September1978]
Abstract
The development of a metastasis Is dependent on aninterplay
between host factors and intrinsic characteristics of malignant
tumor cells. The process of metastasisis highly selective, and the
metastatic lesion representsthe end point of many destructive
events that only a fewcells can survive. Neoplasms, which are
predominantlyheterogeneous, contain a variety of subpopulatlons
ofcells with differing metastatic potential. Furthermore,metastatic
cell variants have been shown to preexist inmurine neoplasms of old
and recent origin. The possibleexistence of highly metastatic
variant cells within a primary tumor suggests that we no longer
should consider aneoplasm to be a uniform entity. Efforts to
designeffective therapeutic agents and procedures againstmalignant
tumors should be directed toward the few butfatal metastatic
subpopulations of cells.
Introduction
Metastasis is defined by Dorland (8) as the transfer ofdisease
from one organ or part to another not directlyconnected with it.
The metastasis of cancer cells is one ofthe most devastating
aspects of neoplasia; it is responsiblefor most therapeutic
failures because patients succumb tomultiple secondary tumor
growths and not necessarily tothe primary tumor. Major advances in
surgical techniques,coupled with advances in general patient care,
have increased the success of the resection of primary
neoplasms.However, in the majority of patients with clinical
cancer,excluding those with skin tumors, metastasis has
alreadyoccurred at the time of diagnosis (53, 58, 59).
Therefore,even the most extensive surgical procedures cannot hopeto
bring about a high “cure―rate. Short of the completeprevention
of cancer, the urgent goal of the oncologistshould be the
prevention or the successful treatment ofmicrometastases.
Additional approaches to the therapy of disseminateddisease may
be forthcoming if the underlying pathogenesisof cancer metastasis
can be clearly understood. The outcome of the metastatic process
depends on the propertiesof both host and tumors, and the balance
of these contributions probably varies in tumor systems (15,
58).
Animal models have proven invaluable in the elucidationof the
host and tumor factors involved in determining theoutcome of cancer
metastasis. The proper utilization ofsuch models may advance
considerably our understanding
I Research sponsored by the National Cancer Institute under
Contract
NO1-CO-75380 with Litton Bionetics, Inc. Presented as the
lecture onAdvancesin Oncologyat the 69thAnnualMeetingof the
AmericanAssociation for CancerResearch,April 1978,Washington,D.
C.
Received May 24, 1978; accepted June 5, 1978.
of the biology of the phenomenon and, therefore, allow
thedevelopment of new approaches to the therapy of disseminated
disease.
We have recently developed one such animal model forqualitative
studies of metastasis. Mice are given i.d. injections, in the
external ear, of 0.05 ml of tumor cell suspension. Three to 4 weeks
later, when the tumors are established, the ear is amputated at its
base and the mice areallowed to survive. Six to 8 weeks later, the
mice are killedand examined for the presence of lymph node or
visceralmetastases. In the model shown in Fig. la, 25,000
viablecells of the B16 melanoma, syngeneic to the C57BL/6mouse,
were injected into the medial surface of the externalear. The ear
and growing tumor were amputated 3 weeksafter s.c. injection. If
the tumor mass contained no metastatic cells, the amputation of the
ear would be curative.Alternatively, if the growing tumor contained
some metastatic cells, which invaded blood vessels and
lymphaticsprior to amputation of the ear, the mice would
eventuallydie of distant tumor growths. As seen in Fig. lb, in
somemice growth of tumor occurred in the lymph nodes regionalto the
ear, i.e., the cervical and submandibular nodes, andin the lungs.
The components of this qualitative assay forthe metastatic
potential of murine neoplasms are shown inFig. lc. This model or
procedure may also be useful fordetermining the metastatic
potential of murine neoplasmsin studies that evaluate the therapy
of spontaneous metastases.
Pathogenesis of Metastasis
The first steps in the pathogenesis of metastasis (Chart 1)are
the local invasion of normal host tissues, penetrationinto
lymphatics or blood vessels by malignant cells, anddetachment. The
actual mechanism for tumor cell invasionremains unclear; it
probably depends on both tumor andhostfactors.
Tumor cells can spread by three major routes. The
firstinvolvesspreadbydirectextensioninwhich a tumor growing in a
body cavity releases cells or fragments that canseed serosal and/or
mucosal surfaces and develop into newgrowths. Two examples are lung
mediastinal tumors thatenter the pleural cavity and malignant
ovarian tumors thatshed cells into the peritoneal cavity. Primary
tumors of thecentral nervous system, although highly invasive,
rarelyproduce metastases in organs outside the nervous system.The
mode of their spread appears to be by direct extensionor via the
cerebrospinal fluid. The second and third routesof spread, which
are the subject of this lecture, are via thelymphatic and
hematogenous compartments of the circulatory system.
Clinical observations have suggested that carcinomasfrequently
spread and grow in the lymphatic system,whereas malignant tumors of
mesenchymal origin spread
SEPTEMBER 1978 2651
Tumor Heterogeneity and the Biology of Cancer Invasion and
Metastasis1
Isaiah J. Fidler
Cancer Biology Program, National Cancer Institute Frederick
Cancer Research Center, Frederick, Maryland 21701
on May 31, 2021. © 1978 American Association for Cancer
Research. cancerres.aacrjournals.org Downloaded from
http://cancerres.aacrjournals.org/
-
I. J. Fidler
Primary MalignantNeoplasm
Vascularization
Interactionwith HostPlatelets,LymphocytesandOther
BloodElements
J
LymphaticsVenulesCapIllaries
Chart 1. The pathogenesis of a metastasis. The outcome of the
sequential steps ofthe process depends on the interaction oftumor
cellswith their host.
Arrestin CapIllaryBedofOrgans
Lung
Metastases
more frequently via the hematogenous route (7). These
twocategories are rather arbitrary, however, because the bloodand
lymph systems are intimately interlinked. The studiesof del Regato
(7), Fisher and Fisher (18-20), Wood (60),Zeidman (61, 62), and
Zeidman and Buss (63) clearly demonstrated that cancer cells may
invade the lymphatics directly or may gain access to them via blood
vessels and thatcancer cells that invade lymphatics can find their
way toblood vessels via venolymphatic anastomoses or by way ofthe
thoracic duct.
The thin-walled venules, like the lymphatic channels,offer
little resistance to penetration by tumor cells and thusprovide the
most common pathway for the entry of tumorcells into the
circulation. In contrast, the arteries, the wallsof which contain
elastic and collagen fibers, are rarelyinvaded by tumor cells.
After infiltrating the vessels, tumorcells can detach and be
carried away passively in thebloodstream or remain localized and
proliferate at the siteof vessel invasion. Frequently, a thrombus
will form aroundactively growing tumor cells that have penetrated
the circulatory system , and it has been suggested by many
investigators (1, 3—5)that a direct correlation exists
betweenthegrowth potential of a malignant tumor and the formation
offibrin around it.
Detachment and embolization of tumor cells, regardlessof whether
their transport is via the lymphatics or blood,are probably
continuous processes (56, 57). Most malignanttumors have a
well-established blood supply with multiplethin-walled vessels, and
Franks (22) has suggested that thedevelopment of metastases begins
immediately with thegrowth of primary malignant cancers. In support
he citesthe large number of clinical cases in which primary
tumorswere surgically removed, and yet the patient succumbed
tometastatic lesions many years later. A sudden change in
venous pressure, such as that occurring during a cough,could
lead to momentary blood turbulence and the releaseof a shower of
emboli (61). Similarly, diagnostic proceduresand surgical trauma
may cause a sudden increase in thenumber of tumor cells released
into the circulation (20).
However, the presence of tumor cells in the circulationdoes not
always lead to a metastasis (15, 47, 53, 58). In anextensive
reviewof the literature on circulating tumor cells,Salsbury (47)
concluded that there is no evidence that themere presenceof tumor
cells in the circulation indicates aworse prognosis than the
absenceof tumor cells.
During circulatory transport, tumor cells can undergo avariety
of interactions, including aggregation with othertumor cells (39),
platelets (24), lymphocytes, and other hostcells (14, 39). Some
tumor cells are thromboplastic andelicit fibrin formation either
during their circulation or soonafter their arrest in capillary
beds (1, 3, 55, 60). If bloodborne tumor cells are aggregated by
homotypic or heterotypic cell interactions or by soluble blood
components intolarge emboli, their success in forming tumors after
theirarrest in the microcirculation should be increased (35).
It has been demonstrated experimentally that larger emboli are
more effective per input tumor cell in implantationand in survival
to form gross tumor colonies after i.v.injection (12).2 Thus,
purely mechanical factors such asembolic size and deformability (as
well as capillary diameterand deformability) should be important in
the implantationprocess (48). The rates at which tumor cells or
their cellemboli pass through capillary beds are not related to
cell oremboli size but instead appear to be related to their
deform
2 M. L. Kripke, E. Gruys, and I. J. Fidler. Metastatic
Heterogeneity of Cells
from an UltravioletLight-inducedMurine Fibrosarcomaof
RecentOrigin,submitted for publication.
CANCERRESEARCHVOL. 382652
Invasion Transport
Extravasatlon AdherenceofTumor Cells
EstablIshmentofMlcreenvlronmentandGrowthInto
on May 31, 2021. © 1978 American Association for Cancer
Research. cancerres.aacrjournals.org Downloaded from
http://cancerres.aacrjournals.org/
-
Tumor Heterogeneity and Metastases
were adapted to tissue culture. The B16 variants, whichgrew in
culture after the first in vivo selection, were established as a
continuous line and were designated B16-F1.Cells from this line
were reinjected into new syngeneicanimals, and 3 weeks later a new
group of lung tumorcolonies was removedand cultured to yield line
B16-F2andso on. With each succeeding cycle in vivo, the ability of
theselected B16 linesto implant, survive, and form lung
tumorsincreased (13). After 10 such selections B16-F10 was obtamed,
which forms significantly more gross lung tumorsper input cell than
does B16-F1 after i.v. or i.c.3 injectioninto syngeneic mice
(17).
Clinical observations of a large number of patients withtumors
of defined histological classification suggest thatthere is a
tendency for primary tumors arising in an organto metastasize to
and grow in particular distant organs (59).In many experimental
tumor systems, selective patterns ofhoming and subsequent growth
into metastases have beendemonstrated. For example, as early as
1947 Cloudman (6)showed differences betweenthe homing patterns of
murinelymphomas and plasmacytomas. Murine thymomas tend
tometastasize to and grow in all the lymphoid organs,
buttransplantable plasmacytomas tend to metastasize to thebone
marrow (43-45). Other tumors that have been reportedto demonstrate
organ specificity for metastasis are melanomas (17, 26, 30),
histiocytomas (9), and fibrosarcomas(52).
Thus, clinical and experimental observations of the predilection
of certain neoplasms for growth in specific organssuggest that the
pattern of metastasis is not random tumorcell distribution but
rather that it reflects properties of thecirculating malignant
cells, host capillary endothelium, andorgan environment. Two
long-standing arguments havebeen advanced to explain the
distinctive patterns of distribution of metastases. In 1928 Ewing
(10) initially suggestedthat metastasis is influenced by purely
“mechanicalfactors,―such as anatomic and hemodynamic factors in
thevasculature. In contrast, in 1889 Paget (42) suggested
the“seedand soil―hypothesis, which states that the
microenvironment of one organ may favor the arrest and subsequent
growth of tumor emboli, while the microenvironmentof another organ
may not.
More recently, a third hypothesis has emerged that proposes that
properties of the tumor cells themselves mayalso influence their
patterns of spread and growth. Aremarkable demonstration of tumor
cells homing to andgrowing in specific organs was first reported by
Kinsey(30),who used the Cloudman melanoma, and then by Sugarbaker
et a!. (52), who used a murine fibrosarcoma. In theirexperiments
the tumors shown to spread from a s.c. site tothe lungs were
implanted s.c. The tumor cells metastasizednot only to the lung but
also to small fragments of lungtissue implanted s.c. They did not
metastasize to otherorgan fragments implanted s.c. , which served
as controls.
In the B16 system, the ability of B16-F10 but not theoriginal
B16-FO to form lung metastases exclusively aftereither i.v. or i.c
injection, even in parabiotically joinedanimals, suggested that the
selection of tumor cells withpreferencefor a particular organ might
be possible. Recent
aTheabbreviationusedis:ic., intracardial.
ability during transcapillary transport (62, 63). Under
certainconditions, the adhesion of tumor cells to endothelium
insusceptible organs could lead to vessel wall damage andthe
subsequent accumulation of neutrophils. Since leukocytes commonly
passthrough the endothelium, tumor cellscould enter the
extravascular tissues by following pathwaysset by leukocytes that
have traversed the vessel wall (55).Alternatively, platelets can
aggregate at the site of tumorcell lodgement and release mediators
that could contributeto vascular spasm, increased endothelial
permeability and,perhaps, increased motility of tumor cells
(28).
The process of extravasation may be facilitated by aspecific
tumor cell product(s) that is chemotactic to othertumor cells but
not to normal cells (41). When malignantcells reach an
extravascular environment, they usually continue to proliferate
(Fig. 2). In malignant lesions, vascularization of the
micrometastases is probably enhanced by atumor angiogenesis factor,
a glycoprotein that stimulatesendothelial cell movement and
division (21).
Metastasis as a Selective Process
It is not possible to give an all-inclusive review of the
vastliterature on the clinical and experimental aspects of
cancermetastasis. Therefore, I have chosen to address only onemajor
issue which has recently received a great deal ofattention and
which I feel has far-reaching consequencesfor our understanding of
cancer biology in general and forour approaches to the therapy of
disseminated disease inparticular.
Specifically, I wish to discuss the possibility that neoplasms
are heterogeneous and contain subpopulations ofcells with differing
metastatic capabilities. We can ask,“Doesthe process of
metastasis represent the randomsurvival of tumor cells, or does it
result from the survivaland growth of a specialized subpopulation
of cells?―Beforeproceeding to consider this question, I shall
give you thehistorical background that led to its formulation.
Several years ago we investigated the fate of circulatingtumor
emboli following the i.v. injection of
[1251J-5-iodo-2'-deoxyuridine-labeled tumor cells. This technique
allows usto determine the distribution and survival of tumor cells
intheir recipient. Injected mice were killed at different times,and
their organs were collected and processedto determinethe number of
viable radioactive tumor cells. The majorityof the injected tumor
cells were arrested initially in thelungs. Tumor cell death began
shortly thereafter. By 24 hrafter injection, only 1% of the
injected tumor cells hadsurvived in host organs, and by Day 14, at
which time tumorcolonies were visible in the lungs, less than 0.1%
of theoriginal cells injected had survived (11). The fact that
themajority of the circulating emboli died and only a
minoritysurvived to yield metastases was not surprising but
raisedthe following question. Did the 0.1% of the emboli surviveat
random or did the surviving cells represent a subpopulation of
cells with properties that enhanced their survival andgrowth?
To answer this question we performed the followingexperiment.
The B16 melanoma was injected i.v. into syngeneic C57BL/6 mice.
Three weeks later pulmonary tumorcolonies were dissected free of
lung tissue, and tumor cells
SEPTEMBER1978 2653
on May 31, 2021. © 1978 American Association for Cancer
Research. cancerres.aacrjournals.org Downloaded from
http://cancerres.aacrjournals.org/
-
Table 1Heterogeneity of the B16 melanoma parent tumor for
experimental metastasis
Median no. of pulmo- No. of animals with extrapulmoSource of
cells° nary metastases nary metastases
I. J. Fidler
experiments by Brunson et a!. (2) have clearly demonstratedsuch
selection. In their studies a brain-preferring B16 melanoma linewas
selectedby repeatedcyclingin vivoofmelanoma cells obtained from the
rare brain metastasesthat developed initially following the i.c.
injection of theparent tumor. Similarly, Nicolson and Brunson (38)
haverecently reported the selection of tumor variants that canhome
to and grow in the adrenal gland or ovary afterapproximately l2in
vivo selections. These selected variantsare of immense importance,
for they may allow us toidentify the determinants responsible for
the preference oftumor growth for a particular organ. Collectively,
the abovedata suggest that the survival of a few tumor cells
thatsubsequently developed into metastases was not a
randomoccurrence but was due to certain unique properties of
thesurviving cells.
Heterogeneity of Murine Malignant Neoplasms
What remained unanswered was the question of whetherthese unique
metastatic cells preexisted in the tumor cellpopulation or whether
they arose during metastasis by aprocess of adaptation to local
environmental conditions. Ifhighly metastatic variant cells could
be shown to preexist inthe parent population, this would support
the suggestionby Nowell (40) that tumor cell variants arise within
developing tumors, are subjected to host selection pressures,
andare responsible for the emergence of new sublines withincreased
malignant potential.
To distinguish between these possibilities, my colleagueMargaret
Kripke and I performed an experiment similar indesign to the
classical fluctuation test devised by Luria andDelbrück (36) to
distinguish between selection and adaptation in the origin of
bacterial mutants. In our study a cellsuspension of the B16
melanoma parent line was divided
into two parts. One portion was used for i.v. injection
intosyngeneic C57BL/6 mice. The other portion was used toproduce 17
clones, which were then also injected i.v. intogroups of C57BL/6
mice. Eighteen days after the tumorcells were injected, the number
of lung metastases in eachrecipient was counted. If the number of
metastatic foci inthe lungs of the mice receiving the cloned
sublines wassimilar to the number of foci seen in mice receiving
theparent line, this would indicate that the parent populationwas
homogeneous and that the metastatic foci probablyresulted from
adaptation during the process of metastasis.Alternatively, if the
cloned sublines gave rise to widelydifferent numbers of lung
colonies, this would suggest thatthe parent tumor was heterogeneous
and that cells of bothhigh and low metastatic potential preexisted
in the parentpopulation.
As shown in Table 1, the study clearly demonstrated thatthe B16
melanoma is heterogeneous and that a few highlymetastatic tumor
cell variants preexisted in the parentalpopulation. There was also
considerable variation amongthe clones in the number and sites of
pulmonary metastasesproduced following i.v. injection. Control
subcloning experiments demonstrated that the variability among
theclones was not generated during the cloning procedure(16). The
extreme degree of heterogeneity observed withthe B16 melanoma is
not surprising considering that thetumor arose in 1954 and has been
transplanted for approximately 10 times the life span of a mouse.
Repeated passages, both in animals and in cell culture, have
providedample opportunity for variant cell types to arise. For
thisreason we wished to determine whether a tumor of a morerecent
origin would also exhibit heterogeneity by the critenon of
metastatic potential. To address this question weused a
fibrosarcoma induced in a C3H mouse by chronicUV irradiation (32).
Following the s.c. injection of cells, this
40.5 (8-131)@'B16 parent line(60)b
Clone 16(10)Clone15(11)Clone 12(9)Clone24 (9)Clone 19(10)Clone7
(10)Clone21 (8)Clone 18(11)Clone 5 (10)Clone6 (9)Clone 17(9)Clone3
(9)Clone 1 (9)Clone 2(10)Clone 13(9)Clone 14 (9)Clone9 (10)
8/60 ovary, 11/60 lymph nodes, 6/60 liver, 4/60 kidney,3/60 gut,
2/60 adrenal
0/101/11 lymph node0/91/9 ovary, 1/9 liver, 1/9 lymph
node0/100/101/8 lymph node0/110/100/90/91/9 lymph node0/90/102/9
ovary, 1/9 liver2/9 ovary6/10 lymph node, 2/10 adrenal, 1/
10 kidney, 2/10 liver
3.556
101317183645.599
150214237254.5260
>500>500
(2-15)(2-20)(0-34)(5-29)(0-42)(0-43)(1-48)(0-91)(2-171)(5—232)
(104—210)(160-450)
(73-321)(7-450)
(50-350)
a C57BL/6miceweregiveninjectionsin the tail veinof
50,000viablesinglecellsandkilled 18 days later. The number of
pulmonary tumor colonies was determined with theaid of a dissecting
microscope.
b Numbers in parentheses, number of mice per group.C Numbers in
parentheses, range.
2654 CANCERRESEARCHVOL. 38
on May 31, 2021. © 1978 American Association for Cancer
Research. cancerres.aacrjournals.org Downloaded from
http://cancerres.aacrjournals.org/
-
Heterogeneityofthe
UV-2237fibrosarcomaparenttumorforexperimentalmetastasisSourceof
ceIls@'Medianno. of pulmo
nary metastasesNo.of extrapulmonarymetastasespbUV-2237
parentline (34)(@1
60.5(17-300)@'Ø@i34Clone
15(10)1(0-9)0/10
-
Table 3Spontaneousmetastasisof UV-2237parent tumor and its
clone
subpopulations
Growth of s.c. tumors@@No.of mice
with metasta- Av. sizeat 2 wkSourceof cells se$a Incidence
(Cumm)ParentUV-2237 2/5 5/5 382Clone46 0/5 0/5 0CIone38 0/5 1/5
0Clone42 0/5 4/5 0CIonel5 0/5 3/5 16Clone18 0/5 5/5 95Clone3l 0/5
5/5 437CIone44 1/5 5/5 389Clone43 1/5 5/5 465Clone 9 1/5 5/5
616Clonel2 2/5 4/5 54Clone 22 2/5 5/5 237Clone30 2/5 5/5 272Clone
47 2/5 5/5 566Clone39 2/5 5/5 853Clone4l 3/5 4/5 10Clone 27 3/5 5/5
401CIone26 3/5 5/5 766Clone 5 3/5 5/5 824CIone33 4/5 5/5 132Clone
25 4/5 5/5 132Clone34 4/5 5/5 245
a C3H mice were given s.c. injections on the flank of 1 x
10'viable cells. The number of mice with metastaseswas determinedat
time of death or 5 months after injection.
b Mice without tumors were observed for 5 months.
absolute correspondence among the three tests. UV-2237clone 12
produced the highest number of experimentallung metastases
following i.v. injection but was only intermediate in its ability
to metastasizefrom a s.c. site. Thus,clone 12 appears to be a clone
that is adept at surviving inthe circulation to form lung
metastasesbut maybedeficientin its invasive capabilities. Clone 12,
however, appears tobe the exception and thus the significant degree
of correspondence among the three tests for the other 20
clonespermits us to conclude that at least for this UV-2237
fibrosarcoma the convenient experimental
metastasisassayapproximates rather well the more tedious
measurementsofspontaneous metastasis.
Summary
In summary I wish to state that the heterogeneous natureof
neoplasms with regard to a large number of characteristics has been
well recognized. Populations of human andanimal neoplasms
frequently demonstrate a great variationin chromosome number and
DNA content (37, 54). Tumorsare known to be heterogeneous with
regard to their antigenic properties (46), immunogenicity (29),
hormone receptors (51), pigment production (25), metabolic
characteristics (31), and growth rate (49), as well as in their
susceptibility to a variety of cytotoxic drugs (23, 27). Not
surprisinglythen, we have now demonstrated that neoplasms are
alsoheterogeneouswith regard to invasion and metastasis,i.e.,that
they contain a variety of subpopulations of cells withdiffering
metastatic potentials.
The development of a metastasis appears to be dependent on an
interplay between host factors and intrinsiccharacteristics of the
tumor cells. The processof metastasis
HeterogenemTable
414'of UV-2237fibrosarcomafor tumor growth andetastasis
following i.v. injection ofcellsNo.of
Time ofdeathNo.ofmicewith (days)
mice with extrapulpulmo-monarySource
ofcellsnary
me- metastatastases sest' MedianRangebParent
223719/19 17/19 2118—26Clone4l3/50/5 15283-152Clone460/51/5
15257-152CIone472/51/5 15255-152Clone384/51/5
15250-152Clonel53/51/5 8636-152CIone424/51/5 6342-152Clone
125/5 2/5 4237-73Clone185/5 3/5 5739-91Clone55/5 3/5
3634-127Clone
315/5 3/5 2727Clone305/5 4/5 4937-81Clone335/5 4/5
4525-82Clone445/5 4/5 3734—44Clone265/5 4/5 2420-24Clone275/5 5/5
3833-38Clone435/5 5/5 3624-58Clone345/5 5/5 3624-36Clone395/5 5/5
3434-92Clone255/5 5/5 3428-36Clbne95/5 5/5 2727-36Clone
224/5 5/5 24 22-34of 1 x 10' viablecells.
I. J. Fidler
a C3H mice were given i.v. injectionsMetastaseswere determined
at autopsy.
b Surviving mice were killed and autopsied on Day 152
afterinjection.
is highly selective, and the metastatic lesion represents theend
point of several destructive events that only a few cellscan
survive.Only a few cells within a primary neoplasmmayactually
invade blood vessels, and of these even fewer willsurvive
transport, can attach firmly to the endothelium ofsmall blood
vessels,will undergo extravasation, will evadehost defenses, and
will grow into tumor foci.
We may conclude the following: (a) the outcome ofmetastasis is
dependent to a large extent on a selectionprocess that favors the
survival and growth of a specialsubpopulation of cells. Therefore,
studies that compareproperties of tumor cells obtained from primary
tumors andthose obtained from their metastases may be helpful
towardthe identification of some of the properties of
metastasticcells; (b) the possible existence of highly metastatic
variantcells within a primary tumor suggeststhat we no longer
canconsider a neoplasmto be a uniform entity. Indeed, effortsto
design effective therapeutic agents and proceduresshould be
directed toward the few but fatal metastaticsubpopulations.
Therapeutic efforts that are directedagainst all neoplastic cells
without regard to their biologicalbehavior in vivo may be
unproductive; (C) tumor variantswith differing metastatic
potentials selected in a variety oftumor systems could be a useful
tool for answering questions regarding the biology of metastasis
and in particularfor testing new therapeutic approaches to
cancer.
Acknowledgments
This presentation would not have been possible were itnot for
the contributions of numerous investigators in thefield of
metastasis. In particular, I wish to acknowledge my
2656 CANCER RESEARCH VOL. 38
on May 31, 2021. © 1978 American Association for Cancer
Research. cancerres.aacrjournals.org Downloaded from
http://cancerres.aacrjournals.org/
-
Tumor Heterogeneity and Metastases
teacher and friend, Dr. Irving Zeidman, for introducing meto and
guiding me in the field of the biology of cancermetastasis.
References1. Baserga, R., and Saffiotti, U. Experimental Studies
on Histogenesis of
Blood-BorneMetastases.Arch. Pathol.,59: 26-34,1965.2. Brunson,
K. E., Beaftie, G., and Nicolson, G. L. Selection and Altered
Properties of Brain-Colonizing Metastatic Melanoma. Nature, 272:
543-545, 1978.
3. Chew, E. C., Josephson, R. L., and Wallace, A. C. Morphologic
Aspectsof theArrestof CirculatingCancerCells.In: L.
Weiss(ad.),FundamentalAspects of Metastasis, pp. 121-150.
Amsterdam: North Holland PublishingCo., 1976.
4. Cliffton, E. E., and Agostino, D. Factors Affecting the
Development ofMetastaticCancer.Effectof Alterationsin Clotting
Mechanism.Cancer,15: 276-283, 1962.
5. Chiffon, E. C., and Grossi, C. E. The Rationale of
Anticoagulants in theTreatmentof Cancer.J. Med.,5:
107-113,1974.
6. Cloudman, A. M. Organophilic Tendencies of Two Transplantable
Tumorsof the Mouse.CancerRae.,7: 585-591, 1947.
7. del Regato, J. A. Pathways of Metastatic Spread of Malignant
Tumors.SeminarsOncol.,4: 33—38,1977.
8. Dorland, W. A. Dorland's Illustrated Medical Dictionary, Ed.
24. Philadelphia: W. B. Saunders Co., 1965.
9. Dunn,T. B. Normaland PathologicAnatomyof the
ReticularTissueinLaboratoryMice,with a Classificationand
Discussionof Neoplasms.J.Natl. Cancer Inst., 14: 1281-1333,
1954.
10. Ewing, J. Neoplastic Diseases,Ed. 3, Chap. 4.
Philadelphia:W. B.SaundersCo., 1928.
11. Fidler, I. J. Metastasis:QuantitativeAnalysisof
Distributionand FateofTumor Emboli Labeled with
“I-5-lodo-2'.deoxyuridine. J. NatI. CancerInst., 45:
733—782,1970.
12. Fidler, I. J. The Relationship of Embolic Homogeneity,
Number, Size andviability to the Incidence of Experimental
Metastasis. European J.Cancer,9: 223-227,1973.
13. Fidler, I. J. Selection of Successive Tumor Lines for
Metastasis. NatureNew Biol., 242: 148-149, 1973.
14. Fidler, I. J. Immune Stimulation-Inhibition of Experimental
Cancer Metastasis.CancerAss.,34: 491-498,1974.
15. Fidler, I. J. Mechanisms of Cancer Invasion and Metastasis.
In: F. F.Baker (ed), Cancer:A ComprehensiveTreatise,Vol. 4, pp.
101-131.New York: Plenum Press, 1975.
16. Fidler, I. J., and Krlpke, M. L. Metastasis Results from
PreexistingVariantCellswithin a
MalignantTumor.Science,197:893-895,1977.
17. Fidler, I. J., and Nicolson, F. L. Organ Selectivity for
Survival and Growthof B16 Melanoma Variant Tumor Lines. J. NatI.
Cancer Inst., 57: 1199-1202, 1976.
18. Fisher, B., and Fisher, E. R. The Interrelationship of
Hematogenous andLymphatic Tumor Cell Dissemination. Surg. Gynecol.
Obstet., 122: 791-798, 1966.
19. Fisher, B., and Fisher, E. R. The Organ Distribution of
Disseminated“Cr-labeledTumor Cells. Cancer Rca., 27: 412-420,
1967.
20. Fisher, E. R., and Fisher, B. RecentObservationson the
ConceptofMetastasis.Arch. Pathol.,83:321-324,1967.
21. Folkman, J. Tumor Angiogenesis. In: F. F. Baker (ad.),
Cancer: AComprehenisve Treatise, Vol. 3, pp. 355-388. New York:
Plenum Press,1975.
22. Franks, L. M. Structure and Biological Malignancy of Tumors.
In: S.Garattini and G. Franchi (ads.), Chemotherapy of Cancer
DisseminationandMetastasis,pp. 71-78.NewYork: RavenPress,1973.
23. Fugi, H., and Mihich, E. Selection for High Immunogenicity
in DrugResistant Sublines of Murine LymphomasDemonstratedby
PlaqueAssay. Cancer Res., 35: 946-952, 1975.
24. Gasic, G. J., Gasic, T. B., Galanti, N., Johnson, T., and
Murphy, S.Platelet-TumorCell Interaction in Mice. The Role of
Plateletsin theSpreadof MalignantDisease.Intern.J.
Cancer,11:704—718,1973.
25. Gray, J. M., and Pierce, G. B. Relationship between Growth
Rate andDifferentiation of Melanomaln Vivo. J. NatI. Cancerlnst.,
32: 1201-1211,1964.
26. Greene,H. S. N., and Harvey,E. K. The
RelationshipbetweentheDissemination of Tumor Cells and the
Distribution of Metastases. CancerRes., 24: 799-811, 1964.
27. Hakannson,L., and Troupe, C. On the Presencewithin Tumors
ofClones That Differ in sensitivity to Cytostatic Drugs. Acta
Pathol.Microbiol.Scand.A,82:32-40,1974.
28. Hilgard,P.TheRoleof BloodPlateletsin
ExperimentalMetastases.Brit.J. Cancer,28:429-436,1973.
29. Killion, J. J., and Kollmorgen, G. M. Isolation of
Immunogenic TumorCellsbyAffinity
Chromatography.Nature,259:674-676,1976.
30. Kinsey, D. L. An Experimental Study of Preferential
Metastasis. Cancer,13: 674-676, 1960.
31. Kirccuta, I., Mustea, I., Rogozaw, I., and Simu, G.
Relations betweenTumor and Metastases.I. Aspectsof the
CrabtreeEffect. Cancer,18:978-984, 1965.
32. Krlpke,M. L. Latency,HistologyandAntigenicityof
TumorsInducedbyUftraviolet Light in Three Inbred Mouse Strains.
Cancer Roe., 37: 1395-1400, 1977.
33. Kripke, M. L., Fidler, I. J., and Gruys, E. Heterogeneity of
MetastaticPotential in Cells. Proc. Am. Assoc. Cancer Res.. 19:
213, 1978.
34. LiII, P. H., and Kripke, M. L. Growth Patterns of
Ultraviolet Light-InducedFibrosarcomas in Subcutaneous, Peritoneal,
and Vascular Compartments of Syngeneic Recipients. Transplantation,
25: 86-87, 1978.
35. Liotta, L. A., Kleinerman,J., and Saidel, G. M. The
Significance ofHematogenousTumor Cell Clumps in
the-MetastaticProcess.CancerRes.,36: 889-894,1976.
36. Luria, S. E., and DelbrOck, M. Mutations of Bacteria from
Virus SensitivIty to virus Resistance. Genetics, 28: 491-51 1,
1943.
37. Makino, 5. The Chromosome Cytology of the Ascites Tumor of
Rats,with SpecialReferenceto the Conceptof the StemlineCell.
Intern.Rev.Cytol., 6: 26-84, 1957.
38. Nicolson, G. L., and Brunson, K. w. Organ Specificity of
Malignant 816Melanomas:In vivo Selection for Organ Preferenceof
Blood-BorneMetastasis.GannMonographCancerRae.,20: 15-24,1977.
39. Nicoison,G.L.,andWinkelhake,J. L.OrganSpecificityof
Blood-BorneTumorMetastasisDeterminedby
CellAdhesion?Nature,255:230-232,1975.
40. Nowell, P. C. The Clonal Evolution of Tumor Cell
Populations. Science,194: 23-28. 1976.
41. Ozaki, T., Yoshida, K., Ushijima, K., and Hayashi, H.
Studies on theMechanismsof Invasion in Cancer. II. In vivo Effects
of a
FactorChemotacticforCancerCells.Intern.J.Cancer,7:93-100,1971.
42.
Paget,S.TheDistributionofSecondaryGrowthinCanceroftheBreast.Lancet,
1: 571-573, 1889.
43. Parks, R. C. Organ-Specific Metastasis of a Transplantable
ReticulumCellSarcoma.J. NatI.CancerInst.,52:971-973,1974.
44. PIlgrim, H. I. The Kinetics of the Organ Specific Metastasis
of aTransplantable Reticuloendothelial Tumor. Cancer Res., 29:
1200-1205,1969.
45. Potter, M., Fahey, J. L., and Pilgrim, H. I. Abnormal Serum
Protein andBone Destruction in Transplantable Mouse Plasma Cell
Neoplasm (Multiple Myeloma). Proc. Soc., Exptl. Biol. Med., 94:
327-333, 1957.
46. Prehn, R. T. Analysis of Antigenic Heterogeneity within
Individual 3-Methylcholanthrene-InducedMouseSarcomas.J.
NatI.CancerInst.,45:1039-1045,1970.
47. Salsbury,A. J. TheSignificanceof the CirculatingCancerCell.
CancerTreat.Rev.,2: 55-72,1975.
48. Sato, H., and Suzuki, M. Deformability and Viability of
Tumor Cells byTranscapillary Passage, with Reference to Organ
Affinity of Metastasisin Cancer. In: L. Weiss (ed.), Fundamental
Aspects of Metastasis,
pp.311-317.Amsterdam:NorthHollandPublishingCo., 1976.
49. Schabel,F. M., Jr. Conceptsfor SystemicTreatmentof
Micrometastases. Cancer, 35: 15-24, 1975.
50. Siegel, 5. Nonparametric Statistics for the Behavioral
Sciences. NewYork:McGraw-HillBookCo., 1956.
51. Sluyser, M., and VanNie, R. Estrogen Receptor Content and
Hormoneresponsive Growth of Mouse Mammary Tumors. Cancer Res., 34:
3253-3257,1974.
52. Sugarbaker,E. V., Cohen,A. M., and Ketcham,A. S. Do
MetastasesMetastasize? Ann. Surg., 174: 161-166, 1971.
53. Sugarbaker, E. V., and Ketcham, A. 5. MechanIsms and
Prevention ofCancer Dissemination: An Overview. Seminars Oncol., 4:
19-32, 1977.
54. SuzukI, N., Frapart, N., Grdina, D. J., Meistnch, N. L., and
Withers, H.R.CellCycleDependencyof Metutatic
LungColonyFormation.CancerRes.,37:3690-3693,1977.
55. Warren, B. A. Environment ofthe Blood-Borne Tumor Embolus
Adherentto Vessel Wall. J. Med., 4: 150-177, 1973.
56. Weiss, L. The Cell Periphery, Metastasis and Other Contact
Phenomena.Frontiers Biol., 7:289-332,1967.
57. Weiss,L. BiophysicalAspectsof the MetastaticCascade.In: L.
Weiss(ad.), Fundamental Aspects of Metastasis, pp. 51-70.
Amsterdam: NorthHollandPublishingCo., 1976.
58. WeIss, L. A Pathobiologic Overview of Metastasis. SemInars
Oncol., 4:5-17, 1977.
59. Willis, R. A. The Spread of Tumors In the Human Body.
London:Butterworths, 1972.
60. Wood, S., Jr. Experimental Studies of the Intravascular
Dissemination ofAscitlc V2 Carcinoma Cells in the Rabbit. with
Special Reference toFibrinogen and Fibrinolytic Agents. Bull
Schweiz. Akad. Med. Wise., 20:92-121, 1964.
61. Zeldman,I. Metastasis:A Reviewof
RecentAdvances.CancerRes.,17:157-162, 1957.
62. Zeidman,I. The Fateof CirculatingTumorCells.I. Passageof
Callsthrough Capillaries. Cancer Res., 21: 38-39, 1961.
63. Zeldman, I., and Buss, J. M. Transpulmonary Passage of Tumor
CallEmboli.CancerRes.,12:731-733,1952.
2657SEPTEMBER1978
on May 31, 2021. © 1978 American Association for Cancer
Research. cancerres.aacrjournals.org Downloaded from
http://cancerres.aacrjournals.org/
-
ic
Fig. 1. Anin vivo assayto determinethe metastaticpotentialof a
transplantableneoplasm.B16melanomacellsare injecteds.c. into
theexternalearof aC57BL/6mouse.Threeweekslaterwhenthe tumor is
established(a),the ear is amputated.Six
weeksthereafter,metastasesin the draining regionallymphnodeaswell
as in the lungsareevident (b). Thecomponentsof this assayare shown
in C.
2658 CANCER RESEARCHVOL. 38
on May 31, 2021. © 1978 American Association for Cancer
Research. cancerres.aacrjournals.org Downloaded from
http://cancerres.aacrjournals.org/
-
Tumor Heterogeneity and Metastases
‘@@
.‘
P1hZ,
p
2d
-1a@@
.@@ @â€Ĩb @‘4@
Fig. 2. Theformationof
anexperimentalpulmonarymetastasis.B16melanomacellswereinjectediv.
into syngeneicmice.Theinjectedmicewerekilledat different time
points thereafter, and their lungs were fixed and serially
sectioned. Tumor cells (arrows, a to i) can be identified by their
melanin content. a,immediate (initial) arrest of tumor cells in the
microvasculature; b, 1 hr postinjection; C, 4 hr postinjection; d,
1 day postinjection; e, 2 days postinjection(note cell division);
f, 4 days postinjection; g, 6 days postinjection; h, 9 days
postinjection; I, 12 days postinjection;j, 18 days postinjection. a
to i, x 450;j,x 100. H & E.
I
,
F
2f.E@
SEPTEMBER 1978 2659
1@@
‘l@.
(f'....JII1p S,@ .1@
.@
i@@r
on May 31, 2021. © 1978 American Association for Cancer
Research. cancerres.aacrjournals.org Downloaded from
http://cancerres.aacrjournals.org/
-
I. J. Fidler
a
@2I
2660 CANCERRESEARCHVOL. 38
on May 31, 2021. © 1978 American Association for Cancer
Research. cancerres.aacrjournals.org Downloaded from
http://cancerres.aacrjournals.org/
-
1978;38:2651-2660. Cancer Res Isaiah J. Fidler MetastasisTumor
Heterogeneity and the Biology of Cancer Invasion and
Updated version
http://cancerres.aacrjournals.org/content/38/9/2651
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
[email protected] 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/38/9/2651To request
permission to re-use all or part of this article, use this link
on May 31, 2021. © 1978 American Association for Cancer
Research. cancerres.aacrjournals.org Downloaded from
http://cancerres.aacrjournals.org/content/38/9/2651http://cancerres.aacrjournals.org/cgi/alertsmailto:[email protected]://cancerres.aacrjournals.org/content/38/9/2651http://cancerres.aacrjournals.org/