Characterization of Two Highly Metastatic Variants of Lewis Lung Carcinoma withDifferent Organ Specificities1
Pnina Brodt2
Department oj Surgery, Division of Surgical Research, McGill University, Montreal, Quebec H3A ¡A4,Canada
ABSTRACT
The biological properties and metastasis of two sublines of the Lewislung carcinoma (3LL) which have maintained a stable pattern of organ-selective metastasis have been studied. Subline M-3LL, a lung-specificvariant which originated from a lung metastasis of the parent line,metastasized only to the lung following injection of lO'-IO'1 tumor cells
The phenomenon of organ-specific metastasis, namely thepreferential metastasis of some neoplastic cells to selectedsecondary sites, has been recognized for many years throughthe study of human and experimental tumors (3). The mechanisms which control target organ selection by the tumor cellsare, however, poorly understood.
Nonspecific factors such as physical proximity of the tumorto a target organ and/or favorable hemodynamic forces (themechanistic view) (4) have been proposed by some. Others havesuggested that specific host and/or tumor-dependent factorsdetermine the site of secondary growth. Among these factors,cell-cell adhesion between the tumor cells and the capillaryendothelium (5) or parenchymal cells (6, 7) of a target organ aswell as organ-specific microenvironmental factors (8) have beenproposed. The available experimental and clinical data suggestthat organ specificity may not be determined by a single hostor tumor cell property but rather that it is influenced by aninterplay of several of these mechanisms (3, 9).
One of the problems which has contributed to the slowprogress to date in elucidating the cellular and molecular basisof these mechanisms has been the scarcity of appropriate experimental tumor models. Studies of organ-specific metastasis
Received 8/6/85; revised 12/5/85; accepted 12/12/85.The costs of publication of this article were defrayed in part by the payment
of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.
and from the Cancer Research Society of Montreal; presented in part at the 75thand 76th Annual Meetings of the American Association for Cancer Research.May 1984 and May 1985(1,2).
2 Medical Research Council of Canada Scholar.
have had to rely mainly on the analysis of: (a) tumor lines whichwere of independent origin and displayed different organ-colonizing potentials in vivo (10) or (b) variant sublines derivedfrom a common tumor and selected for organ-specific metastasis following repeated systemic or intraorgan injection of highdoses of the parent line (11-13). In the former, the study oftumor and host factors related to the organotropic behavior ofthe tumor cells is complicated due to the fact that the tumorlines under study are likely to differ in a variety of tissue andtumor-specific properties unrelated to their patterns of organspecificity. In the latter, on the other hand, tumor cell survivaland growth in the target organs may be the result of systemicor intraorgan inoculation of high doses of tumor cells. Theseconditions may overwhelm and therefore mask local resistantmechanisms operative in the target organ (14).
Line H-3LL was established by the s.c. grafting of a rare spontaneoushepatic metastasis detected in a mouse bearing a 3LL tumor s.c. in theaxillary region. Following implantation of this nodule into a recipientanimal, it gave rise to a local tumor. The tumor was resected, tumorcells were dispersed by enzymatic digestion with trypsin (14), and 5 xIO5 cells were injected into two recipient mice. Numerous hepatic
ORGAN-SELECTIVE METASTASIS OF LEWIS LUNG CARCINOMA VARIANTS
derived from tumor-bearing mice in the axillary region or the flank.For the experiments described, tumor cells from the third implantgeneration and onward were utilized.
Maintenance of Tumor Lines. Single cell suspensions of tumor cellswere obtained from solid s.c. tumors by digestion with 0.02% trypsinin Ca2+- and Mg2+-free PBS-EDTA as we described elsewhere (16).
Dispersed viable cells were enumerated using trypan blue dye exclusion.If required, cells were then cultured in RPMI-FCS. The cells weremaintained in vitro as monolayer cultures for periods not exceeding 2wk prior to their use in the experiments described. Single cell suspensions of the adherent tumor cells were obtained by incubating the cellmonolayer for 5-10 min at 37°Cwith PBS-EDTA. Enzymatic digestion
with 0.02% trypsin was subsequently applied only if required.Tumor Cell Cloning. Tumor cells dispersed from solid tumors were
cultured in vitro in tissue culture flasks (Falcon; 75-cm2 growth area)
for 2 wk prior to cloning. Cell monolayers were fed with fresh medium24 h prior to assay and then dispersed as described above, resuspendedin RPMI-FCS medium, and seeded in 96-well microtiter plates (Falcon;Microtest III) at a concentration calculated to deliver 0.7 cells/well(16). Colonies appeared within 10-14 days later. They were expandedin tissue culture flasks (Corning; 25-cm2 growth area) and maintained
in culture for several weeks (2-3 passages) prior to their injection s.c.or i.v.
Histology. Tissue to be analyzed was fixed in either Bouin's fixative
(lung) or 10% phosphate-buffered formalin (liver, lymph nodes, tumor).Five-urn paraffin sections were prepared and stained with hematoxylinand eosin (17).
Evaluation of Tumor Growth. Tumors growing s.c. were measuredwith calipers. The mean tumor diameter for individual tumors wascalculated from measurements in two planes at right angles. To determine mean tumor diameter for a group of mice, the sum of theindividual measurements was divided by the number of tumor-bearing
tastases when sacrificed. All organs were routinely screened, althoughmetastatic foci were normally detected only in the lung and/or the liver.Lung foci were enumerated following immersion of the lungs in Bouin's
fixative for 24 h. Liver colonies were enumerated immediately followingremoval of the organs. Livers were subsequently fixed in 10% phosphate-buffered formalin for further analyses. Mean diameter of thecolonies was determined as described by Wexler (18).
Grafting of Lymph Nodes. Axillary and/or brachial lymph nodeswere removed aseptically from tumor-bearing mice 15-40 days following the s.c. injection of 2 x 10s tumor cells or the implantation of ametastatic nodule in the axillary' region. They were kept at 4°Cinsterile Hanks' solution until used. Syngeneic recipient mice were anes
thetized with ether, and a small (0.5-1 cm) skin incision was made inthe lower axillary region. Lymph nodes (one per animal) were inserteds.c. through the incision, and the wound was closed with surgicalstainless steel clips (Autoclips; 9 mm; Clay Adams, Inc., Parsippany,NJ). The time of tumor appearance was determined by palpation of theanimals twice weekly, and tumors were measured with calipers asdescribed.
Radiolabeling of Tumor Cells. Tumor cells were radiolabeled by theprocedure we described previously (19) with slight modifications.Briefly, tumor cell cultures were prepared 1 wk prior to assay fromsolid in vivo tumors as described above. The resultant monolayers weredispersed 48 h prior to assay, and cells were seeded in new tissue cultureflasks (Corning; 25-cm2 growth area) at a concentration of 2 x 10*
viable cells/ml in 10 ml of RPMI-FCS medium. Cells were incubatedfor 24 h at 37°Cin a 5% CO2 incubator, at which time FdUrd (SigmaChemical, St. Louis, MO) was added to a final concentration of IO"5M and ['"IJdUrd (5 Ci/mg; Amersham Corp., Oakville, Canada) to afinal concentration of 0.5 /i('i/nil. The cells were incubated for 18
additional h, the monolayers were dispersed with PBS-EDTA, and thecells were washed 4 times in RPMI medium. Tumor cell viability and'"I uptake were then monitored prior to i.v. injection of the cells into
Light microscopy analysis of tissue sections of Tumors M-3LL and H-3LL was undertaken to assess the degree of morphological homology between the variant lines. As can be seenin Fig. 1, the majority of cells of both tumor variants retainedthe characteristic morphology of the parent line (Fig. 1, a-c)originally classified as a "poorly differentiated epidermoid carcinoma" (15). An increased incidence of multi- and uninu-
The highly metastatic clonal lines H-59 and M-27 weresubsequently selected for further analysis. Growth rates of thesesublines were compared following the s.c. injection of 10*tumor
cells into recipient animals. As evidenced in Fig. 2, there wereno significant differences between the incidence or rate ofgrowth of the tumors in the two experimental groups.
Subsequently, the patterns of spontaneous metastasis of thesesublines were studied. Local tumors were derived by either s.c.injection of dispersed tumor cells or by the implantation of
Table 1 Experimental metastasis of Tumor 3LL sublinesSeven female C57BL/6 mice in each group were given injections i.v. of 10'
tumor cells. Sixteen days later, all mice including two untreated control animalswere given injections i.p. of 25 /ig of FdUrd, followed 1 h later with 1 /iCi of 1251-dUrd. Animals were sacrificed 24 h later, and their organs were removed andplaced in Bouin's fixative (lungs) or 10% phosphate-buffered formalin (livers) for
count of isotope uptake in a gamma counter.
Organ examined
LiverTumor
injectedH-3LL
M-3LLNilNo.
ofnodules/
animal"68(44-102)*
0cpm(10-3)/
animal18(9.1-19)
3 (2.8-4.4)2.5 (2.2-2.8)LungNo.
ofnodules/
animal"16
(4-97)39 (15-» 100)'cpm(IO-3)/
animal1.4(0.5-17)
5.3(1-18)0.5
" Results are expressed as median of counts in seven animals.* Numbers in parentheses, range.'The score of s>100 was given when the nodules were too numerous for an
ORGAN-SELECTIVE METASTASIS OF LEWIS LUNG CARCINOMA VARIANTS
f
Fig. 1. Light microscopic appearance of Tumor 3LL and its highly metastatic variants. Cross-sections derived from the following s.c. tumors are shown: a, theparent Lewis lung carcinoma line: b, lung-specific subline M-3LL: and c, liver-homing variant H-3LL. H & E, x 500. A large multinucleated cell (arrow) and a secondlarge cell with a central nucleus and clear cytoplasm (two arrows) typically seen in sections of Tumor H-3LL can be seen (c). A similar cell can also be seen (d, arrow)in a second section of H-3LL. Periodic acid-Schiff reagent, x 500. One-wk-old tissue culture monolayers of M-3LL (e) and H-3LL (/) are also shown, x 425.Arrowheads in /show two large atypical cells seen characteristically in cultures of H-3LL. The same cells can be seen at a higher magnification in g. x 700.
ORGAN-SELECTIVE METASTASIS OF LEWIS LUNG CARCINOMA VARIANTS
Table 2 Blood-borne metastasis of cloned sublines derived from Tumor M-iLLFive animals in each group were given injections in the tail vein of 2.5 x 10*-
5x10* tumor cells. Animals were sacrificed 21-22 days later, and metastaticnodules in the lungs were enumerated following fixation in Bouin's fixative.
" Median of five organs.* Numbers in parentheses, range.
Table 3 Blood-borne metastasis of cloned sublines of Tumor H-3LL
In each group, three to five female C57BL/6 mice were given injections i.v.(tail vein) of 5 x 10' tumor cells. Animals were sacrificed 19 days later, and
" Median of three to five livers.* +, low; ++, moderate; +++, high; ++++, very high.' Numbers in parentheses, range.
2.O
1.6
— 1.2
o>
oTÕ
0.8
0.4¿Õ
10 15 20Time
25days)
30 35 40
Fig. 2. Growth curves of clona! lines M-27 and H-59. Five animals in eachgroup received a s.c. inoculation of IO5tumor cells. All animals developed tumorsby Day 14. O, Tumor M-27; A, Tumor H-59.
individual metastatic nodules isolated from the lungs (M-27)or livers (H-59) of tumor-bearing mice. Animals were sacrificed4-5 wk later when their tumors measured approximately 2 cmin diameter.
While screening tumor-bearing mice for metastatic lesions,we noted that the regional lymph nodes draining the tumor sitewere enlarged in mice bearing Tumors H-3LL or H-59. Thisenlargement which was apparent from the third week onwardafter transplantation of the tumors was not seen in animalsbearing Tumors M-3LL or M-27. Light microscopy analysis ofthin sections derived from several of these enlarged lymphnodes indicated that tumor cells had invaded the nodes (Fig.4).
To measure the incidence and time course of lymphaticmetastasis in tumor-bearing mice, axillary and/or brachialnodes were removed from tumor-bearing mice between 15 and40 days following the s.c. injection of 2 x IO5 tumor cells or
the implantation of individual metastatic nodules in the rightaxillary region. Tumors measured 0.8-2.0 cm at time of lymphnode excision. The excised lymph nodes were implanted intonew recipient mice, and the appearance of tumors was monitored for up to 6 mo.
may have been caused by a direct intralymphatic inoculation oftumor cells (22).
In an attempt to determine whether different potentials toarrest in the vasculature of the lung and liver regulated theorgan-specific patterns of metastasis noted in the present tumormodel, [125I]dUrd-labeled cells of H-59 and M-27 were injected
i.v., and their organ distribution was assessed by monitoringthe retention of ['25I]dUrd in the different organs for up to 24
from tumor-bearing mice. Animals bearing M-27 tumors were sacrificed 26-46 days following the s.c. injection of IO'-5 x 10* dispersed tumor cells or the
enumerated immediately following removal of the organs.* Seven animals were analyzed in each group with H-59 tumors and five in each group with M-27 tumors.' Mean ±SD.d Numbers in parentheses, range.' The difference between this group and that implanted with pulmonary nodules of Tumor M-27 or given injections of dispersed tumor cells of line H-59 was
of a liver derived from an animal bearing Tumor H-59 (1.7 cm) 27 days after the s.c. injection of IO6 tumor cells. Note multiple micro-metastases as well as one large lesion (topright), x 40. c, a higher magnification (x 400)of hepatic micrometastases. Normal hepato-cytes can be seen on the right, d, typical lungderived from animal bearing Tumor M-27 (2.1cm) 37 days after the s.c. implantation of anisolated lung nodule (see "Materials and Methods"). Extensive pulmonary metastasis is evi
Morphologically, the tumors derived from both lines appeared similar with the exception of a higher incidence of giantcells seen in tissue sections and culture monolayers of TumorH-3LL. Subsequently we studied the metastasis of the twohighly metastatic clonal lines H-59 and M-27 derived fromTumors H-3LL and M-3LL, respectively. We found that theseclonal lines displayed the organ selectivity characteristic of theirparental lines. We also noted that pulmonary but not hepaticmetastasis of clone H-59 could be considerably reduced whenanimals were implanted with isolated liver nodules rather than
given injections of dispersed tumor cells.Using [125I]dUrd-labeled tumor cells we noted further that
the arrest and retention of tumor cells in the livers and lungsas measured by I2SIuptake for 24 h after cell inoculation i.v.
The different patterns of metastasis of the tumor lines andour finding that they home to their respective target organsregardless of the site of tumor inoculation (i.e., s.c. in theaxillary region or flank, or tail vein) argue that, in the presenttumor model, specific tumor cell properties and/or altered hostresponses rather than anatomical or so-called "mechanistic"
At present, this relationship between the target organ specificity and the different routes of dissemination apparently utilized by the tumor cells (i.e., lymphatic or hematogeneous) is
ORGAN-SELECTIVE METASTASIS OF LEWIS LUNG CARCINOMA VARIANTS
Fig. 4. Lymphatic metastasis of Tumor H-59. a, axillary lymph nodes and liver of amouse which was given injection of 10' H-59
cells s.c. in the axillary region and sacrificedwhen mean tumor diameter measured 2.4 cm.Black arrow points to the axillary lymph nodewhich was draining the tumor site. Metastasisis also apparent in the contralateral node andin the inguinal lymph node (while arrow), h.low-power view (x 72) of a cross-section of aregional (axillary) lymph node derived fromanimal bearing Tumor H-59 (2.0 cm). Extensive infiltration of the node by tumor cells(Ughi staining areas) is evident, c. high-power(x 400) vie»of the same section showingclusters of lymphocytes (top left) surroundedbv tumor cells.
Female C57BL/6 mice bearing s.c. tumors in the axillary region were sacrificedat different intervals (15-40 days) following tumor implantation. Axillary and/orbrachial lymph nodes were removed and implanted s.c. into new recipient femaleC57BL/6 mice. Animals were then monitored for tumor growth for 6 mo.
Type oftumorM-3LLM-27
H-59Size
of tumors at Incidence of tumorstime of lymph from implants ofnode excision regional lymph
in lungs and livers ofi.v. inoculated miceFifteen C57BL/6 mice were given injections of 10* radiolabeled tumor cells
into the tail vein. Three mice in each group were sacrificed at 20 min and 4 and24 h after inoculation; organs were collected and rinsed twice with 70% ethanol,and I25Iuptake was monitored.
ORGAN-SELECTIVE METASTASIS OF LEWIS LUNG CARCINOMA VARIANTS
Metastasis in the nude mice was again confined to the lung.These findings suggest that, in the present model, T-cells do
not play a major role in the organ specificity of metastasis. Weare currently investigating the importance in this model ofother tumor-inhibitory mechanisms, in particular those mediated by macrophage-like cells (i.e., Kuppfer cells, sinus histio-cytes).
Our findings following the i.v. inoculation of [125I]dUrd-
labeled tumor cells derived from highly metastatic clonal linesof Tumors H-3LL and M-3LL suggest that, in the presentmodel, the observed target organ selectivity is not related todifferent patterns of tumor cell arrest and distribution measurable immediately following tumor inoculation as has beenshown to be the case in other tumor models (28). The resultspoint instead to the importance of other, later mechanismswhich may be organ specific and may determine whether tumorcells could survive and proliferate in the microenvironment ofa particular organ. Similar findings were also described by Hart(3) for organ-specific Tumors M 5076 (liver) and B16-F10(lung). Recent studies reported by Tarin et al. (8) have shownthat soluble factors released by liver fragments in vitro werecytotoxic to some tumors and that this correlated with thepotential of these tumors to metastasize to the liver. In anothercommunication by Mandick and Berger (29), similar conclusions were reached using a somewhat different in vitro assay.In our tumor model, recent results4 have shown that tumor cellsderived from the liver-homing subline H-59 were significantly(3- to 4-fold) more adhesive to monolayers of cultured hepato-cytes than tumor cells derived from the lung-specific sublineM-27. This is in accordance with findings reported in othertumor systems (6, 30). The relationship between the increasedbinding of the tumor cells to the hepatocytes and their abilityto survive and proliferate in the liver is the subject of currentinvestigations in our laboratory.
1. Brodi. P. Characterization of two organ-specific variants of a murine carcinoma. Proc. Am. Assoc. Cancer Res., 25: 250. 1984.
2. Brodt, P. Metastatic patterns of two organotropic variants of the Lewis lungcarcinoma. Proc. Am. Assoc. Cancer Res., 26: 221, 1985.
3. Hart, I. R. "Seed and soil" revisited: mechanisms of site specific metastasis.Cancer Metastasis Rev., /: 5-16, 1982.
4. Weiss, L., Haydock, K., Pickren, J. W., and Lane, W. W. Organ vascularityand metastatic frequency. Am. J. Pathol., 101: 101-114, 1980.
5. Alby, L., and Auerbach, R. Differential adhesion of tumor cells to capillaryendpthelial cells in vitro. Proc. Nati. Acad. Sci. USA, 81: 5739-5743, 1984.
6. Schirrmacher, V., Chiengsong-Popov. R., and Arnheiter, H. Hepatocyte-tumor cell interaction in vitro. I. Conditions for rosette formation andinhibition by anti-H-2 antibody. J. Exp. Med., 151: 984-989, 1980.
7. Middelkoop, O. P., Roos, E.. and Van de Pavert. I. V. Infiltration oflymphosarcoma cells into hepatocyte cultures: inhibition by ambivalent antibodies against liver plasma membranes and lymphosarcoma cells. J. Cell.Sci.. 56:461-470. 1982.
8. Tarin. D. The cell biology of metastatic tumor spread. J. Med. Sci., 12:1071-1073, 1984.
9. Proctor, J. W. Rat sarcoma model supports both "soil seed" and mechanicaltheories of metastatic spread. Br. J. Cancer, 34: 651-654, 1976.
10. Netland. P. A., and Zetter, B. R. Organ specific adhesion of metastatic tumorcells in vitro. Science (Wash. DC). 224: 1113-1115, 1984.
11. Brunson. K. W., Beattie, G., and Nicolson. G. L. Selection and alteredtumour cell properties of brain-colonizing metastatic melanoma. Nature(Lond.), 272: 543-545, 1978.
12. Tao. T.. Matter. A.. Vogel, K., and Burger. M. M. Liver colonizing melanomacells selected from B16 melanoma. Int. J. Cancer, 23:854-857, 1979.
13. Pal. K., Kooper, L.. Timar. J.. Ranjai, J.. and Lapis. K. Comparative studyon Lewis lung tumor lines with "low" and "high" metastatic capacity. I.
Growth rate, morphology, and resistance to host defence. Invasion Metastasis, 5.- 159-169, 1985.
14. Roos, E., and Dingemans, K. P. Mechanisms of metastasis. Biochim. Bio-phys. Acta, 560: 135-166, 1979.
15. Suguira, K., and Stock, C. C. Studies in a tumor spectrum. III. The effect ofphosphoramides on the growth of a variety of mouse and rat tumors. CancerRes., /5: 38-51, 1955.
16. Brodt, P.. Parhar. R., Sankar, P.. and Lala, P. K. Studies on clonal heterogeneity in two spontaneously metastasizing mammary carcinomas of recentorigin. Int. J. Cancer, 35: 265-273, 1985.
17. McManus. J., and Maury. T. Staining Methods: Histologie and Histochem-ical, pp. 55-59. New York: Harper and Row Publishers, Inc., 1960.
19. Brodt. P., Kongshaven, P.. Vargas, F.. and Gordon, J. Increased natural hostresistance mechanisms in B-lymphocyte-deprived mice. J. Reticuloendothel.Soc.. 30: 283-289, 1981.
20. Siegel, S. Nonparametric Statistics for the Behavioural Sciences, pp. 116-126. New York: McGraw-Hill, 1958.
21. Brodt, P., and Segal, S. Differences in immunogeneity between the localtumor 3LL and its metastases-derived line as detected by sensitization withantigen-fed macrophages. Invasion Metastasis, 2: 185-196, 1982.
22. Carr, I. Lymphatic metastasis. Cancer Metastasis Rev., 2: 307-317, 1983.23. Kim, V. Possible avenues for the control of metastatic behavior in breast
cancer. In: V. P. Hollander (ed.), Hormonally Responsive Tumors. NewYork: Academic Press, in press. 1986.
24. Miner, K. M., Klostergaard. J.. Granger, G. A., and Nicolson, G. L. Differences in cytotoxic effects of activated murine peritoneal macrophages andJ774 monocytic cells on metastatic variants of B16 melanoma. J. Nati.Cancer Inst.. 70: 717-724, 1983.
26. Brodt. P., Feldman. M., and Segal, S. Differences in the metastatic potentialof two sublines of tumors 3LL selected for resistance to natural NK likeeffector cells. Cancer luminimi. Immunother., 16: 109-113, 1983.
27. Gorelik. E.. Feldman. M.. and Segal. S. Selection of 3LL tumor sublineresistant to natural effector cells concomitantly selected for increased metastatic potency. Cancer Immunol. Immunother.. 12: 105-109, 1982.
28. Poste, G., and Nicolson. G. L. Modification of the arrest of metastatic tumorcells in the microcirculation after treatment with plasma membrane vesiclesfrom highly metastatic cells. Proc. Nati. Acad. Sci. USA. 77:399-403. 1980.
29. Mandick. II. M.. and Burger. M. M. A novel mechanism for organ specificmetastasis. In: Proceedings of the Bat-Sheva Seminar on Tumor Metastasis:Control Mechanisms, p. 42, 1983.
30. Nicolson, G. L., Dulski, K., Basson. C., and Welch, D. R. Preferential organattachment and invasion in vitro by B16 melanoma cells selected for differingmetastatic colonization and invasive properties. Invasion Metastasis, 5:144-158. 1985.
