The ability of a tumor cell to survive is critical for successful dissemination to sites distant from the primary tumor. Tumor cells must enter blood circulation, resist hemodynamic shear stress of the blood circulation, successfully extravasate, and then migrate through dense tissue stroma to a site favorable for tumor growth. Some tumor cells must therefore be endowed with peculiar abilities to successfully metastasize, whereas others, although capable of forming tumor in specific organs, cannot metastasize. This property has often been associated with the homing ability of a given tumor cell, likely through the expression of organ-specific homing receptors that are critical for the extravasation process. The present work was aimed at establishing the point at which metastatic and nonmetastatic lymphoma cells diverge. Although 164T2 and 267T2 lymphoma cell lines can successfully form thymic lymphoma when injected intrathymically, only the 164T2 clone can efficiently form tumor in kidneys, spleen, and liver after intravenous inoculation. Using the Indium-labeling technique to monitor the homing kinetic of both cell lines, we showed that the critical step for the successful metastasis of the lymphoma cell was determined in the final steps of the disseminating process, namely after homing. These results indicate that, whereas binding of tumor cells to vascular endothelium through specific adhesion mechanisms is a prerequisite for dissemination of tumor cells, the resistance of a tumor cell to the antagonist action of the host and/or its ability to grow tumor occurs only after homing to the target organ.
METASTASIS OR THE formation of secondary tumors at distant sites from the primary tumor is a major cause of mortality. The failure of most current therapies to successfully treat metastases stresses the urgency of understanding the underlying mechanisms.
Metastasis is a multistep process that includes (1) primary tumor growth; (2) release of cancer cells into lymphatic and blood vessels; (3) survival of the tumor cells within the circulation; (4) arrest in the microvasculature of the target organ that involves tumor cell-endothelial cell interactions, resulting in (5) extravasation of tumor cells and (6) invasion of target organs; (7) migration of tumor cells into tissues; and, finally, (8) growth of the tumor at the metastatic site.1,2
For efficient malignancy, cancer cells must be able to accomplish each of these eight steps while escaping surveillance by the immune system.3 Homing of cancer cells from the circulation into target organs has been considered as a major rate-limiting step in hematogenous metastasis. It depends on the capacity of cancer cells to survive within the circulation and to extravasate into target organs, and poor survival in the aggressive blood environment resulted in only few circulating cells able to escape from that environment.4,5 Only cancer cells that have successfully extravasated into target organs can form secondary tumors.2,6 Homing of tumor cells and particularly lymphoma cells shares some similar mechanisms with homing of normal leukocytes.7 The latter process depends on interactions between several pairs of adhesion molecules expressed on leukocytes and endothelial cells.8 Leukocyte transmigration through the endothelium occurs in three steps, namely cell attachment, rolling, and shear-resistant adhesion.9 In each of these steps, the adhesion molecules involved have been defined in vivo.10-13Similarly, tumor cell homing depends on their interactions with endothelial cells and require adhesion molecules. Tumor cells express the same cell adhesion molecules (CAMs) as their normal counterparts and in vitro adhesion assays have shown that tumor cell-endothelial cell interactions can be blocked by anti-CAM monoclonal antibodies (MoAbs), specifically by anti–LFA-1 MoAb.14This type of interaction has been strongly correlated with the metastatic potential of several tumor cells such as hepatomas, melanomas, mammary carcinomas, and lymphomas.15-19 The same is true in vivo, where treatment with anti–LFA-1, -CD44 and -α5β3 MoAbs was shown to block lymphoma metastasis.14,20-22 These observations establish a link between the metastatic potential of tumor cells and their ability to home to target organs.
However, some conflicting results suggest a lack of correlation between the expression of cell adhesion molecule, homing, and the metastatic potential of tumor cells. Indeed, it has been shown that hematogenous spreading as well as peripheral node invasion of lymphoma-derived leukemic cells may occur independently of the expression of the lymphocyte homing receptor, LFA-1, and intercellular adhesion molecule-1 (ICAM-1).23 In addition, data from direct observations of metastasis in vivo using the intravital videomicroscopy technology have shown that most tumor cells entering the circulation extravasate efficiently into tissues independently of their metastatic potential.24 In fact, mammary carcinoma and melanoma cell lines with high and low metastatic potential differ not in extravasation and homing but in migration through the perivascular tissue and subsequent tumor growth.25,26 Furthermore, blocking of integrin function by the disintegrin agent did not affect the extravasation and homing of melanoma cells in vivo but reduced tumor growth.27
In the light of these findings, we sought to determine whether the generally accepted correlation between the extravasation, homing of lymphoma, and their metastatic potential also exists in vivo. To this end, we used two murine lymphoma cell lines, the metastatic 164T2 and the nonmetastatic 267T2 cells. Although both cell lines give rise to a thymic lymphoid tumor after intrathymic inoculation in histocompatible mice, only the 164T2 cells induced massive tumor growth in kidneys, spleen, and liver 6 to 8 weeks postinjection. In vivo migration assays showed that, when injected intravenously, both lymphoma cell lines infiltrated the target organs at a similar rate and with a similar fate, indicating that metastasis formation of these lymphoma cell lines is determined after their homing.
MATERIALS AND METHODS
The mouse thymic lymphoma cell lines 164T2 and 267T2 were established in vitro from an in vivo radiation-induced thymic lymphoma in C57BL/Ka mice using the method of Lieberman et al.28YAC-1 lymphoma cell line was obtained from American Type Culture Collection (Rockville, MD). The H59 cells (kindly provided by Dr Daniel Oth, Institut Armand-Frappier, Québec, Canada) were derived from the Lewis lung carcinoma.29 The cells were cultured in RPMI 1640 supplemented with 10% fetal calf serum, 2 mmol/L glutamine, 10 mmol/L HEPES buffer, and antibiotics.
Flow cytometric analysis.
Cells were stained at 4°C and washed in phosphate-buffered saline (PBS) containing 0.5% bovine serum albumin and 0.2% sodium azide (PBA). Before staining, cells were incubated with 10 μg/mL of human IgG (Sigma, St Louis, MO) for 20 minutes to block nonspecific binding. Fluorochrome- or biotin-labeled MoAbs were then added at appropriate concentrations and incubated for another 20 minutes. Cells were then washed four times with PBA. For indirect staining with streptavidin-Red 670, cells were washed three times after staining with the first MoAb and then incubated for 20 minutes on ice with the fluorescent conjugate. Flow cytometric analysis was performed on an XL flow cytometer (Coulter Electronics, Hialeah, FL).
C57BL/6x129 H-2b histocompatible mice were bred in our animal facility and were maintained under specific pathogen-free conditions.
The in vivo model of lymphoma metastasis.
164T2 and 267T2 lymphoma cells were injected intravenously via the tail vein to histocompatible mice. Animals were observed regularly for clinical signs of lymphoma development, were killed 6 to 8 weeks postinoculation, and examined for the presence of lymphoid tumors. Mice were examined macroscopically, and their kidneys, liver, spleen, and thymus were harvested, weighed, and fixed in 10% formalin for histologic examination.
In vivo migration assays.
The migration of 164T2 and 267T2 lymphoma cell lines was analyzed as previously described.30 Briefly, 10 million cells were labeled with 10 μCi of 111In in 0.5 mL RPMI for 15 minutes at room temperature. The cells were washed four times with RPMI and resuspended in PBS. Each mouse was injected intravenously with 106 cells labeled at 0.5 to 106 CPM. At various times, animals were killed, and kidneys, spleen, liver, and thymus as well as heparinized blood samples were recovered. The total activity in the entire blood volume of the body was computed by measuring radioactivity in 400-μL aliquots of blood and assuming a total volume of 2 mL of blood per mouse. Homing of lymphoma cells was measured by counting radioactivity in each organ using a γ-counter (Gamma-7000).
Lymphoma tumor growth in vivo.
We have previously shown that intrathymic inoculation of 164T2 and 267T2 lymphoma cell lines can induce the formation of thymic lymphoma within 3 to 6 weeks postinjection.31 However, when injected intravenously, we have found that only 164T2 cells but not 267T2 had the ability to induce tumors in peripheral organs in histocompatible C57BL/6 × 129 mice. Tumors were detected in liver, kidneys, and spleen at 6 to 8 weeks after intravenous injection. When 105 cells were injected, 17 of 32 (53%) mice developed solid tumors (Table 1). A score of 100% within the same time interval was obtained when 5 × 105 or 106 164 T2 cells were injected. In contrast, no tumor was detected in animals injected with 267T2 cells at any of the above-mentioned doses. In addition, the resistance of animals to 267T2 cells metastasis was not time-dependent, because no clinical signs associated with tumor development were detected for up to 20 weeks. Macroscopic examination of kidneys of mice injected with 164T2 cells showed that the kidneys were significantly enlarged with atypical pale coloration. Histologic examination showed massive interlobular infiltration by lymphoma cells in the deep cortical and outer medulla areas (Fig 1). Infiltration of the spleen by lymphoma cells induced severe splenomegaly characterized by massive and diffuse infiltration of lymphoma cells. Although no macroscopic signs were observed in liver of mice injected with 164T2 cells, histologic analysis showed massive lymphoma cell infiltration in perivascular stroma surrounding the central vein of hepatic lobules (Fig.1). In contrast, no infiltration by lymphoma cells was detected in any of these organs in mice injected with the 267T2 cells. These results established the metastatic potential of 164T2 in its ability to induce the formation of tumors in kidney, spleen, and liver and the inability of 267T2 cells to do the same.
|No. of Injected Cells .||Mice Injected With .|
|164T2 Cells .||267T2 Cells .|
|1 × 105||17/32||0/20|
|5 × 105||5/5||0/5|
|1 × 106||10/10||0/10|
|No. of Injected Cells .||Mice Injected With .|
|164T2 Cells .||267T2 Cells .|
|1 × 105||17/32||0/20|
|5 × 105||5/5||0/5|
|1 × 106||10/10||0/10|
Mice were considered positive if at least one solid tumor was found, either in the kidney, spleen, or liver. In most of the mice examined, tumor development of lymphoid tumors in kidneys was generally accompanied with severe splenomegaly. Development of lymphoma in kidneys was in all cases symmetrical. Tumor growth in all organs was confirmed by histologic examination.
Phenotypic characterization of lymphoma cell lines.
In light of work showing that LFA-1, VLA-4, and VLA-5 play key roles in lymphoma metastasis,12,18,32,33 the expression of adhesion molecules on both lymphoma cell lines was determined. Flow cytometric analysis showed that both lymphoma cell lines expressed similar levels of LFA-1, ICAM-1, ICAM-2, and α5β1 but did not express α4β1 (data not shown). The 164T2 and 267T2 cell lines were CD3+CD4−CD8− and CD3+CD4−CD8+, respectively.
In vivo migration assays.
The differential capacity of 164T2 and 267T2 lymphoma cells to metastasize despite similar adhesion molecule profiles prompted us to investigate whether it might be caused by a distinct in vivo migration pattern to target organs. For this purpose, the 164T2 and 267T2 lymphoma cell lines were labeled with 111In and cell migration was measured after intravenous injection. The results (Fig 2) indicate that, at 1 hour postinjection, the majority of both lymphoma cells leave the blood circulation to migrate to target organs, because only 7% of 164T2 and 9% of 267T2 were detected in the blood. At that time, similar numbers of 164T2 and 267T2 were detected in the liver, spleen, and kidneys. However, 25.7% of 164T2 cells and only 9.1% of 267T2 were detected in the lungs. At 3 hours postinjection, similar numbers of 164T2 and 267T2 cells were detected in liver, kidney, and spleen. The accumulation of 164T2 and 267T2 cells into the lungs that was observed at 1 hour appears to be transient, because at 3 hours only 9.6% of 164T2 and 1.44% of 267T2 cells were still detected. By 24 hours, the totality of both injected cells had left the circulation and the majority had migrated in the liver (34.5% 164T2 and 31.6% 267T2). In addition, significant numbers of 164T2 and 267T2 cells were also found in the kidneys (3.1% in both cases) and spleen (6% and 10.5%, respectively), and only negligible numbers of both lymphoma cells were retained in the lungs (0.21% 164T2 and 0.36% 267T2). These results indicate that both lymphoma cells, independently of their metastatic potential, had the same capacity to migrate to target organs with the similar fate and at the same rate. It is noteworthy that no homing of either lymphoma cells could be detected in the thymus.
The migratory pattern obtained with 164T2 and 267T2 was specific to the target organs in which 164T2 cells eventually form tumors. Lymphoma YAC-1 and carcinoma H59 cells, which have been shown to preferentially migrate to the lungs and the liver, respectively, were found to home to their target organs and show different pattern from those of 164T2 and 267T2 cells. We found, indeed, that most of YAC-1 and H59 cells accumulated in the lungs of mice, whereas 164T2 and 267T2 lymphoma had a preferential homing to liver (Fig 2,vFig 3).
In the present work, we showed that the metastatic potential of two thymic lymphoma cell lines is regulated at a stage subsequent to their migration into target organs. In vivo migration assays showed that the nonmetastatic 267T2 cells migrated at the same rate and showed similar homing specificity to that of the metastatic 164T2 cells. In both cases, homing occurred between 3 and 24 hours postinjection and almost half of total injected cells had passed from the circulation to target organs within this short time. The homing of both lymphoma cells had the same target organ specificity, yet only the 164T2 cells gave rise to secondary tumors, indicating that homing is not a determinant factor of their metastatic potential.
This conclusion appears to be at variance with those of previous studies that have shown that extravasation and subsequent homing of cancer cells are major rate-limiting events in hematogenous metastasis.4-6,15,34 Of particular relevance to the present work are the reports that the interactions of Raw 117 large cell and AKR lymphomas with endothelial cells correlated not only with their extravasation and homing capacities but also with their metastatic potential.16,17 However, this correlation was based on in vitro adhesion between tumor cells and endothelial cells on the one hand and on the development of secondary tumors after injection into mice on the other hand; the migration of injected cells had not been monitored. The different parameters measured in those studies and in the present work probably explain the apparently conflicting conclusions. It seems that, although interactions between tumor cells and endothelial cells in vivo are a prerequisite for the former's extravasation, in vitro interactions may not always be a reliable indicator of the in vivo situation. Considerable variation in the expression of cell adhesion molecules at the surface of tumor cells and/or in the heterogeneous binding properties of tumor cells to high endothelial venules35-37 can also explain differences in metastatic properties among lymphoma cells. For normal lymphocytes, for instance, it has been shown that the in vitro binding of lymphocytes to high endothelial venules did not reflect homing in vivo.38 Furthermore, the correlation established between the in vitro invasive capacity and the metastatic potential of mammary carcinoma cell lines is not reflected in vivo,24,39indicating that the mechanisms governing cell invasion in vivo depend on additional factors than those governing cell invasion in vitro.
In addition, our results indicate that survival of lymphoma cells within the microcirculation might not be the reliable regulator of hematogenous metastasis that it was thought to be. In previous studies, only a very limited number of cancer cells in the microcirculation reached target organs to form distant tumors. Most of the cells simply died or were destroyed by the host's immune system or by lethal deformation.4,5,40 However, in our case, almost half of the lymphoma cells migrated to their target organs (44.5% of 164T2 and 46.7% 267T2 cells at 24 hours postinjection). We thus conclude that the metastatic potential of the lymphoma cell lines used in this study is not determined by survival and homing but by events occurring after extravasation. In agreement with our work, Koop et al,24using intravital videomicroscopy to monitor metastasis in vivo, showed that more than 80% of the tumor cells entering circulation survived and successfully extravasated and migrated into target organs. In a similar vein, melanoma and mammary carcinoma cell lines of high and low metastatic potential had similar capacities to extravasate and invade target organs; however, they differed in subsequent migration through perivascular tissue and in their tumor growth.25,26Therefore, whenever it was tested, the metastatic potential of cancer cells was shown to be determined by postextravasation events. Whether the differential growth of 164T2 and 267T2 lymphoma cells is due to differential responses to paracrine and/or autocrine growth factors remains to be determined. What other mechanisms could differentially affect metastasizing and nonmetastasizing tumors? Smithson et al41 showed that susceptibility to NK cell lysis affected the metastatic potential of lymphoma cells within target organs. In the case of those investigators, only 5% of total cells were recovered at 20 hours postinjection.33 This is not so in our model, in which half of total 164T2 and 267T2 cells were recovered at 24 hours postinjection, and in vitro cell lysis assay showed that, in contrast to YAC-1 cells, neither lymphoma was sensitive to NK cell lysis (data not shown). These findings suggest that, although NK cell lysis may affect the survival of lymphoma in vivo, it is not sufficient to explain the differential metastatic potential of these lymphoma cells. However, the differential effects of other immune cells such as cytotoxic T lymphocytes (CTLs) and macrophages cannot be ruled out.
Adhesion molecules have been implicated in the metastatic process of several tumor cells, including lymphomas.7,42 Most relevant to our work, LFA-1, CD44, and α5β3 are known to participate in lymphoma metastasis.14,20-22 Based on the fact that lymphoma cells use similar mechanisms of extravasation and migration as their normal counterparts,7 it has been suggested that it is by blocking such mechanisms that MoAbs to these CAMs inhibit metastasis. Because control of lymphoma metastasis appears to occur at the postextravasation level, one must suppose that, in addition to extravasation, adhesion molecules may somehow also be implicated in the control of tumor growth during postextravasation events, a model that is supported by the recent observation that expression of α4 integrins inhibits metastasis formation of lymphoma without affecting homing33 and that VLA-2 adhesion molecule is involved in rhabdomyosarcoma cell metastasis during migration through the perivascular tissue.43 One possibility is that cell-cell interactions between tumor cells and host stroma may regulate the expression of matrix metalloproteinases,31 which are abundantly expressed in cases of malignant lymphomas.44Several indications indeed suggest that active tissue remodeling by matrix metalloproteinases is determinant in the development of tumor growth.45 Postextravasation events involving adhesion molecules and matrix metalloproteinases in the metastatic process of 164T2 and 267T2 cells are now under investigation.
The indium labeling technique is a very well-established method to assess the kinetic distribution of leukocytes and lymphoma cells in patients with malignancy, because 111In-oxine is a cytoplasmic marker with very low spontaneous release and is not significantly incorporated into cell membranes.46-48 It was important, nevertheless, to start our kinetic analysis very early after entry of the cells into circulation while stopping after 24 hours for the following reasons. (1) Tumor cells migrate to their target organ very rapidly, most of them within 24 hours. (2) Although released indium as a consequence of cell mortality is not reused by other cells,46 engulfment of labeled-lymphoma cells by resident macrophages could have interfered with recirculation of tumor cells in long-term study. (3) Previous studies had reported that high concentrations of radioactivity by 111In-oxine-labeling may affect long-term cellular functions of labeled leukocytes.48,49 This latter issue was further addressed by comparing the migratory pattern of our lymphoma cells with that of other tumorigenic cell lines. The specific migratory pattern observed with all the different tumorigenic cell lines used in our study suggest that 111In-oxine-labeled cells preferentially migrated to organs in which they form tumors after intravenous inoculation of unlabeled cells. Moreover, the kinetics obtained in terms of blood clearance of labeled cells were similar to that previously obtained using radiolabeled cells or fluorescently labeled cells visualized by intravital videomicroscopy studies. The stable accumulation of 164T2 and 267T2 lymphoma cells in liver, kidneys, and spleen compared with their transient accumulation into the lungs is in agreement with previous studies showing that liver, spleen, and kidneys are preferred target sites for lymphoma metastasis.7,22,33This pattern of migration has also been observed with circulating lymphoblasts and activated mature lymphocytes that show significant affinity for the liver and lungs, although the retention in the lungs appears to be transient for the most circulating blasts.50The higher retention in the lungs at 3 hours for YAC-1 and H-59 cells compared with that of 164T2 and 267T2 cells could due to their larger size. In fact, cell arrest by size restriction has been documented.51 Alternatively, differential responses among the cancer cells to organ-specific soluble factors and chemoattractants may also explain the differences observed in the migration pattern.
In summary, in this study, we show that metastasis of lymphoma cells is determined subsequently to their migration and invasion of target organs. These results complement those obtained from IVVM studies of carcinoma and melanoma cell metastasis. Together, these results point to postextravasation events as the focus of future investigation on metastasis control.
The authors thank Doris Legault and Claire Beauchemin for excellent technical assistance.
Supported by a grant from the Cancer Research Society of Canada (to E.F.P.). F.A. is supported by a Biochem Pharma Fellowship Award attributed by La Fondation Armand-Frappier.
Address reprint requests to Yves St-Pierre, PhD, Immunology Research Center, Institut Armand-Frappier, PO Box 100, Laval, Québec, Canada H7N 4Z3.
The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. section 1734 solely to indicate this fact.