Colony-stimulating activity from the new metastatic TS/A cell line and its high- and low-metastatic clonal derivatives.

We investigated the presence of colony-stimulating factor (CSF) in supernatants obtained from TS/A, a new metastatic murine cell line, and from its high-and low-metastatic clonal derivatives (E and F clones, respectively). TS/A cells produced a CSF in vitro that induced proliferation and differentiation of murine monocytic and granulocytic progenitors in agar cultures. In TS/A-bearing mice remarkable splenomegaly, blood granulocytosis and thymus depletion were observed along with a stimulatory activity in serum and a strong proliferative activity both in spleen and in bone marrow populations. Conditioned media from E clones showed an in vitro colony-stimulating activity greater than those of F clones. Mice injected subcutaneously with cells of all clones studied showed granulocytosis, splenomegaly and thymus depletion, although to varying degrees. However, no direct correlation between granulocytosis-splenomegaly and the number of spontaneous lung metastases was observed.

Quantitative and qualitative haemopoietic alterations could play an important role in the metastatic process and the study of the production of colony-stimulating factor(s) (CSF) by metastasizing tumours could help to clarify if and when such an influence exists. Therefore, we studied the production of colony-stimulating activity by TS/A line, a new murine spontaneous mammary carcinoma cell line that metastasizes spontaneously to the lungs (Nanni et al., 1983) and by its highand low-metastatic clonal derivatives (Lollini et al., 1984). Data are presented here on the haemopoietic alterations observed during in vivo growth of TS/A Correspondence: P. Nanni Received 14 January 1985; and in revised form 11 April 1985. and clones, on the detection of in vitro produced CSF and on the comparison between this activity and metastatic potential.

Mice
Eight-12 week-old female BALB/cAnNCRIBR mice (hereafter referred to as BALB/c), purchased from Charles River, Calco, Italy, or bred in our facilities were used throughout the study.
Cell lines and clonal derivatives The parental TS/A cell line was derived from a spontaneous mammary adenocarcinoma which arose in a retired breeder BALB/c female (Nanni et al., 1983). Clonal derivatives with different metastatic potential, which have been previously selected and characterized (Lollini et al., 1984), were also used in the present study: E1, E2 and E3 clones are highly metastatic and Fl, F2 and F5, are poorly metastatic. In the experiments here reported, TS/A cells were between the 20th and the 35th in vitro passage and clones were between the 30th and the 45th in vitro passage after cloning.
A cell line derived from the B16 melanoma of C57BL/6 origin (kindly provided by Dr A. Mantovani, Istituto Mario Negri, Milan, Italy) was used as a control in some experiments.
All cell lines and clones were cultured in Dulbecco's MEM supplemented with 2 mM glutamine, 100 U ml-1 penicillin, 100 ug ml-1 streptomycin (referred to as DMEM) and with 10% heat-inactivated foetal calf serum (FCS) (GIBCO, Paisley, Scotland) in a 5% CO2 humidified atmosphere at 370C and were routinely subcultured twice weekly by trypsin-EDTA treatment.

Assays for CSF activity
Assays for detection of CSF production were made on conditioned media obtained as follows: 1-2 x 106 cells were seeded in 75 cm2 flasks and cultured in DMEM +10% FCS. Supernatants were collected 3 days later, centrifuged at 3000rpm for 20min and filtered through 0.45pm Millex filters (Millipore). At the same time, monolayers were processed to determine cell number/flask.
The ability of conditioned media to induce bone marrow proliferation was tested as reported elsewhere (Pessina et al., 1981): 105 bone marrow cells from normal BALB/c mice were cultured in 35 mm Petri dishes with 1 ml semisolid medium (McCoy's 5A medium, 20% FCS [GIBCO], 0.3% Difco agar [DIFCO Laboratories, Detroit, Michigan, USA]) supplemented with conditioned medium at 0.31-40% final concentration. After a 7-day culture, plates were scored microscopically and colonies of .50 cells were counted. Colony morphology was evaluated after staining of the whole culture dish with May Grumwald-Giemsa solution as reported by Konwalinka et al. (1982).
In vivo studies Mice were injected s.c. with 105 or 106 viable cells. During tumour growth the following parameters were examined: tumour volume (calculated as 4/3 i. [(a + b)/4]3, where a = maximal tumour diameter and b = tumour diameter perpendicular to a), or tumour weight at sacrifice; peripheral leukocytes (no. mm 3) and percentage of peripheral lymphocytes, spleen weight, spleen and thymus cell yield.
Spleen cells from TS/A-bearing mice were studied by cytochemical (PAS-, ANAE-and peroxidase stains) and immunofluorescence techniques: surface immunoglobulins (SIg) were visualized with antimouse IgG FITC-conjugated antiserum from Miles; Thy 1 antigen was detected by anti-Thy 1.2 monoclonal antiserum from NEN. Proliferative activity of spleen and bone marrow cells from tumour-bearing mice was tested in a [3H]-TdR uptake assay: 4 x 105 spleen cells or 5 x 104 bone marrow cells (collected from femurs) were plated in each well of Microtest II plates (Nunc, Denmark), incubated for 96h, with a final 4 h pulse with 1 4uCi [3H]-TdR, and then processed as above.
Lung metastases were enumerated as previously described (Nanni et al., 1983) with the aid of a dissecting microscope.

Results
Haemopoietic alterations in TS/A-bearing mice Progressive granulocytosis, splenomegaly, and thymus depletion were observed in BALB/c mice s.c. injected with the TS/A cell line (Table I). Splenomegaly was already detectable 1 week after cell injection: both spleen weight and cell yield reached values which were more than 5-fold greater than control levels by the 30th day after inoculation. At this time, spleen comprised -10% T cells (Thy 1-positive cells), -20% B lymphocytes (SIg positive), -20% mature monocytes (PAS-and ANAE-positive, peroxidase-negative) and 50% hypogranular metamyelocytes and polymorphonuclear cells (PASand peroxidasepartially positive, ANAE-negative). Leukocytosis reached values of >200,000 mm 3 and was mainly due to enlargement of the granulocytic population. The yield of bone marrow cells in TS/A-bearing mice seemed to be unaffected (data not shown).
A stimulating activity on the proliferation of murine bone marrow cells was detected, by means of the [3H]-TdR uptake assay, in serum from 30day TS/A-bearing mice (Table II). The stimulating activity was proportional to the number of s.c. injected TS/A cells and significantly greater than the activity of control serum.
Such a stimulating activity in the serum of TS/Abearing mice could result in alterations in proliferative activity of spleen and bone marrow. Therefore, we evaluated both the spontaneous [3H]-TdR uptake in cells from these organs at different times after TS/A s.c. injection and their susceptibility to the addition of TS/A conditioned medium. Data from 30-day TS/A-bearing mice are shown in Table III. Both spleen and bone marrow cells from tumour-bearing animals showed a higher proliferative activity than cells from controls: such a difference was already evident by I week after   strong stimulating activity on murine bone marrow cells was also detected by means of the [3H]-TdR uptake assay performed in the presence of TS/A conditioned medium: data from a representative experiment are given in Figure Ib, where up to a 2,000-fold stimulation was obtained. Conditioned medium from an unrelated cell line (B16 melanoma) showed no activity (data not shown).
Colony-stimulating activity in TS/A clonal derivatives We had recently derived high-and low-metastatic clones (E and F clones, respectively) from the parental TS/A cell line (Lollini et al., 1984). To investigate whether all clonal derivatives produced CSF and whether a direct relationship occurred between such an activity and metastatic potential, we compared three high-metastatic E clones with three low-metastatic F clones. Conditioned media from all 6 clones were able to  (Figure 2a), but supernatants from E clones induced higher numbers of colonies than supernatants from F clones. A higher in vitro activity by conditioned media from E clones was also shown by means of [3H]-TdR uptake in murine bone marrow cells (Figure 2b). In both assays supernatants from F clones also showed an activity lower than those observed with E clones when the plateau was reached (at concentrations of conditioned medium ranging from 10 to 40% and from 20 to 40% for the colony and the [3H]-TdR uptake assays, respectively). The differences among clones in cell size and in doubling time cannot account for this phenomenon. In all clones examined, tumour-bearing mice showed progressive leukocytosis, mainly due to enlargement of the granulocytic population, even though values of leukocytes among different clones were scattered (Table IV). On the whole, comparison between E and F clones did not show different group patterns. It should be underlined Conditioned medium (%) that a very low leukocytosis was observed in F5 tumour-bearing mice. Animals were then sacrificed when individual mean tumour diameters exceeded 2.2 cm, and leukocytes no. mm3, spleen weight and number of lung metastases were evaluated for each group (Table V). Leukocytosis and splenomegaly did not appear to be directly related to the number of lung metastases: the F2 clone induced the highest leukocytosis and splenomegaly but a very low number of metastases. It is emphasized that, even when comparison among clones was made in conditions of similar tumour dimensions, the leukocytosis induced by the F5 clone (near 24,000 mm-3) remained much lower than that observed with other clones.

Discussion
Qualitative and quantitative haemopoietic alterations could play an important role in the metastatic process. Enhancement of lung colonization has been reported to occur in association with increasing granulocytosis (Milas et   al., 1984) and a correlation between splenomegaly and metastases has been suggested (Sato et al., 1981). We studied in vitro colony-stimulating activity and in vivo haemopoietic alterations in the new murine TS/A cell line, that has been derived from a spontaneous mammary carcinoma and is able to metastasize spontaneously to the lung (Nanni et al., 1983). Moreover, we examined two sets of clones selected from TS/A line, which are all able to metastasize spontaneously but strongly differ in metastatic potential (E clones induce higher numbers of lung metastases than F clones) (Lollini et al., 1984).
TS/A conditioned medium exerted a colonystimulating activity on granulocyte-macrophage progenitors. Moreover, when the TS/A line was injected s.c., progressive spleen enlargement and thymic atrophy as well as a dramatic increase in peripheral granulocytic population were observed. Even though the TS/A line is able to metastasize, spleen was free from metastases and not able tQ induce tumours, when injected into syngeneic animals (data not shown). Serum from TS/Abearing mice was also found to stimulate proliferation of normal murine bone marrow cells. Therefore, a colony-stimulating activity was detected both in TS/A conditioned medium and in serum from TS/A bearing animals. However, the possibility that the alterations we observed in vivo could be due also to an interaction between tumour and host cells cannot be ruled out.
In vitro production of CSF and in vivo occurrence of splenomegaly and granulocytosis have been shown for all the TS/A clonal derivatives examined. We observed a discrepancy between in vitro and in vivo assays. In vitro both agar colony and [3H]-TdR uptake assays seem to indicate that supernatants from E clones have a CSF activity higher than those of F clones. On the contrary, such a pattern was not revealed by in vivo studies: mice injected with F2 cells displayed the strongest leukocytosis and splenomegaly whereas those injected with F5 cells were the least altered and all the other groups ranged in between.
We could not find any correlation between in vivo haematological parameters and other characteristics of TS/A clonal derivatives, such as in vitro doubling time, cell dimensions or in vivo tumour volume. We are currently examining three different hypotheses: either our cell lines produce a second factor which cannot be detected in vitro, or in vitro growth pattern of F clones does not allow a high CSF production, or some interaction with host environment occurs that alters CSF in vivo production.
When the possibility of a correlation between CSF production and metastases is considered, in vitro production of CSF clearly correlates with the relative metastatic capacity of E and F clones, but again in vivo haematologic parameters do not. Moreover, it should be borne in mind that E and F clones probably differ in the early steps of the metastatic process, since they do not give rise to significantly different numbers of lung colonies, when injected intravenously.
In conclusion, we believe that the relationship between CSF production and the metastatic process should be further explored, in particular in relation to late events (such as survival in the blood stream and attachment and growth in target organs) and possible reciprocal interactions between CSFproducing tumour cells and host cells elicited by CSF itself.