Int. J . Cancer: 49, 296302 (1991) 0 1991 Wiley-Liss, Inc.
Publication of the International Union Against Cancer Publication de I'Union lnternationale Contre le Cancer
TUMORICIDAL ALVEOLAR MACROPHAGE AND TUMOR INFILTRATING MACROPHAGE CELL LINES Alicia V. PALLERONI', Luigi VARESIO~, Rosemary B. WRIGHT'and Michael J. BRUNDA' 'Department of Oncology, Roche Research Center, Hofhann-La Roche, Nutley, NJ 07110; and 2National Cancer Institute, Frederick, MD 21701, USA. Continuous alveolar macrophage (AM) and tumorinfiltrated (TIM) cell lines have been generated from C57816J mice by in vitro infection with the 12 retrovirus carrying the v-raf and v-myc oncogens. Four cloned AM cell lines (AMJZC8, AMJZ-C10, AMJZ-CI I, AMJZ-CZO) and 3 cloned TIM cell lines (TIMJZ-C4, TIMJ2-C7 and TIMJZ-C15) were expanded for further characterization. Flow cytometry detected the product of the raf gene in the cytoplasm of all these cell lines. Studies on the tumoricidal properties of these AM and TIM cell lines demonstrated differences in their response to a panel of known macrophage activators. Four of these cell lines (AMJZ-C8, AMJ2-C10, TIMJZ-C7 and TIMJZ-CI 5) were activated following exposure to recombinant murine interferon gamma (rMulFN-y) but not lipopolysaccharide (LPS) or muramyl dipeptide (MDP). AMJZ-CZO was only activated by incubation with rMulFN-y plus LPS. AMJZ-CI I and TIMJZ-C4 are the cell lines that most closely resembled the response pattern of the parental AM and TIM, since they could be activated by either the combination of rMulFN-y plus LPS or rMulFN-y plus MDP. Constitutive expression of MHC-class-ll antigens was low on AMJZ-CI I or TIMJZ-C4 but was increased following exposure to rMulFN-y. Neither cell line secreted substantial amounts of IL-l or TNF but both secreted large amounts of IL-6. Thus these cell lines could be powerful tools to study AM and TIM activation and cytotoxicity.
AM and TIM share many properties with mononuclear phagocytes from other tissues, but they have distinctive structural, metabolic and functional characteristics (Warheit et al., 1988, Brunda et al., 1991). Tumoricidal activity of AM and TIM can be induced by in vitro or in vivo exposure of these cells to certain biological response modifiers (BRM) (Fidler et al., 1982; Brunda et al., 1991). In addition, the cytolytic capacity of activated AM has been correlated with therapeutic effects of BRMs against pulmonary metastases (Fidler et al., 1982). The mechanisms through which AM and TIM exert these activities remain unclear in part due to the difficulty in isolating sufficient numbers of cells from animals through lengthy and laborious procedures. Establishment of cell lines representative of defined AM and TIM populations would greatly assist in the study of these functions. A number of murine peritoneal and AM cell lines have been established by infection with SV40 (Mauel and Defendi, 1971; Cahoon and Mills, 1981; Innocent and Herscowitz, 1989) or by transfection of SV40 DNA into primary cells (Schwarzbaum et al., 1984). In addition, cell lines with macrophage-like functions obtained from virus-induced murine tumors (Raschke et al., 1978; Franz et al., 1985) or from immature hematopoietic precursor cells, which in contrast to mature macrophages, undergo marked proliferation in vitro, have been reported (Blasi et al., 1985; Roberson and Walker, 1988). However, to our knowledge, no tumoricidal murine AM or pulmonary TIM cell lines have been established. Recent work in our laboratory (Brunda et al., 1991) has determined that TIM from pulmonary metastases responded to a panel of BRM to induce the tumoricidal statc in a way similar to the AM response and differed from the pattern obtained with peritoneal macrophages. The purpose of the present study was to explore the feasibility of generating functional murine AM and TIM cell lines. In this report, we describe the characterization and functions of AM and pulmonary TIM cell lines.
MATERIAL AND METHODS
Mice Ten-week-old C57BL/6J female mice were obtained from Jackson Laboratories (Bar Harbor, ME). Cell lines The GG2EE cell line, generated as previously described (Blasi et al., 1985), was grown in DMEM supplemented with HEPES buffer (PH 7.4, 5 mM final concentration), L-glutamine (4 mM), gentamicin (50 pgiml), fungizone (4 pgiml), and 5% FBS. P815, a DBA/2 mastocytoma, was grown as a stationary suspension culture in RPMI-10% FBS while B16 FlO (a malignant melanoma) was grown as an adherent culture. All the cell lines were free of Mycoplasma contamination. Virus preparation The GG2EE cell line was the source of the 32 virus used. Supernatants from exponentially growing cultures of GG2EE were collected and cell debris removed by centrifugation (12,000 g, rotor SS34, Sorvall RC2) for 15 min. After clarification, the supernatants were filtered through a 0.45-pm membrane (Millipore, Bedford, MA) and the 52 virus was pelleted at 100,000 g for 1 hr in a Beckman ultracentrifuge, followed by re-suspension in 3% of the original volume in conditioned medium from syngeneic lung fibroblasts (CM), according to the procedure reported by Soderland and Naum CM) con(1975), using 50% Ham F12 plus 50% CM (F12 taining 2 kg/ml Polybrene (Sigma, St. Louis, MO).
+
Generation of macrophage cell lines AM and TIM were obtained as previously described (Brunda et al., 1991). AM and TIM were stimulated to proliferate in F12 + CM. Macrophage monolayers were periodically trypsinized (3 times a week) prior to viral infection for removal of fibroblasts and other non-monocytic cells. Macrophages were infected approximately 1 month after initiation of culture. Sub-confluent monolayers of AM and TIM proliferating in F12 CM were infected overnight with 1 ml of the viral suspension. Mock infected cultures received no viral suspension but received 2 pg/ml Polybrene. Forty-eight hours after CM was removed and replaced by DMEM infection F12 10% FBS. Clusters of non-adherent rapidly proliferating cells were evident 8 to 10 days after infection. Several AM and TIM cell lines growing in semi-solid medium (0.96% methylcellulose) were obtained and subsequently recloned by limiting dilution. Two clones (AMJ2-C11 and TIMJ2-C4) were extensively characterized while an additional 5 clones (AMJ2-C8, AMJ2-CI0, AMJ2-C20, TIMJ2-C7 and TIMJ2-CIS) were evaluated for tumoricidal activity.
+
+
Non-speci$c esterase staining Cells were sedimented onto glass slides by cytocentrifugation, air-dried and stained for non-specific esterase (a-naphthyl esterase) by the method of Koski et al. (1976), using a kit (Sigma) according to manufacturer's specifications.
Received: May 3, 1991.
297
CYTOTOXIC MACROPHAGE CELL LINES
Phagocytosis AMJ2-C11 and TIMJ2-C4 cells were incubated at 37°C for either 30 min or 3 hr with opsonized 1.8-Nm fluorescent monodispersed carboxylated microspheres (Polysciences, Warrington, PA) at a bead-to-cell ratio of 1 5 . The excess of beads was removed by centrifugation of the suspension on bovine serum cushion at 900 rpm (rotor HG-4L, Sorvall RC3 centrifuge). The cells at the interface were recovered and washed with PBS. The fluorescent bead uptake was analyzed using a Cytofluorograf system 50-H with 215 1 computer (BectonDickinson , Westwood, MA). Detection of cell-sur3cace antigens by flow cytometric analysis The following antibodies were used for flow cytometry analyses: MAb to MAC-1 and MAC-2. (Boehringer, Indianapolis, IN); anti-CD5 (anti-Lyt 1.2) and anti-Thy 1.2 (Cedarlane, Hornby, Canada); anti-Fc receptor NEI-032, and anti-Ly-5 NEI-020 (NEN, Boston, MA). Fluorescein isothiocyanateconjugated (F1TC)-anti-Ig (goat anti-rat and goat anti-mouse Ig) were purchased from Becton-Dickinson, Mountain View, CA and Cappel, Malvern, PA, respectively. Cells were washed twice in a pinocytosis-inhibiting buffer: PBS containing 0.2% sodium azide (SA) and 0.5% BSA, (PBS-SA-BSA). Cell densities were adjusted to 1 X lo7 cells/ml, and 100 kl of the cell suspensions (1 X lo6 cells), were placed in 96-well roundbottom microtiter plates (Costar, Cambridge, MA) mixed with the antibody and incubated for 45 min. All staining procedures were carried out at 4°C. Cells were then washed 3 times with PBS-SA-BSA buffer and 10 p1 of the appropriate FITC-anti-Ig were added. After a 2nd incubation of 45 min, cells were washed 3 times and analyzed using a Cytofluorograf system. Positive staining was determined by comparison with relevant control of non-specific anti-mouse or anti-rat Ig. Detection of intracellular raf by flow cytometry Cells were permeabilized and prepared for immunofluorescence flow cytometry essentially as described by Giardina et al. (1988), except that since these cells grow in suspension culture, the detachment step was omitted. The murine monoclonal antibody (MAb) STEGI 1 generated against a 30-kDa v-raf protein purified from an E. coli expression vector (Giardina et al., 1988) was used. Cytotoxicity assays Macrophage cell lines were incubated with BRM overnight and the tumoricidal activity of the lines was measured in an 18 to 24-hr lllh-release assay as described elsewhere (Brunda et al., 1983). For determination of total incorporated radioactivity, 1-100th volume of 1% SDS (Sigma) in 0.1 N NaOH was added prior to harvesting to the designated wells in order to release the total incorporated radioactivity in the target cells. Results were expressed as percent cytotoxicity calculated by the following formula: % cytotoxicity = [(A - B)/C] X 100 where A = cpm released by target plus effector cells; B = spontaneous cpin released by target cells; and C = total cpm incorporated in target cells. Samples were run in quadruplicate and SD was less than 10% of the mean in all cases. Experiments were performed a minimum of 4 times, with comparable results. Induction of MHC-class-II-antigenexpression Cells were seeded in 75-cm2 flasks in DMEM 5% FBS at a density of 1 x lo5 cell/ml and cultured for 3 days in medium or in the presence of 100 Uiml rMuIFN-y (Genentech, South San Francisco, CA). After the incubation period, cells were washed 3 times with PBS-SA-BSA. Cell densities were ad-
justed to 1 X lo7 cells/ml and 100 k1 were dispensed into 96-well microtiter plates (Costar) and incubated with anti-Ia (Cedarlane). After washes and staining with FITC-goat antimouse Ig, cells were assessed for expression of Ia by flowcytometry analysis. In some experiments, normal AM or TIM were stained with anti-Ia as described above. Cytokine assays Supernatants obtained from AMJ2-ClI or TIMJ2-C4 cells following in vitro culture overnight with BRM as described above were assayed for cytokine activities. In some experiments, BRM-incubated cells were washed, placed in media and incubated for an additional 24 hr. TNF activity was assayed on the basis of its cytotoxicity for L-M mouse fibroblasts (ATCC, Rockville, MD) as determined by an MTT dyereduction assay in the presence of 2 kg/ml Actinomycin D (Sigma) (Green et al., 1984). A reference interim human rTNF standard (National Institute for Biological Standards and Control, Potters Bar, UK) was used to estimate units of bioactivity. IL-1 activity was assayed on the D10.G4 cell line (Kaye et al., 1983). In our IL-1 assay, neither murine TNF-a or IL-6 was active. IL-6 was assayed on B9 cells (Van Snick et al., 1986) and neither IL-1 nor TNF were detected in this assay. No activity in the cytokine assays was obtained from supernatants containing BRM incubated in the absence of macrophages. RESULTS
Generation of AM and TIM cell lines Four cloned AM cell lines (AMJ2-C8, AMJZC10, AMJ2C11 and AMJ2-C20) and 3 cloned TIM cell lines (TIMJ2-C4, TIMJ2-C7 and TIMJ2-CIS) were obtained using the technique described in “Material and Methods”. These cells grew rapidly with a doubling time of 12 to 18 hr. The cell lines were macrophages based on the following criteria: (1) stained nonspecific esterase positive; (2) were phagocytic; and (3) expressed macrophage-associated cell-surface markers (MAC- 1, MAC-2, Fc receptor, Ly-5) but not T-cell (Thy-1, Lyt-1) or B-cell (Ig) markers (data not shown). It has been previously reported that other macrophage cell lines generated by infection with the J2 virus contain high levels of an intracellular raf-related product as detected with the murine MAb STEGI 1 (Giardina et al., 1988). Using this MAb, we have likewise been able to detect intense cytoplasmic staining of the AM and TIM lines (Fig. 1, a, b). High reactivity was observed in cells permeabilized by lysolecithin treatment, but in contrast no appreciable cell-surface staining was detected in unpermeabilized cells. Using identical assay conditions, no intracellular raf-related product was detected in the SV40-immortalized murine AM cell line C3HAMSV40 (Palleroni et al., 1989) (data not shown). These results demonstrate that the macrophage cell lines described in this study did not arise spontaneously but were originated subsequent to infection with the 52 virus. Tumoricidal activities Because killing of tumor cells by macrophages generally requires macrophage activation, we next studied the ability of these lines to become tumoricidal after incubation with a variety of activators. None of the 7 cell lines was constitutively cytotoxic (data not shown). However, as shown in Figure 2a, following incubation with activating agents, the AM cell lines exhibited two distinct patterns of tumoricidal activities. In the case of AMJ2-C8 and AMJZC10, rMuIFN-y alone induced significant levels of cytotoxicity against P8 15 tumor target cells, but no cytotoxicity was observed after incubation of these cells with LPS or MDP. On the other hand, incubation of AMJ2-C11 and AMJ2-C20 with either rMuIFN-y or LPS resulted in low or undetectable levels of cytotoxicity but, in
298
PALLERONI ET AL.
MJ2- C l l ANTI V - R A F
15’+
(B)
A
L C 0 U
N
T 8 1
L
:i
’“+
I L O G GREEN FLUORESCENCE
ANTI V-RAF
ieee
ID)
L c
C 0 U N T 8 1
L O 6 GREEN FLUORESCENCE
1909
TIMIZ-CI #us B
‘1 L
FIGURE1 - Detection of rufin AMJ2-C11 and TIMJ2-C4 cells. Cells were prepared for immunofluorescence as described in “Material and Methods”. Unpermeabilized cells stained with (a) FITC-goat anti-mouse Ig only or (b) anti-ruf MAb STEGI 1 followed by FITC-goat anti-mouse Ig. Permeabilized cells stained with (c) FITC-goat anti-mouse Ig only or (d) anti-ruf MAb STEGI 1 followed by FITC-goat
anti-mouse Ig.
contrast, the combination of rMuIFN-y plus LPS induced very high levels of tumoricidal activity. In addition, AMJ2-C11 cells were also activated by the combination of MDP plus rMuIFN-y, but not by either agent alone. The activation of the AM cell lines was dose-dependent, with AMJ2-Cl1, for example, requiring a minimum of 0.8 U/ml of rMuIFN-y in combination with 5 kg/rnl of LPS to evoke the cytotoxic response. When we compared the cytotoxic activity of these lines with that of normal resident murine AM obtained from C57BL/ 65 mice, we found that AMJ2-C11 cells were the ones that most closely resembled the normal AM, since the 2 populations were similarly activated to become tumoricidal by the
combination of rMuIFN-y plus LPS and the combination of rMuIFN-y plus MDP (Brunda ef al., 1991). A similar activation pattern was observed with TIM (Fig. 2b), since 2 of the lines, TIMJ2-C7 and TIMJ2-Cl5, were cytotoxic to P8 15 cells following incubation with rMuIFN-y alone but not after incubation with LPS or MDP. The 3rd cell line TIMJ2-C4 behaved like the parental TIM (Brunda et al., 1991) for the reason that these cells became tumoricidal only after incubation with combinations of rMuIFN-y plus LPS or rMuIFN-y plus MDP. The kinetics of in vitro induction of the turnoricidal state of AMJ2-C11 cells was then investigated. Cells were incubated
299
CYTOTOXIC MACROPHAGE CELL LINES
r
MDP
IM7 80 L
t
“1 .\&
,
10 0
70
6o
0
200
1
22
2.5 u/ml
200
0.3
22
0.3
2.5 W/ml
6 IFN CAMW + LPS (S ug/ml)
I . . . . . . . .t MDP t . TIMJP-CI
nw2-ci 5 TIMJZ-C~
A-A
)--.
’
lo0
L 200
22
2.5
0.3
-
A-
.-=-m
-
_
-
-
*
=
-
-
=
=
-
I
FIGURE2 - Cytotoxic activity of AM (a)and TIM fb) cell lines against P815 cells. Cells (2.5 X lo5 cellslwell) were incubated with the indicated concentrations of IFN-y, LPS, MDP, IFN-y plus 5 pgiml LPS, LPS plus 5 Ulml IFN-y, and MDP plus 5 Uiml IFN-y for 18 to 24 hr. Cytotoxicity was assayed at an effector-to-target-cellratio of 25:l. The representative data above are from an experiment in which all cell lines were tested for their cytotoxic activity in identical conditions.
300
PALLERONI ET A L .
with medium, rMuIFN-y, or rMuIFN-y plus LPS for periods of 2, 4, 8, 12 and 24 hr, in order to determine the minimum time required for induction of a cytotoxic response after incubation of AMJ2-Cl1 with activating agents. At each time point, activators were removed and cells were exposed to "'InOx-labelled P815 target cells for approximately 20 hr. Medium or rMuIFN-y failed to elicit a cytotoxic response even after prolonged (24-hr) incubation. In contrast, incubation of AMJ2-Cll cells with rMuIFN-y plus LPS for only 8 hr was sufficient to induce tumoricidal activity. This response reached a maximum at 12 hr and was sustained for at least 24 hr (Fig. 3a). To examine the kinetics of cell lysis, AMJ2-CI 1 cells were incubated either with medium, rMuIFN-y , or rMuIFN-y plus LPS for 24 hr. After removal of activators, cells were incubated with l'lInOx-labelled P815 target cells for 2, 4, 8, 24 and 48 hr. We observed that a minimum of 12 hr of contact was required for lysis to occur, while it continued to increase through 48 hr of effector-target contact (Fig. 3b). These findings demonstrate that the kinetics of AMJ2-C 1 1-mediated kill-
A
KINETICS OF ACTIVATION
0-0 A-A
MEDIUM -7
IFN
W-¤ IM-7
+ LPS
3 5 --
Expression of MHC-class-If antigens Since macrophages are capable of functioning as accessory cells (Weinberg and Unanue, 1981), it was important to establish whether the AMJ2-C11 and TIMJ2-C4 cell lines were able to synthesize and express the MHC-class41 antigens on their membranes. By flow-cytometry analysis, approximately 18% of normal AM and less than 5% of TIM were found to express Ia. Consistent with this finding, very low levels of MHC-classI1 antigens were constitutively present on AMJ2-C11 and TIMJ2-C4 cells. In codtrast, as has been reported with other murine macrophages (Wong et al., 1984), high levels of MHCclass-11-antigenexpression were observed following incubation of these cell lines with 100 U/ml of rMuIFN-y (Fig. 4a, b). A lower percentage of TIMJ2-C4 cells expressed the MHCClass-I1 antigens following culture with rMuIFN-y . Thus, these lines express low levels of MHC-class-11 antigens, like the parental macrophage populations from which they are derived, but the basal level of MHC-class41 antigen expression can be significantly augmented by rMuIFN-y. Cytokine secretion The ability of AMJ2-C 1 1 and TIMJ2-C4 to secrete cytokines in response to BRM was then examined (Table I). Neither cell line produced detectable amounts of IL-I in response to any BRM tested on day 1 , and only minimal amounts on day 2. Minimal amounts of TNF were produced on day 1 and none on day 2. In contrast, both AM and TIM produce IL-I and AM secrete large amounts of TNF (Brunda et al., 1991). The inability of these cell lines to produce IL-1 or TNF is not a general failure to secrete cytokines, since both cell lines secrete substantial amounts of IL-6; of note, cells stimulated with LPS plus rMuIFN-y consistently secreted higher amounts of IL-6 than cells stimulated with LPS alone. Thus these cloned cell lines, which maintain the functional cytotoxic properties of AM and TIM, do not have the same functional phenotype with respect to cytokine production.
/-------.
30 -.
i -54
/
I
0
0
4
12
20
16
24
DISCUSSION
T I M E (HRS)
B
KINETICS
OF
KILLING
0-0 A-A
.-B
- 5 4 : : : 0
ing of P8 15 cells are characteristic of macrophage-mediated cytotoxicity in vitro (Taniyama and Holden, 1981).
4
8
12
MEDIUM IFN 7
-
IFN-7
+ LPS
:
:
:
:
:
;
:
:
I
16
20
24
20
32
36
40
44
40
T I M E (HRS)
FIGURE3 - Kinetics of activation and of killing for AMJ2-ClI Cells. (a)AMJ2-ClI Cell (2 X lo5 Celk/Well) Were incubated with either medium%IFN-?I Uiml), Or IFN-y Plus Lps (50 'Im1 + Fg/ml) for varying times. Cytotoxicity was assayed at an effector-totarget-cell ratio of25:1. (b)AMJ2-Cl1 cells (2 X lo5 cells/well) were incubated with either medium, I F N - ~(50 ~ / m l )or , 1m-yplus ~ p (50 Uiml + 5 pgiml) for 24 hr. Cytotoxicity at an effector-to-targetcell ratio of 2 5 1 was assessed at varying times.
The study of AM and TIM has been hampered in large part by the difficulty of obtaining cells in sufficient numbers. Although AM and TIM maintain growth potential in vivo and are capable of substantial proliferation following environmental or mechanical insults (Innocent and Herscowitz, 1989), their ability to grow in culture is severely restricted (Cahoon and Mills, 1981; Innocent and Herscowitz, 1989). Our goal was to generate immortalized AM and TIM cloned cell lines that would maintain stable phenotype and tumoricidal properties that were similar to those of the parental AM and TIM populations. Although several murine AM cell lines, generated by infection with SV40, have been described (Cahoon and Mills, 1981; Innocent and Herscowitz, 1989), these cell lines have not been reported to possess tumoricidal activity. Furthermore, there are no cell lines that have been generated from TIM. Using the recombinant retrovirus 52, carrying that v-raf and v-myc oncogenes, we have been successful in obtaining immortalized macrophage cell lines from AM and TIM parental populations. A very important macrophage function is the ability of these cells to be activated to the tumoricidal state. Cloning of the heterogeneous immortalized AM and TIM popula~ons resulted in cell lines which exhibited 2 distinct patterns of cytotoxic activity. The AMJ2-C8, AMJz-ClO, TIMJ2-J and T1MJ2-C15 response pattern (Fig. lines gave a very s 2). These Ceh were stimulated to become tumoricidal f o k w ing exposure to rMuIFN-y alone. Activation of these cells with the combination of rMuIFN-y plus LPS or rMuIFN-y plus
30 1
CYTOTOXIC MACROPHAGE CELL LINES
iue
188
A
913
26 148 166 188 L O G GREEN FLUORESCENCE
B
9e
266
FIGURE4 - Induction of Ia expression on AMJ2-CI 1 ( a ) and TIMJ2-C4 ( b ) cells following incubation with rMuIFN-y. (- - -) Mediumincubated cells stained with FITC-goat anti-mouse Ig. (-) Medium-incubated cells stained with anti-Ia followed by FITC-goat anti-mouse Ig. (-) IFN-y-treated cells stained with anti-Ia followed by FITC-goat anti-mouse Ig. TABLE I - BRM INDUCTION OF CYTOTOXICITY AND CYTOKINE SECRETION
Day 1 BRM
None rMuIFN-y LPS rMuIFNr + LPS BRM
None rMuIFN-y LPS rMuIFNr + LPS
% cytotoxl'Ityl
It. l 2
TN F3
1L-64
-
AMJZCll
TIMJZC4
AMJ2CI I
TIMJ2C4
AMJZCI 1
TIMJ2C4
AMJ2Cl 1
TIMJ2C4
0 0 0 61
0 0 0 32
Publication of the International Union Against Cancer Publication de I'Union lnternationale Contre le Cancer
TUMORICIDAL ALVEOLAR MACROPHAGE AND TUMOR INFILTRATING MACROPHAGE CELL LINES Alicia V. PALLERONI', Luigi VARESIO~, Rosemary B. WRIGHT'and Michael J. BRUNDA' 'Department of Oncology, Roche Research Center, Hofhann-La Roche, Nutley, NJ 07110; and 2National Cancer Institute, Frederick, MD 21701, USA. Continuous alveolar macrophage (AM) and tumorinfiltrated (TIM) cell lines have been generated from C57816J mice by in vitro infection with the 12 retrovirus carrying the v-raf and v-myc oncogens. Four cloned AM cell lines (AMJZC8, AMJZ-C10, AMJZ-CI I, AMJZ-CZO) and 3 cloned TIM cell lines (TIMJZ-C4, TIMJ2-C7 and TIMJZ-C15) were expanded for further characterization. Flow cytometry detected the product of the raf gene in the cytoplasm of all these cell lines. Studies on the tumoricidal properties of these AM and TIM cell lines demonstrated differences in their response to a panel of known macrophage activators. Four of these cell lines (AMJZ-C8, AMJ2-C10, TIMJZ-C7 and TIMJZ-CI 5) were activated following exposure to recombinant murine interferon gamma (rMulFN-y) but not lipopolysaccharide (LPS) or muramyl dipeptide (MDP). AMJZ-CZO was only activated by incubation with rMulFN-y plus LPS. AMJZ-CI I and TIMJZ-C4 are the cell lines that most closely resembled the response pattern of the parental AM and TIM, since they could be activated by either the combination of rMulFN-y plus LPS or rMulFN-y plus MDP. Constitutive expression of MHC-class-ll antigens was low on AMJZ-CI I or TIMJZ-C4 but was increased following exposure to rMulFN-y. Neither cell line secreted substantial amounts of IL-l or TNF but both secreted large amounts of IL-6. Thus these cell lines could be powerful tools to study AM and TIM activation and cytotoxicity.
AM and TIM share many properties with mononuclear phagocytes from other tissues, but they have distinctive structural, metabolic and functional characteristics (Warheit et al., 1988, Brunda et al., 1991). Tumoricidal activity of AM and TIM can be induced by in vitro or in vivo exposure of these cells to certain biological response modifiers (BRM) (Fidler et al., 1982; Brunda et al., 1991). In addition, the cytolytic capacity of activated AM has been correlated with therapeutic effects of BRMs against pulmonary metastases (Fidler et al., 1982). The mechanisms through which AM and TIM exert these activities remain unclear in part due to the difficulty in isolating sufficient numbers of cells from animals through lengthy and laborious procedures. Establishment of cell lines representative of defined AM and TIM populations would greatly assist in the study of these functions. A number of murine peritoneal and AM cell lines have been established by infection with SV40 (Mauel and Defendi, 1971; Cahoon and Mills, 1981; Innocent and Herscowitz, 1989) or by transfection of SV40 DNA into primary cells (Schwarzbaum et al., 1984). In addition, cell lines with macrophage-like functions obtained from virus-induced murine tumors (Raschke et al., 1978; Franz et al., 1985) or from immature hematopoietic precursor cells, which in contrast to mature macrophages, undergo marked proliferation in vitro, have been reported (Blasi et al., 1985; Roberson and Walker, 1988). However, to our knowledge, no tumoricidal murine AM or pulmonary TIM cell lines have been established. Recent work in our laboratory (Brunda et al., 1991) has determined that TIM from pulmonary metastases responded to a panel of BRM to induce the tumoricidal statc in a way similar to the AM response and differed from the pattern obtained with peritoneal macrophages. The purpose of the present study was to explore the feasibility of generating functional murine AM and TIM cell lines. In this report, we describe the characterization and functions of AM and pulmonary TIM cell lines.
MATERIAL AND METHODS
Mice Ten-week-old C57BL/6J female mice were obtained from Jackson Laboratories (Bar Harbor, ME). Cell lines The GG2EE cell line, generated as previously described (Blasi et al., 1985), was grown in DMEM supplemented with HEPES buffer (PH 7.4, 5 mM final concentration), L-glutamine (4 mM), gentamicin (50 pgiml), fungizone (4 pgiml), and 5% FBS. P815, a DBA/2 mastocytoma, was grown as a stationary suspension culture in RPMI-10% FBS while B16 FlO (a malignant melanoma) was grown as an adherent culture. All the cell lines were free of Mycoplasma contamination. Virus preparation The GG2EE cell line was the source of the 32 virus used. Supernatants from exponentially growing cultures of GG2EE were collected and cell debris removed by centrifugation (12,000 g, rotor SS34, Sorvall RC2) for 15 min. After clarification, the supernatants were filtered through a 0.45-pm membrane (Millipore, Bedford, MA) and the 52 virus was pelleted at 100,000 g for 1 hr in a Beckman ultracentrifuge, followed by re-suspension in 3% of the original volume in conditioned medium from syngeneic lung fibroblasts (CM), according to the procedure reported by Soderland and Naum CM) con(1975), using 50% Ham F12 plus 50% CM (F12 taining 2 kg/ml Polybrene (Sigma, St. Louis, MO).
+
Generation of macrophage cell lines AM and TIM were obtained as previously described (Brunda et al., 1991). AM and TIM were stimulated to proliferate in F12 + CM. Macrophage monolayers were periodically trypsinized (3 times a week) prior to viral infection for removal of fibroblasts and other non-monocytic cells. Macrophages were infected approximately 1 month after initiation of culture. Sub-confluent monolayers of AM and TIM proliferating in F12 CM were infected overnight with 1 ml of the viral suspension. Mock infected cultures received no viral suspension but received 2 pg/ml Polybrene. Forty-eight hours after CM was removed and replaced by DMEM infection F12 10% FBS. Clusters of non-adherent rapidly proliferating cells were evident 8 to 10 days after infection. Several AM and TIM cell lines growing in semi-solid medium (0.96% methylcellulose) were obtained and subsequently recloned by limiting dilution. Two clones (AMJ2-C11 and TIMJ2-C4) were extensively characterized while an additional 5 clones (AMJ2-C8, AMJ2-CI0, AMJ2-C20, TIMJ2-C7 and TIMJ2-CIS) were evaluated for tumoricidal activity.
+
+
Non-speci$c esterase staining Cells were sedimented onto glass slides by cytocentrifugation, air-dried and stained for non-specific esterase (a-naphthyl esterase) by the method of Koski et al. (1976), using a kit (Sigma) according to manufacturer's specifications.
Received: May 3, 1991.
297
CYTOTOXIC MACROPHAGE CELL LINES
Phagocytosis AMJ2-C11 and TIMJ2-C4 cells were incubated at 37°C for either 30 min or 3 hr with opsonized 1.8-Nm fluorescent monodispersed carboxylated microspheres (Polysciences, Warrington, PA) at a bead-to-cell ratio of 1 5 . The excess of beads was removed by centrifugation of the suspension on bovine serum cushion at 900 rpm (rotor HG-4L, Sorvall RC3 centrifuge). The cells at the interface were recovered and washed with PBS. The fluorescent bead uptake was analyzed using a Cytofluorograf system 50-H with 215 1 computer (BectonDickinson , Westwood, MA). Detection of cell-sur3cace antigens by flow cytometric analysis The following antibodies were used for flow cytometry analyses: MAb to MAC-1 and MAC-2. (Boehringer, Indianapolis, IN); anti-CD5 (anti-Lyt 1.2) and anti-Thy 1.2 (Cedarlane, Hornby, Canada); anti-Fc receptor NEI-032, and anti-Ly-5 NEI-020 (NEN, Boston, MA). Fluorescein isothiocyanateconjugated (F1TC)-anti-Ig (goat anti-rat and goat anti-mouse Ig) were purchased from Becton-Dickinson, Mountain View, CA and Cappel, Malvern, PA, respectively. Cells were washed twice in a pinocytosis-inhibiting buffer: PBS containing 0.2% sodium azide (SA) and 0.5% BSA, (PBS-SA-BSA). Cell densities were adjusted to 1 X lo7 cells/ml, and 100 kl of the cell suspensions (1 X lo6 cells), were placed in 96-well roundbottom microtiter plates (Costar, Cambridge, MA) mixed with the antibody and incubated for 45 min. All staining procedures were carried out at 4°C. Cells were then washed 3 times with PBS-SA-BSA buffer and 10 p1 of the appropriate FITC-anti-Ig were added. After a 2nd incubation of 45 min, cells were washed 3 times and analyzed using a Cytofluorograf system. Positive staining was determined by comparison with relevant control of non-specific anti-mouse or anti-rat Ig. Detection of intracellular raf by flow cytometry Cells were permeabilized and prepared for immunofluorescence flow cytometry essentially as described by Giardina et al. (1988), except that since these cells grow in suspension culture, the detachment step was omitted. The murine monoclonal antibody (MAb) STEGI 1 generated against a 30-kDa v-raf protein purified from an E. coli expression vector (Giardina et al., 1988) was used. Cytotoxicity assays Macrophage cell lines were incubated with BRM overnight and the tumoricidal activity of the lines was measured in an 18 to 24-hr lllh-release assay as described elsewhere (Brunda et al., 1983). For determination of total incorporated radioactivity, 1-100th volume of 1% SDS (Sigma) in 0.1 N NaOH was added prior to harvesting to the designated wells in order to release the total incorporated radioactivity in the target cells. Results were expressed as percent cytotoxicity calculated by the following formula: % cytotoxicity = [(A - B)/C] X 100 where A = cpm released by target plus effector cells; B = spontaneous cpin released by target cells; and C = total cpm incorporated in target cells. Samples were run in quadruplicate and SD was less than 10% of the mean in all cases. Experiments were performed a minimum of 4 times, with comparable results. Induction of MHC-class-II-antigenexpression Cells were seeded in 75-cm2 flasks in DMEM 5% FBS at a density of 1 x lo5 cell/ml and cultured for 3 days in medium or in the presence of 100 Uiml rMuIFN-y (Genentech, South San Francisco, CA). After the incubation period, cells were washed 3 times with PBS-SA-BSA. Cell densities were ad-
justed to 1 X lo7 cells/ml and 100 k1 were dispensed into 96-well microtiter plates (Costar) and incubated with anti-Ia (Cedarlane). After washes and staining with FITC-goat antimouse Ig, cells were assessed for expression of Ia by flowcytometry analysis. In some experiments, normal AM or TIM were stained with anti-Ia as described above. Cytokine assays Supernatants obtained from AMJ2-ClI or TIMJ2-C4 cells following in vitro culture overnight with BRM as described above were assayed for cytokine activities. In some experiments, BRM-incubated cells were washed, placed in media and incubated for an additional 24 hr. TNF activity was assayed on the basis of its cytotoxicity for L-M mouse fibroblasts (ATCC, Rockville, MD) as determined by an MTT dyereduction assay in the presence of 2 kg/ml Actinomycin D (Sigma) (Green et al., 1984). A reference interim human rTNF standard (National Institute for Biological Standards and Control, Potters Bar, UK) was used to estimate units of bioactivity. IL-1 activity was assayed on the D10.G4 cell line (Kaye et al., 1983). In our IL-1 assay, neither murine TNF-a or IL-6 was active. IL-6 was assayed on B9 cells (Van Snick et al., 1986) and neither IL-1 nor TNF were detected in this assay. No activity in the cytokine assays was obtained from supernatants containing BRM incubated in the absence of macrophages. RESULTS
Generation of AM and TIM cell lines Four cloned AM cell lines (AMJ2-C8, AMJZC10, AMJ2C11 and AMJ2-C20) and 3 cloned TIM cell lines (TIMJ2-C4, TIMJ2-C7 and TIMJ2-CIS) were obtained using the technique described in “Material and Methods”. These cells grew rapidly with a doubling time of 12 to 18 hr. The cell lines were macrophages based on the following criteria: (1) stained nonspecific esterase positive; (2) were phagocytic; and (3) expressed macrophage-associated cell-surface markers (MAC- 1, MAC-2, Fc receptor, Ly-5) but not T-cell (Thy-1, Lyt-1) or B-cell (Ig) markers (data not shown). It has been previously reported that other macrophage cell lines generated by infection with the J2 virus contain high levels of an intracellular raf-related product as detected with the murine MAb STEGI 1 (Giardina et al., 1988). Using this MAb, we have likewise been able to detect intense cytoplasmic staining of the AM and TIM lines (Fig. 1, a, b). High reactivity was observed in cells permeabilized by lysolecithin treatment, but in contrast no appreciable cell-surface staining was detected in unpermeabilized cells. Using identical assay conditions, no intracellular raf-related product was detected in the SV40-immortalized murine AM cell line C3HAMSV40 (Palleroni et al., 1989) (data not shown). These results demonstrate that the macrophage cell lines described in this study did not arise spontaneously but were originated subsequent to infection with the 52 virus. Tumoricidal activities Because killing of tumor cells by macrophages generally requires macrophage activation, we next studied the ability of these lines to become tumoricidal after incubation with a variety of activators. None of the 7 cell lines was constitutively cytotoxic (data not shown). However, as shown in Figure 2a, following incubation with activating agents, the AM cell lines exhibited two distinct patterns of tumoricidal activities. In the case of AMJ2-C8 and AMJZC10, rMuIFN-y alone induced significant levels of cytotoxicity against P8 15 tumor target cells, but no cytotoxicity was observed after incubation of these cells with LPS or MDP. On the other hand, incubation of AMJ2-C11 and AMJ2-C20 with either rMuIFN-y or LPS resulted in low or undetectable levels of cytotoxicity but, in
298
PALLERONI ET AL.
MJ2- C l l ANTI V - R A F
15’+
(B)
A
L C 0 U
N
T 8 1
L
:i
’“+
I L O G GREEN FLUORESCENCE
ANTI V-RAF
ieee
ID)
L c
C 0 U N T 8 1
L O 6 GREEN FLUORESCENCE
1909
TIMIZ-CI #us B
‘1 L
FIGURE1 - Detection of rufin AMJ2-C11 and TIMJ2-C4 cells. Cells were prepared for immunofluorescence as described in “Material and Methods”. Unpermeabilized cells stained with (a) FITC-goat anti-mouse Ig only or (b) anti-ruf MAb STEGI 1 followed by FITC-goat anti-mouse Ig. Permeabilized cells stained with (c) FITC-goat anti-mouse Ig only or (d) anti-ruf MAb STEGI 1 followed by FITC-goat
anti-mouse Ig.
contrast, the combination of rMuIFN-y plus LPS induced very high levels of tumoricidal activity. In addition, AMJ2-C11 cells were also activated by the combination of MDP plus rMuIFN-y, but not by either agent alone. The activation of the AM cell lines was dose-dependent, with AMJ2-Cl1, for example, requiring a minimum of 0.8 U/ml of rMuIFN-y in combination with 5 kg/rnl of LPS to evoke the cytotoxic response. When we compared the cytotoxic activity of these lines with that of normal resident murine AM obtained from C57BL/ 65 mice, we found that AMJ2-C11 cells were the ones that most closely resembled the normal AM, since the 2 populations were similarly activated to become tumoricidal by the
combination of rMuIFN-y plus LPS and the combination of rMuIFN-y plus MDP (Brunda ef al., 1991). A similar activation pattern was observed with TIM (Fig. 2b), since 2 of the lines, TIMJ2-C7 and TIMJ2-Cl5, were cytotoxic to P8 15 cells following incubation with rMuIFN-y alone but not after incubation with LPS or MDP. The 3rd cell line TIMJ2-C4 behaved like the parental TIM (Brunda et al., 1991) for the reason that these cells became tumoricidal only after incubation with combinations of rMuIFN-y plus LPS or rMuIFN-y plus MDP. The kinetics of in vitro induction of the turnoricidal state of AMJ2-C11 cells was then investigated. Cells were incubated
299
CYTOTOXIC MACROPHAGE CELL LINES
r
MDP
IM7 80 L
t
“1 .\&
,
10 0
70
6o
0
200
1
22
2.5 u/ml
200
0.3
22
0.3
2.5 W/ml
6 IFN CAMW + LPS (S ug/ml)
I . . . . . . . .t MDP t . TIMJP-CI
nw2-ci 5 TIMJZ-C~
A-A
)--.
’
lo0
L 200
22
2.5
0.3
-
A-
.-=-m
-
_
-
-
*
=
-
-
=
=
-
I
FIGURE2 - Cytotoxic activity of AM (a)and TIM fb) cell lines against P815 cells. Cells (2.5 X lo5 cellslwell) were incubated with the indicated concentrations of IFN-y, LPS, MDP, IFN-y plus 5 pgiml LPS, LPS plus 5 Ulml IFN-y, and MDP plus 5 Uiml IFN-y for 18 to 24 hr. Cytotoxicity was assayed at an effector-to-target-cellratio of 25:l. The representative data above are from an experiment in which all cell lines were tested for their cytotoxic activity in identical conditions.
300
PALLERONI ET A L .
with medium, rMuIFN-y, or rMuIFN-y plus LPS for periods of 2, 4, 8, 12 and 24 hr, in order to determine the minimum time required for induction of a cytotoxic response after incubation of AMJ2-Cl1 with activating agents. At each time point, activators were removed and cells were exposed to "'InOx-labelled P815 target cells for approximately 20 hr. Medium or rMuIFN-y failed to elicit a cytotoxic response even after prolonged (24-hr) incubation. In contrast, incubation of AMJ2-Cll cells with rMuIFN-y plus LPS for only 8 hr was sufficient to induce tumoricidal activity. This response reached a maximum at 12 hr and was sustained for at least 24 hr (Fig. 3a). To examine the kinetics of cell lysis, AMJ2-CI 1 cells were incubated either with medium, rMuIFN-y , or rMuIFN-y plus LPS for 24 hr. After removal of activators, cells were incubated with l'lInOx-labelled P815 target cells for 2, 4, 8, 24 and 48 hr. We observed that a minimum of 12 hr of contact was required for lysis to occur, while it continued to increase through 48 hr of effector-target contact (Fig. 3b). These findings demonstrate that the kinetics of AMJ2-C 1 1-mediated kill-
A
KINETICS OF ACTIVATION
0-0 A-A
MEDIUM -7
IFN
W-¤ IM-7
+ LPS
3 5 --
Expression of MHC-class-If antigens Since macrophages are capable of functioning as accessory cells (Weinberg and Unanue, 1981), it was important to establish whether the AMJ2-C11 and TIMJ2-C4 cell lines were able to synthesize and express the MHC-class41 antigens on their membranes. By flow-cytometry analysis, approximately 18% of normal AM and less than 5% of TIM were found to express Ia. Consistent with this finding, very low levels of MHC-classI1 antigens were constitutively present on AMJ2-C11 and TIMJ2-C4 cells. In codtrast, as has been reported with other murine macrophages (Wong et al., 1984), high levels of MHCclass-11-antigenexpression were observed following incubation of these cell lines with 100 U/ml of rMuIFN-y (Fig. 4a, b). A lower percentage of TIMJ2-C4 cells expressed the MHCClass-I1 antigens following culture with rMuIFN-y . Thus, these lines express low levels of MHC-class-11 antigens, like the parental macrophage populations from which they are derived, but the basal level of MHC-class41 antigen expression can be significantly augmented by rMuIFN-y. Cytokine secretion The ability of AMJ2-C 1 1 and TIMJ2-C4 to secrete cytokines in response to BRM was then examined (Table I). Neither cell line produced detectable amounts of IL-I in response to any BRM tested on day 1 , and only minimal amounts on day 2. Minimal amounts of TNF were produced on day 1 and none on day 2. In contrast, both AM and TIM produce IL-I and AM secrete large amounts of TNF (Brunda et al., 1991). The inability of these cell lines to produce IL-1 or TNF is not a general failure to secrete cytokines, since both cell lines secrete substantial amounts of IL-6; of note, cells stimulated with LPS plus rMuIFN-y consistently secreted higher amounts of IL-6 than cells stimulated with LPS alone. Thus these cloned cell lines, which maintain the functional cytotoxic properties of AM and TIM, do not have the same functional phenotype with respect to cytokine production.
/-------.
30 -.
i -54
/
I
0
0
4
12
20
16
24
DISCUSSION
T I M E (HRS)
B
KINETICS
OF
KILLING
0-0 A-A
.-B
- 5 4 : : : 0
ing of P8 15 cells are characteristic of macrophage-mediated cytotoxicity in vitro (Taniyama and Holden, 1981).
4
8
12
MEDIUM IFN 7
-
IFN-7
+ LPS
:
:
:
:
:
;
:
:
I
16
20
24
20
32
36
40
44
40
T I M E (HRS)
FIGURE3 - Kinetics of activation and of killing for AMJ2-ClI Cells. (a)AMJ2-ClI Cell (2 X lo5 Celk/Well) Were incubated with either medium%IFN-?I Uiml), Or IFN-y Plus Lps (50 'Im1 + Fg/ml) for varying times. Cytotoxicity was assayed at an effector-totarget-cell ratio of25:1. (b)AMJ2-Cl1 cells (2 X lo5 cells/well) were incubated with either medium, I F N - ~(50 ~ / m l )or , 1m-yplus ~ p (50 Uiml + 5 pgiml) for 24 hr. Cytotoxicity at an effector-to-targetcell ratio of 2 5 1 was assessed at varying times.
The study of AM and TIM has been hampered in large part by the difficulty of obtaining cells in sufficient numbers. Although AM and TIM maintain growth potential in vivo and are capable of substantial proliferation following environmental or mechanical insults (Innocent and Herscowitz, 1989), their ability to grow in culture is severely restricted (Cahoon and Mills, 1981; Innocent and Herscowitz, 1989). Our goal was to generate immortalized AM and TIM cloned cell lines that would maintain stable phenotype and tumoricidal properties that were similar to those of the parental AM and TIM populations. Although several murine AM cell lines, generated by infection with SV40, have been described (Cahoon and Mills, 1981; Innocent and Herscowitz, 1989), these cell lines have not been reported to possess tumoricidal activity. Furthermore, there are no cell lines that have been generated from TIM. Using the recombinant retrovirus 52, carrying that v-raf and v-myc oncogenes, we have been successful in obtaining immortalized macrophage cell lines from AM and TIM parental populations. A very important macrophage function is the ability of these cells to be activated to the tumoricidal state. Cloning of the heterogeneous immortalized AM and TIM popula~ons resulted in cell lines which exhibited 2 distinct patterns of cytotoxic activity. The AMJ2-C8, AMJz-ClO, TIMJ2-J and T1MJ2-C15 response pattern (Fig. lines gave a very s 2). These Ceh were stimulated to become tumoricidal f o k w ing exposure to rMuIFN-y alone. Activation of these cells with the combination of rMuIFN-y plus LPS or rMuIFN-y plus
30 1
CYTOTOXIC MACROPHAGE CELL LINES
iue
188
A
913
26 148 166 188 L O G GREEN FLUORESCENCE
B
9e
266
FIGURE4 - Induction of Ia expression on AMJ2-CI 1 ( a ) and TIMJ2-C4 ( b ) cells following incubation with rMuIFN-y. (- - -) Mediumincubated cells stained with FITC-goat anti-mouse Ig. (-) Medium-incubated cells stained with anti-Ia followed by FITC-goat anti-mouse Ig. (-) IFN-y-treated cells stained with anti-Ia followed by FITC-goat anti-mouse Ig. TABLE I - BRM INDUCTION OF CYTOTOXICITY AND CYTOKINE SECRETION
Day 1 BRM
None rMuIFN-y LPS rMuIFNr + LPS BRM
None rMuIFN-y LPS rMuIFNr + LPS
% cytotoxl'Ityl
It. l 2
TN F3
1L-64
-
AMJZCll
TIMJZC4
AMJ2CI I
TIMJ2C4
AMJZCI 1
TIMJ2C4
AMJ2Cl 1
TIMJ2C4
0 0 0 61
0 0 0 32
