Spontaneous hybridization of macrophages and Meth A sarcoma cells

SPONTANEOUS HYBRIDIZATION OF MACROPHAGES AND METH ASARCOMA CELLSLill-Tove R. BUSUND

1*, Mette K. KILLIE2, Kristian BARTNES

3, Randi OLSEN1 and Rolf SELJELID

2

1Department of Clinical Pathology, Institute of Medical Biology, University of Tromsø, Tromsø, Norway2Department of Experimental Pathology, Institute of Medical Biology, University of Tromsø, Tromsø, Norway3Department of Cardiovascular Surgery, Tromsø University Hospital, Tromsø, Norway

We present evidence of hybridization between Meth Asarcoma cells and syngeneic as well as semigeneic peritonealmacrophages. The resultant hybrids are characterized bymorphology, membrane markers, ploidy, chromosomal con-tent and functional features. Briefly, after a few days of co-culture, cells appeared with morphology intermediate be-tween the 2 original cell types. Typical macrophage surfacemolecules appeared in the hybrids. Meth A cells were labeledwith red fluorescence and macrophages with green fluores-cence. After 4 days in vitro, hybrids with yellow fluorescenceappeared. Macrophages from BALB.K mice (H-2 Kk) werecocultivated with Meth A cells from BALB/c mice (H-2 Kd).The semigeneic hybrids displayed both specificities, as dem-onstrated by flow cytometry. The hybrids appeared moder-ately phagocytic, less so than the macrophages and markedlymore so than the essentially nonphagocytic Meth A cells. Thehybrids had a mean number of 76 chromosomes, as opposedto 53 in the Meth A cells and 40 in the macrophages. Themacrophage DNA index was set at 1; Meth A cells werefound to have an index of 1.6 in G1 phase, and the hybrids hada 2.6 index. The hybrids grew more slowly in vitro than MethA cells, but grew faster in vivo.© 2002 Wiley-Liss, Inc.

Key words: macrophages; Meth A sarcoma; cell hybridization

The current understanding of tumor biology has been largelydetermined by concepts and ideas borrowed from parasitology:transformed tumor cells are considered as something foreign thatother cells of the body—in particular the immune cells—shalldefend us against. On this conceptual basis, both the presence ofthe immune cells in tumor tissue as such and their properties andinteractions have been interpreted in terms of a “fight againstcancer.”

The published observations on tumor-associated macrophages(TAMs) neither support nor contradict this understanding of cel-lular interactions in tumor tissue. TAMs have been isolated andstudied in vitro without any unequivocal conclusions about theirfunction in tumor tissue having been reached. An understanding ofthe in vivo relevance of TAMs has also been approached byadministering compounds that inhibit macrophage function,1 byadoptive transfer of macrophages2 and by gene transfer.3 Resultsfrom these studies, together with the functional data on TAMsstudied in vitro, rather than supporting a role for the macrophagein the defense against malignant growth, indicate that macro-phages—at least in some tumors and at some steps of progres-sion—may provide a necessary support for optimal neoplasticproliferation.

The idea that macrophages may stimulate the formation oftumor stroma—including vasculature—through the action of theirnumerous secretory products is readily acceptable. The idea thatmacrophages may also contribute crucially to the central propertyof malignant growth (invasion) is less obvious. Both we4 andothers5 have advocated this possibility by describing phenomenathat suggest hybridization between tumor cells and macrophages invitro and by indicating that such hybrids would probably have, tothe fullest extent, all the properties of a malignant tissue: relentlessgrowth, invasion and stimulation of a supporting stroma.

We describe here our observation on the cocultivation of MethA tumor cells with mouse macrophages in vitro, with the conclu-

sion being that true hybrids are being formed, hybrids with in-creased growth potential in vivo.

MATERIAL AND METHODS

AnimalsAdult female (BALB/c � C57BL/6) F1 (CB6) mice (H-2d/b),

BALB.K (H-2k/k) and BALB.K mice (H-2k/d) were purchasedfrom Charles River (Sulzfeld, Germany), kept under specificpathogen-free housing conditions and used in experiments 1 weekafter arrival, at 6 weeks of age.

CellsMethyl-cholantrene-induced BALB/c (H-2d) sarcoma Meth-A

cells6 were kept in long-term culture without feeder cells. Meth Acells are MHC class I-positive but do not express class II mole-cules constitutively nor after interferon-� (IFN-�) treatment.7 Thecells tested negative for cytomegalovirus and lactate dehydroge-nase elevating (LDH) virus. The BALB/c B-cell lymphoma A208

was obtained from the American Type Culture Collection (Rock-ville, MD). Cell lines were grown in RPMI-1640 with 10% FCSsupplemented with penicillin (105 IU/l), streptomycin (0.1 g/l),HEPES buffer (20 mM) and 2-mercaptoethanol (5 � 10�5 M), pH7.4, at 37°C in humidified air containing 5% CO2. Peritoneal cellsfrom BALB/c or BALB.K (H-2k/k) were collected by lavage withice-cold PBS. After 2 hours in vitro, nonadherent cells werewashed off, and the remaining adherent cells represented almostexclusively macrophages since �98% phagocytosed latex beadsand expressed F4/80 (fluorescein isothiocyanate monoclonal anti-body [FITC MAb] F4/80, Serotec, Oxford, England).9 The cellswere cultivated serum-free in RPMI-1640 for 24 hr, and eventualremaining fibroblasts were washed off. Meth A tumor cells wereadded to pure cultures of peritoneal macrophages at a ratio of 1:10.After 4 days of coculture under standard culture conditions, hybridcells emerged. No agents were added to facilitate cell hybridiza-tion. The nonadherent Meth A cells were washed off, and theremaining culture of macrophages and semiadherent hybrids werekept on ice until the cells were scraped off the culture dish with aTeflon policeman and washed 3 times before limiting dilutioncloning.

Limiting dilution cloningThe hybrid cells were seeded in 96-well, V-bottomed Petri

dishes at an average concentration of 0.1 cells/well. Limitingdilution cloning of syngeneic hybrids resulted in 1 positive out of

Grant sponsor: Norwegian Cancer Society; Grant sponsor: Aakres Foun-dation.

*Correspondence to: Oncology Research Centre, Level 2, Clinical Sci-ences Building, Prince of Wales Hospital, High Street, Randwick NSW2031, Australia. Fax: �61-2-9382-2629. E-mail: [email protected]

Received 26 September 2001; Revised 16 November 2001; Accepted 19November 2001

Published online 29 January 2002

Int. J. Cancer: 98, 573–581 (2002)© 2002 Wiley-Liss, Inc.DOI 10.1002/ijc.10249

Publication of the International Union Against Cancer

FIGURE 1.

574 BUSUND ET AL.

96 wells, and semigeneic hybrids gave 2 positive wells out of 96.The cloned cell cultures were expanded and the cloning proceduresrepeated twice with 3 and 2 positive syngeneic clones out of 96wells, and with 2 and 1 positive semigeneic clones, respectively.One of each of these final syngeneic and semigeneic clones weresubsequently expanded and used in the consecutive experiments orkept frozen for 2–4 months.

Tumor inoculationAfter washing, 107 Meth A cells or cloned hybrid cells sus-

pended in 50 �l PBS were injected subcutaneously in the left flankof mice in groups of 4 mice. Mean tumor diameters (mean oftransversal and longitudinal diameters), and tumor volumes weremeasured daily. Tumor volumes were measured by immersingdissected tumors in water and measuring the increase in totalvolume. In the transplantation experiments into syngeneic andsemi- and allogeneic recipients, 106 Meth A cells or cloned hybridcells were inoculated subcutaneously in 50 �l PBS in groups of 4,

and the percentage of animals with tumor growth was recorded(tumor take).

Fluorocytometric analysisCells (5 � 105) in 50 �l PBS supplemented with 5% mouse

serum (Sigma, St. Louis, MO) were incubated with staining re-

FIGURE 2 – Membrane staining. Meth A tumor cells (a, red fluorescence), peritoneal macrophages (b, green fluorescence) and enlarged, yellowcells emerged after coculture of red Meth A cells and green macrophages for 4 days (c, d). Red and green filter are shown in (c) and (d),respectively. The dyes are aliphatic fluorescent chromophores that irreversibly label the lipids in the cell membrane.

FIGURE 3 – Membrane markers. Percentage of hybrid cells express-ing the macrophage markers F4/80 (open bars), MOMA 2 (hatchedbars) and Mac 1 (solid bars). The expression was assessed using flowcytometry on cells stained with fluorescent monoclonal antibodies.Cloned hybrids were cultured in vitro for 14 days before the experi-ment. The results are expressed as mean of 6–8 individual experi-ments � SD. *, p � 0.05 (hybrids versus macrophages and Meth Acells).

FIGURE 1 – Morphology of hybrids. (a) Scanning electron micros-copy of Meth A tumor cell attached to adherent macrophage. The cellswere cocultured under standard conditions for 2 days. (b) Scanningelectron microscopy of Meth A cells and macrophages grown incoculture for 4 days. Small, round, nonadherent Meth A cells (arrow1) and large, adherent macrophages (arrow 2) are visible in betweenapparent semiadherent hybrid cells with long foot processes that aredominating the culture. (c) The long foot processes seem to explorecontacts with both macrophages and other hybrid cells. Transmissionelectron microscopy of a cloned hybrid cell shows vacuoles of varyingsizes and untidily arranged microvilli on the surface (d), multilobu-lated nuclei with marginated chromatin (e) and aggregates of mem-brane-enclosed tubuli between nuclear lobuli (f). The apparent hybridcells show a morphology intermediate between the macrophages andthe Meth A cells.

575MACROPHAGES HYBRIDIZE WITH METH A SARCOMA CELLS

agents (10 �g/ml) in V-bottomed wells (Nunc, Naperville, IL) for30 min at 4°C. Between steps, samples were washed 3 times inice-cold PBS with 0.1% albumin (PBSA). After fixation with 2%paraformaldehyde, the fluorescence of 104 cells/sample was re-corded with an Epics Profile II flow cytometer equipped with a 488nm argon laser (Coulter Electronics, Luton, UK). Expression ofmacrophage markers was assessed using the rat IgG2b MAbsanti-F4/80 (Serotec),10 anti-MOMA-2 (Serotec),11 and anti-Mac-1(Pharmingen, San Diego, CA),12 biotin-conjugated sheep F(ab)2-anti-rat Ig fragments (Boehringer-Mannheim, Mannheim, Ger-many) and r-phycoerythrin-streptavidin (Jackson ImmunoRe-search, West Grove, PA).

MHC class I expression was assessed by staining with thefollowing antibodies: FITC-conjugated CTKk (mouse IgG2a anti-H-2 Kk) and K3 (mouse IgM anti-H-2 Kd) combined withbiotin-conjugated LO-MM-9 (rat IgG2a anti-mouse IgM) (all fromSerotec) and r-phycoerythrin-streptavidin (Jackson ImmunoRe-search). MHC class I expression was determined in double-stain-ing experiments. Cells were recorded as double positive when theirred and green fluorescence exceeded that of 99% of cells inparallel samples incubated with FITC-conjugated IgG2a and IgM, monoclonal murine isotype control Ab instead of class I-specificreagents. The percentage of positive cells in single staining wasdetermined by cumulative subtraction of a background histogramobtained from samples in which the primary reagent was replacedby an isotype control Ab (FITC-labeled when appropriate) asdescribed13 using the Epics Elite software (Coulter Electronics).

Membrane staining with fluorescent dyesThe fluorescent dyes PKH26 Red cell linker kit and the com-

plementary labeling PKH2 Green cell linker kit (Sigma) weremaintained as a 40 �M stock solution in 95% ethanol and dilutedto 0.2 �M in endotoxin-free PBS (�0.005 ng/ml endotoxin byLimulus Amebocyte Lysate assay; Cape Cod Associates, WoodsHole, MA). The labels are aliphatic fluorescent chromophores, andthe methodology allows for irreversible labeling of the lipids in thecell membrane rather than surface protein labeling. Before cocul-ture, Meth A cells and peritoneal macrophages were separatelystained for 5 min with PKH 26 Red and PKH 2 Green, respec-tively. The cells were washed 4 times in cold medium, mixedtogether and cocultured under standard conditions.

Ploidy measurementsViable cells were stained with propidium iodide (Sigma) and

analyzed within 8 hr after staining on a FACScan flow cytometer(Becton Dickinson, Mountain View, CA).

Scanning and transmission electron microscopyAfter overnight fixation in McDowell’s fixative, pH 7.4 (4%

formaldehyde, 1% glutaraldehyde in a phosphate buffer with os-molarity of 320 mOsm), the cells were scraped off the culture dishwith a Teflon policeman and pelleted. Then they were postfixed in1% OsO4, dehydrated in a graded series of ethanol washes, criticalpoint dried in a dryer from Balzer Union (Bal-Tec, Lichtenstein),mounted on aluminum stubs before coating with gold using aPolaron SEM Coating Unit E5000 (Polaron, Waterford, England)and embedded in ethylether/Araldite (Serva, Heidelberg, Ger-many). Ultrathin sections were cut on a Reichert Ultracut Sultramicrotome, (Reichert, Vienna, Austria), using a diamondknife (Diatome, Switzerland), mounted on Formvar-coated coppergrids and contrasted with unranyl acetate and Reynolds lead ci-trate. The sections were examined in a Jeol 1010 TransmissionElectron Microscope (Jeol, Tokyo, Japan).

Immunogold staining in transmission electron microscopyFor preembedding immunolabeling, the cell cultures were

washed in phosphate buffer, pH 7.2, after fixation in 8% glutaral-dehyde, blocked in 10% FCS, incubated with primary antibodydiluted in the blocking agent, washed in phosphate buffer andincubated in a protein A-gold complex (University of Utrecht,Utrecht, the Netherlands) before embedding and sectioning asdescribed above.

Chromosome analysisMetaphase spreads were prepared from cultures with colcemid

(0.5 �g/ml for 3 hr) by standard techniques and stained withGiemsa. Metaphase spreads of 40–80 per sample were photo-graphed and counted to determine the chromosome mode.

Statistical analysisAnalyses of differences between the groups were determined

with Student’s t-test. Differences were considered statisticallysignificant when p � 0.05.

FIGURE 4 – Membrane markers. Macrophages from BALB.K (Kk;hatched bars) mice hybridized with Meth A cells derived fromBALB/c (Kd; solid bars) mice showed hybrid cells with MHC class Iantigens from both strains, H-2 Kk (Biotin) and H-2 Kd (FITC),respectively. Flow cytometry of cloned allogeneic hybrids cultured invitro for 2 months. The results are expressed as mean of 6–8 individ-ual experiments � SD.

FIGURE 5 – Phenotype. Transmission electron microscopy showsmicrovilli on the surface of a cloned allogeneic hybrid cell cultured invitro for 2 months and stained with gold-labeled anti-H-2 Kk antibod-ies.

576 BUSUND ET AL.

RESULTS

MorphologyFigure 1a shows a scanning electron microscopy of a Meth A

tumor cell attached to an adherent macrophage. The cells werecocultured under standard conditions for 2 days. After 4 days ofcoculture, apparent hybrids emerged. As can be seen in Figure 1b,the hybrids are semiadherent, flattened, star-shaped cells with 2–6extremely elongated foot processes, and they dominate the cellculture. The cell diameter of the nonadherent Meth A cells (arrow1) is approximately 10 �m, whereas the lengths of the footpro-cesses in hybrids are up to 100 �m. The hybrids show no contactinhibition with respect to each other, and they grow in a 3-dimen-sional pattern. To some extent they show contact inhibition withregard to the adherent macrophages (arrow 2), which are visible inbetween the hybrid cells. However, the long foot processes seemsto explore contacts with both macrophages and other hybrid cells(Fig. 1c). At higher magnification, an extensive membrane rufflingis visible (not shown). In transmission electron micrographs (Fig.1d), untidily arranged microvilli can be seen on the surface ofcloned hybrid cells. Vacuoles of varying sizes, some in closecontact with the cell membrane, can be seen. The nuclei aremultilobulated, with marginated chromatin (Fig. 1e). Betweennuclear lobuli, aggregates of membrane-enclosed tubuli can beseen (Fig. 1f). The apparent hybrids conclusively show a morphol-ogy intermediate between the macrophages and the Meth A cells.

Membrane stainingThe plasma membrane of Meth A tumor cells were labeled in

vitro with red fluorescence and analyzed with standard fluores-cence microscopy (Fig. 2a), peritoneal macrophages were labeledwith green fluorescence (Fig. 2b). After 4 days of coculture,yellow, enlarged and semiadherent hybrid cells emerged in thecultures (Fig. 2c,d). The yellow color of the plasma membrane ofhybrids indicates cell fusion between red tumor cells and greenmacrophages. Yellow hybrids also appeared in a separate experi-ment in which tumor cells and macrophages were stained greenand red, respectively. When a red filter was used during fluores-cence microscopy, the hybrids exhibited green staining (Fig. 2c).When a green filter was used, the hybrids exhibited red staining(Fig. 2d).

Membrane markersApproximately 40% of the cloned hybrids, grown in vitro for 14

days prior to the experiment, were positive for the macrophagemarkers F4/80, MOMA 2 and Mac 1. This percentage is about halfof the expression of the parental peritoneal macrophages (Fig. 3).

A slight reduction after 14 days in culture could be seen (data notshown). However, after 2 months in culture, significant expressionof the macrophage markers could still be detected (data notshown).

Macrophages from BALB.K (H-2 Kk) mice were coincubatedwith BALB/c-derived Meth A cells (H-2 Kd). Since the BALB.Kgenome is identical to that of BALB/c except for MHC locus,hybrids between these strains will express the allogeneic markerH-2 Kk only if they have parts of the BALB.K genome. Cellsemerged in the coculture with identical morphology as hybridsbetween BALB/c macrophages and Meth A cells and cloned asdescribed previously. As can be seen from Figure 4, the semige-neic hybrid cells expressed class I allomorphs encoded in theBALB/c genome (Kd) as well as in that of BALB.K (Kk). Figure5 shows microvilli on the surface of the allogeneic hybrids stainedpositive with gold-labeled anti-H-2 Kk. As the hybrids were grownin vitro for 2 months before labeling with anti-H-2 Kk, thischaracteristic is probably a result of an inherited gene coding forH-2 Kk.

Transplantation of syngeneic and semi- and allogeneicrecipients

By transplantation of allogeneic tumors, immunoreactionsagainst allogeneic barriers can be utilized to demonstrate expres-sion of non-self histocompatibility antigens on transplanted cells.As can be seen from Table I, there was 100% tumor take byinoculation of BALB/c hybrid cells in homozygous BALB/c ani-mals, whereas no tumors could be recorded by inoculation of thesame hybrids in homozygous BALB.K animals. Following inoc-ulation of BALB/c � BALB.K hybrid cells in homozygousBALB/c or BALB.K animals, there was no tumor take. After 5weeks, 100% tumor take was recorded after inoculation ofBALB/c hybrids and BALB/c � BALB.K hybrids in heterozygous(BALB/c � BALB.K)F1 animals. Inoculation of Meth A cells inhomozygous BALB.K produced no tumor growth, whereas inoc-ulation in homozygous BALB/c mice and heterozygous (BALB/c � BALB.K) mice produced 80% tumor take after 5 weeks.Conclusively, these observations are compatible with hybridiza-tion of macrophages and tumor cells.

DNA contentThe Meth A cells are aneuploid, with a DNA index of 1.6 in G1

phase (Fig. 6), whereas the macrophages are diploid (DNA in-dex � 1). The hybrid cells were subtetraploid, with an averageDNA index of 2.6, which is exactly the sum of average Meth Acells and macrophages. Mean chromosome number was 76 in

TABLE I – TUMOR TRANSPLANTATION INTO SYNGENEIC AND SEMI- AND ALLOGENEIC RECIPIENTS1

CellsRecipient strain

BALB/c BALB.K (BALB/c � BALB.K)F1

Meth A (BALB/c)3 weeks � 0 �5 weeks �� 0 ��

Hybrid (BALB/c � BALB.K)3 weeks (�) (�) ��5 weeks 0 0 ���

Hybrid (BALB/c)3 weeks �� (�) ��5 weeks ��� 0 ���

1No tumor growth was recorded after inoculation of cloned BALB/c hybrids in homozygous BALB.Kanimals, or after inoculation of cloned BALB/c � BALB.K hybrids in homozygous BALB/c or BALB.Kanimals. One hundred percent tumor take was recorded after inoculation of BALB/c hybrids as well asafter inoculation of BALB/c � BALB.K hybrids in heterozygous (BALB/c � BALB.K) F1 animals.Inoculation of Meth A cells in BALB.K produced no tumor, whereas inoculation of Meth A in BALB/cand (BALB/c � BALB.K) F1 animals gave 80% tumor take after 5 weeks. Meth A cells (106) or clonedhybrid cells suspended in 50 �l PBS were injected subcutaneously in the left flank of mice in groups of4. The results are expressed as mean of three individual experiments. 0, no tumor take; (�), small tumorinitially that disappeared after 3–4 days; �, 50% tumor take; ��, 80% tumor take; ���, 100% tumortake.

577MACROPHAGES HYBRIDIZE WITH METH A SARCOMA CELLS

hybrids, as opposed to 53 in Meth A cells and 40 in macrophages(Fig. 7). For comparison, hybrids between the A 20 lymphoma cellline and syngeneic macrophages (hybridized under the same con-ditions as Meth A and macrophages) revealed a chromosome countof 66, whereas the A 20 cells exhibited a count of 53.

Growth in vitro and in vivoWhen cloned hybrid cells and Meth A cells were cultured in

vitro and counted daily, no significant cell death was recorded.Compared with Meth A cells, the hybrids exhibited significantlyreduced growth in vitro after 4 days in culture (Fig. 8). Bymeasuring mean tumor diameter after subcutaneous inoculation ofMeth A tumor cells or cloned hybrid cells (Fig. 9), no significantdifference in size between Meth A tumors and hybrid tumors couldbe found. However, by measuring tumor volume (Fig. 10), it wasdemonstrated that the hybrids grew significantly faster than theMeth A. This indicates a deep and spherical growth of hybridtumors, whereas Meth A tumors grew more superficially.

PhagocytosisIn contrast to Meth A cells, the cloned hybrid cells phagocy-

tosed latex beads (Fig. 11) but not to the same extent as peritonealmacrophages. The hybrids needed a longer time to phagocytosebeads and exhibited a reduced total capacity compared with mac-rophages (Fig. 12).

DISCUSSION

There are numerous published reports of apparent fusion be-tween transformed cells and normal tissue cells in vitro. Malignantcells obviously have highly variable phenotypes, and it has beeninsufficiently demonstrated that the strange novel cells appearingunder coculture were not merely another display of a bizarremalignant phenotype. In a limited number of cases, actual fusionhas been documented. Even more rarely has the formation of truehybrids been proved, i.e., the formation of cells with a genomerepresenting the sum of the 2 parental genomes.

Hybridization with normal somatic cells has mainly been at-tempted in order to study the effect of the normal genome on thecorresponding malignant genome.14 With this aim, fusion betweentransformed cells and fibroblasts as well as between lymphocytesand epithelial cells has been attempted.15–17 Hybrids produced byfusion of different types of transformed cells have also beenreported.18

Most of these studies have led to the conclusion—challenged bysome19—that the normal genome suppresses malignancy and thatfull malignancy was only restored after the hybrid cell had accom-plished a segregation or loss of chromosomes, presumably parts ofthe normal genome containing antioncogenic sequences.

It is obvious that hybrids produced by fusion of malignant cellsand normal somatic cells could be of great interest from an entirelydifferent perspective. It is conceivable that essential properties ofmalignant tumors could be based on a combination of properties oftransformed cells and normal tissue cells: relentless growth andreduced apoptosis on the one hand, and invasion, stroma formationand deregulation of immune processes on the other. Hybrids be-tween malignant cells and tissue macrophages would appear es-pecially interesting in this perspective. Macrophages are present inall tissues of the body, they are “the masters of the connectivetissue,” they regulate formation of stroma and microvasculature,they take part in the fine tuning of specific as well as nonspecific

FIGURE 6 – Ploidy. (a–c) Meth A cells are aneuploid, with a DNAindex of 1.6 in G1 phase, whereas the macrophages are diploid (DNAindex � 1). The cloned hybrid cells are subtetraploid, with an averageDNA index of 2.6, exactly the sum of the average values for the MethA and macrophage DNA index. Viable cells were stained with pro-pidium iodide and analyzed by flow cytometry. The experiments wererepeated twice with the same results.

578 BUSUND ET AL.

immune responses—and they are highly invasive. Munzarova andcoworkers5 have particularly advocated this idea. Others have alsoconsidered the possibility that fusion of malignant cells withstroma cells may provide tumor cells with significantly novelproperties.20,21 Rachkovsky et al.22 and Chakraborty et al.23 havereported enhanced metastatic potential of hybrids between murinemelanoma cells and macrophages.

Ever since we started to study the interaction of tumor cells andmacrophages in vitro,24 we have repeatedly observed phenomenathat suggest fusion between the 2 cell types: In coculture ofnonphagocytic Meth A tumor cells and prelabeled (plastic micro-spheres) macrophages, aneuploid cells with intracellular micro-

spheres appeared after a few days. Recently we have studied thisphenomenon in more detail and made observations that unequiv-ocally demonstrate hybridization, as reported here: (i) After a fewdays of coculture cells appeared with morphology intermediatebetween the two original cell types. (ii) Typical macrophage sur-face molecules appeared in the hybrids (antigens for F 4/80,MOMA 2, Mac 1). (iii) Meth A cells were labeled with redfluorescence (PKH 26, Sigma) and macrophages with green fluo-rescence (PKH 2), and after 4 days in vitro hybrids with yellow

FIGURE 7 – Chromosome count. Mean chromosome count was 76 incloned hybrids (solid circles), as opposed to 53 in Meth A cells (opencircles). For comparison, hybrids between the A 20 lymphoma cell lineand syngeneic macrophages (hybridized at the same conditions asMeth A and macrophages) revealed a mean chromosome count of 66,whereas the A 20 cells showed a count of 53. For macrophages, thechromosome count was 40, as expected for nontransformed murinecells. The results are expressed as mean of 8 individual experiments �SD.

FIGURE 8 – Growth in vitro. The cloned hybrids (solid circles)showed significantly reduced growth in vitro after 4 days in culturecompared with Meth A cells (open circles) (p � 0.05). The cells wereseeded in culture wells (105 cells/ml), grown under standard conditionsand counted daily. The results are expressed as mean of 8 individualexperiments � SD.

FIGURE 9 – Growth in vivo: diameter. No significant difference inmean tumor diameter could be found between Meth A tumors (opencircles) and cloned hybrid tumors (solid circles). Meth A cells orcloned BALB/c hybrid cells (107 suspended in 50 �l PBS) wereinjected subcutaneously in the left flank of mice. There were 4 animalsin each group, and the results are expressed as mean of 4 individualexperiments � SD.

FIGURE 10 – Growth in vivo: volume. By volume, the cloned hybridtumors (solid circles) were significant larger than Meth A tumors(open circles). Tumor volumes were measured by immersing dissectedtumors in water and measuring the increase. Meth A cells or clonedBALB/c hybrid cells (107 suspended in 50 �l PBS) were injectedsubcutaneously in the left flank of mice. There were 4 animals in eachgroup, and the results are expressed as mean of 4 individual experi-ments � SD.

579MACROPHAGES HYBRIDIZE WITH METH A SARCOMA CELLS

fluorescence appeared. (iv) Macrophages from BALB.K mice (H-2Kk) were co-cultivated with Meth A cells from BALB/c (H-2 Kd),and their hybrids displayed both specificities, as demonstrated byflow cytometry. (v) The hybrids appeared moderately phagocyticmless so than the macrophages, but markedly more than the essen-tially nonphagocytic Meth A cells. (vi) The hybrids had a meannumber of 76 chromosomes, as opposed to 53 in the Meth A cellsand 40 in the macrophages. (vii) The macrophage DNA index wasset at 1; Meth A cells were found to have an index of 1.6 in G1phase, and the hybrids 2.6. (viii) Hybrids grew more slowly in vitrothan Meth A cells, but faster in vivo.

The flattened, elongated and adherent morphology of the hybridcells indicates a differentiation of the neoplastic cells toward amacrophage-like phenotype. Foot processes in general are pseu-dopodia, blunt projections on the cell surface, used by cells forphagocytosis and locomotion. The prominent foot processes inhybrid cells have an unknown significance. They could be atypicallocomotor organs or they may serve to explore contact with othercells.

Time lapse studies of the hybrids indicate that the foot processesmay be organs for both locomotion and contact with other cells(data not shown). Hybrids also show numerous prominent, regu-larly arranged microvilli, which are usually found on the freesurface of cells whose principal function is secretion or absorption.Numerous intracellular vacuoles in hybrids indicate metabolicallyactive cells. The yellow color of the hybrids that emerged in thecocultures after separate staining of Meth A cells and macrophagesdemonstrates fusion between red and green membranes. The pos-

sibility of dye transfer between cells in coculture was ruled out byincubation of unstained macrophages and tumor cells in superna-tant from stained cells. No transfer of staining could be detectedunder these conditions. Phagocytosis of stained Meth A cells bymacrophages was not observed. The presence of membrane anti-gens on cloned hybrids clearly demonstrates the contribution of themacrophage phenotype, as all 3 macrophage-markers tested werefound on the cell surface of the hybrids. The presence of H-2Kd/H-2 Kk on hybrids demonstrates both macrophage phenotypeand hybridization of H-2 Kd-positive tumor cells and H-2 Kk-positive macrophages. The hybrids were grown in vitro for 2months before they were stained with the antibodies, which clearlydemonstrates that the membrane antigens have been transcribedand processed by the hybrids, not merely transferred from theparental cells.

Growth of cloned hybrids (H-2 Kd/H-2 Kk) in F1 heterozygousmice (BALB/c � BALB.K F1) indicates tolerance against H-2 Kd

and H-2 Kk antigens in tumors. The lack of growth of this clone inBALB/c and BALB.K mice indicates immunologic responseagainst H-2 Kk and H-2 Kd antigens, respectively. This findingindicates sufficient expression of H-2 Kk on hybrid cells to initiatetumor rejection. The percentages of H-2 Kk-positive hybrids aswell as peritoneal macrophages (Fig. 4) probably represent tech-nical underrepresentations.

We conclude that true hybrids between Meth A sarcoma cellsand macrophages in vitro have been observed. We also report thatthe hybrid cells (produced in vitro) give rise to real tumors wheninjected into syngeneic mice. These tumors actually grow fasterthan tumors produced after injection of Meth A cells, an observa-tion that may suggest an important contribution of macrophageproperties to the overall capacity of the hybrid tumors in vivo.

We have observed hybrid cells in tumors established afterinoculation of pure Meth A cells (data not shown). This findingindicates that hybridization also occurs in vivo and is not an invitro artifact. The observations will be described in detail in asubsequent article.

The mechanisms operating during the combined expression ofMeth A/macrophage properties of the hybrids are so far notknown. The simplest explanation is that the 2 genomes are ex-

FIGURE 11 – Phagocytosis: scanning electron microscopy. Phagocy-tosis of latex beads by cloned hybrid cell as seen by scanning electronmicroscopy.

FIGURE 12 – Phagocytic capacity. In contrast to Meth A cells, thehybrid cells phagocytosed latex beads. Compared with macrophages(open circles), the hybrids (solid circles) exhibited reduced phagocyticcapacity. The cells were grown under standard culture conditions andgiven excessive amounts of monodisperse latex beads (4 �m in diam-eter). Intracellular beads were counted at indicated time points byphase contrast microscopy. The results are expressed as mean of 8individual experiments � SD.

580 BUSUND ET AL.

pressed relatively independently. There is no indication that tumorsuppressor genes or similar genes in the normal macrophagegenome have in any way counteracted the expression of themalignant phenotype, under the conditions present in our study.

ACKNOWLEDGEMENTS

We thank Mr. P. Prydz and Mrs. R. Skogstad for technicalassistance.

REFERENCES

1. Heuff G, Oldenburg HS, Boutkan H, et al. Enhanced tumor growth inthe rat liver after selective elimination of Kupffer cells. Cancer Im-munol Immunother 1993;37:125–30.

2. Mantovani A, Ming WJ, Balotta C, et al. Origin and regulation oftumour associated macrophages: the role of tumour-derived chemo-tactic factor. Biochim Biophys Acta 1986;865:59–67.

3. Heike Y, Sone S, Yano S, et al. M-CSF gene transduction in multi-drug-resistant human cancer cells to enhance anti-P-glycoprotein an-tibody-dependent macrophage-mediated cytotoxicity. Int J Cancer1993;54:851–7.

4. Seljelid R, Busund LTR. The biology of macrophages. II Inflamma-tion and tumors. Eur J Hematol 1994;52:1–12.

5. Munzarova M, Laucrova L, Kovarik J, et al. Fusion-induced malig-nancy? A preliminary study (A challenge to today’s common wis-dom). Neoplasma 1992;39:79–86.

6. DeLeo AB, Shiku H, Takahashi T, et al. Cell surface antigens ofchemically induced sarcomas of the mouse. I. Murine leukemia virusrelated antigens and alloantigens on cultured fibroblasts and sarcomacells: description of a unique antigen on Balb/c Meth A sarcoma. JExp Med 1977;146:720–34.

7. Harada M, Matsuzaki G, Yoshikai Y, et al. Autoreactive and heatshock protein 60-recognizing CD4� T-cells show antitumor activityagainst syngeneic fibrosarcoma. Cancer Res 1993;53:106–11.

8. Kim KJ, Kanellopoulus-Langevin C, Merwin RM, et al. Establish-ment and characterization of BALB/c lymphoma lines with B cellproperties. J Immunol 1979;122:549–54.

9. Szu-Hee L, Starkey PM, Gordon S. Quantitative analysis of totalmacrophage content in adult mouse tissues. J Exp Med 1985;161:475–89.

10. Starkey PM, Turley L, Gordon S. The mouse macrophage-specificglycoprotein defined by monoclonal antibody F4/80: characterization,biosynthesis and demonstration of rat analogue. Immunology 1987;60:117–22.

11. Kraal G, Rep M, Janse M. Macrophages in T and B cell compartmentsand other tissue macrophages recognized by monoclonal antibodyMOMA-2. An immunohistochemical study. Scand J Immunol 1987;26:653–61.

12. Sanchez-Madrid F, Simon P, Thompson S, et al. Mapping of antigenic

and functional epitopes on the � and subunits of two related mouseglycoproteins involved in cell interactions, LFA-1 and Mac-1. J ExpMed 1983;158:586–602.

13. Overton WR. Modified histogram subtraction technique for analysisof flow cytometry data. Cytometry 1988;9:619–26.

14. Harris H, Klein G. Malignancy of somatic cell hybrids. Nature 1969;224:1314–6.

15. Harris H, Branwell ME. The suppression of malignancy by terminaldifferentiation: evidence from hybrids between tumour cells and ker-atinocytes. J Cell Sci 1987;87:383–8.

16. Benedict WF, Weissman BF, Mark C, et al. Tumorigenicity of humanHT 1080 fibrosarcoma � normal fibroblast hybrids: chromosomedosage dependency. Cancer Res 1984;44:3471–9.

17. Giarcomoni D. Tumorigenicity and intracisternal A particle expres-sion of hybrids between murine myeloma and lymphocytes. CancerRes 1979;39:4481–4.

18. Miller FR, McInerney D, Rogers C, Miller BE. Spontaneous fusionbetween metastatic mammary tumor subpopulations. J Cell Biochem1988;36:129–36.

19. Ephrussi B, Davidson RL, Weiss MC. Malignancy of somatic cellhybrids. Nature 1969;224:1314–5.

20. Larizza L, Schinnmacher V. Somatic cell fusion as a source of geneticrearrangement leading to metastatic variants. Cancer Metast Rev1984;3:193–222.

21. Baetselier de P, Roos E, Verschueren H, et al. Sequestration ofmetastatic properties via somatic cell fusion. Implication for tumorprogression in vivo. In: Chandra P, ed. New experimental modalitiesin the control of neoplasia. New York: Plenum Press, 1986. 41–55.

22. Rachkovsky M, Sodi S, Chakraborty A, et al. Melanoma � macro-phage hybrids with enhanced metastatic potential. Clin Exp Metast1998;16:299–312.

23. Chakraborty A, Sodi S, Rachkovsky M, et al. A spontaneous murinemelanoma lung metastasis comprised of host � tumor hybrids. CancerRes 2000;60:2512–9.

24. Kaplan G, Seljelid R. Striking differences in the interactions betweenmacrophages and 2 closely related tumor lines. In: Keith J, McBrideB, Stuart A, eds. The macrophage and cancer. Edinburgh: Edinburgh,1977. 347–63.

581MACROPHAGES HYBRIDIZE WITH METH A SARCOMA CELLS