Phospholipid scramblase 1 (PLSCR1) is an IFN-inducible, endofacial plasma membrane protein that has been proposed to mediate accelerated transbilayer movement of plasma membrane phospholipids in cells exposed to elevated cytoplasmic [Ca 2+]. The marked transcriptional up-regulation of this gene by IFN in a wide variety of cell types suggested that PLSCR1 might also contribute to biological effects associated with IFN. To study the potential contribution of cellular PLSCR1 to the antiproliferative and tumor-suppressive activities of IFN, PLSCR1 cDNA was stably expressed in the human ovarian cancer cell line HEY1B, and the growth of these cells was compared with matched vector transfected controls both in vitro and in vivo. Whereas we detected no difference in either growth rate or morphology between PLSCR1-transfected cells and vector controls during in vitro culture in serum, when these cells were implanted s.c. into athymic nude mice, we observed a marked suppression of tumor development from cells transfected to express elevated levels of PLSCR1. Tumors from the PLSCR1-transfected cells were greatly reduced in size, showed increased infiltration of leukocytes and macrophages, and appeared to undergo differentiation to a more uniform and spindle-shaped morphology that markedly contrasted the highly undifferentiated and pleiomorphic cell shape normally observed for HEY1B cells in vitro or for vector-transfected control HEY1B cells both in vitro and in vivo. These data suggest that the up-regulation of PLSCR1 expression in tumor cells exposed to IFN may contribute to the observed tumor-suppressive action of this cytokine.
The IFNs are a family of pleiotropic cytokines that are responsible for providing vertebrates with innate immunity against a wide range of viruses and other microbial pathogens (1) . In addition, IFNs regulate cell proliferation, apoptosis, and immune responses, properties that underlie their uses in the treatment of cancer. IFNs alter patterns of gene expression in cells by binding to specific cell surface receptors and activating Janus kinase/signal transducers and activators of transcription pathways. The antiproliferative activity of IFN has been related to an accumulation of cells at the G1-S boundary of the cell cycle (2) . Cell cycle factors regulated by IFNs include c-myc, RB, cyclin D3 and cdc25A (3, 4, 5, 6) . In addition, the IFN-inducible proteins PKR and RNase L are implicated in both the antiproliferative and apoptotic activities of IFN (7, 8, 9) . Recently, the induction of tumor necrosis factor-related apoptosis-inducing ligand by IFNs has also been linked to the proapoptotic activities of IFNs (10 , 11) . Similarly, several genes have been cloned which, when down-regulated, suppress the growth-inhibitory or apoptotic activities of IFN-γ. The technical knockout strategy led to the cloning of five novel genes for IFN-γ death-associated proteins, DAP-1 to -5, and the identification of two other genes encoding thioredoxin and cathepsin D protease (12 , 13) . Furthermore, DNA microarray approaches identified several apoptotic genes to be IFN regulated, including BAK and fas (14 , 15) . The immune-regulatory role of IFN-γ in the antitumor action is related to the up-regulation of antigen processing and presentation genes including MHC class I and MHC class II expression (14, 15, 16) . Despite progress in this area, with a few exceptions little is known about the roles of specific IFN-stimulated genes in the control tumor proliferation in vivo.
PLSCR13 is a member of a newly identified family of endofacial membrane proteins that has been proposed to mediate accelerated transbilayer movement of plasma membrane phospholipids in response to elevated cytoplasmic [Ca2+] (17, 18, 19, 20, 21) . PLSCR1 has been putatively implicated in the de novo movement of PS and other aminophospholipids to the plasma membrane outer leaflet after cellular injury, and it may contribute to cell surface PS exposure during early stages of programmed cell death (19 , 20 , 22) . The cell surface exposure of these aminophospholipids is known to promote activation of plasma complement and coagulation proteases, and to promote clearance of cells by the reticuloendothelial system (23) . We reported recently that the PLSCR1 gene is highly inducible by IFN-α or IFN-β, and to a lesser extent by IFN-γ (14 , 24) . Elevated expression of plasma membrane PLSCR1 after IFN induction was not accompanied by any detectable increase in PS exposed at the cell surface, nor was increased plasma membrane PL scramblase activity observed, suggesting possible alternative function(s) for this protein (24 , 25) . PLSCR1 has also been reported recently to be a substrate of cellular protein kinases, including protein kinase Cδ, an IgE receptor-linked tyrosine kinase, and the proto-oncogene c-Abl tyrosine kinase, although the biological implication of these various interactions also remains to be clarified (22 , 26 , 27) . Three additional members of the PLSCR gene family and their orthologues in other species have been identified recently, although transcriptional regulation by IFN has only been observed for PLSCR1 (21) .
A possible role for PLSCR1 in regulating tumor cell proliferation in vivo was suggested by the isolation of an NH2-terminal truncated form of mouse PLSCR1 (TRA1) from a mouse leukemogenic cell line, Mm-P, cDNA library (28 , 29) . Expression of the TRA1 cDNA resulted in the conversion of nonleukemogenic Mm-S1 cells into leukemogenic cells. To directly determine the effect of PLSCR1 expression on tumor growth in vivo, we have expressed PLSCR1 in a human ovarian carcinoma cell line, HEY1B (30) , and compared rates of tumor cell growth in vitro and in vivo. Our data suggest that expression of PLSCR1 affects both size and morphology of tumors that develop in vivo without detectable changes in either proliferation or cell morphology during growth in vitro.
MATERIALS AND METHODS
The human ovarian carcinoma cell line HEY1B (a gift from Dr. Alexander Marks, University of Toronto, Toronto, Canada and Yan Xu, Cleveland Clinic) was cultured in RPMI 1640 supplemented with streptomycin-penicillin and 10% heat-inactivated fetal bovine serum (30) . Cell numbers were determined by counting in a hemocytometer. The in vitro growth rates were determined in the presence and absence of 1000 units/ml of IFN-α2a (Roche).
Subcloning and Monitoring Expression of Human PLSCR1 in HEY1B Cells.
The human PLSCR1 cDNA was subcloned under the control of a cytomegalovirus promoter in plasmid vectors pcDNA3 (Invitrogen) and pIRES-hyg (Clontech). The PLSCR1 cDNA was inserted into pcDNA3 using BamHI and XhoI, and into pIRES-hyg using BamHI and NotI. Transfection of the HEY1B cells was followed by selection in medium containing G418 or hygromycin. Expression of PLSCR1 protein was confirmed by Western blotting with murine mab 4D2 specific for human PLSCR1 (24 , 27) , and protein loading was confirmed with a mab against β-actin (Roche). Cell extracts with 200 μg of protein were fractionated on SDS-10% polyacrylamide and proteins transferred to ImmobilonTM-P membrane (Millipore) blocked with 5% nonfat powdered milk in PBS containing 0.1% (v/v) Tween 20, and incubated with antibody 1 h. Membranes were then washed with PBS containing 0.1% (v/v) Tween 20 and incubated with goat antimouse antibody tagged with horseradish peroxidase (Life Technologies, Inc.) for 1 h and developed with enhanced chemiluminescence substrate (Amersham).
Tumor Growth in Vivo.
HEY1B cells (106 cells/site) were injected s.c. into the flanks of groups of four to six BALB/c nude mice (Taconic, Germantown, NY). A single flank inoculation site was used for each mouse. Tumors occurred at all of the sites of inoculation in controls and experimental groups. Tumor volumes in vivo were monitored every 3–4 days, and volumes of the excised tumors were both determined by measuring with a caliper (h × w × d).
Preparation of Cell Cultures for Microscopy.
HEY1B cells transfected with pIRES-hyg/PLSCR1 and pIRES-hyg (vector control) were subcultured on glass coverslips (Fisher) in RPMI containing 10% fetal bovine serum. After 48 h, the cells were treated 30 min with 2% formaldehyde (Polysciences Inc., Warrington, PA) and stained with H&E.
Preparation of Tumor Tissue for Microscopy.
Tumors were excised into Tissue-Tek optimal cutting temperature O.C.T. compound (VWR), and snap frozen by immersing in isopentane chilled with liquid nitrogen. Frozen samples were thin sectioned using a cryostat and mounted on glass slides. Sections were either stained immediately or stored at −70°C until use. For H&E staining, the excised tumors was fixed with 4% paraformaldehyde and processed for paraffin embedding and thin sectioning.
Frozen sections were fixed and permeabilized by sequential immersion in acetone (10 min) and methanol (10 min) at −20°C. After incubation with Biotin/Avidin Block reagents (Vector, Burlingame, CA), the fixed tissue sections were incubated overnight at 4°C with 5 μg/ml of biotin-conjugated mab 4D2 (against human PLSCR1; Ref. 24 ), washed twice in PBS, and stained by 30-min room temperature incubation in 1 μg/ml fluorescein-conjugated streptavidin (Vector). After extensive washing in PBS, the stained sections were mounted in SlowFade (Molecular Probes, Eugene, OR) for fluorescence microscopy.
Light microscopy of H&E-stained cell cultures and tissue sections was performed using an Axioplan 2 microscope and images recorded with a color digital CCD camera (SV Micro color 80155) interfaced to a computer workstation (AxioVision 2.0 software; Carl Zeiss, Thornwood, NY). Fluorescence microscopy of immunostained tissue used a Zeiss fluorescein filter set, a black and white digital CCD camera (Hamamatsu C4742-95-12), and the pseudo-color processed images recorded.
Expression of PLSCR1 in the Human Ovarian Carcinoma Cell Line HEY1B.
To directly measure the effects of PLSCR1 in the absence of other IFN-induced proteins, we constitutively expressed human PLSCR1 cDNA in the human ovarian cancer cell line HEY1B. Stable expression of PLSCR1 was obtained using two different mammalian expression plasmids, pcDNA3 and pIRES-hyg, by selection in G418 or hygromycin, respectively (Fig. 1)⇓ . Only 1 of ∼50 clones transfected with pcDNA3/PLSCR1 expressed significantly elevated amounts of PLSCR1 as determined in Western blots probed with mab specific for human PLSCR1 (mab 4D2). The clone, S48, expressed ∼4-fold more PLSCR1 than the parental cells (Fig. 1A⇓ , Lane 3). This compared with a ∼10-fold increase in PLSCR1 levels obtained with IFN-α treatment of the cells (Fig. 1A⇓ , Lanes 4–6). In contrast, using pIRES-hyg, we obtained about 4–10-fold over basal levels of PLSCR1 in every one of eight clones that were analyzed. Plasmid pIRES-hyg/PLSCR1 expresses a bicistronic mRNA under the control of a cytomegalovirus promoter. The first open reading frame is for PLSCR1 followed by an IRES and a hygromycin B phosphotransferase sequence. Therefore, after transfection of the HEY1B cells and selection, expression of PLSCR1 was tightly coupled to hygromycin resistance. Accordingly, the pool of selected cells showed levels of PLSCR1 protein equivalent to the IFN-induced level (Fig. 1B⇓ , compare Lanes 2 and 3). By performing experiments on the pool of stably transfected cells, each expressing similar levels of PLSCR1, it was possible to reduce possible contributions attributable to nonspecific differences between clonal cell lines.
PLSCR1 Does Not Affect Cell Growth Rates or Morphology in Vitro.
To determine the possible involvement of PLSCR1 in the antiproliferative activity of IFN-α, cell growth rates were measured in the absence or presence of IFN-α (1000 units/ml). Comparing the clonal cells lines containing vector pcDNA3, the cell growth rates of the vector control (V24) and PLSCR-transfected cells (S48) were nearly identical. Also, whereas IFN caused a modest (∼2-fold) reduction in cell growth rates, there was no difference between V24 and S48 cells (data not shown). Similar results were seen with the pooled cells contain pIRES-hyg with or without PLSCR1 (Fig. 2)⇓ . Cells transfected to express elevated levels of PLSCR1 grew at nearly the same rate as the vector control cells, and IFN-α (1000 units/ml) only slightly reduced growth rates of both cell types. These results show that PLSCR1 does not affect in vitro growth rates and that the HEY1B cells are relatively insensitive to the antiproliferative action of IFNα. The appearance of the PLSCR1-transfected and control cells was also similar (Fig. 3)⇓ .
PLSCR1 Suppresses Tumor Growth in Vivo.
To determine the possible effect of PLSCR1 expression on tumor growth in vivo, the PLSCR1-transfected and control cells were injected s.c. into the flanks of nude mice. Cells transfected with either the empty plasmid (pcDNA3) clone V24 or pcDNA3/PLSCR1 clone S48 showed a dramatic difference in tumor sizes at 3 weeks after implantation (Fig. 4)⇓ . Tumor growth monitored in groups of six mice per cell type showed that the clone V24 controls cells grew ∼8-fold the rate of the PLSCR1-transfected S48 cells (Fig. 5)⇓ . At 24 days after implantation, the tumors were excised and measured. An identical 8-fold difference in tumor sizes was thus confirmed, supporting the accuracy of measuring tumors in vivo with the use of calipers (Fig. 6)⇓ .
To rule out clonal variation as the cause of the antitumor effect, we separately implanted groups of four mice with pools of transfected cells containing pIRES-hyg with or without a PLSCR1 cDNA insert (four independent experiments were done with similar findings). The tumor results obtained from the uncloned pooled cells (transfected with PLSCR1 cDNA) were similar to those obtained for the S48 clone, thus eliminating clonal differences as the basis for the antitumor activity of PLSCR1 (Fig. 7)⇓ . There was a potent suppression of tumor growth in the PLSCR1-transfected, pooled HEY1B cells. At 22 days after implantation the PLSCR1-transfected tumors was 1178 ± 444 mm3, whereas the HEY1B cells transfected with empty vector grew to 178 ± 119 mm3.
Expression of PLSCR1 Is Maintained in Vivo.
Between 32 and 45 days after implantation the growth rates of the PLSCR1-transfected HEY1B cells increased, possible because of a selection for more rapidly growing cells (Fig. 7)⇓ . To confirm that expression of PLSCR1 was maintained in vivo, tumors were excised after 3 weeks of growth, sectioned, and stained for PLSCR1 antigen (Fig. 8)⇓ . Whereas PLSCR1 was clearly observed in the cells containing the PLSCR1 plasmid, endogenous PLSCR1 expression in the HEY1B tumors from the empty vector control cells was barely detectable. In addition, Western blot assays on tumor extracts using antibody to PLSCR1 showed that expression was maintained through day 46 (data not shown).
PLSCR1 Causes Neutrophil and Macrophage Infiltration into the Tumors and Morphological Changes in the Tumor Cells.
At 3 weeks after implantation there was a dramatic difference in appearance between tumors transfected with PLSCR1 cDNA and the control tumors. The tumors containing either empty plasmid consisted of large, poorly differentiated cells with a pleiomorphic polyhedral shape (Fig. 9A)⇓ . By contrast, the PLSCR1-transfected tumor cells appeared to have undergone a more uniform, spindle-shaped transition of many (but not all) cells, suggesting a more differentiated state (Fig. 9B)⇓ . Additionally, examination of the excised tumor nodules at 14 days revealed that those developed from PLSCR1-transfected cells showed a multifocal and deeply invading inflammatory infiltrate, consisting primarily of polymorphonuclear leukocytes with macrophages near the vascular complexes. By contrast, the tumors developed from vector transfected controls showed a relatively minor inflammatory infiltrate, which was generally restricted to the tumor perimeter (Fig. 10)⇓ .
We have shown that the amount of PLSCR1 expressed in the plasma membrane impacts the tumorigenic potential of a human ovarian carcinoma cell line, HEY1B. Thus, our findings are consistent with a report showing that expression of an NH2-terminal truncation of a murine PLSCR1 (TRA1) in murine monocytic cell lines correlated with leukemogenic potential. The TRA1 cDNA was able to transform nonleukemogenic cells into leukemogenic cells as determined by studying survival after transplantation into syngenic or athymic mice (28 , 29) . However, there was no significant difference in the rate of cell proliferation in culture between monocytic cell lines with high and low leukemogenic properties (28) . Similarly, we found no significant difference in the in vitro growth rates of the PLSCR1-transfected cells and control cells (Fig. 2)⇓ . In sharp contrast, there was a dramatic difference in the ability of the PLSCR1-transfected and control cells to form tumors after being implanted into nude mice (Figs. 4⇓ 5⇓ 6⇓ 7)⇓ . The tumor-suppressor phenotype was observed in polyclonal pools of PLSCR1-transfected tumor cells, as well as for individual clonal lines, thus ruling out differences in growth rate among individual clonal cell lines per se as a cause.
The mechanism by which elevated expression of PLSCR1 might exert a tumor-suppressive function in vivo remains unresolved. As was noted, we observed no effect of PLSCR1 expression on mitotic index, growth rate, or morphology during in vitro cell culture. Furthermore, despite the putative role of PLSCR1 in remodeling plasma membrane phospholipids to expose PS at the cell surface, we did not observe an increase in cell surface PS in the PLSCR1-transfected HEY1B cells grown in tissue culture, as determined by either annexin V or factor Va binding assays, or as measured by expression of membrane catalytic activity for the prothrombinase reaction (data not shown). Consistent with these negative results, we failed to detect previously any increase in cell surface expression of PS in HEY1B cells when IFN was used to induce elevated expression of plasma membrane PLSCR1 from the endogenous gene (24) . By contrast to these negative results in vitro, when implanted s.c., the PLSCR1-transfected HEY1B cells when compared with vector controls showed marked reduction in rates of tumor growth that was accompanied by morphological changes in the PLSCR1-transfected tumor cells, which suggested a more differentiated state, plus evidence for infiltration of the tumor by leukocytes and other inflammatory cells. This raises the possibility that under conditions of cell growth in vivo, the level of PLSCR1 expression potentially alters proliferation either by modulating cell response to one or more growth factors or cytokines, or by promoting necrotic or apoptotic cell death. In this context it is of interest to note that PLSCR1 has been reported to be a substrate of protein kinase Cδ, and thereby elevate PS exposure in cells undergoing apoptosis (22) , as well as a substrate of the IgE receptor tyrosine kinase (26) and the c-Abl tyrosine kinases (27) , implying a possible role of PLSCR1 in one or more intracellular signaling pathways regulating cell proliferation or apoptosis.
We thank Lilin Li, Hongfan Peng, and Huiqin Nie for technical assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 Supported by Grants CA89132 (to R. H. S. and P. J. S.), CA6220 (to R. H. S.), HL36946, HL63819 (to P. J. S.), and HL61200 (to T. W.) from the Department of Health and Human Services of the United States Public Health Service. This is manuscript number 14534-MEM from the Scripps Research Institute.
↵2 To whom requests for reprints should be addressed, at Department of Cancer Biology, NB40, The Lerner Research Institute, The Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195. Phone (216) 445-9650; Fax: (216) 445-6269; E-mail:
↵3 The abbreviations used are: PLSCR1, phospholipid scramblase 1; PS, phosphatidylserine; mab, monoclonal antibody; IRES, internal ribosome entry site.
- Received August 28, 2001.
- Accepted November 13, 2001.
- ©2002 American Association for Cancer Research.