- Original Articles
- Open Access
An Essential Role for Macrophage Migration Inhibitory Factor (MIF) in Angiogenesis and the Growth of a Murine Lymphoma
© Picower Institute Press 1999
- Accepted: 15 January 1999
- Published: 1 March 1999
Macrophage migration inhibitory factor (MIF) has been shown to counterregulate glucocorticoid action and to play an essential role in the activation of macrophages and T cells in vivo. MIF also may function as an autocrine growth factor in certain cell systems. We have explored the role of MIF in the growth of the 38C13 B cell lymphoma in C3H/HeN mice, a well-characterized syngeneic model for the study of solid tumor biology.
Materials and Methods
Tumor-bearing mice were treated with a neutralizing anti-MIF monoclonal antibody and the tumor response assessed grossly and histologically. Tumor capillaries were enumerated by immunohistochemistry and analyzed for MIF expression. The effect of MIF on endothelial cell proliferation was studied in vitro, utilizing both specific antibody and antisense oligonucleotide constructs. The role of MIF in angiogenesis also was examined in a standard Matrigel model of new blood vessel formation in vivo.
The administration of anti-MIF monoclonal antibodies to mice was found to reduce significantly the growth and the vascularization of the 38C13 B cell lymphoma. By immunohistochemistry, MIF was expressed predominantly within the tumor-associated neovasculature. Cultured microvascular endothelial cells, but not 38C13 B cells, produced MIF protein and required its activity for proliferation in vitro. Anti-MIF monoclonal antibody also was found to markedly inhibit the neovascularization response elicited by Matrigel implantation.
These data significantly expand the role of MIF in host responses, and suggest a new target for the development of anti-neoplastic agents that inhibit tumor neovascularization.
The pathogenesis of cancer is complex, involving cellular transformation and proliferation, stromal support responses such as angiogenesis, and evasion of host immune defenses (1–3). Inflammatory cytokines have been recognized in many instances to play an important role in the development and growth of tumors (4). Interleukin-1 (IL-1), IL-2, hematopoietic colony-stimulating factors (CSFs), platelet-derived growth factor (PDGF), and transforming growth factor β1 (TGF-β1) each can act as an autocrine growth factor for tumor cells, and antibodies to these cytokines inhibit tumor cell proliferation in model systems (5). Growth factors such as TGF-β1 and chemokines such as IL-8 also have been shown to have angiogenic activity and to promote the development of the supporting vas-culature which is necessary for solid tumor growth (6–8).
Macrophage migration inhibitory factor (MIF) was originally described 30 years ago as a T cell-derived factor that inhibited the random migration of macrophages in vitro (9,10). In more recent studies, MIF also has been found to play an essential regulatory role in macrophage activation and in mitogen- and antigen-driven T cell proliferation (11,12). MIF is produced by these cells in response both to glucocorticoids and inflammatory stimuli, and acts to counterregulate the immunosuppressive effects of glucocorticoids on macrophage and T cell cytokine production (11–13). Circulating levels of MIF increase as a consequence of various systemic inflammatory conditions and neutralizing anti-MIF antibodies suppress delayed-type hypersensitivity reactions, antigen-specific T cell activation, and the toxic response to septicemia (11,14–16).
In the present study, we report that the mediator MIF plays an essential role in the formation of new blood vessels. The administration of neutralizing anti-MIF antibodies to mice was found to significantly retard the growth of the 38C13 B cell lymphoma, and reduced tumor growth was associated with a marked reduction in tumor angiogenesis. In vivo, MIF was found to be expressed predominantly by tumor endothelium and in vitro, microvascular endothelial cells were found to secrete MIF protein and to require MIF to proliferate. Finally, MIF was observed to be necessary for the outgrowth of new vessels in an in vivo model of angiogenesis utilizing Matrigel implantation. These data assign a previously unexpected role for MIF in the angiogenic response, and in the resultant growth of certain neoplasms.
38C13 B Lymphoma Growth In Vivo
38C13 B lymphoma cells (provided by J. D. Kemp, Department of Pathology, University of Iowa) were collected from exponential growth phase culture in RPMI 1640 medium containing glutamine (300 µg/ml), sodium pyruvate (110 µg/ml), 2-mercapto-ethanol, (5 × 10–5 M), HEPES (10 mM, pH 7.2), and 10% heat inactivated fetal calf serum (FCS), and then washed twice and resuspended in phosphate-buffered saline (PBS) (1 × 106 cells/ml). Following the methods of Kemp et al. (17), groups of five C3H/HeN female mice (20–25 g, Harlan, Indianapolis, IN) were shaven on the upper flank and 0.05 ml of the 38C13 cell suspension (5 × 104 cells) was injected i.d. with a 1-ml syringe and 27-gauge needle. In the initial tumor outgrowth experiments, mice received an i.p. injection of 0.3 ml PBS, IgG1 isotype control antibody (0.5 mg), or a purified anti-MIF monoclonal antibody (MAb) (0.5 mg) (13) 1 hr after 38C13 cell injection and then every 24 hr for 7 days. The anti-MIF mAb was produced as mouse ascites, precipitated with NH4SO4 and purified by anion exchange chromatography on FPLC (HiTrapQ, Pharmacia, Uppsala, Sweden). The lipopolysaccharide (LPS) content of anti-MIF and control antibodies was determined to be 0.05 fg/ng protein by the Limulus amoebocyte lysate assay. Anti-MIF mAb was >95% pure as determined by Coomassie blue staining/SDS-PAGE. In the established solid tumor experiments, the tumors were allowed to grow for 5 days to a mean weight of 50 mg before treatment was begun. Mice then received an i.p. injection of 0.3 ml PBS, IgG1 isotype control antibody (0.5 mg), or an anti-MIF mAb (0.5 mg) every 12 hr for 4 days. Tumor size was determined with Vernier calipers according to the following formula: weight (mg) = (width, mm)2 × (length, mm)/2 (18). In addition to examining tumor weights in situ, we excised initial outgrowth tumors and directly measured their wet weights: PBS, 671.4 ± 50.6 mg; IgG1, 693.4 ± 110.9 mg; anti-MIF, 205 ± 62.5 mg (p < 0.05). Anti-MIF monoclonal antibody was produced as mouse ascites, precipitated with ammonium sulfate, and purified by anion exchange chromatography on FPLC (HiTrapQ, Pharmacia). Statistical significance was assessed by two sample t-tests (assuming unequal variances) (19).
Tumors were excised from euthanized mice, fixed in neutral buffered 4% formalin, sectioned, and processed for immunohistochemical analyses. To assess vascularization, the deparaffinized sections were incubated with an anti-CD31 mAb (1:50 dilution) (clone MEC 13.3) (Pharmingen, San Diego, CA) or an IgG2a isotype control (Pharmingen). Sections then were incubated with an alkaline phosphatase-linked secondary antibody and developed with new fuchsin (Dako, Corporation, Carpinteria, CA) as substrate. To assess MIF and vWF protein expression, peroxidase-blocked (3% H202) sections were incubated with an affinity-purified, monoclonal anti-MIF antibody and, following three washes in PBS/0.05% Tween-20, the bound antibody was visualized using the universal LSAB-2 horseradish peroxidase kit according to the manufacturer’s instructions (Dako) (15). The sections were stained with 3-amino-9-ethylcarbazole (for anti-MIF) or diaminobenzidine (for anti-vWF) as chromogenic substrate. For double-immunostaining, anti-MIF stained sections were washed and then labeled with anti-vWF Ab, incubated with an alkaline phosphatase-linked secondary Ab, and developed with new fuchsin (Dako) as substrate. Control sections incubated with an isotype control or without primary antibody showed no immunoreactivity.
Microvascular Endothelial Cell Proliferation
Human microvascular endothelial cells (primary, fourth passage) (Clonetics, San Diego, CA) were cultured in 96-well flat bottom plates (5 × 103 cells/well) with 100 µl Endothelial Cell Growth Medium (Clonetics) (diluted 1:5 with RPMI 1640) supplemented with 1% heat-inactivated FCS. Endothelial cell cultures were >95% pure as demonstrated by flow cytometry for Factor VHI related antigen (clone no. F3520, Sigma Chemical Co., St. Louis, MO). Cells were incubated for 20 hr alone as control or in the presence of IgG1 isotype control (Sigma) or neutralizing anti-MIF mAb (25–200 µg/ml). In separate experiments, cells were transfected for 20 hr by the lipofectin method (Gibco, Gaithersburg, MD) with the following phosphorothioate oligonucleotides: S-MD?:5′-GCC-ATC-ATG-CCG-ATG-TTC — AT- 3′; AS -MIF: 5′ — ATG-AAC — ATC — GGC-ATG-ATG-GC-3′ (designed to span the MIF translation start site) (10 µg/ml) (Oligo’s, etc., Wilsonville, OR) (20). Transfection efficiency studies were conducted with several concentrations of anti-MIF anti-sense oligonucleotides (MIF was estimated by Western blot analysis and a UMAX scanner; data not shown). The proliferative activity was measured by the incorporation of [3H]thymidine (4 µCi/ml) (DuPont, Boston, MA) into DNA over the last 16 hr of incubation as measured by liquid scintillation counting. Data are expressed as the mean ± SD (n = 3).
MIF Protein Expression
In Vivo Angiogenesis Assay
An in vivo angiogenesis assay using Matrigel was performed as previously described (21). Briefly, female BALB/c mice (≥6 months old; Jackson Laboratories, Bar Harbor, ME,) were injected subcuta-neously (s.c) with 0.5 ml liquid Matrigel (Collaborative Biomedical Products, Bedford, MA) carefully mixed with aFGF (1 ng/ml; R&D Systems, Minneapolis, MN) and heparin (64 units/ml) (and/or monoclonal antibodies, 25 µg/ml) near the abdominal rnidline. The negative control animals (no angiogenesis) were injected with Matrigel containing heparin (64 U/ml) alone. For the antibody studies, mice were injected 30 min prior to Matrigel injection and every other day during the study with 500 µg purified anti-MIF monoclonal antibody (or control IgG) i.p. in PBS. Mice were sacrificed 8 days after the Matrigel injection and the Matrigel plugs consisting of the animals’ tissues (overlying skin and peritoneal lining) were recovered by dissection. The plugs were fixed in 10% neutral buffered formalin, cleared, paraffin embedded, and sectioned at 5 µm. Sections were evaluated using Masson’s Trichrome stain and von Willebrand Factor staining for endothelial cells (Dako). In addition, 100 µg of the Matrigel plug was used for hemoglobin analysis using the Drabkin reagent kit 525 (Sigma).
Inhibition of B Cell Lymphoma Growth In Vivo
Our initial experiments were aimed at defining a potential role for MIF in tumorigenesis. The 38C13 B cell lymphoma of C3H/HeN mice is a well-characterized syngeneic model for the study of solid tumor biology (16, 22–24). C3H/HeN mice were administered an i.d. injection of 50,000 38C13 tumor cells and then treated at 1-day intervals with a neutralizing anti-MIF mAb, control IgG1, or PBS (13). After 7 days, the tumors were measured with Vernier calipers and their weights calculated. As shown in Figure 1, anti-MIF antibodies significantly reduced the initial outgrowth of the 38C13 B cell lymphoma in mice compared to controls. The specificity of the anti-MIF mAb was confirmed by Western blot analysis of 38C13 B cell lymphoma tissue (Fig. 1C, Lane b).
The proliferation of cultured 38C13 B cell lymphoma cells was unaffected by either anti-MIF mAb (at an established neutralizing concentration of 100 µg/ml) (11) or MIF-specific anti-sense phosphorothioate oligonucleotides: (control IgG1 mAb: 44324 ± 6775 cpm; anti-MIF mAb: 46674 ± 4371 cpm, n = 4, p > 0.5; MIF-specific sense oligo: 62399 ± 3548 cpm; MIF-specific anti-sense oligo: 61884 ± 7377 cp, n = 4, p > 0.5.) Additionally, murine rMIF at concentrations shown previously to be maximally bioactive (13) had no effect on 38C13 cell proliferation in vitro [PBS control 57433 ± 2219 cpm; rMIF (10 ng/ml): 60190 ± 7377 cpm, n = 4, p > 0.5]. Taken together, these observations suggest that MIF plays a role in tumor outgrowth in vivo.
Reduction of Tumor Vascularization
Inhibition of Microvascular Endothelial Cell Proliferation In Vitro
Inhibition of Angiogenesis In Vivo
The present study further expands the emerging role of MIF in biological responses by showing that this mediator plays a critical role in the initial outgrowth and vascularization of a mouse B cell lymphoma. That MIF is required for the vascularization process complements recent observations showing that MIF mRNA and protein are expressed in dermal capillaries and inflammatory lesions in vivo (33). A strict dependence of solid tumor growth on neo-vascularization has been established by a number of model studies in vivo (26–28). This process involves several sequential steps, including (1) degradation of capillary basement membrane, (2) endothelial cell migration and proliferation, and (3) formation of a new lumen (34). In the present study, MIF was found to be expressed in tumor-associated endothelium and to be necessary for microvascular endothelial cell proliferation in vitro and for neovascularization in a Matrigel model in vivo. MIF thus shares certain of the proangiogenic properties of vascular endothelial cell growth factor (VEGF), transforming growth factors (TGFs), acidic (aFGF) and basic fibroblast growth factors (bFGF), and other mediators that have been investigated in recent years (35). Interestingly, acidic FGF has been found in preliminary studies to induce MIF protein expression by endothelial cells in vitro (C. Metz, personal communication). Taken together with the observation that anti-MIF neutralizes acidic FGF-mediated effects in vivo, these observations suggest that MIF functions as a “downstream” mediator of acidic FGF.
Certain of the known biological properties of MIF suggest that its role in the tumor angiogenic response may be unique. MIF circulates constitutively in plasma, and has been shown to be released from immune cells by both proinflammatory stimuli and by glucocorticoids (12–14). Once released in an inflammatory setting, MIF has the capacity to override certain of the antiinflammatory and immunosuppressive effects of glucocorticoids (13). Tumors such as lymphomas are frequently sensitive to the cytolytic effects of steroids, and an additional mechanism of action of MIF may be to override these effects and promote tumor growth (36–38). Accordingly, anti-MIF may serve to neutralize the MIF expressed within tumor sites and augment the anti-tumor effect of endogenous glucocorticoids. Further studies of this regulatory phenomenon, particularly as it may pertain to glucocorticoid-respon-sive, human lymphomas, will be necessary to evaluate the potential anti-tumor role of this mechanism versus the strict effects of anti-MIF on tumor angiogenesis.
An additional feature of MIF action in tumor growth that is still to be explored may be in the generation of tumor-specific helper or cytotoxic T cell responses. MIF has been shown to be required for antigen-specific, T cell activation in vivo (11). TH2 cells produce higher amounts of MIF upon mitogenic stimulation than TH1 cells, and MIF is essential for the development of the humoral response (11). In vivo, anti-MIF mAbs may act to neutralize TH2 cell-derived MIF and promote a shift toward increased TH1 effector cell activity and consequent macrophage cytolysis of tumor cells. Further analysis of MIF’s role in the establishment of TH1/TH2 cell responses during tumorigenesis may provide important insight into the mechanisms underlying host immunity against tumors.
In conclusion, these findings support the concept that MIF plays a critical role in angiogenesis and in the establishment of certain solid tumors. Elucidation of the precise mechanism of action of MIF during the host angiogenic and immune response to tumors may contribute new information on the complex, regulatory pathways that govern neoplastic growth. The recent determination of the crystal structure of MIF, which shows a unique trimeric α/β structure, together with the discovery of an enzymatic activity for MIF should assist in the development of potential inhibitors of MIF that may find clinical utility as anti-neoplastic agents (39,40).
This work was supported by grants from the NIH (AI42310-01A1) and the Arthritis Foundation.
- Nicholson GL. (1987) Tumor cell instability, diversification, and progression to the metastatic phenotype: from oncogene to oncofetal expression. Cancer Res. 47: 1473–1487.Google Scholar
- Folkman J. (1986) How is blood vessel growth regulated in normal and neoplastic tissue? Cancer Res. 46: 467–473.PubMedGoogle Scholar
- Muul LM, Spiess PJ, EP Director, Rosenberg SA. (1987) Identification of specific cytolytic immune responses against autologous tumor in humans bearing malignant melanoma. J. Immunol. 138: 989–995.PubMedGoogle Scholar
- Pugh-Humphreys RGP, Woo J, Thomson AW. (1991) Cytokines and their receptors as potential therapeutic targets. In: The Cytokine Handbook. Academic Press, New York, pp. 357–386.Google Scholar
- Lang RA, Burgess AW. (1990) Autocrine growth factors and tumourigenic transformation. Immunol Today 11: 244–249.View ArticlePubMedGoogle Scholar
- Barnard JA, Lyons RM, Moses HL. (1990) The cell biology of transforming growth factor beta. Biochim. Biophys. Acta 1032: 79–87.PubMedGoogle Scholar
- Roberts AB, Thompson NL, Heine U, Flanders C, Sporn MB. (1988) TGF-beta: possible roles in carcinogenesis. Br. J. Cancer 57: 594–600.View ArticlePubMedPubMed CentralGoogle Scholar
- Arenberg DA, Kunkel SL, Polverini PJ, Glass M, Burdick MK, Strieter RM. (1996) Inhibition of IL-8 reduces tumorigenesis of human non-small cell lung cancer in SCID mice. J. Clin. Invest. 97: 2792–2802.View ArticlePubMedPubMed CentralGoogle Scholar
- David JR. (1996) Delayed hypersensitivity in vitro: its mediation by cell free substances formed by lymphoid cell-antigen interaction. Proc. Natl. Acad. Sei. U.S.A. 56: 72–77.View ArticleGoogle Scholar
- Bloom BR, Bennet B. (1966) Mechanism of a reaction in vitro associated with delayed-type hypersensitivity. Science 153: 80–82.View ArticlePubMedGoogle Scholar
- Bacher MB, Metz CN, Calandra T, et al. (1996) An essential role for macrophage migration inhibitory factor in T-cell activation. Proc. Natl. Acad. Sci. U.S.A. 93: 7849–7854.View ArticlePubMedPubMed CentralGoogle Scholar
- Calandra T, Bernhagen J, Mitchell RA, Bucala R. (1994) The macrophage is an important and previously unrecognized source of macrophage migration inhibitory factor. J. Exp. Med. 179: 895–1902.View ArticleGoogle Scholar
- Calandra T, Bernhagen J, Metz CN, et al. (1995) MIF as a glucocorticoid-induced counter-regulator of cytokine production. Nature 377: 68–71.View ArticlePubMedGoogle Scholar
- Bernhagen J, Calandra T, Mitchell RA, et al. (1993) MIF is a pituitary-derived cytokine that potentiates lethal endotoxaemia. Nature 365: 756–759.View ArticlePubMedGoogle Scholar
- Bernhagen J, Bacher M, Calandra T, et al. (1996) An essential role for MIF in the tuberculin delayed-type hypersensitivity reaction. J. Exp. Med. 183: 277–282.View ArticlePubMedGoogle Scholar
- Calandra T, Spiegel LA, Metz CN, Bucala R (1998) Macrophage migration inhibitory factor (MIF) is a critical mediator of the activation of immune cells by exotoxins of gram-positive bacteria. Proc. Natl. Acad. Sci. U.S.A. 95: 11383–11388.View ArticlePubMedPubMed CentralGoogle Scholar
- Kemp JD, Cardillo T, Stewart BC, et al. (1995) Inhibition of lymphoma growth in vivo by combined treatment with hydroxyethyl starch deferoxamine conjugate IgG monoclonal antibodies against the transferrin receptor. Cancer Res. 55: 3817–3824.PubMedGoogle Scholar
- Taetle R, Rosen F, Abramson I, Venditti J, Howell S. (1987) Use of nude mouse xenografts as preclinical drug screens: in vivo activity of established chemotherapeutic agents against melanoma and ovarian carcinoma xenografts. Cancer Treatment Rep. 71: 297–304.Google Scholar
- Zar JH. (1984) Biostatistical Analysis. Prentice-Hall, Englewood Cliffs, NJ.Google Scholar
- Waeber G, Calandra T, Roduit R, et al. (1997) Insulin secretion is regulated by the glucose-dependent production of islet β-cell macrophage migration inhibitory factor (MIF). Proc. Natl. Acad. Sei. U.S.A. 94: 4782–4787.View ArticleGoogle Scholar
- Bernhagen J, Mitchell RA, Calandra T, Voelter W, Cerami A, Bucala R. (1994) Purification, bioactivity, and secondary structure analysis of mouse and human MIF. Biochemistry 33: 14144–14155.View ArticlePubMedGoogle Scholar
- Passanti A, Taylor RM, Pili R, et al. (1992) A simple, quantitative method for assessing angiogenesis and antiangiogenic agents using reconstituted basement membrane, heparin, and fibroblast growth factor Lab. Invest. 67: 519–528.Google Scholar
- Bergman Y, Haimovich J. (1977) Characterization of a carcinogen-induced murine B lymphocyte cell of C3H/eB origin. Eur. J. Immunol. 7:413–417.View ArticlePubMedGoogle Scholar
- Campbell MJ, Esserman L, Levy R. (1988) Immunotherapy of established murine B cell lymphoma. Combination of idiotype immunization and cyclophosphamide. J. Immunol. 141: 3227–3233.PubMedGoogle Scholar
- Basham TY, Palladin MA, Badger CC, Bernstein ID, Levy R, Merigan TC. (1988) Comparison of combinations of interferons with tumor specific and nonspecific monoclonal antibodies as therapy for murine B cell and T cell lymphomas. Cancer Res. 48: 4196–4200.PubMedGoogle Scholar
- Kaminski MS, Kitamura K, Maloney DG, Levy R. (1987) Idiotype vaccination against murine B cell lymphoma. Inhibition of tumor immunity by free idiotype protein. J. Immunol. 138: 1289–1296.PubMedGoogle Scholar
- Folkman J, Cotran R. (1976) Relation of vascular proliferation to tumor growth. Int. Rev. Exp. Pathol. 16: 207–248.PubMedGoogle Scholar
- Gimbrone MA, Leapman SB, Cotran R, Folkman J. (1972) Tumor dormancy in vivo by prevention of neovascularization. J. Exp. Med. 136: 261–276.View ArticlePubMedPubMed CentralGoogle Scholar
- Folkman JK, Watson D, Ingber D, Hanahan J. (1989) Induction of angiogenesis during the transition from hyperplasia to neoplasia. Nature 339: 58–61.View ArticlePubMedGoogle Scholar
- Vecchi AC, Garlanda C, Lampugnani MG, et al. (1994) Monoclonal antibodies specific for endothelial cells of mouse blood vessels. Their application in the identification of adult and embryonic endothelium. Eur. J. Cell Biol. 63: 247–254.PubMedGoogle Scholar
- Xie Y, Muller WA. (1993) Molecular cloning and adhesive properties of murine platelet/endothelial adhesion molecule-1. Proc. Natl. Acad. Sci. U.S.A. 90: 5569–5573.View ArticlePubMedPubMed CentralGoogle Scholar
- DeLisser HM, Newman PJ, Albeda SM. (1993) Platelet endothelial cell adhesion molecule (CD31). Curr. Top. Microbiol. Immunol. 184: 37–45.PubMedGoogle Scholar
- Buck CA, Baldwin HS, DeLisser H, et al. (1993) Cell adhesion receptors and early mammalian heart development—an overview. C R Acad. Sci. III 316: 849–859.Google Scholar
- Bacher MB, Meinhardt A, Lan HY, et al. (1997) Migration inhibitory factor expression in experimentally induced endotoxemia. Am. J. Pathol. 150: 235–246.PubMedPubMed CentralGoogle Scholar
- Whalen GF, Zetter BR. (1992) Angiogenesis. In: Wound Healing, W. B. Saunders, Philadelphia, pp. 77–95.Google Scholar
- Folkman J, D’Amore PA. (1996) Blood vessel formation: what is its molecular basis? Cell 87: 1153–1155.View ArticlePubMedGoogle Scholar
- Fiorentino MV. (1991) Lymphomas in the elderly. Leukemia 5: 79–85.PubMedGoogle Scholar
- Goss PE. (1992) New perspectives in the treatment of non-Hodgkin’s lymphoma. Semin. Oncol. 19: 23–29.PubMedGoogle Scholar
- Zackheim HS. (1994) Treatment of cutaneous T-cell lymphoma. Semin. Dermatol. 13: 207–215.PubMedGoogle Scholar
- Sun HW, Bernhagen J, Bucala R, Lolis E. (1996) Crystal structure at 2.6-A resolution of human macrophage migration inhibitory factor. Proc. Natl. Acad. Sei. U.S.A. 93: 5191–5196.View ArticleGoogle Scholar
- Rosengren E, Bucala R, Aman P, et al. (1996) The immunoregulatory mediator MIF catalyzes a tau-tomerization reaction. Mol. Med. 2: 143–149.PubMedPubMed CentralView ArticleGoogle Scholar