Phospholipase A2-mediated inflammation induces regression of malignant gliomas

ELSEVIER

CANCER LETTERS Cancer Letters 102 (1996) l-6

Phospholipase AZ-mediated inflammation induces regression of malignant gliomas David H. Goddarda, John S. Bomalaskib, Stanley Lipperc, Robert G.L. Shor@, Mike A. Clarkd%* “The Long Islund College Hospital, Brooklyn, NZ and SUNY-Stony Brook, Stony Brook, NY. USA bMedical College of Pennsylvania, and Hahnemnn University, Philadelphia, PA, USA ’ Winthnp-University Hospital, Mineola, NY and SUNY-Stony Brook, Stony Brook, NL USA dEnnzon,Inc., 20 Kingsbridge Road, Piscataway, NJ 08854, USA Received 23 October 1995; accepted 10 January 1996

Abstract An ideal form of cancer therapy is the harnessing of innate immunity to eradicate spontaneously arising clones of malignant cells. To date, attempts to develop effective immunotherapies have met with limited success. Prostaglandins and leukotrienes, collectively known as eicosanoids, are important mediators of immune and inflammatory responses. Harnessing these compounds could be a method to treat cancers. Eicosanoids are formed after cleavage of fatty acids from phospholipids by phospholipase enzymes. We have previously described, characterized and cloned a naturally occurring mammalian activator of phospholipase AZ. Injection of a 24 amino acid peptide from this phospholipase A2 activating protein (PLAP), resulted in induction of an acute inflammatory response, and a concomitant regression of gliomas in rats. Administration of 5OOpg of this protein resulted in a 50% decrease of the tumor mass within 72 h. Tumor regression coincided with a greater than twentyfold increase in levels of prostaglandin Ez (PGE2) and leukotriene B4 (LTB,), and a marked infiltration of natural killer (NK) cells. These data suggest that activation of phospholipase AZ and modulation of the eicosanoid biosynthetic pathway may provide a novel therapeutic strategy for the successful treatment of malignant tumors of the nervous system. Keywords:

Phospholipases; Eicosanoids; Gliomas; Inflammation; Natural killer cells

1. Introduction Central to the immune surveillance theory, first proposed by Lewis Thomas [l] more than 30 years ago, is the role of the immune system in the identifi-

* Corresponding

8852950.

author. Tel.: +I 908 9804991; fax: +I 908

cation and eradication of spontaneously arising clones of malignant cells. Cytotoxic T lymphocytes (natural killer, NK; and lymphokine activated killer, LAK) cells play a vital role in this process. Recent data indicate that functions of these cells may be altered by arachidonic acid metabolites, known collectively as eicosanoids [2,3]. Phospholipase A2 (PLA,) enzyme activation is believed to be the ratelimiting step in the eicosanoid biosynthetic pathway, and several naturally occurring activators of PLA2

0304-3835/96/$12.00 0 1996Elsevier Science Ireland Ltd. All rights reserved PII: SO304-3835(96)04142-O

2

D.H. Goddard et al. /Cancer Letters 102 (1996) l-6

have been described [3-81. Venoms, particularly those from bees and wasps, incite intense local inflammatory reactions through the generation of eicosanoids. Of those identified, the bee venom peptide melittin is the most comprehensively characterized [91. We have identified, characterized, and cloned a mammalian protein that is structurally and funcsimilar to melittin, tionally termed PLAP (phospholipase A2 activating protein) [lo]. High concentrations of PLAP are found in joint effusions in patients suffering from rheumatoid arthritis, a disorder characterized by intense inflammatory and immune responses [I 11. This protein is at least a thousand fold more potent than melittin in activating PLA*, but 1s not cytotoxic even at high concentrations [ 121. We have synthesized a 24 amino acid peptide within PLAP that contains the PLA2 activating domain of the molecule [lo]. Injection of this peptide or native PLAP into rabbit joints, or subcutaneous tissue, induces an intense inflammatory reaction [II]. Malignant gliomas of the central nervous system are associated with poor patient survival as they frequently arise in surgically inaccessible sites and tend to be resistant to both radiation and chemotherapy [15]. Median survival, even in treated patients, is estimated to be no more than about 18 months 1151. Since it has been suggested that parenteral administration of venoms may enhance immunologic responses to cancer chemotherapies [ 161, we were interested to know whether PLAP had the capacity to induce inflammatory necrosis in tumor tissues. To address this question, we utilized a rat model in which glioma cells were cultured in subcutaneous air pouches. Such pouches have long been utilized to study in vivo events associated with inflammation [ 181. Furthermore, the choice of this in vivo culture system was based on earlier studies demonstrating that formation of the air pouch membrane was associated with expression of transforming growth factor p, and basic fibroblast growth factor, polypeptide growth factors known to provide crucial mitogenic stimuli for normal and transformed cells [19]. In addition, these pouches allow ready access to the fluids that surround and bath the tumors enabling the investigator to identify the infiltrating cells and fluid constituents.

2. Materials and methods 2. I. Induction of air pouches Air pouches were formed on the dorsum of female Wistar rats (average weight 200 g; average age 8 weeks, purchased from Taconic Labs) by injection of 5 ml of sterile air, according to previously described methods [18]. Pouch inflation was maintained by re-injection of air every 45 h. The pouches were allowed to mature for 14 days prior to the start of the experiments. 2.2. Tumor formation Exponentially growing C6 glioma cells, obtained from the American Type Culture Collection (Bethesda, MD), cultured in Ham’s F12 medium supplemented with horse serum and fetal calf serum (15% and 2.5%; v/v) were injected into the rat air pouches (1 X 10’ cells/animal). Tumors were allowed to grow for 14 days before injection of the peptide. 2.3. Synthesis of PUP peptide and tumor injection PLAP peptide (ESPLIAKVLTI’EPPIITPVRRT) and PLAPP’O- peptide (ESPLIAKVLTTEIITPVRRT) were synthesized according to described methods [10,14]. Peptides (lOOpug or 200pglml final concentration) were bound overnight to agarose beads (Affigel blue; BioRad, Melville, NY) which had been extensively washed in sterile deionized water and resuspended in sterile saline (0.9%; w/v). This procedure results in a slow release of the peptide from the beads into the surrounding tissues (data not shown). Prior to injection of the peptide, fluids were aspirated from the centers of the tumors. Tumors were injected with 5 ml of the PLAP-agarose, or PLAPpro-agarose suspension, or agarose alone, and animals euthanized at 72 or 144 h. Samples of fluids were collected from the tumors and total white cell counts determined by Coulter counter (Coulter, Hialeah, FL). Fluids were then subjected to centrifugation (10 000 X g for 10 min) to remove cells and other debris, and the supernatants stored at -80°C before analysis. Tumors were removed by dissection. Maximal

D.H. Goddurd et ul. /Cuncer Letters 102 (1996) I-6

tumor wall thickness was measured and samples of the tumors were fixed in formalin, or snap-frozen in liquid nitrogen for histological analysis. 2.4. Hi.ctological analysis Hematoxylin and eosin staining was performed on sections cut from paraffin-embedded tissues that had been fixed in formalin. Immunocytochemistry was performed on sections cut from snap-frozen tissues. Neurofilaments were visualized using antibodies to a 200 kDa glial neurofilament protein (Clone 2F1 1; Dako Corp; Carpinteria, CA). NK-cells were visualized by staining with MAb NRK-lP, a type II transmembrane protein expressed on the surface of rat

3

NK-cells and large granular lymphocytes (Clone 3.2.3; Endogen Products, Cambridge, MA). Primary antibodies were visualized using appropriate biotinlabelled secondary antibodies, and Avidin-BiotinComplex (ABC ELITE Kit, Vector Laboratories, Burlingame, CA). 2.5. Eicosanoid assays Eicosanoid were determined in samples of supernatants from tumor fluids using ELISA kits which measure PGE, and LTB,. These were purchased from Cayman Chemical Company (Ann Arbor, MI), and used according to the instructions of the manufacturer.

Fig. 1. Effects of PLAP treatment on morphologic characteristics of C6 glial tumors grown in rat air pouches. (a-c) Micrographs of hematoxylin and eosin stained sections. (d-f) Glial tumors stained using antibodies to 200 kDa glial neurofilament protein. (g-i) Micrographs of glial tumors stained with MAb to NRK-IP, a type II transmembrane protein expressed on the surface of rat NK-cells and large granular lymphocytes. (n,d,g) Representative sections cut from untreated tumors. (b,c,h) Representative sections cut from tumors treated with PLAP. tc.f.ii Sections cut from tumors treated within an inactive PLAPP’O- peptide.

4

D.H. Goddard et al. /Cancer Letters 102 (1996) 1-6

2.6. Analysis of data Data were analyzed by using Student’s t-test. 3. Results Within 14 days of inoculation, C6 glioma cells proliferate to form large tumors. The capacity of PLAP to induce necrosis was determined by injection of PLAP coupled to agarose beads, or agarose beads alone, into the tumors. As a further control, tumors were injected with PLAPp”‘- peptide, a synthetic PLAP peptide in which elimination of the two internal proline amino acids results in a loss of PLA, stimulatory activity [lo]. The histology of the control (Fig. la,d,g), PLAP (Fig. lb,e,h), and PLAPP’O- (Fig. lc,f,i) peptide0

20 ,-

7

; ~~~1Agarose BZB PLAPProD

PLAP

I

Day 6 Fig. 2. The time course of PLAP-induced tumor regression. C6 ghomas were grown in rats as described in Fig. 1 for 2 weeks. (a) The results of experiments in which rats were treated with agarose, PLAP or PLAPp”- (5OOrg per animal). Tumors were excised on days 3 and 6. Maximum tumor thickness was measured and the results shown represent mean + SEM in each treatment group (n = 8 per group). As shown, and compared to PLAPpr’-, there was a more than 50% regression in tumor size 3 days after a single treatment with PLAP (*P c 0.0001, f-test; one-tailed), and a greater than 75% reduction in tumor size at day 6 (**P < 0.001, r-test; one-tailed). In several of the PLAP-treated animals, there appeared to be an almost total regression of the tumor, with only the air pouch membrane remaining.

250

500

1000

PLAP DOSE (ug ) Fig. 3. Dose response of tumor regression in response to treatment with PLAP. Rats were treated with PLAP on day I and the tumors were excised on day 4 following a single injection of PLAP. Maximum tumor thickness was measured and the results shown represent the mean f SEM for each concentration of PLAP (n = 8 animals in each treatment group). Compared with agarose, PLAP at a dose of 5OOpg per animal induced a 50% regression in tumor thickness (*P < 0.01, **P < 0.001, f-test; one-tailed). Increased tumor necrosis was not observed after administration of a higher dose of peptide (lOOO~g), and lower doses of peptide (25Opg) had no measurable effect on tumor thickness.

treated tumors are shown. The morphology of the hematoxylin and eosin stained tumors injected with agarose (Fig. la), or PLAPP’O- (Fig. lc), were indistinguishable from untreated animals (data not shown). In contrast, injection of PLAP caused dramatic necrosis of the tumors (Fig. 1b). Confirmation that PLAP induced necrosis of glial cells was demonstrated by loss of neurofilament staining in PLAPtreated tumors (Fig. le), compared with either PLAPP’* peptide (Fig. If), or controls (Fig. Id). Staining of tumor tissues with vital dyes identified a mononuclear cell infiltrate, consisting of a predominance of cytotoxic T cells with an NK phenotype (Fig. lg,h,i). More NK-cells were present in PLAP-treated tumors than either PLAPP’O- or control tumors.

D.H. Goddurd et al. /C’u~cer Letters 102 (1996) l-6 Table

I

5

PLAP-induced regression of tumors was both time and dose dependent (Figs. 2 and 3). Fluid present in the necrotic centers of the tumors was examined and found to contain large numbers of mononuclear cells and eicosanoids. Treatment with PLAP, but not PLAPP’“- induced an increase in numbers of lymphocytes, and amounts of PGE,, and LTBJ (Table I )> indicating that PLAP-induced tumor necrosis is accompanied by modulation in tissue levels of eicosanoids. These findings are consistent with the observation that tumor growth is accompanied by an indolent intlammatory response.

formation, PLAP treatment also resulted in tumor infiltration by NK-cells (Fig. I). These cells are well known inhibitors of metastatic growth and appear to block tumor cell growth by interacting with cytokines such as IL- 12 [24-261. Taken together, these data indicate that intervention with PLAP appears to be another example of NK-cell interaction with soluble mediators resulting in regulation of cell growth. While there is continued interest in the potential role of recombinant toxins as therapeutic agents, overall results have been disappointing [27]. Reasons advanced for the failure of trials of adoptive immunotherapy include the concept that functional behavior of cytotoxic T-lymphocytes may be modulated by local levels of eicosanoids [28,29]. Thus, in patients with metastatic breast cancer, failure of adoptive cellular immunotherapy with LAK cells has been attributed to an over-production of PGEz by mononuclear cells 131. Accumulating data indicate that NK-cell cytotoxicity is suppressed by local levels of PGE,, and enhanced by local levels of LTB4 [28,29]. Our study provides novel experimental evidence showing that PLAP treatment dramatically augments the levels of eicosanoids, enhances NK-cell tumoritidal activity, and causes inflammatory necrosis in a glioma model of malignancy. PLAP may ultimately provide an effective treatment for human malignant tumors of the central nervous system.

4. Discussion

References

Summary 01. PLAP-induced gliomn effusions

changes

in eicosanoid

profiles

in

Treatment

Cells( x IO”/ml) PGE? (pg/ml) LTB4 (pgiml)

Agarosc

PLAP

PLAPP’O-

l.Yt0.2 2277 f 987 27 k Cl

2.7 f 1.3* SlOS6+ lS258* 17042 l306*

I .6 + 0.3 8748-c6205 ll7?38

Results shown represent means f SD for samples obtained from animals in each treatment group (II = 8 animals per group). All samples were assayed in duplicntc. Fluids were also assayed for PGFz,,. thromhoxane HZ, 8-isoprostane, LTC+ and LTE,. and no significant diffcrenccs were observed (data not shown). *I’ < 0 OS t-test. one-tallcd compared to ngnrose PLAPP’“-

This study demonstrates that injection of PLAP, the only known naturally occurring human PLA, activator, induces regression of malignant gliomas. The mechanism by which tumor regression occurs involves PLA? activation, and subsequent production of PGE, and LTB+ The more specific biochemical mechanisms underlying PLAP-induced tumor cytolysis are yet to be elucidated. However, we do know that PLA, regulation is important in the regulation of cellular homeostasis. For example, if mitochondrial respiration is damaged by PLA2 activation, cell death occurs [20,21]. Free fatty acids, such as arachidonic acid, the precursor of PGEz and LTB4, also inhibits protein synthesis [22] and mitogenic responses [ 231. In addition to PLA, activation and eicosanoid

Thomas. L. (19.59) In: Cellular and Humoral Aspects of the Hypersensitivity State, pp. S29-532. Editor: H.S Lawrence, Cnssel Publishing. London. Leung. K.H. (1989) Inhibition of human NK-cell and LAKcell cytotoxicity and differentiation by PGE?. Cell Immunol.. 12.7. 384-195. Baxevanis. C.N.. Reclos, G.J.. Gritzapis, A.D., Dedousis. G.V.Z., Missitzis, 1. and Pnpamichnil. M. (1993) Elevated prostaglandin E2 production by monocytes is responsible for the depressed levels of natural killer and lymphokineactivated killer cell function in patients with breast cancer. Cancer, 72.49 I -SO I, Clark, M.A., Littlejohn, D., Conway. T.M.. Mong. S.. Steiner, S. and Crooke, S.T. (1986) Lcukotriene Da treatment of bovine aortic endothelinl cells and murine smooth muscle cells in culture results in an increase in phospholipaae A2 activity. 1. Biol. Chem. 261. 10713-1071X. Clark, M.A., Chen. M.J.. Crooke, S.T. and Born&ski, J.S. (1988) Tumor necrosis factor (cnchectin) induces phosphol-

6

[6] [7]

[8]

[9] [lo]

[ll]

[12]

[13]

[14]

[15] [16]

[17]

D.H. Goddard et al. /Cancer Letters 102 (1996) IL.5

ipase A2 activity and synthesis of a phospholipase A2activating protein in endothelial cells. Biochem. J., 250, 125-132. Aderem, A.A., Scott, W.A. and Cohn, Z.A. (1986) Evidence for sequential signals in the induction of the arachidonic acid cascadein macrophages.J. Exp. Med., 163, 139-154. Bunch, R.M., Luini, A. and Axelrod, J. (1986) Phospholipase A2 and phospholipase C are activated by distinct GTPbinding proteins in responseto al-adrenergic stimulation in FRTLS thyroid cells. Proc. Natl. Acad. Sci. USA, 83, 72017205. Burch, R.M. and Axelrod, J. (1987) Dissociation of Bradykinin-induced prostaglandin formation from phosphatidylinositol turnover in Swiss 3T3 tibroblasts: evidence for G-protein regulation of phospholipase AZ. Proc. Natl. Acad. Sci. USA, 84.63746378. Dempsey, C.E. (1990) The actions of melittin on membranes. Biochim. Biophys. Acta, 1031, 143-161. Clark, M.A., Gzgur, L.E., Conway, T.M., Disposito, J., Crooke, S.T. and Bomalaski, J.S. (1991) Cloning of a phospholipase A2 activating protein. Proc. Natl. Acad. Sci. USA, 88,5418-5422. Bomalaski, J.S., Fallon, M., Turner, R.A., Crooke, S.T., Meunier, P.C. and Clark, M.A. (1990) Identification and isolation of phospholipase A2 activating protein (PLAP) in rheumatoid arthritis: induction of eicosanoid synthesis and the induction of the inflammatory responsein joints injected with PLAP in viva. J. Lab. Clin. Med., 116, 814825. Steiner, M.R., Bomalaski, J.S. and Clark, M.A. (1993) Responsesof purified phospholipases A2 to phospholipase A2 activating protein (PLAP) and melittin. Biochim. Biophys. Acta, 1166, 124-130. Bomalaski, J.S., Baker, D.G., Brophy, L., Resurrection, N.V., Spilberg, I., Munian, M. and Clark, M.A. (1989) A phospholipase A2 activating protein (PLAP) stimulates human neutrophil aggregation and release of lysosomal enzymes, superoxide and eicosanoids.J. Immunol., 142, 39573962. Bomalaski, J.S., Ford, T., Hudson, A.T. and Clark, M.A. (1995) Phospholipase AZ-activating protein induces the synthesis of IL-l and TNF in human monocytes. J. Immunol., 154,4027-403 1. Komblith, P.L. and Walker, M. (1988) Chemotherapy for malignant gliomas. J. Neurosurg., 68, 1-17. Braganca, B.M. and Hospattankar, A.V. (1978) Potentiating action of cobra venom cytotoxin on the antitumour effects of an alkylating agent (melphalan). Eur. J. Cancer, 14, 707712. Walker, M.D., Green, S.B., Byar, D.P. et. al. (1980) Randomized comparisons of radiotherapy and nitrosoureas for

the treatment of malignant glioma after surgery. N. Engl. J. Med., 303, 1323-1329. [18] DeBrito, F.B., Moore, A.R., Carry, D.G. and Willoughby, D.A. (1987) A time course study of the changesthat occur to the subcutaneousmodel of synovium following polyarthritogen. Br. J. Exp. Pathol., 68, 559-567. 1191Goddard, D., Bruno, G., Baker, D. and Carsons, S. (1994) Cytokine and extracellular matrix expression in the maturing rat air pouch. Arthritis Rheum., 37, 191. [20] Higuchi, M., Shirotani, K., Higashi, N., Toyoshima, S. and Osawa,T. (1992) Damage to mitochondrial respiration chain is related to phospholipase A2 activation caused by tumor necrosisfactor. J. Immunother., 12, 4149. [21] Hugosson, M., Andreu, D., Boman, H.G. and Glaser, E. (1994) Antibacterial peptides and mitochondrial presequences affect mitochondrial coupling, respiration and protein import. Eur. J. Biochem., 223, 1027-1033. [22] Rotman, E.I., Brostrom, M.A. and Brostrom, CO. (1992) Inhibition of protein synthesis in intact mammalian cells by arachidonic acid. J. Biochem. (Tokyo), 282.487495. [23] Glasgow, W.C., Afshari, C.A., Barrett, J.C. and Eling. T.E. (1992) Modulation of the epidermal growth factors mitogenie response by metabolites of linoleic and arachidonic acid in Syrian hamster embryo Bbroblasts: differential effects in tumor suppressor gene (+) and (-) phenotypes. J. Biol. Chem., 267.10771-10779. [24] Aste-Amezaga, M., D’Andrea, A., Kubin, M. and Trinchieri, G. (1994) Cooperation of natural killer cell stimulatory factor/interleukin-12 with other stimuli in the induction of cytokines and cytotoxic cell-associated molecules in human T and NK-cells. Cell. Immunol., 156, 480-492. [25] Hill, L.L., Perussia, B., McCue, P.A. and Korngold, R. (1994) Effect of human natural killer cells on the metastatic growth of human melanoma xenografts in mice with severe combined immunodeficiency. Cancer Res., 54,736-770. [26] Trinchieri, G. (1994) Recognition of major histocompatibility complex class I antigens by natural killer cells (Review). J. Exp. Med., 180.417-421. [27] Pastan, I., Chaudhary, V. and FitzGerald, D.J. (1992) Recombinant toxins as novel therapeutic agents. Annu. Rev. Biochem., 61.331-354. [28] Cifone, M.G., Botti, D., Festuccia, C., Napolitano, T., DelGrosso, E., Cavallo, G., Chessa, M.A. and Santoni, A. (1993) Involvement of phospholipase A2 activation and arachidonic acid metabolism in the cytotoxic functions of rat NK-cells. Cell Immunol., 148, 247-258. [29] Vaillier, D., Daculsi, R., Gualde, N. and Bezian, J.H. (1992) Effect of LTB4 on the inhibition of natural cytotoxic activity by PGE2.Cell Immunol., 139, 248-258.