Evidence for the presence of an antitumor factor in serum of normal animals

Cancer Letters, 6 (1979) 2S5--240

235

© Elsevier/North-HollandScientificPublishersLtd.

EVIDENCE FOR THE PRESENCE OF AN ANTITUMOR FACTOR IN SERUM OF NORMAL ANIMALS

SAUL GREEN*, MARY ANN CHIASSON**and RAMESH G+SHAH Memorial Sloan-Kettering Cancer Center, 1275 York Avenue. N e w Fork, New York 10021

(U.S.A.) (Received 19 Sevtember 1978) (Accepted ~.0 November 1978)

SUMMARY We have previously reported finding a factor with antitumor activity (TNF, tumor necrosis factor) in extracts of serum from normal mice. The possibility that TNF exists in the blood of normal animals of other species was explored. Horse, mouse, dog, human, sheep, calf, rat and sha~k serums were fraetionated with (NH4)2SO+ and filtered through S-200 sephacryl gel. Proteins of molecular weight 90,000 to 180,000 were pooled, contend,rated and dialyzed. TNF, determined by L-cell assay in vitro and Meth tt assay in vivo found in fractions from mouse, dog and human serum. Agarose electrophoresis of the TNF from mouse and human se~trn indicated the principle components were ~ ~2 globulins. Preparative PAGE indicated that mouse TNF migrated slowly and was made up of at least 4 components while human TNF was a faster moving, monomeric protein.

INTRODUCTION TNF is the name given to a substance found in the serum of Corynebacterium parvum (CP)-inoculated mice, following the injection of bacterial endotoxin [Green, S. and Dubrj~msky, A., unpublished], tt has been purified about 50-fold. It is an a2 globulin and has a molecular weight of about 150,000 [2, Green, S. and Dobrjansky, A., unpublished]. Methods for the detection of this factor include an in vitro assay using mouse L-cells [2, Green, S. and Dobrjansky, A., *Addres~ c~rrespondenee to" Dr. Saul Green, Memorial Sloan-KetteringCancer Center, 1275 York Avenue,~Iew York, N.Y. 10021, U.S.A. **,A portioPnof this work was submitted by MAC in partial fulfillment of the requirements for the degree of Master of Sciencein the Department of Biology, Graduate School of Arts and Sciences, New York University,N.Y., U.S.A. Abbreviations: CP, corynebactenum parvum; PBS, phosphate-buffered ~line; TNP, tumor necrosis factor.

236 unpublished and an in vivo bioassay using female BALB/c mice bearing a methylcholanthrene-induced fibrosarcoma (Moth A [2, Green, S. and D o b ~ m s k y , A., unpublished] ). TNF-like activity has been found in extracts of liver microsomes from CPtreated mice [3] and more recently, in very low, but measurable amounts in microsomes from livers o f normal mice [9]. Recently, TNF has been found in the serum of normal mice [8]. The possibility existed, therefore, that TNF was present in the blood of animals of other species. This communication pre~ents the results of studies exploring this possibility. MATERIALS AND METHODS Mice Young adult female mice used were: outbred CD®-I Swiss and inbred BALB/c Charles River Breedb~g Laboratories, North Wilmington, Massachusetts. Tumor The 3-methylcholanthren=.,.induced sarcoma (Moth A) o f BALB/e origin was passaged in the ascites form in syngeneic mice. Sterilization procedure Glassware and solutions were autoclaved at 120°C for 20 rain. Plastics were gas.sterilized ~ t n ethylene-oxide for 3 h at 120°C. TNF assay The in vivo assay was carried o u t as has been described [Green, S. and Dobrjansky, A., unpublished]. The in vitro assay with L-cells (NCTC 929 [2] ) was modified as follows: L-cells were grovm as monolayers in a culture medium consisting o f McCoy's 5A, non-essential amino acids, L.glutamine, vitamins, 10% heat-inactivated fetal calf serum, penicillin (II)0 units/ml) and streptomycin ( I 0 0 mg/ml). A mixture of I ml of the culture medium containing 105 L-cells and 0.1 mi of the serially diluted test material was incubated in petri dishes (LUX) at 37°C for 72 h fa humidified 8% CO: in air, Each dilution was set up in triplicate. The TNF titer was deflated as the number o f milligrams of protein in the TNF fraction which resulted in a 50% inhibition in L-cell growth, compared with controls grown under the above conditions. Serum

sources

Animals w e r e : hybrid sheep and mongrel dogs from the Sloan-Kettering ~ulimals Facility, Woodside, L.L, New York; CD-1 female mice and CD female :rats were from the Charles River Breeding Laboratories, North WiLmington, Massachusetts. Horse serum (INo. 14-403, lot 8753) was from Microbiological Associates, Bethesda, Maryland; fetal calf sertml was from the Flow Laboratories; Rockvil~e, Maryland; nurse shark serum was a giff; from Dr. L.W. Clem of

237 the University of Florida, Gainesville, Florida; human serum and plasma w e r e from Gibco, No. A476016, Grand Island, New York and The New York Blood Center, 310 E. 67th St., New York, New York, respectively.

Chemicals and reogents Sephacryl S-200 (superfine) was from Pharmacia, Upsala, Sweden; McCoy's 5A, with added penicillin and streptomycin was from the Sloan-Kettering Institute media Laboratory. Other assays Protein was determined by the method of Lowry et al. [5]; agarose electrophoresis was as described in the Coming Instruction Manual [6]. PreparaL~ion o f the G.200 H fraction from all s,erum. All procedures were carried out at 4~C. One hundred milliliters of serum was sterilized by filtration through a sterile 0.22/~zn millipore filter and was diluted with 100 mI of sterSe phosphate-buffered saline (PBS). Forty nine grams of solid (NH4)2SO4 was added slowing with stirring. The pH was maintained at 7.0 by addition of NH4OH. The precipitated protein was removed by centrifugation at 195,000 X g for 15 rain and 49 g o£ (NH4)2SO4 was added to the supernatant solution with stirring. The insoluble material [35--70% saturation with respect to (NH4):SO4] was collected by centrJfugation, taken up in a small volume of sterile water, dialyzed agair~ststerile PBS containing penicillin and streptomycin until free of ~NH4):SO4 and diluted with sterile PBS to a final protein of 500-700 mg/ml. SEPHACRYL-200 (SUPERFINE)

t :; ~t

T

x

I'

E.

TkF.Rkdl~ '[ ~vlno Ser~lrr,Glbucnln

.J

c.l ~, SOybe~ Trypsll MouGe$orurn "~ It~hibitc~

GP-E.dv

2 -

Bluo O4xfmrl

"C~ Cylo:nomo

I

I ~',1 I00

300

200

ML

Fig. 1. Typical clution pattern of proteins in a 35--70% fraction from serum of CP-endotoxin-treated mice after filtration through a S-200 Sephacry~colun~ ( ). Calibration oI~the column with pu:rcproteins of known molecular weight (. . . . . ). Shaded region in the area containing TNF activity.

238 Gel filtration of the 35--7(}% fraction was carried out on a coltmm of Sephacryl S-200 (superfine), previously equilibrated against sterile PBS. The bed height was 85 cm, the eluant was PBS and the flow rate was 15 nd]h. The gel column was first calibrated for determination of molecular weights with dextran blue (void volume), purified a[dolase (I60,000), bovine serum albumin (67,000), soybean trypsin inhibitor (21,600), and cytochrome c [12,500 (Fig. 1)]. Tenmilliliter aliquots of the ~ialyzed 35--70% fraction from each test serum were filtered through the column ~md the protein which was eluted between 195 and 236 n~ of PBS was pooled, concentrated by lyophilization and dialyzed against 0.15 M NaCI until free of phosphate. The concentrated material, called G-200 II, was sterilized by filtration through a sterile Nalgen~ 0.22 #m filter. As previously reported [L,8], when serum from CP-primed, endotoxin-treated mice (TNF-rich serum) w~s used, all the TNF activity w~.s present in this fraction (Fig. 1). With serum from normal anin,als as the starting material, the recovered G.200 II fraction was further concentrated by 20 h centrifugation at 195,000 x g [ 2 ] . RESULTS AND DISCUSSION A G-200 II fraction was p]~epared from 100-ml atiquots of sterile serum from various species (see Methods) and concentrated by ultracent~i/ugation. TNF-Hke antivity was determined by L-cell assay and bioassay with Meth A tumors in I~,ALB/c mice (Table 1). Witt~ the L-cell assay, horse, mouse, dog and human TABLE I TNF-LIKR ACTIVITYIN Go200 II FRACTIONS FROM SERUMSOF NORMAL ANIMALS* Species

Horse Mouse Dog Human Sheep Calf (fetal) Rat Shark

TNF-assay In vitro

In rive

mg Protein giving 50% growth inhibition

wg Protein mouse

(-)--( ÷ + + ) necrosis

0.21 0.94 1.20 1.35 8.40 Less than 20% gro~vth inhibition at 10 mg Less than 20% growth inhibition at 1.0mg Less than 20% growth inhibition at 1.0 mg

74 67 68 80 nd** nd

212" J,~* ½"** 3/3*" ~h*+*~/~** nd nd

99

2/2-

nd

nd

*See Materials ~nd Methods. **nd, not done.

239 serum bad significant antitumor activity while sheep, calf, rat and shark did not inhibit L-cell growth. In the bioassay mouse, dog, and human had good TNF activity at about the same concentration of protein/mouse. Horse had no measurable tumor necrosis inducing activity. We have previou,,ily found that up to 100 pg of our preparation of E. coil ~.ndotoxin (see Materials) was not toxic for L-cells in culture but that about 50 #g of this endotoxin given i.v. induced necrosis in Meth A tumors in BALB/c mice [9]. Some information can be. deduced from these observations: First, that a positive effec,t in both the L-cell assay and the Meth A assay indicates the presence of TNF-like activity rather than endotoxin-like activity. Second, that a positive effect in the L-cell assay and a negative effect in the Meth A assay indicates the possibility of the expression of general growth inhibitory, not necessarily antitumor activity. Third, that antitumor activity shown in the Meth A system in vivo with no corresponding anti-L-cell activity is suggestive of the presence of endotoxin contamination, probably incurred during isolation and concentration procedures. If these deductions are correct, the results in Table 1 ,demonstrate the presence of TNF-like activity in the serum of normal animals of at least 3 different species. AGAROSE CP-ENO0 MOUSE

7" ~. ,',= a~ A~.B POLYACRYLAM IDE

A

~

NORMAL MOUSE

)"

,8 a= a I ALB

A

~ ~ZI i

NORMAL HUMAN

Y

~I !

B ~

,8 az ", At~B

a

• '

:,:,xl

B ~, I I

C

(+)

........

C ~:::~;:~:~I ~~~:I

C

I

Fig. 2. Agarose and polyaeryhmide electrophoresis of 13-200II fraction from CP-endotoxizltreated and normal mouse serum and froiL serum of normal humans.

240

Electrophoretic [i] pi.~ofi]esof G-200 Ii fractions from serum of CP-endotoxin-t.reatedmice, norm.~Llmice and normal humans are compared in Fig. 2. A major l~,~akwhich migrates in the region of the a i-a2 globulins is clearlydemonstrated ~n all 3 samples (Fig. 2A). Analysis ( ) of these 3 samples on prep~ative ,?AGE [7,2] revealsclose similaritybetween the proteins in the G-200 II ,fractionIrom the CP-endotoxin-treated and normal mice. Preparative P A G E electrophoresis of each of the 3 G-200 iIpreparations was carried out. Unstained gelswere s~icedinto 3 segments (most cathodic-A, B, and most anodic-C) pooled and the proteins in each group were extracted, concentrated and tested for TNF.. All the TNF-E~ke activitywss found in the extract of the B segments of the acrylamide gels on wh;.ch mouse G-200 II had been run. In contrast,TNF.like activity was found in extracts from the C segments of gels on which human G-200 iI was run. Experiments to purify T N F from G-200 II from normal h u m a n serum ~re underway along with a determination of the specificityof h u m a n T N F ~or human tumors carried in nu/nu mice or growing in tissue culture. ACKNOWLEDGIEMENTS This researc~lL was s u p p o r t e d in part b y Public Health Service grant C A 0 8 7 4 8 f r o m t h e Nati~nal Cancer Ir~stitu~e, by grant BC225, f r o m t h e American Cancer S o c i e t y and b y f u n d s f r o m the Martin Elk League for Cancer Research and t h e Cele Butwin F o u n d a t i o n , Inc. REFERENCES 1 Coming Agaros,~ E]eetro~'horesis Systems: Operations Manual No, 470166. Coming 490, San Anton~io ltd., Palo ~Mto, Calif. 2 Earle, W.R. (19~tB)Production of malignancy in vitro. IV. The mouse fibroblast cultures and changes seen in living eclls. J. Nail. Cancer. Inst., 4, 165--212. 3 Green, S., Dobrjm:~ky, A., Carswell, E.A., Kasscl, R.L., Old, L.J., Fiore, N. and Schwartz, M.K. (1976) Partial .~urlfie~lLionof a serum factor that causes necrosis of tumors. Proc. Natl. Aead. Sei. U;!;A,,73, 3~1--385. 4 Green~S., Dobrjan!;ky, A., Chi~sson, M A., Carsweli, E.A., Schwartz, M.K, and Old, L.J. (19'77) Corynebas~erium par~nm as the priming agent in the production of tumor necrosis factor in ~he mouse. J. Na~l. Cancer Inst., 59, 1519--1522. 5 Green, S., Dobrjansky, A., C~iasson, M.A., CaP#well, E.A., Helson, L., Schwartz, l~.K. and Old, L.J. (1976) t;ecrosis of Meth A Tumors by a factor from liver of C. parvumendotoxin-treated n~ice. Proe. Am. Assoc. Cancer Res., 17, 84. 6 Lowry, O.H., Roseb~'ough, N.J., Parr, A.L. and Randall, R.J. (1951) Protein measurement with the Folin I~ht.molreagent. J. B~iol.Chem., 193, 265--275. 7 Ornstein, L. (1964) ~'#is~'..El,~ctrophoresi~. Ann. N.Y. Acad. Sei., 121,321~323. 8 Shah, R.G., Green, S. ar,~dMoore, M.A.S. (1978) Colony stimulating and inhibiting activities in mouse sert~m after C. parvum-endotoxin treatment. J., Reticuloendothelial Soc., 23, 29--41.