(CANCER RESEARCH 46, 2208-2213, May 1986]

Autocrine Tumor Growth Regulation and Tumor-associated Hypoglycemia inMurine Melanoma B16 in Vivo1

S. Vuk-Pavlovic, E. C. Opara,2 S. Levanat, D. Vrbanec, and K. Pavelic

Departments of Pharmacology fS. V-PJ and Endocrine Research Unit [E. C. OJ, Mayo Foundation, Rochester, Minnesota 55905; Rugjer Boskovic Institute [S. L, D.V.], Bijenicka 54, 41000 Zagreb, Croatia, Yugoslavia; and Grace Drug Cancer Center [K. P.], Roswell Park Memorial Institute, Buffalo, New York 14628

ABSTRACT

A substance immunochemically cross-reactive with insulin (SICRI)appears in melanoma B16 growing in diabetic and nondiabetic CS7BL/6mice. Progression of tumor size is paralleled by the increase of SICRIlevels in the serum of both diabetic and nondiabetic animals; this increasecorrelates with a decreased concentration of circulating glucose and anelevated concentration of growth hormone in blood. Melanoma B16 grownunder serum-free culture conditions secretes SICRI into the medium.Affinity-purified SICRI stimulates glucose uptake by rat epididymaladipocytes and competes with radiolabeled insulin for binding to thesecells. Low concentrations of SICRI enhance growth of cultured melanomaB16 cells, whereas high concentrations of this substance have inhibitorygrowth effects on these cells. Porcine insulin, human insulin-like growthfactors I and II, human growth hormone, platelet-derived growth factor,epidermal growth factor, and fibroblast growth factor have negligibleinfluence on growth of melanoma B16.

INTRODUCTION

Some tumors have been found to produce growth factorswhereby they stimulate their own growth and maintain transformed pheno type (1). Recently, we described production andsecretion of substances which cross-react with insulin-specificantibodies in several human tumors. The high frequency ofassociation of SICRI3 with these diseases in humans (2-6)

prompted us to look for a suitable experimental model whichwould enable us to study the production, activity, and structureof SICRI in vivo. In many respects, murine melanoma B16exhibits SICRI-associated properties analogous to those foundin human tumors (2-6) and therefore was chosen for furtherstudy.

Growth of melanoma B16 in normal and alloxan-diabeticC57BL/6 mice is accompanied by (a) an increase in levels ofSICRI in blood, (¿>)progressive hypoglycemia in diabetic andnondiabetic animals, and (<•)the increase of circulating growth

hormone concentrations. Correlations of tumor mass, SICRIconcentrations, and glucose in blood as indicators of tumormetabolism in a phenomenological mathematical model enabled us to demonstrate an operative autocrine system in vivoand to quantitate the extent to which B16 stimulated its owngrowth (7). We speculated that SICRI, growth hormone and,possibly, somatomedins were probable candidates for beinghumoral mediators in an autocrine/endocrine growth self-incitement mechanism (4, 7).

Received 6/17/85; revised 1/2/86; accepted 1/17/86.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 inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1This research was partially supported by National Cancer Institute Programgrant CA-13038 and a Biomedicai Research Support Group grant from RoswellPark Memorial Institute.

2 WHO fellow.3The abbreviations used are: SICRI, substance immunochemically cross-re

active with insulin: EOF, epidermal growth factor, PCS, fetal calf serum; FGF,fibroblast growth factor; IGF I and II, insulin-like growth factors I and II,respectively; PDGF, platelet-derived growth factor; RIA, radioimmunoassay;HPLC, high-performance liquid chromatography; POE, polyoxyethylene 4000;PBS, phosphate-buffered saline (20 mM potassium phosphate-0.15 M NaCl, pH7.0).

Here, we report the dynamics of SICRI production and itsresultant effects in vivo. Moreover, by in vitro experiments, weprovide indirect evidence for mechanistic associations of SICRIconcentrations, blood glucose levels, and tumor volume in vivo.Our results show that melanoma B16 could be a suitable modelto study autocrine self-regulation in neoplasms and possibly oftumor-associated hypoglycemia and its hypoglycemia-relatedsecondary effects.

MATERIALS AND METHODS

Animals and Tumor. Two and one-half-mo-old male CS7BL/6 mice,weighing 22-28 g each, were kept five per cage and fed standard pelletedfood and water ad libitum. They were obtained either from RugjerBoskovic Institute, Zagreb, Yugoslavia, or from The Jackson Laboratory, Bar Harbor, ME.

Melanoma B16, originally obtained from Holt Radium Institute,Manchester, England, has been maintained at the Rugjer BoskovicInstitute since 1975 by s.c. inoculations of suspensions containing 2 xIO6tumor cells into flanks of C57BL/6 mice. Alternatively, melanoma

B16 obtained from the Division of Cancer Treatment Tumor Repository, Frederick Cancer Research Center, Frederick, MD, was used.Three diameters (A,B, and Q of tumor prolate ellipsoids were measuredand the tumor volume was calculated as (/lx/fxc)ir/6.

Diabetes was induced by a single i.v. injection of alloxan, 100 or 120mg/kg (Merck, Darmstadt, Federal Republic of Germany or Sigma, St.Louis, MO). Diabetes induction was confirmed by the apparent disappearance of circulating insulin (measured by radioimmunoassay) andelevation of glucose concentrations in blood [from 5.5 ±1.5 (SD)-31.0±5.0 HIM].

Cultivation of Melanoma B16. Melanoma B16-BL6 cells were maintained in a monolayer in RPMI 1640 medium (Gibco, Grand Island,NY) supplemented with 10% FBS (Gibco) and 20 mM 4-(2-hydroxy-ethyl)-l-piperazineethanesulfonic acid. Cells plated at the density of 1x IO4cells/ml of medium were grown in wells of 24-well plates (Fisher,

Rochester, NY) or plastic Petri dishes (35 x 10 mm; Fisher).In some cases, soft agar was added to the medium in order to establish

cell colonies (8, 9). Warm 0.5% agar, 2.5 ml (Gibco) in Eagle's minimum essential medium (Gibco) containing 10% FCS or Nu-Sérum(Collaborative Research, Lexington, MA) were placed into the 35- x10 mm Petri dishes and solidified at room temperature. Cells weresuspended in 0.85 ml of Eagle's minimal essential medium with 10%

FCS (plating layer) containing 0.3% agar. Triplicate cultures wereincubated at 37°Cfor 5 days in water-saturated air supplemented with

5% CO2.Cells were inspected and colonies were counted with xlOO or x200

using a Nikon inverted phase contrast microscope. Clusters of 30 ormore cells were defined as colonies. Cultures were established also froma B16 melanoma grown s.c. in a female C57BL/6 mouse. A cellsuspension was prepared by passing tumor tissue through a nylon meshand three times through a 21-gauge needle and cells were grown inserum-free transferrin-supplemented (5 mg/liter; Sigma) Iscove's modified Dulbecco's medium (Sigma). The medium was changed every

second day. The first medium change was discarded and the subsequentones were used as the source of SICRI.

Stimulation of DNA synthesis by growth factors was measured incells grown in Costar (Cambridge, MA) 96-well clusters (75,000 cells/well) by measuring [/weiA>>/-3H]thymidine(specific activity, 2 Ci/mmol;

1 mCi/ml; Research Products International, Mt. Prospect, IL) incorporation. After 24-h incubation of the cells with various growth factors,

2208

on April 28, 2021. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

AUTOCRINE TUMOR GROWTH AND TUMOR-ASSOCIATED HYPOGLYCEMIA

[3H]thymidine, 0.2 ¿iCi/2n\ was added to each well. Cells were harvested

after 6 h using a Bélicocell harvester. Radioactivity of incorporatedthymidine was measured by use of 3a70 scintillation cocktail (ResearchProducts International, Mount Prospect, IL) and Packard Tri-Carbliquid scintillation spectrometer. Protein concentrations were measuredby the Lowry method (10).

Growth Factors. Growth factors were obtained commercially: porcineinsulin from Novo (Copenhagen, Denmark) or Eli Lilly (Indianapolis,IN); IGF I from AMgen (Thousand Oaks, CA); IGF II, PDGF, andEGF from Collaborative Research (Lexington, MA); and FGF fromSeragen (Boston, MA). Before use, these substances were dissolved inserum-free RPMI 1640 medium and pressure sterilized through 0.22-

tim pore filters (Millipore, Bedford, MA).Measurement of Glucose, SICRI, and Growth Hormone Concentra

tions. Glucose levels were measured either by the use of commercialGlucose oxidase-Peroxidase kits (Biodynamics, Indianapolis, IN) or bythe orthotoluidine method (11). The levels of SICRI were measured bycommercial insulin-specific radioimmunoassay kits (Phadebas, Uppsala, Sweden and Boris Kidric Institute, Vinca, Yugoslavia) or by kitsassembled from commercial antibodies (Cappel, Mähern.PA) and A,4-monoiodotyrosine insulin (360 //Ci/mg; Eli Lilly, Indianapolis, IN).SICRI concentrations are measured by insulin-specific RIA and therefore are relative, expressed as insulin equivalents. In order to excludeRIA artifacts due to degradation of radiolabeled tracer, fixed amountsof '2SI-labeledinsulin were incubated with conditioned media or purifiedSICRI (10-500 Munits/ml) for 12 h at room temperature (RIA conditions). At all SICRI concentrations, more than 90% of radioactivitywas always precipitated by 8.5% trichloroacetic acid. Moreover, afterincubation, I25l-labeled insulin waseluted from the 1-125 HPLC columnat the same time as the fresh I25l-labeled insulin. These experimentsshow that I25l-labeled insulin is not degraded in the course of the RIA

procedure. The failure of SICRI preparations to increase the amountof I25l-labeled insulin precipitated by 14% POE (after 12 h incubation)demonstrated the absence of insulin-binding materials in these samples.Moreover, RIA performed with 18% POE and human IgG (1 mg/ml)instead of the second antibody gave results consistent with those obtained with the standard double antibody procedure. Growth hormoneconcentrations were determined by commercial RIA kits (Serono, Sai-

uggia, Italy).Purification of SICRI. Tumor tissue from diabetic animals was ho

mogenized in nine parts of PBS. The homogenate was subsequentlycentrifuged (20,000 x g for 30 min at 4°C).The supernatant was

fractionated by 2% (w/v) increments of polyoxyethylene 4000 (FisherScientific, Pittsburgh, PA) between 0 and 24%. After each addition ofPOE, samples were centrifuged, and pellets were dissolved in PBS andassayed by insulin-specific RIA. Protein concentrations in these solutions were measured either by the method of Lowry et al. (IO) or by theBio-Rad (Richmond, CA) protein assay.

The redissolved POE precipitate (between 10 and 14% POE) orserum-free tissue culture medium conditioned by tumor cells and con

centrated by use of Amicon UM2 membrane, was passed through acolumn of purified guinea pig anti-porcine insulin antibodies immobilized on Affi-Gel 10 (Bio-Rad) according to the manufacturer's instruc

tions. The column was extensively washed with PBS and the boundmaterial was eluted with 1.0 M acetic acid. The eluate was neutralizedimmediately with 2 M sodium hydroxide and dialyzed against PBS inSpectrophor dialysis bags (molecular weight cutoff = 1000; SpectrumMedical Industries, Los Angeles, CA).

HPLC of dissolved 10-14% POE precipitate or affinity-purifiedSICRI in PBS or l M acetic acid was performed at the flow rate of 1ml/min on a 7.5- x 250-mm protein analysis 1-125 column (molecularweight cutoff = 80,000; Waters Associates, Milford, MA) calibratedwith the Bio-Rad gel filtration standard. The fractions in PBS wereassayed by RIA directly and those in l M acetic acid only after removalof acid by freeze-drying and reconstitution with the RIA buffer.

The apparent relative molecular mass of native SICRI was estimatedby gel filtration both of unpurified tumor extract and of the redissolved10-14% POE precipitate on a 0.9- x 50-cm Sepharose 6B column. Asthe mobile phase, PBS containing bovine serum albumin (0.3 mg/ml)at a flow rate of 10 ml-cm"2-h"1 was used. Blue dextran, human IgG,

human serum albumin, bovine I2sl-labeled insulin, and potassium fer-

ricyanide were used for column calibration.Glucose Uptake and Binding of SICRI and Insulin by Rat Epididymal

Adipocytes. Adipocytes were isolated from epididymal adipose tissueof male Sprague-Dawley rats (Holtzmann, Madison, WI) weighing120-180 g each and fed ad libitum. The cell suspension was preparedby digestion with Clostridium histolyticum collagenase, 0.5 mg/ml(Millipore, Freehold, NJ) in Krebs-Ringer bicarbonate buffer containing bovine serum albumin, 20 mg/ml (Armour, Kankakee, IL) for 45min at 37°C(12). The resulting cell suspension was passed through

lOOO-jitnand 250-;um nylon meshes, centrifuged at 400 x g for 1 minand washed three times in the albumin-supplemented Krebs-Ringer-bicarbonate buffer. Finally, adipocytes were washed and resuspendedin the assay buffer consisting of 10 m\i 4-(2-hydroxyethyl)-l-pipera-zineethanesulfonic acid-135 mM NaCl-4.8 mM KC1-2.5 HIMCaCl2-1.7HIM MgSO.t-1.9 mM NaH2PO4-bovine plasma albumin, 20 mg/ml-bacitracin, 0.5 mg/ml (Sigma, St. Louis, MO)-3 mM glucose, pH 7.4,and incubated at 37°Cin 95% O2/5% CO2 gas mixture. A small portion

of cells was fixed in 2% OsO< in collidine HCl-bufter and counted byuse of a Coulter counter, Model ZB (Coulter Electronics, Hialeah, FL)(13). Routinely, 3-4 x IO5 living cells/ml were used, both for glucose

uptake and insulin/SICRI binding experiments.Glucose uptake by adipocytes was measured by the disappearance of

glucose from the incubation medium (14). Cells were incubated in 16-x 100-mm Nunc (Roskilde, Demmark) polyethylene test tubes for 2 hat 37°Cin the presence or absence of SICRI or insulin. After centrif-

ugation, 20 /¡Iof the assay medium from each tube was mixed with 2ml of the Glucose oxidase-Peroxidase reagent and the absorbance wasmeasured at 505 nm following 10 min incubation at 37'C. The amount

of glucose taken up by adipocytes was calculated as the difference ofglucose concentrations in assay mixtures and in cell-free assay buffer.

SICRI and insulin binding to adipocyte receptors was measured bycompetition of SICRI and insulin with 125I-labeledinsulin (15). Duplicate samples of adipocytes were incubated in 0.5-ml aliquots of assaybuffer containing I25l-labeled insulin, 0.3 ng/ml and various concentra

tions of unlabeled insulin or SICRI. Nonspecific binding was definedas I25l-labeled bound to adipocytes in the presence of unlabeled insulin,100 ng/ml. Incubations were carried out with gentle shaking at 20°C

for 1 h. Duplicate 0.2-ml aliquots of each assay mixture were centrifuged through 0.2 ml of dinonylphthalate (BDH, Dorset, England) ina Beckman Microfuge (Palo Alto, CA). The lower aqueous phase withunbound insulin was drained through a hole at the bottom of the tubewhile radioactivity bound to cells located with upper phase was countedusing a Searle (Des Plaines, IL) gamma counter.

RESULTS

Growth of Melanoma B16 in Normal and Alloxan-DiabeticMice. We monitored tumor volume and SICRI, glucose, andgrowth hormone concentrations in the blood of melanoma Bio-bearing mice as a function of time after tumor inoculation. Theresults are shown in Fig. 1. The tumor volume increase (Fig.la) was accompanied by an elevation of SICRI concentrationin blood (Fig. lé),by depression of peripheral glucose levels(Fig. le), and by an increase of circulating growth hormoneconcentrations (Fig. Id). We correlated different data sets fromFig. 1 by a least squares fitting procedure. The relationshipbetween SICRI concentration in blood (S) and tumor volume(V) can be expressed by the empirical equation 5=19 exp(1.6x IO"4 V) with r = 0.74, the relationship between S and

circulating glucose concentration (G) by G = 5.3 exp(-1.0 X10~2S) with r = -0.81, and between G and the level of growth

hormone in blood (H)byH= 83 exp(-0.22 G) with r = -0.59(7). When we performed an experiment similar to that shownin Fig. 1 in alloxan-diabetic mice, we noted that tumor growthwas similarly accompanied by a decrease of glucose concentration in blood (Fig. 2), although the growth rate and transplant-ability were lower compared to normoinsulinemic hosts. No

2209

on April 28, 2021. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

AUTOCRINE TUMOR GROWTH AND TUMOR-ASSOCIATED HYPOGLYCEMIA

0 10 20 30Time After Tumor Inoculation/Days

Fig. 1. Melanoma B16 volume (a) and concentrations of circulating SICRI(/i). glucose (<•).and growth hormone (ti) in CS7BL/6 mice as a function of timeafter s.c. inoculation of 2 x 10* tumor cells into the flank. Points, mean ±SD(bars) groups of 10-12 animals. >/('. fiunits; ml., ml.

15 20 25Time After Tumor Inoculation/Days

Fig. 2. Melanoma B16 volume (a) and blood glucose concentrations (b) innondiabetic (O) and alloxan-diabetic animals (•)as a function of time after tumorinoculation. Points, mean ±SD (bars) of 10 animals.

reversal of diabetes symptoms was observed in sham-inoculatedalloxan-diabetic mice.

Purification of SICRI. Twenty-fold SICRI enrichment can beobtained by polyoxyethylene fractionation. Supernatants obtained by centrifugation of tumor homogenates (usually 25-day-old tumors were used) contained approximately 10 /¿unitsSICRI/1 mg of protein. By extrapolating SICRI concentrationsin these supernatants to actual tumor volumes, we calculated

that the highest SICRI concentration in the tumor was 4700iiunits/g. Approximately 60% of measurable SICRI in thehomogenate supernatant was precipitated between 10 and 14%POE and for further experiments we used the material precipitated between these two POE concentrations. In none of thetumor extracts was there any growth hormone detected.

In serum-free supernatants of 4- to S-day-old B16 cultures,SICRI up to 30 milliunits/1 could be measured. Specific activityof up to 4.2 milliunits/mg protein was measured in these mediaafter dialysis and concentration. SICRI, affinity purified fromthese supernatants, had a specific activity of 110 milliunits/mgprotein. It was concentrated by use of an Amicon UM-2 membrane to the concentration of 2 units/1 and used for tissueculture experiments and chromatography. Thus, the affinity-purification step yielded 26-fold purification. The HPLC resolution of POE-precipitated SICRI and porcine insulin on the I-

125 column are shown in Fig. 3a. Whereas native SICRI inPBS was eluted in the void volume of the column, in l M aceticacid it was coeluted with insulin. This change of elution positionwas due obviously to dissociation of larger SICRI assemblies.It is interesting that in PBS, native insulin was retarded by theI-125 column.

By use of a calibrated Sepharose 6B column, we found themolecular weight of POE-precipitated SICRI to be approximately 120,000 (Fig. 3, h and inset). The same molecular weightvalue was obtained by use of culture-derived SICRI and acalibrated Spherogel-TSK 3000 SW HPLC column.

Insulin and SICRI Effects on Rat Adipocytes. Both insulinand SICRI enhanced glucose uptake and they both bound torat epididymal adipocytes to a similar extent (Fig. 4). The effectof SICRI affinity purified from tumor tissue and from concentrated conditioned culture media was similar. The results ofthese bioassays were variable and depended on the metabolicstatus and age of the rats. However, qualitatively they were well

250

6 8 10 12 14 18 20 30 40 50 60Elution Time/min Elution Vol./fraction

Fig. 3. Molecular sieving chromatography of SICRI and insulin, a, HPLCelution profiles from the I-125 column (7.S x 2SO mm) of 100 ¡Aredissolvedmelanoma B16 extract, 4800 Mnnits/m I (24 mg protein/ml) [precipitate obtainedby precipitation between 10 and 14% POE (•)and of pure porcine insulin (O) inPBS ( ) and 1.0 M acetic acid ( )]. The flow rate was 1 ml/min. The columnwas calibrated with thyroglobulin (('„).ovalbumin (<>.-().myoglobin (Mb), and

vitamin B,2 (lin), b, chromatogram of melanoma B16 extract (200 *tl of 4800Munits/ml redissolved 10-14% POE precipitate) on a Sepharose 6B 0.9- x SO-cmcolumn developed with PBS containing bovine serum albumin, 0.3 mg/ml at aflow rate of 10 ml cm~2 h~'. The column was calibrated with blue dextran (('„,void volume), human IgG, human serum albumin (USA). I2sl-labeled bovineinsulin ('"/-/). and potassium ferricyanide ((',. total volume). Inset, plot of A.»[= (V, —V*)/(y, —K>),where V, stands for the elution volume] versus log apparentrelative molecular mass of human serum albumin, IgG, and SICRI derived fromdata in h.

2210

on April 28, 2021. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

AUTOCRINE TUMOR GROWTH AND TUMOR-ASSOCIATED HYPOGLYCEMIA

200 -

50 IOO 25O 50O 1000[SICRI, Insulin]/^U-mLM

Fig. 4. Glucose uptake by (a) and '"I-labeled insulin binding to (h) ratepididymal adipocytes as a function of concentration of insulin (O) and affinity-purified SICRI (•).iti', /limits: ml., ml.

reproducible.In Vitro Stimulation of Melanoma B16 Proliferation by

SICRI. In order to assess the effects of SICRI on melanomaBl6 cells, we measured in vitro cell proliferation (either as totalprotein accumulation or [meiA>»/-3H]thymidine*incorporation)

as a function of SICRI concentration under various conditions.SICRI effects on cell proliferation were determined in serum-

supplemented B16-BL6 cultures, but in serum-free cultures,SICRI almost fully compensated for the growth promotingactivities of serum, even at the low concentrations of SICRI(0.5 ftunits/ml). However, in the presence of SICRI, 0.5 /junits/ml the absolute proliferation rate was almost the same as therate in fully supplemented SICRI-free medium (data not

shown).The growth rates of the B16-BL6 cell line as a function of

SICRI and insulin concentrations were compared (Fig. 5). Theeffects of these two substances were strikingly different. WhileSICRI produced a dose-response effect typical for mitogens(16), porcine insulin did not exert an effect except for a slightinhibition of growth at the highest insulin concentrations, although this insulin is physiologically active in mice in vivo (17).From the mechanistic point of view, it is imperative to determine whether the growth promotion can be ascribed only toSICRI. Therefore, we compared the effects of affinity-purifiedSICRI on growth of B16-BL6 cells with analogous effects ofthe redissolved 10-14% POE precipitate. As can be seen in theinset of Fig. 5, within the limits of experimental accuracy, theresults are indistinguishable showing that most of the mitogenicactivity in these preparations is associated with SICRI.

SICK 1-stimulated [3H]thymidine incorporation was obtained

when concentrated and dialyzed (through a molecular weightcutoff = 1000 membrane) tissue culture medium conditionedby melanoma B16 was used (Fig. 6). Here strong stimulationof thymidine incorporation was followed by inhibition of incorporation at higher SICRI concentrations. SICRI affinity purified from these conditioned media exerted similar effects inB16 cell cultures (Fig. 6).

aW TI

I

l

Fig. 5. Total protein content [percentage of control ±SD (torvi] of melanomaB16-BL6 cells as a function of affinity-purified SICRI (•)and insulin (O) in theRPMI 1640 medium in water-saturated air with 5% CO2 at 37*C. Ten thousand

cells were suspended per plate and incubated with SICRI or insulin for 24 h. Theprotein content of the cells was measured after 3 more days of incubation. Totalprotein content of cells from each well was determined by the method of Lowryfi al. (10). Inset, total protein content of B16-BL6 cells as a function of concentration of SICRI added as redissolved 10-14% POE precipitate (•)or affinity-purified preparation (O). /if,', /limits; ml., ml.

i8

O I 2 4 8 16 32 64 128[SICRI, Insulin]/^U-mL'1

Fig. 6. |/«cr/ir/'1111hyinidinc incorporation by melanoma B16 cells in serum-free transferrin-supplemented Iscove's modified Dulbecco's medium as a function

of concentration of SICRI added either as concentrated and dialyzed conditionedmedium (•),or as affinity-purified preparation (O), or of insulin (A), Seventy-five thousand B16 cells were incubated in 0.2 ml (ml.) of SICRI or insulincontaining medium for 24 h when 0.2 /iCi ['I Ijiliymidinc was added. Cells were

incubated for 6 more h and harvested, /if, /¿units.

Effects of Growth Hormone, IGF I, IGF II, PDGF, EGF, andFGF on Melanoma B16 Proliferation in Vitro. If growth hormone and/or somatomedins enhance tumor growth, these substances could be indirect SICRI -dependent mediators of tumorgrowth autostimulation. Therefore, we examined melanomagrowth response (measured as appearance of colonies in soft

2211

on April 28, 2021. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

AUTOCRINE TUMOR GROWTH AND TUMOR-ASSOCIATED HYPOGLYCEMIA

agar) to growth hormone (0-50 ng/ml) to the somatomedins,insulin-like growth factor I (0-25 nig/nil), insulin-like growthfactor II (0-16 ng/ml), platelet-derived growth factor (0-16 ng/ml), epidermal growth factor (0-16 ng/ml), and fibroblastgrowth factor (0-10 ng/ml). The maximum effects of differentgrowth factors on melanoma cells grown in plastic under identical conditions did not differ significantly from control valuesshowing that among these growth factors only SICRI had aprominent effect.

DISCUSSION

We have recently demonstrated that SICRI concentrationsin blood of melanoma Bio-bearing mice correlate with thevolume of this tumor and that autostimulation contributes tomelanoma growth (7). In this report, we show that SICRI actsas the mediator of this response and that it may be capable ofaffecting glucose transport in the host.

SICRI production and secretion in diabetic mice with B16melanoma and especially its secretion into serum-free culture

media demonstrated that this entity was indeed produced bythe melanoma B16. This B16 tumor secretes SICRI into blood,and the circulatory SICRI level was closely correlated withtumor volume. This feature distinguishes SICRI from othertumor-produced growth factors which are predominantly contained ;>/.v/fH(18).

Molecular Size of SICRI. We have been using the term"substances immunochemically cross-reactive with insulin" for

tumor-associated insulinoids because at this time we do notknow their genetic and structural relationships to "inuminoreactive insulin," the preproinsulin gene products. However,

the molecular sieving data reported in this paper shed somelight on the gross size of SICRI allowing comparison withinsulin.

Under nondenaturing conditions, the relative mean molecular mass of SICRI is 120,000 as in non-Hodgkin's lymphoma

(5). Acid denaturation, however, reduces the molecular weightvalue of SICRI to that of monomeric insulin suggesting thatunder physiological conditions SICRI is an appregate. Theprecise relationship between tumor-produced SICRI and pancreatic insulin remains to be determined by other methods.However, functional differences between SICRI and insulinindicate that the two entities differ to a certain extent.

SICRI-induced Hypoglycemia and Elevated Growth-HormoneLevels in Blood. Tumor-associated elevation of circulatory SICRI concentrations was accompanied by marked depression ofglucose levels both in diabetic and nondiabetic animals. Although we did not have enough purified material to study theeffects of i.v.-administered SICRI on circulatory glucose levels,we believe that SICRI-stimuIated glucose uptake by rat adipo-cytes confirmed the relationship between elevated SICRI levelsand glucose concentration in blood. If this is true, it would beimportant to explore whether fluctuations in circulatory SICRIconcentrations (6) are related to hypoglycémieepisodes inmalignant diseases (19) and whether elevated levels of growthhormone in blood (4) are induced by fluctuations of glucoseconcentrations in circulation (20).

Mitogenic Effects of SICRI. The redissolved precipitate ofthe tumor extract obtained by POE fractionation (10-14%POE) and the affinity-purified SICRI isolated from this POEpreparation exerted similar mitogenic action on B16-BL6 cellsin culture. The same pertains to the action of serum-free me

dium conditioned by the Bl 6 cells and to SICRI affinity purifiedfrom this source. These different SICRI preparations exerted

maximum effects at somewhat different concentrations. Thiscan be due (a) to possibly different sensitivities (21) of twodifferent melanoma cells lines to SICRI, (b) to differing potencies of intracellular SICRI isolated from tumor tissue and ofsecreted SICRI, (c) to the presence of other growth-promotingsubstances in crude POE precipitate and (d) to combinationsof the first three factors.

Growth of melanoma B16 cells in culture was stimulatedonly by SICRI and was not increased by insulin, IGF I, IGF II,growth hormone, EGF, PDGF, and FGF. This exclusive susceptibility to stimulation by SICRI is surprising because it wasshown for other systems that the more responsive a cell line invitro is to one growth factor the more responsive it is to all ofthem (21).

Melanoma B16 as a Model of Autocrine Stimulation andTumor-associated Hypoglycemia. Most reports on autostimulation mechanisms in tumors deal with in vitro action of tumor-derived growth factors. Therefore, studies of tumor autostimulation //; vivo are of particular interest. For these studies,melanoma B16 appears to be a good model because: (a) it is aneasily measurable, fast-growing solid tumor which can be grownin serum-free media for comparative in vitro studies; (b) itproduces the growth-promoting SICRI, (and possibly othergrowth factors) and releases it in blood. As long as more specificassays are not available, SICRI can conveniently be measuredby insulin-specific RIA; and (c) melanoma B16 growth andSICRI secretion can be partially manipulated by transplantingthe tumor from normoinsulinemic to diabetic hosts. This treatment initially will slow down both tumor proliferation andSICRI secretion (22). Furthermore, study of SICRI-associatedhypoglycemia might provide insight into some of the origins ofhypoglycemias which often accompany tumors in humans (19).

However, for the purpose of such studies, experimental tumors in their high passages might not be the best objects,because during the first few passages cell populations undergoextensive selection (23). In spite of that, SICRI-secreting tumors might be relatively convenient because by transplantingthem into diabetics, a controlled selection pressure can beapplied (22). Such systems might be useful for further studiesof mechanisms of clona 1selection in tumors and tumor-inducedderepression of latent genes.

ACKNOWLEDGMENTS

We thank Drs. F. G. Prendergast and R. J. Bernacki for generousand friendly support. We thank Drs. C. Porter and M. Ip for helpfulsuggestions, Drs. W. S. Brimijoin and J. E. Gerich for use of theirfacilities, Dr. H. L. Moses and James Chamberlain (AMgen) for thegenerous gifts of IGF I, E. L. Webster and S. Trafalski for typing, andP. J. Callahan for artwork.

REFERENCES

1. Sporn, M. B., and Todaro, G. L. Autocrine secretion and malignant transformation of cells. N. Engl. J. Med-, 303: 878-880, 1980.

2. Baltic, V., Levanat, S., Petek, M., Bratic-Mikes, V., Pavelic, K., and Vuk-Pavlovic, S. Elevated levels of substances immunologically cross-reactivewith insulin in blood of patients with malignant and nonmalignant pulmonarytissue proliferation. Oncology (Basel), 42: 174-178,1985.

3. Pavelic, K., Bolanca, M., Vecek, N., Pavelic, J., Marotti, T., and Vuk-Pavlovic, S. Carcinomas of the cervix and corpus uteri in humans: stage-dependent blood levels of substance(s) immunologically cross-reactive withinsulin. J. Nati. Cancer Inst., 68: 891-894, 1982.

4. Pavelic, K., Odavic, M., Pekic, B., Hrsak, I., and Vuk-Pavlovic, S. Correlationof substance(s) immunologically cross-reactive with insulin, glucose andgrowth hormone in Hodgkin lymphoma patients. Cancer Lett., 17: 81-86,1982.

5. Pavelic, K., and Vuk-Pavlovic, S. C-peptide does not parallel increases ofserum levels of substances cross-reactive with insulin in non-Hodgkin Km

2212

on April 28, 2021. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

AUTOCRINE TUMOR GROWTH AND TUMOR-ASSOCIATED HYPOGLYCEMIA

phoma patients. Blood, 61:925-928, 1983.6. Pavelic, Lj., Pavelic, K., and Vuk-Pavlovic, S. Human mammary and bron

chial carcinomas. In vivo and in vitro secretion of substance immunologicallycross-reactive with insulin. Cancer (Phila.), 53:2467-2471, 1984.

7. Bajzer, Z., Pavelic, K., and Vuk-Pavlovic, S. Growth self-incitement inmurine melanoma K 16: A phenomenological model. Science (Wash. DC),225:930-932,1984.

8. Bradley, T. R., and Metcalf, D. The growth of mouse bone marrow cells invitro. Aust. J. Exp. Biol. Med. Sci., 44: 287-299, 1966.

9. Pluznik, S., and Sacks, L. The cloning of normal "mast" cells in tissueculture. J. Cell. Comp. Physiol., 66: 319-324, 1965.

10. Lowry, O. H., Rosebrough, N. J., Fair, A. L., and Randall, R. J. Proteinmeasurement with folin phenol reagent. J. Biol. Chem., 193:265-275,1951.

11. Hyvarinen, A., and Nikkila, E. A. Specific determination of blood glucosewith »loiu¡ilinc. Clin. Chim. Acta, 7:140-143, 1962.

12. Rodbell, C. Metabolism of isolated fat cells. I. Effects of hormones on glucosemetabolism and lipolysis. J. Biol. Chem., 239:375-380, 1964.

13. Hirsch, J., and Gallian, F. Methods for the determination of adipose cell sizein man and animals. J. Lipid Res., 9:110-119, 1968.

14. Opara, E. C, Wrigglesworth, J. M., and Perry, M. C. Time course of reversalof insulin-stimulated glucose uptake by anti-insulin antibody. Biochem. Sue.

Trans., 10:26-27, 1982.15. Gammeltoft, S., and Gliemann, J. Binding and degradation of I29l-labeled

insulin by isolated rat fat cells. Biodi im. Biophys. Acta, 520:16-32, 1973.16. Ham, R. <... and McKeehan, W. L. Media and growth requirements. Methods

Enzymol., 5«:44-93, 1979.17. LeMarchand-Bnistel, Y., Jeanrenaud, B., and Freychet, P. Insulin binding

and effects in isolated soleus muscle of lean and obese mice. Am. J. Physiol.,234: E348-E358, 1978.

18. Nickel), K. A., Halper, J., and Moses, H. L. Transforming growth factors insolid human malignant neoplasms. Cancer Res., 4.1: 1966-1971, 1983.

19. Papaioannou, A. N. Tumors other than insulinomas associated with hypo-glycemia. Surg. Gynecol. Obstet., 123:1093-1109, 1966; Cell, 33:931-938,1983.

20. Glick, S. M. Normal and abnormal secretion of growth hormone. Ann. NYAcad. Sci., 148:471-478, 1968.

21. Kaplan, P. L., and Ozanne, B. Cellular responsiveness to growth factorscorrelates with a cell's ability to express the transformed phenotype.

22. Pavelic, K. Aplastic carcinoma in diabetic mice: hypoglycemia-suppressedproliferation rate and insulin synthesis by tumor cells. J. Nati. Cancer Inst.,62:139-141,1979.

23. Steele, G. G. Growth Kinetics of Tumors. Oxford, England: Clarendon Press,1977.

2213

on April 28, 2021. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1986;46:2208-2213. Cancer Res   S. Vuk-Pavlovic, E. C. Opara, S. Levanat, et al.  

in VivoHypoglycemia in Murine Melanoma B16 Autocrine Tumor Growth Regulation and Tumor-associated

  Updated version

  http://cancerres.aacrjournals.org/content/46/5/2208

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/46/5/2208To request permission to re-use all or part of this article, use this link

on April 28, 2021. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from