(CANCER RESEARCH 49, 3499-3505. July I. 1989]

Hyaluronic Acid-stimulating Activity in Sera from the Bovine Fetus and fromBreast Cancer Patients1

Margaret Decker, Ernest S. Chiù,Charles Dollbaum, Ali Moiin, Jackson Hall, Rex Spendlove, Michael T. Longaker,and Robert Stern2

Departments of Pathology [M. D., E. S. C.,A.M., J. //., R. S.J, and Surgery (M. T. L.J, School of Medicine, University of California, San Francisco, California 94143;Peralta Cancer Research Institute, Oakland, California fC. D.J; and Hyclone Laboratories [R. Sp.J, Logan, Utah 84321

ABSTRACT

The .vint-qua non of malignancy is the ability of tumor cells to migrateand invade surrounding tissue. There are many substances that have beendescribed that enhance cell motility and hyaluronic acid is prominentamong these. Hyaluronic acid is a high molecular weight alternatingdisaccharide polymer found in abundance in extracellular matrices whenever rapid cell proliferation or tissue regeneration and repair occur. Itcreates a permissive environment for cell motility during embryogenesis,and high levels of hyaluronic acid also correlate with increased tumorcell invasion and aggressiveness. Little is known about the regulation ofhyaluronic acid production, either in normal tissue or in malignancy. Inthis study, we characterize a hyaluronic acid-stimulating activity in fetalcalf serum and describe a similar activity in the sera of breast cancerpatients. The stimulating activity was measured by placing aliquots oftest substance on fibrosarcoma cells. These indicator cells, which synthesize copious quantities of hyaluronic acid, respond to stimulation in atime- and dose-dependent fashion. The fetal calf serum hyaluronic acid-stimulating activity is maximum early in gestation and then falls rapidlyto essentially no activity at term. This activity was partially purified from120-day fetal calf serum by concanavalin A-Sepharose affinity and ionexchange chromatography and is accounted for by a glycoprotein with amolecular weight of 150,000 on gel filtration under native conditions. Thesera of breast cancer patients with measurable burden of disease alsocontained hyaluronic acid-stimulating activity, which was not present innormal serum donors or in breast cancer patients without evidence ofdisease. The production of this stimulating activity may contribute to thedevelopment of the malignant phenotype by inducing hyaluronic acid-rich microenvironments that are permissive to tumor cell invasion andmétastases.

INTRODUCTION

HA3 has a key role in the structure and organization of the

ECM. It is found in high concentrations during embryogenesisand whenever rapid tissue proliferation and regeneration occur(1). It has been proposed to be involved in the detachmentprocess of the cell cycle that allows cells to move (2). A burstof HA synthesis occurs prior to mitosis (3-6), enabling cells tobecome dissociated from neighboring cells and their ECM inpreparation for division. HA also inhibits cell differentiation(7, 8), creating an environment that instead promotes cellproliferation. HA is also concentrated in the environment oftumor cells (9) and striking increases in serum HA are found

Received 12/29/88; revised 3/27/89; accepted 3/31/89.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 (his fact.

1Supported by Grant PO1-CA44768 from the National Cancer Institute (NIH.DHHS); by intramural grants from the Research Evaluation and AllocationCommittee of the School of Medicine, the Academic Senate, and the FetalTreatment Program, Department of Surgery of the University of California, SanFrancisco; and by a HEDCO Foundation fellowship.

1To whom requests for reprints should be addressed, at The Department of

Pathology, School of Medicine, University of California. San Francisco. CA94143-0506.

'The abbreviations used are: HA. hyaluronic acid; HASA, HA-stimulating

activity; Con A, concanavalin A; CPC, cetylpyridinium chloride: CS. calf serum;DME. Dulbecco's modification of minimal essential media; ECM, extracellular

matrix; FCS, fetal calf serum; GAGs. glycosaminoglycans; NBCS. newborn calfserum: TGFfi, transforming growth factor-ii.

in some patients with disseminated neoplasms ( 10). In a numberof experimental animal tumor systems, high levels of HAcorrelate with tumor aggressiveness (11-17).

HA synthesis is elevated in chick fibroblasts transformed byavian sarcoma viruses (18) or treated with tumor promoters(19). Elevated levels of HA synthesis have also been observedin fibroblasts following cocultivation with tumor cells (20) orcultured with conditioned media from a highly invasive humanbladder carcicoma cell line, HCV-29T (9).

The ability of HA to create a permissive environment for cellmovement has been widely documented in the embryologicalliterature and an important parallel may exist between theseobservations and the migration of epithelial tumor cells throughhost stroma in the process of invasion and the formation ofmétastases.Little is known of the mechanisms by which HAproduction is regulated, although tissue levels are known to bemodulated during morphogenesis (1). Current evidence suggests that fibroblasts are induced to produce HA by a factorsecreted by tumor cells (paracrine-like stimulation) or by thefibroblasts themselves (autocrine-like stimulation). In this studywe characterize a factor that has HA-stimulating activity infetal calf serum and describe a similar activity in the serum ofbreast cancer patients.

MATERIALS AND METHODS

Materials. D-IUo-'HJGIucosamine, 44.8 Ci/mmol, was from New

England Nuclear Corporation. DME and RPMI 1640 were preparedin the University of California, San Francisco, Cell Culture Facilityfrom GIBCO powders.

FCS and CS were from the Hyclone Corporation; NBCS was fromGIBCO, as were the two fetuin preparations, one obtained by the Spiromethod (21) and the other by the Deutsch method (22). Rabbit, goat,horse, and human sera were purchased from Cooper Biomédical.Thetype XIV bacterial protease (pronase E) from Streptomyces gríseaswasa Sigma product, as was the CPC, chondroitin sulfate Type A, heparin,and glucosamine-HCI. Ascorbic acid and hyaluronidase from Streptomyces hyaluronlyticus were from Calbiochem. Tris was from Schwarz/Mann, and the metaphosphoric acid from Sargent-Welch. Optifluorscintillation fluid was a product of Packard. Prosil 28, an organic silanefor siliconization of the glass fiber filters, was purchased from PCRInc., Gainesville, FL. TGF/3 was a Collaborative Research product. Theglass microfiber filters (GF/A) and the DEAE-cellulose (DE52) werepurchased from Whatman. Heparin Sepharose, Con A-Sepharose,Sephadex G-25, and the Sephacryls S-200, 300, and 400 were obtainedfrom Pharmacia. Molecular weight markers were purchased fromBioRad.

Collection of Human Serum Specimens. Five cc of blood was collectedfrom breast cancer and control patients after informed consent wasobtained. The blood was collected in an EDTA-treated tube from aperipheral venipuncture and the serum was collected and stored at—20°Cuntil it was used in the standard HASA assay described below

at 10% concentration.Cells and Cell Culture. Human newborn foreskin fibroblasts were

obtained from the cell culture facility of the University of California,San Francisco. Unless otherwise indicated, these cells were cultured inDME-H21 with 10% FCS, 1% penicillin-streptomycin, 1% fungizone.

3499

Research. on September 5, 2021. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

HYALURONIC ACID-STIMULATING FACTOR IN FETAL AND CANCER SERUM

at 37'C in an atmosphere of 5% CO2 and 95% air. The rat fibrosarcoma

cells were obtained from Dr. Merton Bernfield (Stanford, CA). Thisfibrosarcoma cell line, which synthesizes copious quantities of HA, wasoriginally derived from a 3-methylcholanthrene-treated Lewis rat (23).The fibrosarcoma cells were maintained routinely in RPMI 1640 with10% PCS and were divided 1:3 every second day. However, for experiments involving measurement of HA-stimulating activity, cells weretransferred to the DME-H16 and were cultured as described for skinfibroblasts except that 5% CS was used. DME-H16 was utilized becauseof the lower glucose concentrations. In separate experiments not shownhere, we demonstrated that the ['Hjglucosamine radiolabel was incor

porated more efficiently into HA in the presence of lower glucose levels.Measurement of HA Synthesis and HASA. Fibrosarcoma cultures

were seeded routinely at a density of 2.5 x IO5 cells, except whereindicated, in wells of 35-mm diameter (Costar). Cells are 50% confluentat this density. Experiments were conducted during the logarithmicstage of growth and were generally concluded when cells were approximately 80-90% confluent. Cells were allowed to attach for 18-24 hand the medium was then replaced with 1.5 ml fresh medium containingthe appropriate additions or materials to be assayed. After 24 h, [1,6-'Hlglucosamine was added to a final concentration of 50 mCi/ml unless

otherwise indicated. Following a 24-h labeling period, the medium wasremoved and pooled with two 0.5-ml washes of the cell layer withcalcium- and magnesium-free phosphate-buffered saline. Cell layerswere scraped with a rubber policeman. Final sample volume followingrinses was 2.5 ml. Sodium azide was added to these samples to a finalconcentration of 0.02%.

To solubilize GAGs from the above samples, proteins, includingproteoglycan core proteins, were digested with 5 mg/ml pronase, whichwas freshly prepared each time in 0.2 M Tris (pH 8.0), 0.02% sodiumazide, and predigested at 37'C for 1 h. A 0.40 ml aliquot of each sample

and an equal amount of the pronase-containing solution (10 mg/ml)were combined in microfuge tubes and maintained at 37'C for 24 h.The protease digestion was terminated by boiling samples at 100'C for

10 min. Making pinholes in each sealed microfuge tube was a necessaryprecaution at this step.

To degrade HA, a solution of hyaluronidase at a concentration of100 units/ml was prepared in 0.1 M sodium acetate and 0.05 M sodiumchloride, pH 5.0. From each pronase-digested sample, three 100-jilaliquots were combined with 25 n\ (2.5 units) of the hyaluronidasesolution. Three additional KM)//I aliquots for controls were run inparallel, but were combined with buffer only. All six preparation weremaintained at 37'C for 18 h. This digestion was terminated by boilingthe samples at 100'C for 10 min.

Undigested GAGs were separated from the free hexoses and theunincorporated label by CPC precipitation. The GAG carriers included20 iil of 5% chondroitin sulfate and 10 ^1 of 1.5% HA mixed in 40 mMNaCl and 0.02% sodium azide. The GAG precipitation was achievedby the addition of 40 pi of 7.5% CPC dissolved in 40 mM NaCl and0.02% sodium azide. The precipitation step was carried out at 37°Cfor

45 min.Precipitated GAGs were collected by washing samples onto What

man GF/A filters, 2.4 mm, which had been placed onto a 10-wellvacuum-assisted manifold filtration apparatus (Hoefer Scientific). Withthe vacuum on, the filters were prerinsed with 1 ml of a 1% Prosilsolution followed by 1 ml of 0.1 mg/ml unlabeled glucosamine in 1.5%CPC, 0.04 M sodium chloride, and 0.02% sodium azide. The precipitated samples were then transferred quantitatively to the filters with1.5% CPC. The filters were rinsed with 30 ml of 1.5% CPC. Finally,the filters were placed into scintillation vials and allowed to dry in a55°Coven. Optifluor was then added and levels of radioactive label

were determined.Cell Counts. The duplicate cultures for cell counts were initiated at

the same time and grown under conditions identical to their radiola-beled counterparts. A solution of 0.05% trypsin and 0.02% EDTA wasused and cells were incubated for 1 min to lift cells from off the culturedishes. Following dilution, cells were counted in duplicate using ahemacytometer.

Statistical Analysis. The mean and standard deviations were calculated for three samples at each experimental point. The standard

deviation of the level of HA synthesis at each of the experimental pointspresented is the sum of the standard deviation of the three nonhyalu-ronidase-treated samples plus the standard deviation of the three hya-luronidase-treated samples.

RESULTS

An assay was developed for measuring HA-stimulating activity. This assay depends on the response of an appropriateindicator cell to serum HA-stimulating activity by synthesizinghyaluronic acid in a dose- and time-dependent manner. A ratfibrosarcoma cell line that had previously been shown to synthesize copious amounts of HA fulfills these criteria (24).

We compared the production of HA in this fibrosarcoma linewith normal fibroblasts and showed that both cell types respondto PCS in a dose-dependent manner (Fig. 1). However, the HAproduced by normal cells is much lower than the fibrosarcomacells and is sensitive to cell density (data not shown). Thefibrosarcoma line was selected as a suitable model for assay ofHA-stimulating activity in all subsequent experiments. Levelsof activity were routinely expressed on a per-cell basis to eliminate any mitogenic differences between the various sera used.

A serum activity that stimulates HA synthesis has long beenrecognized (20, 25). A number of different commercial serawere tested using our fibrosarcoma-based assay (Fig. 2). PCShad the highest level of activity. NBCS had approximately 40%of the activity of PCS, and CS less than 15% of the activity ofFCS. We observed that different lots of NBCS had widelydifferent levels of activity. Preparations of NBCS are derivedfrom calves between the ages of 1 to 30 days. Widely differentlevels of activity may be a reflection of the average age of thecalves.

18-1

16-

14 -

12-

o 10 -

5S

CO-• 6 -

4 -

2 -

FIBROSARCOMACELLS

FIBROBLASTS

112 16 20

SERUM (%)

Fig. 1. Levels of HASA with varying percentage serum concentration of PCSin the standard reaction mixture using rat fibrosarcoma cells and human foreskinfibroblasts. The levels are indicated in the 1.5-ml final reaction mixture. Theassay and processing procedure are outlined under "Materials and Methods."

3500

Research. on September 5, 2021. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

HYALURONIC ACID-STIMULATING FACTOR IN FETAL AND CANCER SERUM

12-co

10-_JLUü

8-ios

6~XI

4-u<£

2-

0-II<<• 6 <¿/ ^ -^/ / f /

SERUM (10%)

Fig. 2. Comparison of levels of HASA in various animal sera. All sera were10% (v/v) in the final 1.5-ml reaction mixture using rat fibrosarcoma cells as theindicator cells.

g 10-

PCSEXPERIMENT2

PCSEXPERIMENT 1

PCSEXPERIMENT3

4 8 12 16 20 24 28 32

SERUM (%)

Fig. 3. Levels of HASA as a function of serum concentration in the standardreaction assay. Three separate experiments using FCS are shown utilizing variouslevels of serum. The same lot of FCS was used throughout. Calf serum and NBCSare also compared.

A more careful examination of HASA activity as a functionof FCS, NBCS, and CS concentration was made, from 0 to30% (Fig. 3). A slight decrease in synthesis was observedconsistently at low levels of FCS, between 1 and 5% comparedto controls incubated in the absence of serum (Fig. 3, Experiment 1). A substance that inhibits the production of HA maybe present. This inhibition was not overcome until more than5% FCS was utilized. In three separate experiments, synthesisof HA responded to FCS in a linear fashion, between 5 and25%. Activity was inhibited when greater than 25% FCS wasutilized (Fig. 3, Experiment 3). Both NBCS and CS gave muchlower levels of activity, with no stimulation of HA synthesis asserum concentrations were increased.

Because inhibition had been observed with FCS at levelsbelow 5% and above 25%, we postulated that low levels ofactivity in NBCS and CS might be due to a similar inhibition.Combination experiments were performed to test this hypothesis. Neither 7.5% NBCS nor CS inhibited the activity whencombined with 15% FCS (Fig. 4). Some stimulation of FCS

10 -

9 -

8 -

7 -

O

5 -

4 -

3 -

2 -

1 -

-ñññ¡i*PCS

NBCS CS PCS FCSNBCS15%7.5% 7.5% 15% 15% 7.5%

& & &NBCS CS CS7.5% 7.5% 7.5%

SERUM AND SERA COMBINATIONS

Fig. 4. Levels of HASA in various sera and sera combinations to determine ifinhibitory substances are present in NBCS and CS.

total activity was observed; however, NBCS and CS activitieswere not additive.

Commercial FCS is normally pooled from a large number ofcalves of varying fetal ages. By volume, most serum in pooledFCS is contributed by the largest calves, late in gestation, whichwe have shown have low levels of HASA (Fig. 5). Therefore,we postulated that the activity detected in pooled FCS wascontributed by small volumes of sera from early gestationalembryos, which contained high levels of HASA. Instead ofutilizing pooled FCS, sera from individual calves was obtained,fetal age being assessed by crown-rump length. Such sera wereassayed at 5 and 15% (v/v) (Fig. 5). The bovine gestationalperiod is 270 to 275 days. Serum from a calf close to delivery(270 days) had activity intermediate between pooled FCS andCS. Serum from a 180-day fetal calf was only slightly higher inactivity than pooled FCS. However, activity rose rapidly at 150days and was highest at 120 days, the earliest period tested.Plasma derived from an additional 107-day calf fetus wasconsiderably lower than the 120-day fetal calf serum. Thissuggested that peak activity might occur near 120 days ofgestation.

We wished to establish whether the HASA was a platelet-derived product. Plasma from individual calves was assayed

3501

Research. on September 5, 2021. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

HYALURONIC ACID-STIMULATING FACTOR IN FETAL AND CANCER SERUM

52

48

44

40

36

32

or 28

5 24o.OÕ 20

en

16

12

120 d

Table 1 Effect of chemical and physical treatment on HA-stimulating activity inFBS

Treatment Activity (%)

5 10

Serum (%)

15

Fig. 5. Levels of HASA in PCS from individual calves at various gestationalages. Fetal age was determined from crown-rump size. Assays were performedusing 5 and 15% (v/v) in the standard reaction mixture.

(data not shown). The fetal plasma had levels of activity comparable to the corresponding sera.

We previously tested a number of polypeptide factors fortheir ability to stimulate HA synthesis in our system, includingacid and basic fibroblast growth factors, insulin-like growthfactor II, platelet-derived growth factor, nerve growth factor,TGF/3 and TGFa, combinations of TGFa and TGF0, andepidermal growth factor plus TGF0. No stimulation was observed in the range from 0.1 to 100 ng/ml when tested in thepresence of 5% PCS (data not shown).

TGF/3 is present in fetal serum at approximately 5 ng/ml(26). This factor stimulates the expression of a number ofextracellular macromolecules (27-33). However, TGF/3 had nostimulatory effect on the indicator fibrosarcoma cells.

Fetuins are serum globulins that constitute nearly the totalglobulins in the fetus of ungulates and make up 40 to 50% oftotal proteins. We examined each of two commercial fetuinpreparations, the Spiro (21) and Deutsch (22) preparations, todetermine if they contained HASA. Compared to controlsassayed in the presence of 5% FCS, neither fetuin preparationat 1 and 2% w/v had substantial activity. Cells did not survivein the presence of 2% fetuin derived from the Spiro procedure.A 15% increase in activity was observed using 1% of theDeutsch preparation of fetuins, but activity decreased at 2%.Fetuins do not appear to contain HASA.

The HASA from FCS was examined in greater detail (Table1). Activity was protease sensitive and nondialyzable. Activitywas also stable to heating to 90°C for 10 min, but not to acid

Control (10% FCS)Dialyzed for 48 h against DM EHeat, 56'C for 30 minHeat, 90°Cfor 10 minPronase digestion, JVC for 2 h (pro-

nase was predigested at 37°Cfor 1 h)

Acid (pH 2.5 with HC1)Acid (pH 4.5 with metaphorphoric acid)Acid (pH 4.0 with metaphosphoric

acid plus heparin, 100 /ig/ml)Alkali, pH 10withNH4OH

100909585

0

22115

or alkali treatment at pH 2.5 to 10.0, respectively. Milder acidtreatment was also examined. Metaphosphoric acid, 0.5 M, wasadded in a dropwise manner to serum at 0°C.Precipitation of

protein occurred between pH 5.0 and 4.5. Following centrifu-gation, the supernatant was passed through a column of Seph-adex G-25 with DME. Only 21% of activity remained, compared to control sera processed in parallel. The presence ofheparin is protective of fibroblast growth factor during exposureto acid (34). No such protection of the HASA was observedwhen metaphosphoric acid treatment was repeated in the presence of heparin (100 Mg/ml). Activity did not bind to heparinSepharose but did bind to Con A-Sepharose, suggesting thatthe activity was a mannose-containing glycoprotein.

On the basis of these experiments a sequential series of stepswas devised to purify the HASA. Two hundred ml of 120-day-old FCS was passed through a Sephadex G-25 column (l m x2.5 cm i.d.) to remove serum lipids, using 20-mM sodiumphosphate buffer, pH 7.0, with 0.1 M NaCl. The protein peakfrom this column was placed on a Con A-Sepharose column(20 x 1.5 cm i.d.) and washed with a 20-mM sodium phosphatebuffer, pH 7.0, now containing l M NaCl. When the first peakof protein was eluted, the same buffer additionally containing0.1 M a-D-mannopyranoside was applied. A second peak ofprotein was eluted (Fig. 6). Fractions (8 ml) were pooled asindicated by the horizontal arrows, desalted on a column ofSephadex G-25 equilibrated with DME-H16, and assayed inthe standard fibrosarcoma HASA assay.

As shown in Fig. 6, the first protein peak contained inhibitor

8 16 24 32 40 48 56 64 72 80

Fraction no (8 ml)

Fig. 6. Con A-Sepharose column chromatography of 120-day fetal calf serumfollowing delipidation by passage through a Sephadex G-25 column in 20 nmNaPO4, pH 7.0, with 0.1 M NaCl. After application, the Con A column waswashed with 20 m\i NaPO4. containing l M NaCl, with a flow rate of 10 ml/h.After the peak of protein had been eluted. the same buffer, now containing O.I Mn-D-mannopyranoside was applied (fraction 52), which eluted the second proteinpeak. Fractions were pooled as indicated by the horizontal bars and dialyzedagainst DME-H16. The HASA was determined by adding 150 n\ to the standardfibrosarcoma reaction mixture. Boxes, levels of activity; solid line, protein concentration as determined by absorbance at 280 nm.

3502

Research. on September 5, 2021. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

HYALURONIC ACID-STIMULATING FACTOR IN FETAL AND CANCER SERUM

activity that decreased HA deposition below the levels of controlreactions containing only the 5% CS. The H AS A was containedin the second peak.

The Con A-Sepharose-peak two was pooled as indicated bythe arrows and dialyzed against initial buffer and placed on acolumn of DEAE-cellulose (10 x 1 cm i.d.). A liner gradientwas applied derived from 125 ml each of 20 HIM sodiumphosphate, pH 6.3 (initial buffer), and 0.4 M sodium phosphate,pH 5.3 (final buffer). Fractions of 8 ml were collected (Fig. 7).Every fourth fraction was assayed following a G-25 desaltingstep in DME-H16. Aliquots of 150 /tl were used. The peak ofactivity appeared between fractions 36 and 40. These fractionswere pooled and an aliquot was examined by molecular sievechromatography. A single peak of Ha-stimulating activity wasobserved under nondenaturing conditions on Sephacryl S-200column chromatography. Based on plots of the log of molecularweights and elution positions from the column using variousprotein standards, a native relative molecular size of 150,000was obtained (Fig. 8).

We considered that the value of Mr \ 50,000 was perhapsartifactual if the glycoprotein-associated activity was partiallyexcluded and partially included. Molecular sieve chromatography was repeated on similar sized columns of Sephacryl S-300and S-400. The same apparent size of Mr 150,000 was obtained(data not shown).

Because of previous data that linked HA to the malignantphenotype, sera from breast cancer patients were screened forHASA. Sera from seven women with active breast cancer werecollected and high levels of hyaluronic acid-stimulating activitywere observed in all patients (Fig. 9, left). Increases in the levelof activity correlated roughly with increased burden of disease.In one patient in whom HASA levels were serially measured,activity increased as her measurable disease increased (specimens 1,3; Fig. 9, left). Four patients who had been treated forbreast cancer and who had no evidence of disease, had no serumactivity (Fig. 9, middle).

Little or no activity was present in most sera of normalcontrol women (Fig. 9, right). Sera from 13 randomly selectednormal women were assayed for hyaluronic acid-stimulatingactivity. Low activity, less than 2 x IO4 cpm/150 /il activity,

was detected in 11 of them. Two women had moderatelyelevated activity, above 2.0 x IO4 cpm/150 n\. One of these

Fraction no (8 ml)

Fig. 7. DEAE-cellulose (DE-52) column chromatography of Con A-Sepharosebinding material. Fractions indicated by the horizontal area were pooled, dialyzed,and applied to the column. After washing with initial buffer, a linear gradientformed of 125 ml each of 20 min NaPO4, pH 6.3 (initial buffer) and 0.4 M NaPO4.pH 5.3 (final buffer) was applied. The column was eluted at a flow rate of 10 ml/h. Every fourth fraction was assayed by dialyzing against DME-HI6 and adding150 p\ to the standard reaction mixture (1.5 ml total volume).

210

200

150

120110

100

50

20

10

100 200 300 400

Elution volume (ml)

Fig. 8. Molecular sieve chromatography of the HASA obtained from theDEAE-cellulose chromatography. Fractions 36-40 were pooled and placed onthe Sephacryl S-200 column, which was preincubated with DME-H16. Fractionsof 8 ml were collected at a flow rate of 10 ml/h, and every fourth fraction wasassayed directly by using 150 fil in the standard reaction mixture. The peakfraction is indicated by the arrow. In a separate Chromatographie run, the elutionpositions of molecular weight standards were determined by absorbance at 280nm. Protein standards were thyroglobulin, M, 600,000 (M, 210,000, the exclusionvolume); 7-globulin, M, 150,000; ovalbumin, M, 42,700; myoglobin, M, 17,500;and vitamin B,2, M, 1,350.

"normal" controls had had two breast biopsies in the past year

with a diagnosis of severe dysplasia and atypia in each.

DISCUSSION

Despite the importance of HA in basic biological and pathological processes, the regulation of HA deposition is notclearly understood. Modulation of HA synthesis by some lowmolecular weight compounds has been reported. Insulin (35),prostaglandins (36, 37), sodium butyrate (38), interleukin 1(39), and somatomedins (40, 41) can stimulate HA production,whereas thyroid hormone (42) and cortisol (43, 44) have beenshown to be inhibitory.

More recently, fibroblast growth factor has been shown tostimulate HA synthesis in early chick embryo limb buds (45).We have tested both acid and basic fibroblast growth factor inour system and observed no stimulation of HA synthesis. Furthermore, the serum factor we describe showed none of thephysical characteristics of acidic or basic fibroblast growthfactor (i.e., molecular weight, acid and heat lability), and thereare no published reports of fibroblast growth factor in fetal andadult serum.

Comparable levels of HASA were found in plasma and serumin the present experiments. This suggests that the activity wasnot a platelet-derived factor. However, identification of a platelet factor that stimulates HA synthesis from canine and humansources indicates that species differences may exist (46).

It has long been recognized that a factor in serum stimulatesHA synthesis (20, 25). A glycoprotein with 150,000 molecularweight serum factor was isolated from calf serum by Tomidaand coworkers (25, 47). Our data suggest that calf serum is notan optimal source for purification of HASA. The calf serumfactor, like ours, binds to Con A-Sepharose and its activity is

3503

Research. on September 5, 2021. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

HYALURONIC ACID-STIMULATING FACTOR IN FETAL AND CANCER SERUM

Scrum SampleS0rum Sample n Simple

Fig. 9. Sera were obtained from control and patient populations and stored at —70°Cuntil assayed. Sera were added to the standard 1.5-ml culture assay at 10%

(v/v) as described in Fig. 1. Left, breast cancer patients; middle, breast cancer patients in remission with no evidence of disease clinically; and right, normal controls,women ranging from ages of 23 to 50.

sensitive to periodate oxidation, suggesting it is a mannose-containing glycoprotein.

The origin of the HASA may be tumor cells themselves.Indeed, stimulation of HA synthesis in fibroblasts by tumorextracts has been previously reported (48). Recently, conditioned media from a highly invasive human bladder carcinomacell line (HCV-29T) was shown to stimulate HA synthesis infibroblasts (9). No characterization of the conditioned mediafactor is described. A membrane-bound tumor cell factor whichstimulates HA synthesis in fibroblasts has also been identified(49).

The HASA in the serum of breast cancer patients showssimilar biological activity to the fetal serum but has not as yetbeen characterized. Such experiments are currently in progress.If these functionally similar molecules are physically related,this factor (which may be one of a family of factors) mightrepresent examples of oncofetal proteins. In normal development, this HASA would induce hyaluronic acid deposition,which is critical to normal embryonic development. In malignancy, a similar induction might promote cell motility andcreate a permissive environment for invasion and metastasis.

Various factors produced by mesenchymal cells that areinvolved in cell mortality and gel invasion in vitro have recentlybeen identified. These include a scatter factor, which is afibroblast-derived modulator of epithelial cell mobility (50), atumor cell autocrine motility factor (51), and a factor derivedfrom tumor-associated fibroblasts that promotes collagen gelinvasion by normal fibroblasts (52). An additional fibroblast-produced paracrine factor that modulates tumor cells' biochem

ical behavior has been described (53, 54). This may represent anew family of paracrine and/or autocrine regulators that modulate stromal-epithelial interactions. Each of these proteins,functioning as autocrine or paracrine factors, may play criticalroles in such normal processes as embryogenesis and woundhealing and when they lead to faulty epithelial-mesenchymal

interactions, may contribute to carcinogenesis and malignantprogression.

HASA, even though it is measurable in the circulation, doesnot appear to induce all fibroblasts in an organism to producehigh levels of HA. We have shown that in fibroblasts fromnormal patients, basal levels of HA deposition are lower thanin stromal cells isolated from the breast tumor patients (datanot shown). Additionally, many cultured breast cancer-derivedfibroblasts are stimulated to produce hyaluronic acid in response to TGF/3 under conditions in which hyaluronic acidaccumulation by normal fibroblasts is uniformly inhibited (55).

Thus, perhaps only fibroblasts that are "competent," either by

way of genetic changes or other autocrine or paracrine-likestimulations, can respond to the circulating or local hyaluronicacid-stimulating activity.

Herein we have described the existence of a glycoprotein infetal serum and a functionally similar factor in the serum ofbreast cancer patients that stimulates the production of HA. Inbreast cancer patients this activity may correlate with burdenof disease, but its sensitivity in detecting minimal disease isunknown. We have found this HA-stimulating activity in thesera of a lung cancer patient as well as in the conditioned mediafrom a lung and a prostate cancer cell line (data not shown).This demonstrates that the factor is not specific for breastcancer but may be common to all carcinomas. We are in theprocess of examining sera from patients with various types andstages of malignancy to determine the accurate distribution ofthe factor as well as to determine its physical characteristics.

ACKNOWLEDGMENTS

The authors thank Mary Williams and nurses at University ofCalifornia, San Francisco, Oncology Clinic for specimen accrual, andA. Howlett for review of the manuscript. We also wish to thank DavidGeller for editing and preparing the manuscript.

ADDENDUM

While this manuscript was in preparation, another communicationappeared describing a hyaluronate stimulatory factor: Knudson, W.,and O'Toole, B. P. Membrane association of the hyaluronate stimula

tory factor from LX-1 human lung carcinoma cells. J. Cell. Biochem.,38: 165-177, 1988.

REFERENCES

1. O'Toole. B. P. Glycosaminoglycans in morphogenesis. In: E. D. Hay (ed.).Cell Biology of the Extracellular Matrix, pp. 259-294. New York: PlenumPress, 1982.

2. Turley, E. A., and Terranee, J. Localization of hyaluronate and hyaluronate-binding protein on motile and non-motile fibroblasts. Exp. Cell Res., 16: 17-28, 1984.

Ì.O'Toole, B. P., Jackson, G., and Gross, J. Hyaluronate in morphogenesis:

inhibition of chondrogenesis in vitro. Proc. Nati. Acad. Sci. USA, 69: 1384-1386, 1972.

4. Tomida, M., Koyama, H., and Ono, T. Hyaluronate acid synthetase incultured mammalian cells producing hyaluronic acid: oscillatory changeduring the growth phase and suppression by 5-bromodeoxyuridine. Biochim.Biophys. Acta, 338: 352-363, 1974.

5. Mian, N. Analysis of cell-growth-phase-related variations in hyaluronatesynthase activity of isolated plasma-membrane fractions from cultured Im

3504

Research. on September 5, 2021. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

HYALURONIC ACID-STIMULATING FACTOR IN FETAL AND CANCER SERUM

man skin fibroblasts. Biochem. J., 237: 333-342, 1986. 31.6. Brecht, M., Mayer, U., Schlosser, E., and Prehm, P. Increased hyaluronate

synthesis is required for fibroblast detachment and mitosis. Biochem. J., 239:445-450, 1986. 32.

7. Kujawa, M. J., Pechak, D. G., Fiszman, M. Y., and Caplan, A. I. Hyaluronicacid bonded to cell culture surfaces inhibits the program of myogenesis.Develop. Biol., 113: 10-16, 1986.

8. Kujawa, M. J., and Tepperman, K. Culturing chick muscle cells on glycosa- 33.minoglycan substrates: attachment and differentiation. Develop. Biol., 99:277-286, 1983.

9. Knudson, W., and Pauli, B. U. Malignant bladder tumors use host hyaluronate deposition for directed invasion. J. Cell Biol., 105: 181a, 1987. 34.

10. Manley, G., and Warren, C. Serum hyaluronic acid in patients with dessem-inated neoplasm. J. Clin. Pathol., 40:626-630, 1987. 35.

11. Ozzello, L., Lasfargues, E. Y., and Murray, M. R. Growth-promoting activityof acid mucopolyusaccharides on a strain of human mammary carcinomacells. Cancer Res., 20: 600-604, 1960. 36.

12. Takeuchi, J., Mitsuko, S., Sato, E., Shamoto, M., Miura, K., and Nakagaki,S. Variations in glycosaminoglycan components in breast tumors. CancerRes., 56:2133-2139, 1976.

13. Chandrasekran, E. V., and Davidson, E. A. Glycosaminoglycans of normal 37.and malignant cultured human mammary cells. Cancer Res., 39: 870-880,1979.

14. O'Toole, B. P., Biswas, C., and Gross, J. Hyaluronate and invasiveness of 38.the rabbit V¡carcinoma. Proc. Nati. Acad. Sci. USA, 76: 6299-6303, 1979.

15. Angello, J. C., Danielson, K. G., Anderson, L. W., and Hosick, H. L. 39.Glycosaminoglycan synthesis by subpopulations of epithelial cells from amammary adenocarcinoma. Cancer Res., 42: 2207-2210, 1982.

16. Angello, J. C., Hosick, H. L., and Anderson, L. W. Glycosaminoglycan 40.synthesis by a cell line (CI-SI) established from a preneoplastic mousemammary outgrowth. Cancer Res., 42: 4975-4978, 1982.

17. Rimata, K., Honma, Y., Okayama, M., Oguri, K., Hozumi, M., and Suzuki,S. Increased synthesis of hyaluronic acid by mouse mammary adenocarci- 41.noma cell variants with high metastatic potential. Cancer Res., 43: 1347-1354, 1983. 42.

18. Ishimoto, N., Temin, H. M., and Strominger, J. L. Studies of carcinogenesisby avian sarcoma viruses. II. Virus-induced increase in hyaluronic acidsynthetase in chicken fibroblasts. J. Biol. Chem., 241: 2052-2057, 1966. 43.

19. Ullrich, S. J., and Hawkes, S. P. The effect of the tumor promoter phorbolmyristate acetate (PMA) on hyaluronic acid (HA) synthesis by chick embryofibroblasts. Exp. Cell Res., 148: 377-386, 1983. 44.

20. Knudson, W., Biswas, C., and O'Toole, B. P. Interactions between human

tumor cells and fibroblasts stimulate hyaluronate synthesis. Proc. Nati. Acad.Sci. USA, 81: 6767-6771, 1984. 45.

21. Spiro, R. G. Studies on fetuin, a glycoprotein of fetal serum. I. Isolation,chemical composition, and physiochemical properties. J. Biol. Chem., 235:2860-2869, 1960. 46.

22. Deutsch, H. F. Fetuin: The mucoprotein of fetal calf serum. J. Biol. Chem.,208: 669-678, 1954. 47.

23. Biesecker, J. L. Cellular and humoral immunity after allogenic transplantation in the rat. I. Cellular and humoral immunity as measured by a '"( ï

cytotoxicity assay after allogenic tumor and renal transplantation. Trans- 48.plant, 75: 298-307, 1973.

24. Underbill, C. B., and O'Toole, B. P. Transformation-dependent loss of the 49.hyaluronate-containing coats of cultured cells. J. Cell. Physiol., 110: 123-128, 1982.

25. Tomida, M., Koyama, H., and Ono, T. A serum factor capable of stimulating 50.hyaluronic acid synthesis in cultured rat fibroblasts. J. Cell Physiol., 91:323-328, 1976.

26. Masui, T., Wakefield, J. F., Lechner, M. A., LaVeck, M. A., Sporn, M. B., 51.and Harris, C. C. Type ßtransforming growth factor is the primary differentiation-inducing serum factor for normal human bronchial epithelial cells.Proc. Nati. Acad. Sci. USA, 83: 2438-2442, 1986. 52.

27. Ignotz, R. A., and Massagué,J. Transforming growth factor-/} stimulatesexpression of fibronectin and collagen and their incorporation into theextracellular matrix. J. Biol. Chem., 262:4337-4345, 1986. 53.

28. Roberts, A. B., Sporn, M. B., Assoian, R. K., Smith, J. M., Röche,N. S.,Wakefield, L. M., Heine, U. I., Liotta, L. A., Falanga, V., Kehrl, J. H., andFauci, A. S. Transforming growth factor type-beta: rapid induction of fibrosisand angiogenesis in vivo and stimulation of collagen formation in vitro. Proc. 54.Nati. Acad. Sci. USA, 83: 4167-4171, 1986.

29. Chen, J-K., Hoshi, H., and McKeehan, W. L. Transforming growth factorbeta specifically stimulates synthesis of proteoglycan in human adult arterialsmooth muscle cells. Proc. Nati. Acad. Sci. USA, 84: 5287-5291, 1987. 55.

30. Ignotz, R. A., and Massagué,J. Cell adhesion protein receptors as targetsfor transforming growth factor ßaction. Cell, 5/: 189-197, 1987.

Bassols, A., and Massagué,J. Transforming growth factor ßregulates theexpression and structure of extracellular matrix condroitin/dermatan sulfateproteoglycans. J. Biol. Chem., 263: 3039-3045, 1988.Penttinen, R. P., Kobayashi, S., and Bornstein, P. Transforming growthfactor ßincreases mRNA for matrix protein both in the presence and absenceof changes in mRNA stability. Proc. Nati. Acad. Sci. USA, 85: 1105-1108,1988.Rossie, P., Kargenty, G., Roberts, A. B., Roche, N. S., Sporn, M. B., andDeCrombrugghe, B. A nuclear factor 1 binding site mediates the transcrip-tional activation of a type I collagen promoter by transforming growth factor-0. Cell, 52:405-414, 1988.Gospodarowicz, D., and Cheng, J. Heparin protects basic and acidic FGFfrom inactivation. J. Cell. Physiol., 128:475-484, 1986.Bashey, R. I., and Fleischmajer, R. Increased synthesis of hyaluronic acid byinsulin in embryonic chick skin. Proc. Soc. Exptl. Biol. Med., 145: 18-20,1974.Murota, S-i., Abe, M., and Otsuka, K. Stimulatory effect of prostaglandinson the production of hexosamine-containing substances by cultured fibroblasts: (3) Induction of hyaluronic acid synthetase by prostaglandin F2„.Prostaglandins, 14:983-991, 1977.Sisson, J. C., Castor, C. W., and Klavons, J. A. Connective tissue activation.XVIII. Stimulation of hyaluronic acid synthetase activity. J. Lab. Clin. Med.,96: 189-197, 1980.Smith, T. J. n-Butyrate inhibition of hyaluronate synthesis in cultured humanfibroblasts. J. Clin. Invest., 79: 1493-1497, 1987.Bronson, R. E., Bertolami, C. N., and Siebert, E. P. Modulation of fibroblastgrowth and glycosaminoglycan synthesis by interleukin-1. Collagen Reí.Res.,7:323-332, 1987.Castor, C. W., Miller, J. W., and Walz. D. A. Structural and biologicalcharacteristics of connective tissue activating peptide (CTAP-III): a majorhuman platelet derived growth factor. Proc. Nati. Acad. Sci. USA, 80: 765-769, 1983.Nevo, Z. Somatomedins as regulators of proteoglycan synthesis. Connect.Tissue Res., 10: 109-113, 1982.Smith, T. J., Murata, Y., Horowitz, A. L., Philipson. L., and Refetoff, S.Regulation of glycosaminoglycan synthesis by thyroid hormone in vitro. 3.Clin. Invest., 70: 1066-1073, 1982.Moscatelli, D., and Rubin, H. Hormonal control of hyaluronic acid production in fibroblasts and its relation to nucleic acid and protein synthesis. J.Cell Physiol., 91: 79-88, 1976.Smith, T. J. Dexamethasone regulation of glycosaminoglycan synthesis incultured human skin fibroblasts. Similar effects of glucocorticoid and thyroidhormones. J. Clin. Invest., 74: 2157-2163, 1984.Munaim, S. I., Klagsbrun, M., and O'Toole, B. P. Fibroblast growth factor

activity in the early stages of development of chick embryo limb buds. J. CellBiol.,/05:275a, 1987.Myers, S. L. Canine platelet factors stimulate glycosaminoglycan synthesisby synovial cell and organ cultures. Connect. Tissue Res., 14: 41-49, 1985.Tomida, M., Koyama, H., and Ono, T. Effects of adenosine 3':5'-cyclic

monophosphate and serum on synthesis of hyaluronic acid in confluent ratfibroblasts. Biochem. J., 162: 539-543, 1977.Knudson, W., Biswas, C., and O'Toole, B. P. Stimulation of glycosaminoglycan production in murine tumors. J. Cell. Biochem., 25: 183-196, 1984.Knudson, W., and O'Toole, B. P. Properties of a membrane-bound tumor

cell factor which stimulates hyaluronate synthesis in fibroblasts. J. Cell Biol.,70/:339a, 1985.Stoker, M., Gherardi, E., Perryman, M., and Gray, J. Scatter factor is afibroblast-derived modulator of epithelial cell mobility. Nature (Lond.), 327:239-242, 1987.Liotta, L. A., Mandler, R., Murano, G., Katz, D. A., Gordon, R. K., Chiang,P. K., and Schiffmann, E. Tumor cell autocrine motility factor. Proc. Nati.Acad. Sci. USA, 83: 3302-3306, 1986.Schorr, S. L., Schor, A. M., Grey, A. M., and Rushton, G. Foetal and cancerpatient fibroblasts produce an autocrine migration-stimulating factor notmade by normal adult cells. J. Cell Sci., 90: 391-399. 1988.Adams, E. F., Newton, C. J., Tait, G. H., Braunsberg, H., Reed, M. J., andJames, V. H. T. Paracrine influence of human breast stromal fibroblasts onbreast epithelial cells: secretion of a polypeptide which stimulates reductive17/3-oestradiol dehydrogenase activity. Int. J. Cancer, 42: 119-122, 1988.Adams, E. F., Newton, C. J., Braunsberg, H., Shaikh, N., Ghilchik, M., andJames, V. H. T. Effects of human breast fibroblasts on growth and 17/3-estradiol dehydrogenase activity of MCF-7 cells in culture. Breast CancerRes. Treat., //: 165-172, 1988.Stern, R., Huey, G., Rutka, J., Hall, J., and Smith, H. S. Hyaluronic aciddeposition in response to TGF/3 distinguishes normal from breast cancer-derived fibroblasts. J. Cell Biol., 707: 593a, 1989.

3505

Research. on September 5, 2021. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

1989;49:3499-3505. Cancer Res   Margaret Decker, Ernest S. Chiu, Charles Dollbaum, et al.   Fetus and from Breast Cancer PatientsHyaluronic Acid-stimulating Activity in Sera from the Bovine

  Updated version

  http://cancerres.aacrjournals.org/content/49/13/3499

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/49/13/3499To request permission to re-use all or part of this article, use this link

Research. on September 5, 2021. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from