Increased Levels of Fibroblast Growth Factor-Like Activity in Urine from Patients with Bladder Or Kidney Cancer1

Home , FGF3

[CANCER RESEARCH 48, 2083-20S8, April IS, 1988] Increased Levels of Fibroblast Growth Factor-like Activity in Urine from Patients with Bladder or Kidney Cancer1

Gerald W. Chodak,2 Verne Hospelhorn, Sheila M. Judge, Ruth May forth, Hartmut Koeppen, and Joachim Sasse

Departments of Surgery/Urology [G. W. C., V. H., S. M. ]., and R. M.Â¡and Pathology [H. K.J, The University of Chicago and the Pritzker School of Medicine, Chicago, Illinois 60637, and the Department of Surgical Research fJ. SJ, Children's Hospital Medical Center, Boston, Massachusetts 021 IS

ABSTRACT factors may be useful as a marker for this disease. In this paper, we show that men with bladder or renal tumors can be distin Growth factor activity was partially purified from human renal tumors guished from normal men by the presence of higher levels of a and a human bladder cancer cell line by heparin-Sepharose chromatog- heparin-binding, angiogenic-like growth factor in their urine raphy. This activity stimulated bovine capillary endothelial cell prolifer ation and DNA synthesis in BALB/c 3T3 cells. Partially purified growth which is similar to basic fibroblast growth factor. factor preparations from these tumors contained a protein with an ap proximate molecular weight of 17,000 which was recognized by a poly- MATERIALS AND METHODS clonal antiserum raised against a peptide fragment of basic fibroblast growth factor (FGF). This growth factor activity appears to be related to Preparation of Tumor Extracts basic fibroblast growth factor. Measurement of FGF-like activity in 50 urine samples from 32 adult Fresh tumor tissue was obtained through the Department of Pathol males showed that 55% (6 of 11 ) of the urine samples from patients with ogy at the University of Chicago from patients undergoing a radical bladder cancer and 100% (7 of 7) of the urine samples from patients with nephrectomy for adenocarcinoma of the kidney. Tissue was also ob kidney cancer contained activity equivalent to more than 20 ng of basic tained from physicians at the University of Iowa and the University of FGF/h of urine production. In contrast, only 6% (2 of 32) of the urine Texas Medical Center. Samples were frozen immediately and stored at â€”70*Cuntil use. Crude tumor extracts were prepared by thawing the samples from controls, patients with a benign disease, or patients with a history of bladder or kidney cancer contained this level of growth factor tumors, homogenizing them in a blender, and then treating the homog- enate with 2 mg/ml of collagi-naso in 10% phosphate-buffered saline activity. These results suggest that patients with bladder or kidney cancer for 3 h at 37*C. Tumor digests were filtered through 110-pm nylon release an FGF-like factor into urine which may be used as a marker for mesh and sterile filtered through 0.45-Â¿imfilters, and the extracts were these tumors. stored at -70*C.

INTRODUCTION Establishment of Human Bladder Tumor Cell Line (HB-1)

The development of new capillaries plays a fundamental role Tumor tissue fragments were obtained from a male patient undergo in the growth of solid tumors (1). Because tumor-induced ing a transurethral resection of a bladder tumor. The fragments were cut into 1 mm3 pieces using forceps and scissors and placed in RPMI angiogenesis is caused by diffusible factors (2), bodily fluids 1640 (Gibco Laboratories, Grand Island, NY). The pieces were placed surrounding tumors may contain increased levels of angiogenic in two flasks (25 and 75 cm2) in RPMI 1640/10% FBS (Gibco)/10% activity. For example, significantly more angiogenic activity has NCTC 109 (Gibco)/1% penicillin and streptomycin and incubated at been detected in cerebrospinal fluid obtained from patients with 37Â°C. Outgrowths from the expiants were cloned by limiting dilution brain tumors (3), in aqueous humor from patients with eye to remove contaminating normal fibroblasts. Subconfluent monolayers tumors (4), and in urine from patients with bladder cancer (5) of cloned tumor cells were trypsinized, harvested, and resuspended in in comparison to healthy controls and patients with benign RPMI 1640/10% FBS. Athymic nude BALB/c mice (Frederick Cancer disease. Research Institute, Bethesda, MD) were given injections s.c. of 1 to 2 One of the factors that stimulates angiogenesis is basic FGF3 x 10* cells. When the diameter of the tumor reached 2 cm (approxi (6). This protein belongs to a family of proteins that have been mately 10 wk after the cells were injected), the tumor was removed for identified in a wide variety of normal and malignant tissues (7). histolÃ³gica! examination and readaptation to tissue culture. Basic FGF is a 146 amino acid, single chain protein (8) that is Preparation of Ant Â¡-KGFPolyclonal Antiserum cationic, binds to heparin with high affinity, and stimulates the growth of capillary endothelial cells (7). An 11-residue peptide fragment (Arg-Ile-His-Pro-Asp-Gly-Arg- For the last few years, we have studied the relationship Val-Asp-Gly-Val) corresponding to positions 33 to 43 of basic fibro between carcinoma of the bladder and the angiogenic growth blast growth factor (11) was synthesized by solid phase methods (12, factor content in urine (9, 10). Bladder cancer is a recurring 13) using an automated Applied Biosystems 430-A peptide synthesizer. disease, and detection usually requires invasive tests which are The synthetic peptide was conjugated to KLH (Boehringer Mannheim, Indianapolis, IN) according to previously described methods (14) using performed at regular intervals. If the presence of bladder tumors maleimidohexanoyl-iY-hydroxysuccinimide ester as a cross-linking is associated with elevated urinary growth factor levels, these agent. The peptide-carrier conjugate was dialyzed against PBS and stored frozen at -20Â°C. Received 4/23/87; revised 10/23/87, 12/28/87; accepted 1/20/88. The costs of publication of this article were defrayed in part by the payment Polyclonal antibodies were produced by giving 8-lb male New Zea of page charges. This article must therefore be hereby marked advertisement in land White rabbits injections at multiple dorsal i.d. sites of 0.5 mg of accordance with 18 U.S.C. Section 1734 solely to indicate this fact. KLH peptide conjugate emulsified in complete Freund's adjuvant and 1This work supported by NIH Grant CA 33145, the Cancer and Urology boosted at 4- to 6-wk intervals. Research Endowment Fund of the University of Chicago, the Cancer Research Foundation, and a gift from the Johnson & Johnson Corporation. 2To whom requests for reprints should be addressed, at Box 403, the University Collection and Storage of Urine Samples of Chicago, 5841 S. Maryland Ave., Chicago, IL 60637. 1The abbreviations used are: FGF, fibroblast growth factor, FBS, fetal bovine Voided urine was collected and pooled from either healthy males or serum; KLH, keyhole limpet hemocyanin; DM EM, Dulbecco's modified essential male patients with histologically proven Stage B transitional cell car medium; CRF, crude retinal factor, BCE, bovine capillary endothelial; TGF-a, transforming growth factor-Â«;EGF, epidermal growth factor, PDGF, platelet- cinoma of the bladder. The urine was centrifugea at 627 x g for 10 min and then stored at â€”70*C. Individual urine samples were arbitrarily derived growth factor, bFGF, basic fibroblast growth factor, PBS, phosphate- buffered saline. collected from hospitalized male patients, healthy male volunteers, and 2083

Downloaded from cancerres.aacrjournals.org on October 2, 2021. © 1988 American Association for Cancer Research. HBROBLAST GROWTH FACTOR-LIKE ACTIVITY IN HUMAN URINE male patients visiting the outpatient Urology Clinic. Patients were EDTA and seeding the cells in growth medium.Stimulation of cell selected from a single ward in the hospital which is shared by General proliferation was measured by seeding 10,000 to 15,000 capillary Surgery and Urology. To determine the collection period for each endothelial cells/well in a 24-well plate containing 0.5 ml/well of the sample, the time of last urination was recorded, and all subsequent growth medium without CRF. The cells were incubated at 37*C over samples were collected for a minimum period of 6 h. At the end of the night in 10% CO... and the following day the media were replaced with collection period, the time of the patient's last urination was recorded, fresh media also lacking CRF. Test samples (25 to 75 M!)were added and the last specimen was included in the total collection. Each speci to duplicate wells and incubated at 37Â°Cfor72 h. The media were then men was kept on ice during collection and then centrifuged at 627 x g removed, the cells were trypsinized, and the number of cells in each for 30 min to remove cells and debris. The samples were stored sample was determined by counting aliquots of the cell suspension in a individually at -70Â°C until use. Grossly bloody specimens were ex Model Zf Coulter Counter. cluded from analysis (these represented less than 5% of the samples). Each assay included a negative control (cells plus culture medium The age, diagnosis, date of collection, current medications, stage of with no added sample) and a positive control (cells plus culture medium disease (for cancer patients), total volume per sample, and duration of plus 25 n\ of CRF). The results of an experiment were accepted as valid sample collection were recorded for each patient. The stage of each only if CRF induced at least a 50% increase in cell number compared tumor was assessed by radiological methods and/or by histopathologi- to the negative control. cal assessment following biopsy or surgical excision. In some cases, multiple samples were tested from the same patient Western Blot Analysis of Partially Purified Growth Factor to determine interassay variability. To avoid biasing the results, repli Growth factor activity from bladder and kidney tumor extracts was cate samples from the same patient were averaged, and the average partially purified by heparin-afflnity Chromatography. Samples were value was combined with the results from other patients in the same electrophoresed on polyacrylamide gels according to standard proce diagnostic group. dures (19). The gels were overlayed with 0.22-^im-pore nitrocellulose Bio-Rex 70 Cation Exchange Chromatography membranes (Schleicher and Schuell, Keene, NH), and protein transfers were performed at 25 V overnight at 4'C. Prior to immunoblotting, the Crude renal tumor extract was applied to a 1- x 10-cm column membranes were incubated with 3% nonfat dried milk in PBS for 30 containing Bio-Rex 70 (200 to 400 mesh; Bio-Rad) equilibrated with min to prevent nonspecific binding. Reagents were added as previously 0.1 M NaCl/0.01 M Tris-HCl (pH 7.0). Growth factor was eluted with described using anti-FGF (peptide) antibodies (20). AH antibody dilu a 200 ml gradient of 0.1 to 1.0 M NaCI in Tris buffer at a flow rate of tions were performed in PBS containing 1% milk and 0.1% Tween-20. 30 ml/h as previously described (15). Ten Â¿ilofevery third fraction were tested in the 3T3 cell assay. Measurement of Heparin-binding Growth Factor Activity in Individual Human Urine Samples Heparin-Sepharose Chromatography Each urine sample was thawed and centrifuged at 855 x g for 20 min at 4Â°C,and the supernatant was adjusted to 50 itiM Tris (pH 7.0). Tumor Extracts. Partially purified extracts from human renal cell carcinoma or the human bladder cancer cell line, UBI, were applied Each sample was mixed with 0.3 g of heparin-Sepharose on an orbital to 8-ml heparin-Sepharose columns (10x1 cm; Pharmacia) in 0.01 M shaker overnight at 4"(.'. The next morning, the mixture was centrifuged Tris, pH 7.0, at 4'C. The column was washed with 0.01 M Tris/O. I M at 213 x g for 10 min, and the urine was decanted. The heparin- NaCI (pH 7.0). FGF-like growth factors were eluted by either of two Sepharose was transferred to 1- x 5-cm columns (Bio-Rad) and washed methods: initially, a 0.1 to 3.0 M NaCI gradient in 0.01 M Tris (pH 7.0) with 30 ml of 0.6 M NaCl/0.01 MTris (pH 7.0), and growth factor was was used to elute the column. Once we established that tumor-derived, eluted with 2.2 ml of 3.0 M NaCl/0.01 M Tris (pH 7.0). In previous heparin-binding growth factor activity was consistently eluted by 1.2 to studies, we found that this volume was sufficient to elute 90% of growth 1.4 M NaCI, subsequent columns were eluted step-wise with 0.6 M factor activity from the column.4 Two-mi aliquots of the 3.0 M salt NaCI, followed by 1.5 M NaCl/0.01 M Tris (pH 7.0). eluate were transferred to microconcentrator tubes (Amicon, Danvers, Pooled Urine Samples. To determine if urine contained heparin- MA) and centrifuged for 1 h at 4"C. The samples were desalted twice binding growth factor activity, urine samples were thawed, pooled, and by adding 2 ml of 5 mM Tris/0.01 M NaCI to the tubes and centrifuging then mixed overnight with 5 to 8 ml of heparin-Sepharose at 4"( ' on a for 1 h. The final volume of the desalted, concentrated samples was shaker. After decanting the urine, the gel was poured into a 1- x 10-cm adjusted to 170 nl. Fifty /.Â¡\ofeach sample were tested in triplicate in column and washed with 30 ml of 0.6 M NaCI/0.01 M Tris (pH 7.0). the 3T3 cell assay. Growth factor activity was eluted with a 0.1 to 3.0 M NaCI gradient in 0.01 M Tris (pH 7.0). One-mi fractions were collected and stored at -70*C until analyzed. RESULTS Characterization of Human Bladder Tumor Cell Line, IIH-I. Analysis of Growth Factor Activity To determine if human bladder cancer contains bFGF-like 3T3 Cell Assay. Growth factor activity in column eluates was assessed growth factor activity, a human cell line was established. HB-1 by measuring the incorporation of | '11|ili>inuline by quiescent mono- cells grown in nude mice produced tumors that were histologi- layer cultures of BALB/c 3T3 cells as described previously (16). Each cally identical to the original tumor. Both the histology of the plate included control cultures that were fed DMEM containing 4.5 g tumor and the morphology of the tumor cells as determined by of glucose/liter, 50 units/ml of penicillin, 50 Mg/ml of streptomycin, electron microscopy were consistent with transitional cell car 1% /-glutamine, and 10% calf serum. The column eluates were tested cinoma (data not shown). in duplicate by adding 5 or loglio each well. The incorporation of | 'I I|tl>>iirnlim-in the control sample was subtracted from the value for To establish the human origin of the HB-1 cell line, we each test sample. One unit of growth factor activity was defined as one- analyzed the reaction between species-specific monoclonal an half the amount of | 'I I|th\ mulini- incorporation obtained when 50 /il tibodies and the HB-1 cells by fluorescence-activated cell sort of calf serum were added to control cultures. ing. The HB-1 cells reacted strongly with the monoclonal Capillary Endothelial Cell Proliferation Assay. Bovine capillary endo- antibody BBM.l, which recognizes human, but not mouse, /32- thelial cells (clone 76-23, kindly provided by J. Folkman) were grown microglobulin (21), and with antibody W6/32, which recognizes in gelatin-coated dishes as previously described (17) in DMEM/10% a common HLA-A,B,C framework determinant (22). These calf serum/penicillin (50 units/ml)/streptomycin (50 itg/ml)/l% glu- findings indicate that HB-1 cells are of human origin. The tamine/CRF (1 ^1/ml) (18). CRF has been shown to contain acidic tumor cells also reacted strongly with monoclonal antibody FGF which is a mitogen for BCE cells (18). Confluent cultures were passaged each week by detaching the cells with 0.25% trypsin-0.01% 4 Unpublished data. 2084

Downloaded from cancerres.aacrjournals.org on October 2, 2021. © 1988 American Association for Cancer Research. FIBROBLAST GROWTH FACTOR-LIKE ACTIVITY IN HUMAN URINE 9A7, an antibody which is thought to recognize an antigen I 2 present on many human transitional carcinomas (23). Growth Factor Activity in HB-1 Human Bladder Carcinoma Cells and Adenocarcinoma of the Kidney. Previous studies have shown that angiogenic fibroblast growth factors produced by tumors can be partially purified by heparin-affinity chromatog- â€”43,000 raphy (7). To isolate growth factor activity from HB-1 cells, crude extracts (6 mg of protein/ml) of these cells were applied to a heparin-Sepharose affinity column, and protein was eluted â€”25,700 stepwise with increasing concentrations of NaCl. The eluates were tested for growth factor activity in the 3T3 and BCE cell â€”18,400 assays. Although many growth factors stimulate 3T3 cells, only â€”14,300 angiogenic factors are known to stimulate BCE cells. The 1.5 M NaCl eluate contained both 3T3 cell and BCE cell growth factor activity. The 0.6 M eluate stimulated 3T3 cells but not â€” 6,200 BCE cells. Because the protein content of crude kidney tumor extract was high (30 mg of protein/ml), it was necessary to partially 3,000 purify growth factor activity by cation-exchange chromatography on Bio-Rex 70 prior to heparin-Sepharose chromatography A (7). The 3T3 and BCE cell growth factor activities coeluted from Bio-Rex 70 with 0.5 to 0.6 M NaCl in agreement with previous findings (7). Growth factor activity was further purified by chromatography on heparin-Sepharose using a 0.1 to 3.0 M NaCl gradient. The 1.2 to 1.4 M NaCl eluate contained growth factor activity that stimulated both 3T3 and BCE cells (Fig. 1). Western Blot Analysis of FGF-related Proteins in Partially 18.4kD- Purified Preparations from MB-1 Cells and Renal Tumors. To characterize the heparin binding growth factor extracted from I4.3KD- bladder and kidney tumors, we determined the immunoreactiv- ity of these factors with anti-bFGF antiserum. The specificity of the anti-bFGF antiserum was demonstrated by Western blot analysis of recombinant bovine basic FGF (supplied by Amgen I 234567 Corp.) and CRF. Both tumor extracts and bFGF contained a Fig. 2. Western blot of growth factor preparations from bladder tumor cells, protein with approximate molecular weight of 17,000 which kidney tumors, and basic FGF. A, Lane 1: recombinant human bFGF (ISO ng); strongly reacted with the anti-FGF (peptide) antiserum (Fig. Lane 2: prestained protein markers; ovalbumin (M, 43,000); chymotrypsinogen (M, 25,700); 0-lactogIobulin (M, 18,400); lysozyme (M, 14,300); bovine trypsin 2/4). This reaction could be blocked by prior incubation of the inhibitor (M, 6,500); insulin, a and b chains (M, 3,000). In B, kidney tumor factor was fractionated on Bio-Rex 70 prior to heparin-Sepharose fractionation. Lanes 1, 4, and 7:prestained protein markers: Â¿Mactoglobulin(M, 18,400) and lysozyme (M, 14,300). Lanes 2 and 3: heparin-binding growth factor(s) from HB-I bladder tumor cells (total protein, 41 and 82 /jg. respectively). Lanes 5 and 6 contain kidney tumor growth factor that was partially purified on Bio-Rex 70 and heparin- Sepharose (total protein, 10 and 21 ^g, respectively).

antiserum with the FGF peptide that was used as the original antigen. The antiserum did not recognize acidic FGF present in CRF. Thus, the antiserum is specific for basic FGF. Western blot analysis of growth factor activity partially pu S rified by heparin-affinity chromatography from HB-1 bladder eo 5 tumor cells and human kidney tumors showed that both of < these preparations contained a protein with an approximate molecular weight of 17,000 that binds anti-bFGF (peptide) o antibody (Fig. 2B). co I- Growth Factor Activity in Human Urine. We previously pos n tulated that angiogenic factors produced by bladder tumors were released into urine (8). To determine if urine contained a heparin-binding growth factor similar to the bFGF-like factor found in HB-1 bladder cells and kidney tumors, we analyzed human urine samples. Ten liters of urine that were pooled from 1234567 Fraction Number eight patients with bladder cancer were tested for growth factor Fig. 1. Heparin-binding growth factor activity in renal tumor extracts. Human activity after gradient elution from heparin-Sepharose. The kidney tumors were homogenized, adsorbed onto Bio-Rex 70, and eluted with elution profile of urine showed that 3T3 cell growth factor NaCl. Fractions containing growth factor activity were applied to heparin-Seph activity eluted at 1.2 to 1.5 M NaCl similar to the tumor arose columns which were then eluted with an NaCl gradient. Fractions were tested for the ability to stimulate BCE cell proliferation (O) and [3H|thymidine extracts. Urine also contained a second peak of 3T3 cell growth incorporation into BALB/c 3T3 cells (â€¢). factor activity that was eluted by 1.8 to 2.2 M salt (Fig. 3). Only 2085

cancer in one patient and kidney cancer in a second patient by collecting two samples several months apart from both of these patients. The volume of both tumors increased during the collection period. The growth factor activity in each sample was averaged with the entire group. Three samples were tested over several months from another patient who had a history of bladder cancer, during which time he remained free of disease. The result from each sample was averaged with the entire group X 3.0 X of samples. The mean level of growth factor activity in urine X from patients with bladder or kidney cancer was equivalent to X 65 Â±20 ng of bFGF-like activity/h of urine production (Table 2.0 Â¿ 2). Urine from all 7 patients with kidney cancer and 6 of 11 (55%) of the urine samples from patients with bladder cancer contained activity equivalent to more than 20 ng of bFGF-like 1.0 o activity/h of urine production (Fig. 4). The lowest levels of growth factor activity were found in urine from patients with

l l l l i l l l the smallest tumors. 2 3 4 5 6 7 8 9 IO M 12 13 FGF-like activity was measured in urine samples from three FRACTION NUMBER other groups of patients which included healthy male controls, Fig. 3. Gradient elution profile of heparÃn-bindinggrowth factor activity. Ten patients with a benign disease, or a history of bladder or kidney liters of urine from 4 patients with bladder cancer were pooled, concentrated 10- fold, and applied to a heparin-Sepharose column. Growth factor was eluted with cancer (Table 2). Six voided urine samples were tested from a 0.1 to 3.0 M NaCl gradient. Fractions were pooled sequentially in groups of four patients who had a history of bladder cancer. None of three, dialyzed, and tested for 3T3 cell growth factor activity. these patients had evidence of a tumor at the time of urine collection. The mean age of these patients was 64. The mean Table 1 Heparin-binding growth factor activity in pooled urine level of growth factor activity in urine samples from these latter Heparin-binding growth fac four patients was equivalent to 3 Â±3 ng of bFGF-like activity/ tor activity (units) h of urine which was significantly lower than the mean level of Urine source 3T3 cells" BCE cells* heparin-binding growth factor activity in urine from bladder Bladder cancer patients 53 192 and kidney cancer patients (P < 0.005). All of these samples Healthy controls 13 24 contained less than the equivalent of 20 ng of bFGF-like activ Â°One unit of 3T3 cell growth factor activity is defined as one-half the stimu lation of (3H]thymidine incorporation obtained by treatment of the cells with O.OS ity/h of urine production (Fig. 4). ml of calf serum. Fourteen samples from 5 healthy individuals, whose mean * One unit of BCE growth factor activity is defined as one-half the cell number age was 29, contained significantly less heparin-binding growth obtained by treatment of BCE cell cultures with 0.025 ml of CRF. factor activity [3 Â±O(SD) ng of bFGF-like activity/h of urine production] than urine from patients in the bladder/kidney the growth factors present in the 1.2 to 1.5 M salt eluate cancer group (Table 2). The coefficient of variation for replicate stimulated BCE cell proliferation. samples from these five healthy patients ranged from 3 to 8%. To determine whether the amount of heparin-binding growth All 14 urine samples contained less than the equivalent of 20 factor activity in urine was affected by the presence or absence ng of bFGF-like activity/h of urine production (Fig. 4). of a tumor, we compared the amount of 3T3 and BCE cell We also tested 12 urine specimens from 11 patients with growth factor activity in 4 liters of urine that were pooled from benign disease. The diagnoses of these patients included two male patients with histologically proven bladder cancer Crohn's disease (n = 3), benign prostata- hypertrophy, neph- versus the amount in 4 liters of urine from four healthy males. rolithiasis, testicular pain, benign renal cyst, inguinal hernia, The urine pooled from patients with bladder cancer contained benign renal tumor, cecal adenoma, peritonitis, and chronic 4.5-fold and 8-fold more 3T3 cell and BCE cell growth factor renal failure. The mean and median ages of these patients were activity, respectively, than the urine from healthy men (Table 57 and 66, respectively. The mean level of heparin-binding 1). growth factor activity in urine from these individuals was equiv The large difference in growth factor activity between pooled alent to 8 Â±3 ng of bFGF-like activity/h of urine production urine from patients with bladder cancer and healthy males prompted us to measure heparin-binding growth factor activity (Table 1). Only 2 of 12 (17%) of the samples contained more in individual urine samples. The amount of FGF-like growth Table 2 Heparin-binding growth factor activity in urine from adult males factor in each urine sample was determined by comparison to factor ac- (ng a titrÃ¢tioncurve of CRF which contains acidic FGF (18). After of Age tivity (ng bFGF/ bFGF/ these experiments had been completed, recombinant bFGF DiagnosisBladder samples18 urine)046-72range h hurine)0-250 became available, and we subsequently compared the activities or kidney Â±20* 3Â±3Â°8 of pure bFGF and CRF in the 3T3 cell growth factor assay. History of bladder or 61214Growth58-6923-84 0-150-32 kidney cancer The results obtained from this experiment enabled us to express Benign disease Â±3" our previous data in ng of bFGF. Healthy male controlsNo. 23-3965 3Â±(FRange 0-10 FGF-like activity was measured in 18 urine samples collected " Heparin-binding growth factor activity was determined in a 5-h urine sample by measuring the ability of 3 M NaCl eluates from heparin-Sepharose columns to from 9 male patients with bladder cancer and 3 with kidney stimulate incorporation of [3H]thymidine into confluent BALB/c 3T3 cells. cancer. The mean and median ages of the 12 men were 60 and * Mean Â±SE. ' Significantly different from bladder/kidney cancer group at P < 0.005 (Stu 58, respectively. Two of the patients with kidney cancer and 3 dent's t test). of the patients with bladder cancer had metastatic disease at J Significantly different from bladder/kidney cancer group at P < 0.02 (Stu the time of urine collection. We followed the progress of bladder dent's t test). 2086

Downloaded from cancerres.aacrjournals.org on October 2, 2021. © 1988 American Association for Cancer Research. FIBROBLAST GROWTH FACTOR-LIKE ACTIVITY IN HUMAN URINE unlike the tumor-derived growth factors, acidic FGF does not cross-react with anti-bFGF antiserum. TGF-a and EGF are two other angiogenic factors (27) that have been found in urine (28, 29). However, both of these factors have lower molecular weights (approximately 6,000) (30, 31) and lower affinity for 300- heparin-Sepharose (15) than the tumor-derived factors. The bladder and kidney tumor-derived growth factors also differ from the cationic growth factor, PDGF, in affinity for heparin- a Sepharose and ability to stimulate BCE cells; PDGF is eluted C from heparin-Sepharose by 0.6 M salt (15), and it does not â€¢f 200- stimulate BCE cell proliferation. Furthermore, the subunits of PDGF, which are approximately the same molecular weight as the tumor growth factor (32), lack growth factor activity for 3T3 cells. Thus, bladder and kidney tumor-derived growth factors share many of the features of bFGF, and they do not a 100- resemble any other known growth factors. Urine from patients with bladder or kidney cancer also con tained growth factor activity that is similar to FGF. Like the bladder tumor-derived growth factor, urinary growth factor activity had a strong affinity for heparin, and it exhibited mitogenic activity in the BCE and BALB/c 3T3 cell growth factor assays. However, it is premature to conclude that the urinary growth factor is identical to bFGF. In contrast to the growth factors found in bladder and kidney tumors, the urine- 20-- derived growth factor was not recognized by bFGF antiserum â€¢ti; on Western blot analysis. The absence of a reaction may have Bladder Ranal Benign History Normal been due to insufficient protein because of losses during con Cancer Cancer Disease of Controls centration. It is also conceivable that the urinary factor is Bladder or Kidney related, but not identical, to basic FGF. Recently, a gene that Cancer encodes a protein with only partial homology to bFGF was Fig. 4. Heparin binding growth factor activity in individual urine samples. Five-h urine collections were absorbed onto heparin-Sepharose, and the gel was isolated from human bladder cancer cells (33), suggesting that poured into columns. The columns were washed with 0.6 M NaCl, and growth bladder tumors produce more than one class of FGF-related factor was eluted with 3.0 M salt. The samples were tested in duplicate for the molecules, only one of which may be released into urine. ability to stimulate [JH]thymidine incorporation in BALB/c 3T3 cells. The results The large difference in heparin-binding growth factor activity are expressed as ng of basic FGF produced per h. in pooled urine from controls and patients with bladder cancer than the equivalent of 20 ng of bFGF-Iike activity/h of urine led us to measure the level of growth factor in individual urine samples. We found that the differences in growth factor activity production (Fig. 4). Overall, 72% (13 of 18) of the urine detected in pooled samples were also observable in the individ specimens collected from patients with bladder or kidney cancer ual samples. Seventy-two % of the urine samples from patients contained more than the equivalent of 20 ng of bFGF-like with bladder or kidney cancer contained elevated levels (more growth factor/h of urine collection, whereas only 7% (2 of 32) than the equivalent of 20 ng of bFGF-like activity/h of urine of the urine samples from men lacking tumors contained this collection) of growth factor activity, whereas, only 7% of the level of growth factor activity. samples from controls, patients with a benign disease, or a history of bladder cancer contained this high level of activity. DISCUSSION These results are consistent with the hypothesis that angiogenic factors produced by bladder and kidney cancer are released into In previous studies, we demonstrated that human and mouse urine. These data together with data obtained from previous transitional cell carcinoma possesses angiogenic activity (24, studies (9, 15) demonstrate that urine from patients with blad 25). However, the substances responsible for this effect were der cancer contains increased amounts of a heparin-binding not completely characterized. In this study, we have shown that endothelial cell growth factor and a BCE cell motility factor a protein present in extracts of human bladder tumor cells and (9) compared to controls. Both properties are possessed by kidney tumors possesses live of the properties that are charac angiogenic factors. Whether the same molecule is responsible teristic of the angiogenic factor, basic FGF. These include (a) for stimulation in both assays remains to be proven. These similar affinity for heparin-Sepharose, (b) the ability to stimu results suggest that patients with either bladder or kidney cancer late BCE cell proliferation, (c) stimulation of DNA synthesis may be monitored by measuring changes in the urinary level of in BALB/c 3T3 cells, (d) molecular weight of approximately FGF-iike growth factor during the course of their disease. 17,000, and (Ã¨)recognition by anti-basic FGF (peptide) anti Obviously, other sources of urinary growth factor would have bodies. These results show that the tumor-derived growth fac to be excluded. tors are related to basic FGF. Amino acid analysis data will be It is unlikely, however, that this finding is specific for patients required, however, to establish the identity of this protein. with these two tumors. FGF produced by other tumors or These data indicate that the heparin-binding growth factors benign tissues could be released into serum, filtered through found in bladder and kidney tumors are probably unrelated to the kidney, and thus be detected in urine. Our preliminary data other known growth factors. For example, acidic FGF has a suggest that several clinical conditions are associated with high 50% sequence similarity with basic FGF (8, 11, 26). However, levels of basic FGF in urine. For example, high levels of FGF- 2087

bovine brain derived class I heparin-binding growth factor. Science (Wash. like activity were detected in urine from some patients with DC), 235: 945-951, 1986. catheters in the bladder. In addition, the menstrual cycle may 12. Merrifield, R. B. Solid phase peptide synthesis. I. The synthesis of a tetra- influence urinary excretion of bFGF, as suggested by marked peptide. J. Am. Chem. Soc., 85: 2149-2154, 1963. variation in levels among young healthy women.5 Furthermore, 13. Sakakibara, S. Chemistry and Biochemistry of Amino Acids, Peptides, and Proteins, Vol. 1, pp. 51-85. New York, NY: M. Dekker, 1971. urine from some patients with cancers outside the urinary tract 14. Ishikawa, Â£.,Imagawa, M.. Hashida, S., Yoshitake, S., Hamaguchi, Y., and Ueno, T. Enzyme labelling of antibodies and their fragments for enzyme may also contain elevated levels of this factor. The impact of immunoassay and immunohistochemical staining. J. Immunoassay, 4: 209- these conditions on urinary bFGF levels is being studied in 327, 1983. more detail. 15. Chodak, G. W., Shing, Y., Borge, M., Judge, S. M., and Klagsbrun, M. Presence of heparin binding growth factor in mouse bladder tumors and In summary, we have shown that carcinoma of the bladder urine from mice with bladder cancer. Cancer Res., 46: 5507-5510, 1986. and kidney contains a heparin-binding growth factor which is 16. Klagsbrun, M., Longer, R., Levenson, R., Smith, S., and Lillehei, C. The stimulation of DNA synthesis and cell division in chondiocytes and 3T3 cells similar to basic FGF. This factor may partly account for the by a growth factor isolated from cartilage. Exp. Cell Res., 105:99-108, 1977. angiogenic activity of these tumors. Moreover, this study shows 17. Folkman, J., HÂ¡indolisciliId,C. C., and Zetter, B. R. Long-term culture of capillary endothelial cells. Proc. Nati. Acad. Sci. USA, 76:5217-5221, 1979. that urine from patients with bladder or kidney cancer contains 18. D'Amore, P. A., and Klagsbrun, M. Endothelial cell mitogens derived from more FGF-like activity than urine from healthy individuals. the retina and hypothalamus: biochemical and biological similarities. J. Cell Biol., 99: 1545-1549, 1984. 19. Laemmli, W. K. Cleavage of structural proteins during the assembly of the ACKNOWLEDGMENTS head of bacteriophage lambda. Nature (Lond.), 227: 680, 1970. 20. Kurstack, E. Enzyme Immunodiagnosis, p. 48. Boca Raton, FL: Academic We thank Dr. Michel Boileau and Dr. Richard Williams for provid Press, Inc., 1986. 21. Bordsky, F. M., Bodmer, W. F., and Parham, P. Characterization of a ing kidney tumor tissue; Dr. Michael Klagsbrun for helpful discussions; monoclonal anti-b2-microglobulin antibody and its use in the genetic and Maria Rivera, Tom Weyrich, and Paul Keller for technical assistance; biochemical analysis of major histocompatibility antigens. Eur. J. 1mimmi il.. and Diane Torrey for clerical assistance. 9:536-545, 1979. 22. Barnstable, C. J., Bodmer, W. F., Brown, G., Galfre, G., Milstein, C., Williams, A. F., and Ziegler, A. Production of monoclonal antibodies to Group A erythrocytes, HLA, and other human cell surface antigensâ€”new REFERENCES tools for genetic analysis. Cell, 14: 9-20, 1978. 23. Messing, E. M., Bubbens, J. E., Whitmore, K. E., deKernion, J. B., Nestor, 1. Knighton, D., Ausprunk, !>.. Tapper, I)., and Folkman, J. Avascular and M. S., and Fahey, J. L. Murine hybridoma antibodies against human tran vascular phases of tumor growth in the chick embryo. Br. J. Cancer, 35:347- sitional carcinoma-associated antigens. J. Urol., 132:167-172, 1984. 356, 1977. 24. Chodak, G. W., Haudenschild, C., Gittes, R. F., and Folkman, J. Angiogen 2. Greenblatt, M., and Shubik, P. Tumor angiogenesis: transfilter diffusion esis activity as a marker of neoplastic and preneoplastic lesions of the human studies in the hamster by the transparent chamber technique. J. Nati. Cancer bladder. Ann. Surg., 792: 762-771, 1980. Inst., 41: 111-124, 1968. 25. Chodak, G. W., and Summerhayes, I. Detection of angiogenesis activity in 3. Lopez-Poussa, S., Ferrier, L., Vick, J. M., and Domenech-Mateu, J. Angi malignant bladder tissue and cells. Invest. Urol., 132: 1032-1035, 1984. ogenic activity in the CSF in human malignancies. Experientia (Basel), 37: 26. Esch, F.. Ueno, N., Baird, A., Hill, F., Denoroy, L., Ling, N., Gospodarowicz, 413, 1981. D., and Guillemin, R. Primary structure of bovine brain acidic fibroblast growth factor (FGF). Biochem. Biophys. Res. Commun., 133: 554-562, 4. Tapper, D., Langer, R., Bellows, A. R., and Folkman, J. Angiogenesis capacity as a diagnostic marker for human eye tumors. Surgery (St. Louis), 1985. 86: 36-40, 1979. 27. Schreiber, A. B., Marjorie, E., and Derynck, R. Transforming growth factor- 5. Shahabuddin, S., Arnold, F., Costello, C. B., and Kumar, S. Tumor angio alpha: a more potent angiogenic mediator than epidermal growth factor. genesis factor in urologica! cancers. Br. J. Urol., 56: 490-494, 1984. Science (Wash. DC), 232:1250-1254, 1986. 6. Auerbach, R. Angiogenesis-inducing factors. Lymphokines, 4: 69-87, 1981. 28. Starkey, R., Cohen, S., and Orth, D. N. Epidermal growth factor: identifi 7. Klagsbrun, M., and Shing, Y. Heparin-affinity chromatography of anionic cation of a new hormone in human urine. Science (Wash. DC), 189: 800- and cationic capillary endothelial cell growth factors: analysis of hypothala- 802, 1975. mus-derived growth factors and fibroblast growth factors. Proc. Nati. Acad. 29. Twardzik, D. R., Kimball, E. S., Sherwin, S. A., Roncholis, J. E., and Todaro, G. J. Comparison of growth factors functionally related to epidermal growth Sci. USA, Â«2;805-809, 1985. factor in the urine of normal and human tumor-bearing athymic mice. Cancer 8. Abraham, J. A., Mergia, A., Whang, J. L., TurnÃ³lo, A., Friedman, J., Res., 45: 1934-1939, 1985. Hjerrild, K. A., Gospodarowicz, D., and Fiddes, J. C. Nucleotide sequence 30. Marquardt, H., Hunkapiller, M. W., Hood, L. E., and Todaro, G. J. Rat of a bovine clone encoding the angiogenic protein, basic fibroblast growth transforming growth factor type 1: structure and relation to epidermal growth factor. Science (Wash. DC), 233: 545-548, 1986. factor. Science (Wash. DC), 223: 1079-1082, 1984. 9. Chodak, G. W., Scheiner, C. J., and Zetter, B. R. Urine from patients with 31. Cohen, S., and Carpenter, G. Human epidermal growth factor: isolation and transitional cell carcinoma stimulates migration of capillary endothelial cells. chemical and biological properties. Proc. Nati. Acad. Sci. USA, 72: 1317- N. Engl. J. Med., 305: 869-874, 1981. 1321, 1975. 10. Chodak, G. W., Klagsbrun, M., and Shing, Y. W. Partial characterization of 32. Antoniades, H. N. Human platelet-derived growth factor (PDGF): purifica a cell motility from human urine. Cancer Res., 45: 690-694, 1985. tion of PDGF-I and PDGF-II and separation of their reduced subunits. Proc. 11. Strydom, D. J., Harper, J. W., and Lobb, R. R. Amino acid sequence of Nati. Acad. Sci. USA, 75:7314-7317, 1981. 33. Marx, J. L. Oncogene action probed. Science (Wash. DC), 237: 602-603, 5Unpublished results. 1987.

2088

Downloaded from cancerres.aacrjournals.org on October 2, 2021. © 1988 American Association for Cancer Research. Increased Levels of Fibroblast Growth Factor-like Activity in Urine from Patients with Bladder or Kidney Cancer

Gerald W. Chodak, Verne Hospelhorn, Sheila M. Judge, et al.

Cancer Res 1988;48:2083-2088.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/48/8/2083

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

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/48/8/2083. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.