Cyst Fluid Proteases LEO KESNER, WANSHANG YU, AND H. LEON BRADLOW a Department of Biochemistly State University of New York Health Science Center at Brooklyn 450 Clarkson Avenue Brooklyn, New York 11203 "The Rockefeller University 1230 York Avenue New York, New York 10021
INTRODUCTION Cyst fluid obtained by aspiration from women with gross cystic breast disease has been shown to contain a number of unusual proteins.'-5 We have determined that cyst fluid contains a number of active proteases.6 The major protease fraction (HD1) has chymotryptic cleavage characteristics and contains about two-thirds of the total protease activity. This fraction either corresponds to or is tightly bound to the progesterone binding protein (GCDFP-24) when pooled cyst fluid is fractionated according to the scheme outlined by Haagensen.'
MATERIALS AND METHODS Afinity Column Chromatography HDl (100 pL) was layered on columns (0.7 X 10 cm) containing a 10 ml bed of each of the following supports: Sepharose CL-6B-200(Sigma);p-aminobenzamidinemethyl agarose (Sigma); aprotinin-agarose (Sigma); and 6-aminocaproyl-~-tryptophan ester-Sepharose 4B (Pierce). TES buffer (25 mM, pH 7.5) was added, and 2 mL fractions were collected. Elution with 0.5 M NaCl was begun after tube 6. Protease activity was measured on 200 pL aliquots from each tube using ''C-albumin as substrate.6
PLASMA INHIBITION OF CYST FLUID PROTEASE Fresh plasma (EDTA-citrate anticoagulant) was washed with 10 volumes in 3 buffer exchanges (25 mM TES, pH7.5) on a stirred cell using a 10 kDa cutoff membrane 198
KESNER et al.: CYST FLUID PROTEASES
199
(Spectra-Por). Cyst fluid was treated in a similar fashion. Plasma and cyst fluid were diluted 1/5 with buffer and the incubation mixture consisted of 0.5 mL of dilute plasma and 0-0.5 mL volumes of dilute cyst fluid in a total volume of 2 mL. Under these conditions, the plasma proteins acted as the substrate. After 60 min at 37T, 1 mL 15% trichloroacetic acid (TCA) was added. After centrifugation in the cold, the supernates are diluted 1:l with 1% sodium dodecyl sulfate (SDS) to remove the turbidity caused by cyst fluid lipids. The absorbance at 280 nm was read, and the absorbance from a similar tube to which TCA was added at zero time was subtracted from the values obtained with the incubated samples.
DIFFUSION PLATE ASSAY FOR CYST FLUID PROTEASE ACTIVITY Preparation of Coomassie blue-bovine serum albumin substrate: to 4 mL of a bovine serum albumin solution (10 mg/mL in 100 mM sodium phosphate, pH 7.0 buffer) was added 1.2 mL Coomassie brilliant blue (10 mg/mL in buffer). After 30 minutes, 200 pL TCA (70% in water) was added. The precipitate was washed several times with water to remove excess color and TCA, and it was then dissolved in 4 mL of buffer. Preparation of diffusion plates: 300 mg of agar were dissolved in 20 mL of 50 mM sodium phosphate buffer, pH 7.4. This was done with gentle heating (below 90'C). Seven milliliters buffer and 3 mL of the Coomassie blue-albumin complex solution were added, and the warm mixture poured into plastic diffusion plates and allowed to gel. Ten-microliter wells were punched with a pipet tip, and 5 p L cyst fluid added. The plates were covered and then kept at room temperature for 15-20 hours. The reaction was stopped by adding an overlay of 3% acetic acid, which was then discarded after an hour. The diameter of the ring formed was a measured of the protease activity in the fluid. A calibration curve was generated by using several concentrations of trypsin (0-100 pg/mL) and plotting the log of concentration vs. ring diameter.
RESULTS AND DISCUSSION Afinity Column Separations We have previously reported that HD1 cross-reacts with a polyclonal antibody, prepared by Dr. Haagensen, to GCDFP-24.6 Two possibilities exist: (1) that the protease activity is intrinsic to GCDFP-24 protein; (2) that the protease and GCDFP24 are tightly bound and not separated by the many purification steps used to prepare them. A number of affinity columns have been found to be effective for the isolation of protease enzymes.' Chromatography of HD1 on a few of these columns yielded the protease profiles 1. All of the protease activity was retained by the column material shown in FIGURE when p-aminobenzamidine or aprotinin was the ligand, but eluted with the solvent front when columns containing D-tryptophan methyl ester or underivatized Sepharose
ANNALS NEW YORK ACADEMY OF SCIENCES
200
6B were used. The bound protease could be recovered from the p-aminobenzamidine and aprotinin columns by eluting with 0.5 M NaC1. If the protease activity in HD1 resided in a protein that readily dissociated from GCDFP-24, it should have been retained on the aprotinin or p-aminobenzamidine columns while the GCDFP-24 should have appeared in the effluent. This did not happen. Mast cell preparations have been shown to contain the serine protease tryptase which is inhibited by the endogenous associated protein trypstatin.8 Both are released in a fixed ratio to histamine after immunoglobulin E (IgE) dependent activation. The tryptase-trypstatin complex is eluted as a single peak, in a 1:l complex, on a TSK G3000SW column. However when the complex was applied to an aprotinin-Sepharose column, the tryptase was retained and trypstatin (3600 M,) was recovered in the effluent. We have repeated this experiment using HD1. HDl was added to both aprotinin and benzamidine columns where they were bound. The effluent was lyophilized and tested for inhibitory properties against the HD1 released by NaCl elution after the excess salt was removed from the HDI by dialysis. No inhibitor activity was detected (data not shown).
Plasma Inhibition of Cyst Fluid Protease Activity
About 10% of all plasma proteins are protease inhibitors? Two of the major plasma protease inhibitors were tested for their ability to inhibit cyst fluid protease activity. Although a,-macroglobulin was a good inhibitor of HD 1, a,-protease inhibitor was not.6 Since human plasma contains many other protease inhibitors, it was important to see how well plasma proteins inhibited total cyst fluid protease activity.
-
Sephorose 0-Try 0---0 Benzomidine Aprotinin 0--0
%E -- I I 0
u)
0.5M NoCl
3U
3 0
I
z n
20
0
10
0 I
3
5
7
9
II
TUBE NUMBER FIGURE 1. Affinity column chromatography of HDl protease from breast cyst fluid.
KESNER et al.: CYST FLUID PROTEASES
0.2
20 1
7
E C 0 0.15-
a0 insulin > casein, trypsin cleaved platelets). TABLE casein > insulin > albumin, and calpain cleaved casein > albumin > insulin.
ANNALS NEW YORK ACADEMY OF SCIENCES
202
Semiquantitative Estimation of Cyst Fluid Protease Activity Plate diffusion methods for detection of protease activity are commercially available (Bio-Rad 5OO-OOll-OG34). In general, these rely upon an agar-imbedded protein as substrate into which several small wells containing the protease are placed. The solution in the wells is allowed to incubate for several hours, and the plates are then treated with acetic acid which causes the unhydrolyzed protein to appear opaque while the hydrolyzed protein is clear. Cyst fluid, with its high protein, lipid, and pigment content, does not show up well under these circumstances. However, the protease activity in a large number of cyst fluid samples or fractions from a column can be easily visualized if a Coomassie blue-albumin complex is used as the substrate as described in the Methods section.
TABLE 1. Relative Reactivity
High K HD1 High Na HD1 Trypsin Calpain a
of "'C-Proteins Cleaved by each Protease"
I4C-Albumin
I4C-Insulin
I4C-Casein
16 18 1
9 12
1 1 6 14
4
3 1
Activity was determined as cpm of C14 not precipitated by TCA.
SUMMARY AND CONCLUSIONS The precise origin of breast cyst fluid remains obscure. Molina has presented evidence that type I1 cysts (high Na/K ratio) may be transudative, that is, partly derived from plasma elements which enter through gap junctions, while Type I cysts (high K/Na ratio) are primarily secretory." In transudative cysts, plasma protease inhibitors may be present, but the balance between protease and its inhibitors may fluctuate as a result of as yet undetermined circumstances. An imbalance between the protease activity of cyst fluid and its inhibitors may be involved in the pathogenesis of breast gross cystic disease. Accumulation of protein fragments with resistant bonds would produce an elevated oncotic pressure causing a shift of fluid into the cyst capsule. Albumin is a good substrate for the protease, which may account for its low concentration in cyst fluid. The major protease fraction closely corresponds to the progesterone binding protein (GCDFP-24) described by Haagensen. Affinity columns containing aprotinin or benzamidine ligands retain the protease which can then be eluted with 0.5 M NaC1. The HD1 protease and progesterone binding protein are either tightly complexed or are the same protein. Cyst fluid is a complex mixture of biomolecules. If the progesterone binding protein is a protease, many questions must be answered concerning the influence of cyst fluid steroids, lipids, anions, and cations on enzyme action. Determination of the amino acid sequence of
KESNER ef al.: CYST FLUID PROTEASES
203
HD1 may help elucidate the source of the enzyme and its relationship to other tissue proteases. Human plasma contains inhibitors of this protease activity. When pooled, dialyzed plasma was mixed with pooled, dialyzed cyst fluid, the ratio of plasma/cyst fluid at which all activity was inhibited was 6 / 1. A comparison of the rate of cleavage of three ''C-protein substrates shows that cyst fluid proteases cleave in a characteristic manner, distinct from either trypsin or calpain. A simple method for semiquantitative estimation of protease activity in cyst fluid is described which utilizes prestained Coomassie blue-albumin containing agarose gel plates. All cyst fluids tested had protease activity but showed variability in their ability to cleave I4C-albumin by a factor of 4. There is much direct and indirect evidence that proteases are involved in the cancer process. In view of the higher than normal incidence of breast cancer in women who have had gross cystic breast disease, the possibility exists that an imbalance between these proteases and their inhibitors is somehow involved.
REFERENCES 1.
2. 3.
4.
5.
6. 7.
HAAGENSEN, D. E., G. MAZOUJIAN, W. G. DILLEY,C. E. PEDERSON, S. J. KISTER& S. A. WELLS.1979. Breast gross cystic disease fluid analysis. Isolation and radioimmunoassay for a major component protein. I. Nat. Cancer Inst. 62: 239-247. ZANGERLE, P. F., F. SPYRATOS, V. LEDOUSSAL, G. NOEL,K. HACENE,J. C. HENDRICK, 1986. Breast cyst fluid proteins and breast cancer. Ann. J. GEST& P. FRANCHIMONT. N.Y. Acad. Sci. 464: 331-349. COLLETTE,J., J-C. HENDRICK, J. M. JASPER& P. FRANCHIMONT. 1986. Presence of alactoalbumin, epidermal growth factor, epithelial membrane antigen and gross cystic disease fluid protein (15,000 daltons) in breast cyst fluid. Cancer Res. 46: 3728-3733. GAIRARD, B., C. M. GROS,C. KOEL& R. RENAUD.1983. Proteins and ionic components in breast cyst fluids. In Endocrinology of Cystic Breast Disease. A. Angeli, H. L. Bradlow & L. D. Dogliotti, Eds.: 191-195. Raven Press. New York, N.Y. YAP,P. L., W. R. MILLER,M. M. ROBERTS,R. J. CREEL,B. FREEDMAN, C. L. MIRTLE, 1984. Protein concentration in fluid from E. A. D. PRYDE& D. B. L. MCCLELLAND. gross cystic disease of the breast. Clin. Oncol. 10: 35-43. KESNER,L., W. Yu, H. L. BRADLOW, C. N. BREED& M. FLEISHER.1988. Proteases in cyst fluid from human gross cyst breast disease. Cancer Res. 48: 6379-6383. WILCHEK,M., T. MIRON& J. KOHN.1984. Affinity chromatography. Methods Enzymol. 104 3-55.
8.
KIDO,H., N. FUKUSEN & N. KATUNAMA. 1985. Chymotrypsin- and trypsin-type serine proteases in rat mast cells: properties and functions. Arch. Biochem. Biophys. 239 436-443.
HARPEL,P. C. & M. S. BROWER.1983. a,-Macroglobulin: an introduction. Ann. N.Y. Acad. Sci. 421: 1-8. 10. MOLINA,R.,X. FILELLA,M. HERSANZ,M. PRATS, A. VELASCO,G. ZANON,M. J. MARTINEZ-OSABA & A. M. BALLESTA. 1989. Biochemistry of cyst fluid in fibrocystic disease of breast: approach to classification and understanding of the mechanism of formation. Ann. N.Y. Acad. Sci. (This volume.) 9.
INTRODUCTION Cyst fluid obtained by aspiration from women with gross cystic breast disease has been shown to contain a number of unusual proteins.'-5 We have determined that cyst fluid contains a number of active proteases.6 The major protease fraction (HD1) has chymotryptic cleavage characteristics and contains about two-thirds of the total protease activity. This fraction either corresponds to or is tightly bound to the progesterone binding protein (GCDFP-24) when pooled cyst fluid is fractionated according to the scheme outlined by Haagensen.'
MATERIALS AND METHODS Afinity Column Chromatography HDl (100 pL) was layered on columns (0.7 X 10 cm) containing a 10 ml bed of each of the following supports: Sepharose CL-6B-200(Sigma);p-aminobenzamidinemethyl agarose (Sigma); aprotinin-agarose (Sigma); and 6-aminocaproyl-~-tryptophan ester-Sepharose 4B (Pierce). TES buffer (25 mM, pH 7.5) was added, and 2 mL fractions were collected. Elution with 0.5 M NaCl was begun after tube 6. Protease activity was measured on 200 pL aliquots from each tube using ''C-albumin as substrate.6
PLASMA INHIBITION OF CYST FLUID PROTEASE Fresh plasma (EDTA-citrate anticoagulant) was washed with 10 volumes in 3 buffer exchanges (25 mM TES, pH7.5) on a stirred cell using a 10 kDa cutoff membrane 198
KESNER et al.: CYST FLUID PROTEASES
199
(Spectra-Por). Cyst fluid was treated in a similar fashion. Plasma and cyst fluid were diluted 1/5 with buffer and the incubation mixture consisted of 0.5 mL of dilute plasma and 0-0.5 mL volumes of dilute cyst fluid in a total volume of 2 mL. Under these conditions, the plasma proteins acted as the substrate. After 60 min at 37T, 1 mL 15% trichloroacetic acid (TCA) was added. After centrifugation in the cold, the supernates are diluted 1:l with 1% sodium dodecyl sulfate (SDS) to remove the turbidity caused by cyst fluid lipids. The absorbance at 280 nm was read, and the absorbance from a similar tube to which TCA was added at zero time was subtracted from the values obtained with the incubated samples.
DIFFUSION PLATE ASSAY FOR CYST FLUID PROTEASE ACTIVITY Preparation of Coomassie blue-bovine serum albumin substrate: to 4 mL of a bovine serum albumin solution (10 mg/mL in 100 mM sodium phosphate, pH 7.0 buffer) was added 1.2 mL Coomassie brilliant blue (10 mg/mL in buffer). After 30 minutes, 200 pL TCA (70% in water) was added. The precipitate was washed several times with water to remove excess color and TCA, and it was then dissolved in 4 mL of buffer. Preparation of diffusion plates: 300 mg of agar were dissolved in 20 mL of 50 mM sodium phosphate buffer, pH 7.4. This was done with gentle heating (below 90'C). Seven milliliters buffer and 3 mL of the Coomassie blue-albumin complex solution were added, and the warm mixture poured into plastic diffusion plates and allowed to gel. Ten-microliter wells were punched with a pipet tip, and 5 p L cyst fluid added. The plates were covered and then kept at room temperature for 15-20 hours. The reaction was stopped by adding an overlay of 3% acetic acid, which was then discarded after an hour. The diameter of the ring formed was a measured of the protease activity in the fluid. A calibration curve was generated by using several concentrations of trypsin (0-100 pg/mL) and plotting the log of concentration vs. ring diameter.
RESULTS AND DISCUSSION Afinity Column Separations We have previously reported that HD1 cross-reacts with a polyclonal antibody, prepared by Dr. Haagensen, to GCDFP-24.6 Two possibilities exist: (1) that the protease activity is intrinsic to GCDFP-24 protein; (2) that the protease and GCDFP24 are tightly bound and not separated by the many purification steps used to prepare them. A number of affinity columns have been found to be effective for the isolation of protease enzymes.' Chromatography of HD1 on a few of these columns yielded the protease profiles 1. All of the protease activity was retained by the column material shown in FIGURE when p-aminobenzamidine or aprotinin was the ligand, but eluted with the solvent front when columns containing D-tryptophan methyl ester or underivatized Sepharose
ANNALS NEW YORK ACADEMY OF SCIENCES
200
6B were used. The bound protease could be recovered from the p-aminobenzamidine and aprotinin columns by eluting with 0.5 M NaC1. If the protease activity in HD1 resided in a protein that readily dissociated from GCDFP-24, it should have been retained on the aprotinin or p-aminobenzamidine columns while the GCDFP-24 should have appeared in the effluent. This did not happen. Mast cell preparations have been shown to contain the serine protease tryptase which is inhibited by the endogenous associated protein trypstatin.8 Both are released in a fixed ratio to histamine after immunoglobulin E (IgE) dependent activation. The tryptase-trypstatin complex is eluted as a single peak, in a 1:l complex, on a TSK G3000SW column. However when the complex was applied to an aprotinin-Sepharose column, the tryptase was retained and trypstatin (3600 M,) was recovered in the effluent. We have repeated this experiment using HD1. HDl was added to both aprotinin and benzamidine columns where they were bound. The effluent was lyophilized and tested for inhibitory properties against the HD1 released by NaCl elution after the excess salt was removed from the HDI by dialysis. No inhibitor activity was detected (data not shown).
Plasma Inhibition of Cyst Fluid Protease Activity
About 10% of all plasma proteins are protease inhibitors? Two of the major plasma protease inhibitors were tested for their ability to inhibit cyst fluid protease activity. Although a,-macroglobulin was a good inhibitor of HD 1, a,-protease inhibitor was not.6 Since human plasma contains many other protease inhibitors, it was important to see how well plasma proteins inhibited total cyst fluid protease activity.
-
Sephorose 0-Try 0---0 Benzomidine Aprotinin 0--0
%E -- I I 0
u)
0.5M NoCl
3U
3 0
I
z n
20
0
10
0 I
3
5
7
9
II
TUBE NUMBER FIGURE 1. Affinity column chromatography of HDl protease from breast cyst fluid.
KESNER et al.: CYST FLUID PROTEASES
0.2
20 1
7
E C 0 0.15-
a0 insulin > casein, trypsin cleaved platelets). TABLE casein > insulin > albumin, and calpain cleaved casein > albumin > insulin.
ANNALS NEW YORK ACADEMY OF SCIENCES
202
Semiquantitative Estimation of Cyst Fluid Protease Activity Plate diffusion methods for detection of protease activity are commercially available (Bio-Rad 5OO-OOll-OG34). In general, these rely upon an agar-imbedded protein as substrate into which several small wells containing the protease are placed. The solution in the wells is allowed to incubate for several hours, and the plates are then treated with acetic acid which causes the unhydrolyzed protein to appear opaque while the hydrolyzed protein is clear. Cyst fluid, with its high protein, lipid, and pigment content, does not show up well under these circumstances. However, the protease activity in a large number of cyst fluid samples or fractions from a column can be easily visualized if a Coomassie blue-albumin complex is used as the substrate as described in the Methods section.
TABLE 1. Relative Reactivity
High K HD1 High Na HD1 Trypsin Calpain a
of "'C-Proteins Cleaved by each Protease"
I4C-Albumin
I4C-Insulin
I4C-Casein
16 18 1
9 12
1 1 6 14
4
3 1
Activity was determined as cpm of C14 not precipitated by TCA.
SUMMARY AND CONCLUSIONS The precise origin of breast cyst fluid remains obscure. Molina has presented evidence that type I1 cysts (high Na/K ratio) may be transudative, that is, partly derived from plasma elements which enter through gap junctions, while Type I cysts (high K/Na ratio) are primarily secretory." In transudative cysts, plasma protease inhibitors may be present, but the balance between protease and its inhibitors may fluctuate as a result of as yet undetermined circumstances. An imbalance between the protease activity of cyst fluid and its inhibitors may be involved in the pathogenesis of breast gross cystic disease. Accumulation of protein fragments with resistant bonds would produce an elevated oncotic pressure causing a shift of fluid into the cyst capsule. Albumin is a good substrate for the protease, which may account for its low concentration in cyst fluid. The major protease fraction closely corresponds to the progesterone binding protein (GCDFP-24) described by Haagensen. Affinity columns containing aprotinin or benzamidine ligands retain the protease which can then be eluted with 0.5 M NaC1. The HD1 protease and progesterone binding protein are either tightly complexed or are the same protein. Cyst fluid is a complex mixture of biomolecules. If the progesterone binding protein is a protease, many questions must be answered concerning the influence of cyst fluid steroids, lipids, anions, and cations on enzyme action. Determination of the amino acid sequence of
KESNER ef al.: CYST FLUID PROTEASES
203
HD1 may help elucidate the source of the enzyme and its relationship to other tissue proteases. Human plasma contains inhibitors of this protease activity. When pooled, dialyzed plasma was mixed with pooled, dialyzed cyst fluid, the ratio of plasma/cyst fluid at which all activity was inhibited was 6 / 1. A comparison of the rate of cleavage of three ''C-protein substrates shows that cyst fluid proteases cleave in a characteristic manner, distinct from either trypsin or calpain. A simple method for semiquantitative estimation of protease activity in cyst fluid is described which utilizes prestained Coomassie blue-albumin containing agarose gel plates. All cyst fluids tested had protease activity but showed variability in their ability to cleave I4C-albumin by a factor of 4. There is much direct and indirect evidence that proteases are involved in the cancer process. In view of the higher than normal incidence of breast cancer in women who have had gross cystic breast disease, the possibility exists that an imbalance between these proteases and their inhibitors is somehow involved.
REFERENCES 1.
2. 3.
4.
5.
6. 7.
HAAGENSEN, D. E., G. MAZOUJIAN, W. G. DILLEY,C. E. PEDERSON, S. J. KISTER& S. A. WELLS.1979. Breast gross cystic disease fluid analysis. Isolation and radioimmunoassay for a major component protein. I. Nat. Cancer Inst. 62: 239-247. ZANGERLE, P. F., F. SPYRATOS, V. LEDOUSSAL, G. NOEL,K. HACENE,J. C. HENDRICK, 1986. Breast cyst fluid proteins and breast cancer. Ann. J. GEST& P. FRANCHIMONT. N.Y. Acad. Sci. 464: 331-349. COLLETTE,J., J-C. HENDRICK, J. M. JASPER& P. FRANCHIMONT. 1986. Presence of alactoalbumin, epidermal growth factor, epithelial membrane antigen and gross cystic disease fluid protein (15,000 daltons) in breast cyst fluid. Cancer Res. 46: 3728-3733. GAIRARD, B., C. M. GROS,C. KOEL& R. RENAUD.1983. Proteins and ionic components in breast cyst fluids. In Endocrinology of Cystic Breast Disease. A. Angeli, H. L. Bradlow & L. D. Dogliotti, Eds.: 191-195. Raven Press. New York, N.Y. YAP,P. L., W. R. MILLER,M. M. ROBERTS,R. J. CREEL,B. FREEDMAN, C. L. MIRTLE, 1984. Protein concentration in fluid from E. A. D. PRYDE& D. B. L. MCCLELLAND. gross cystic disease of the breast. Clin. Oncol. 10: 35-43. KESNER,L., W. Yu, H. L. BRADLOW, C. N. BREED& M. FLEISHER.1988. Proteases in cyst fluid from human gross cyst breast disease. Cancer Res. 48: 6379-6383. WILCHEK,M., T. MIRON& J. KOHN.1984. Affinity chromatography. Methods Enzymol. 104 3-55.
8.
KIDO,H., N. FUKUSEN & N. KATUNAMA. 1985. Chymotrypsin- and trypsin-type serine proteases in rat mast cells: properties and functions. Arch. Biochem. Biophys. 239 436-443.
HARPEL,P. C. & M. S. BROWER.1983. a,-Macroglobulin: an introduction. Ann. N.Y. Acad. Sci. 421: 1-8. 10. MOLINA,R.,X. FILELLA,M. HERSANZ,M. PRATS, A. VELASCO,G. ZANON,M. J. MARTINEZ-OSABA & A. M. BALLESTA. 1989. Biochemistry of cyst fluid in fibrocystic disease of breast: approach to classification and understanding of the mechanism of formation. Ann. N.Y. Acad. Sci. (This volume.) 9.
