Decrease in Rabbit Bladder Mucosal Glycoprotein after Oophorectomy

0022-5347 /84/1322-0380$02.00/0 Vol. 132, August Printed in U.S.A.

THE JOURNAL OF UROLOGY

Copyright© 1984 by The Williams & Wilkins Co.

DECREASE IN RABBIT BLADDER MUCOSAL GL YCOPROTEIN AFTER OOPHORECTOMY KENNETH WEISMAN, HUGH CALLAHAN, HARRY S. COOPER, ROBERT W. FRITZ AND S. GRANT MULHOLLAND* From the Departments of Urology, Biochemistry and Pathology, Thomas Jefferson University Hospital, Philadelphia, Pennsylvania

ABSTRACT

Hormonal manipulation has been shown to result in less efficient bladder clearance of bacteria. We describe the use of a double antibody technique to semiquantitatively demonstrate the diminution of bladder glycoproteins from the transitional epithelium in oophorectomized rabbits. Rabbit bladder glycoprotein was isolated and used to immunize Swiss-Webster mice. Bladders of normal and oophorectomized rabbits were sequentially stained with mouse antirabbit sera and fluorescein ated goat antimouse antibody. A significant loss of bladder epithelial glycoprotein was evident in oophorectomized rabbits as compared to controls. This study, utilizing a semiquantitative immunologic staining technique, suggests an intact glycoprotein layer is important in the bladder defense mechanism. The transitional cell epithelium of the bladder surface normally interacts with, and eliminates, potential pathogens. Thus, an inoculum of 109 organisms may be eliminated with a 99.9 per cent efficiency in the absence of urinary stasis. 1 It has been suggested that the 1st line of defense against large, as well as small but continuously administered inoculations of bacteria would have to be a relatively nonspecific mechanism, such as the antibacterial properties of urine, the washout effect or the antibacterial action of vesicle mucosa. 2- 5 In addition, there appears to be an antiadherence property of the urothelium which is as yet incompletely understood. We have demonstrated that bacterial adherence to the bladder can be influenced by alteration of the glycoprotein layer after treatment with acid6 or a reducing agent. 7 Furthermore, resistance to bacterial adherence can be restored by application of an exogenous glycoprotein, heparin. 8 This glycoprotein layer is presumably secreted by the transitional epithelium. Mastroianni and other investigators have shown that the secretion of mucin by oviductal mucosa in the rabbit is reduced by 70 per cent after oophorectomization. 10 We have recently demonstrated that oophorectomization correlated wen with a decreased ability to clear bacteria from the bladder. 9 We present evidence that oophorectomization also leads to thinning of the glycoprotein layer covering the urothelium. MATERIALS AND METHODS

Animals. Female New Zealand white rabbits (2 to 3 kg.) were purchased. Oophorectomies were performed as previously described. 9 Swiss-Webster mice were also obtained and maintained on a normal diet which included acidified water. Isolation of glycoprotein (GPl) fraction. Bladders were removed from groups of 10 to 12 animals. The mucosal layers were dissected from the muscularis, suspended in 0.15 M NaCl and stirred gently at room temperature for 3 hours. The suspension was dialyzed vs. water overnight and then centrifuged (1200 g, 30 min.) and the supernatant lyophilized. The average yield of crude extract was 15 mg./bladder. Reextraction of the 1200 g pellet with 0.1 M NaCl did not yield a significant amount of material. The crude extract was dissolved in 0.1 M phosphate buffered saline, pH 7.2 (PBS) and filtered through a column of ACA 34 (1.5 cm. x 72 cm.) (LKB). One ml. fractions were Accepted for publication March 19, 1984. Supported by the Dr. Ralph and Marion Falk Research Fund. * Requests for reprints: Dept. of Urology, Thomas Jefferson University Hospital, 1025 Walnut St., Rm. 1112, Philadelphia, PA 19107.

collected and absorbances determined at 220 nm. and 280 nm. The excluded peak from this column was subsequently chromatographed on ACA 22 column (0.9 cm. X 55 cm.). Fractions eluting just after the void volume were pooled and designated GPl. This material amounted to 10 per cent of the crude extract. Antisera. Antisera to mouse immunoglobulins (GAMig) were raised in goats. Normal mouse serum was obtained from nonimmunized mice in our colony. Rabbit serum albumin (RSA) was coupled to sepharose 4BCL by the cyanogen bromide method. 11 Goat antirabbit serum albumin was obtained from Paul Maurer, Jefferson Medical College. For the mice immunizations, GPl (2 mg./ml. PBS) was emulsified with an equal volume of complete Freunds adjuvant, and 50 µl. injected into each hind footpad. The animals were bled from the retro-orbital plexus after 2 weeks and then before every booster immunization (10 µg. in incomplete Freund's adjuvant approximately every 4 weeks). Enzyme linked immunosorbent assay (ELISA). Microtiter plates were coated overnight (0.2 ml./well) with purified GPl (1 µg./ml.) diluted in .01 M NaHCOs buffer pH 9.6 with .02 per cent NaN 3 • The plate was incubated at 4C overnight on a rocking platform and then washed 10 times with 0.05 M PBS, pH 7.4 containing 0.02 per cent NaN 3 0.05 per cent Tween 20 to avoid nonspecific absorption (PBS-Tween). Antiserum dilutions were prepared in 0.05 M PBS pH 7.4 and then added to each antigen coated well. As a control, antiserum was also added to wells which had been coated with rabbit serum albumin. The plates were rocked at 37C for 60 min., again washed 10 times with PBS-Tween, and excess alkaline phosphatase conjugated goat antimouse immunoglobulin in PBS-Tween was added. Enzyme substrate was prepared by dissolving 1 tablet (5 mg.). of p-nitrophenyl phosphate (Sigma) in 5 ml. of 10 per cent diethanolamine buffer which had been warmed to room temperature. Diethanolamine buffer was made in the following way: 97 ml. of diethanolamine (Mallinkrodt), 100 mg. ofMgCb · 6H 2 0 and 0.2 gm. NaN 3 were dissolved in 800 ml. of water. This buffer was stored at 4C in the dark. The enzyme substrate reaction was monitored by absorption at 405 nm. after 15, 30 and 45 minutes in a Titertech Multiscan. Chemical and physical analyses. Ultraviolet and visible ab. sorption data were obtained on either a Cary 14 or Varian 635 spectrophotometer. Carbohydrate content was estimated colorimetrically from reducing sugar values as measured by the

380

381

BLADDER GLYCOPROTEIN

Park-Johnson method12 after hydrolysis (llOC, 1 NHCl) for 12 hours, with a glucose standard. Analytical ultracentrifugation experiments were performed in a Beckman Model E analytical ultracentrifuge at 20C. Preparation of bladder staining. Six 3 kg. female New Zealand white rabbits were oophorectomized and 6 were kept as controls. Ninety days postoperatively, the animals were sacrificed. Bladders were removed and fixed in formalin, embedded in paraffin and slides were prepared in the usual manner. Specimens were then stained using a double antibody technique (indirect immunofluorescence). After deparaffinization and rehydration slides were treated with varying dilutions of antiglycoprotein sera and incubated at room temperature for 20 minutes, washed with PBS, incubated with fluoresceinated goat antimouse IgG (Litton Bionetics) for 20 minutes at room temperature, and saline washed. Sections were examined using a Leitz-Ortholux epifluorescent microscope with a KP500 exciter filter and K510 barrier filter. They were scored for intensity of staining and extent of epithelium stained. Scoring of slides. After staining, Kodachromes were taken of the specimens. Two observers who were familiar with the staining technique but unaware of the hormonal status of the animal viewed each Kodachrome in random order. Observers were asked to arbitrarily rate each slide for overall intensity of staining on a scale from O to 10. Judgments as to the intensity and extent of the mucosa stained were made by our pathologist upon 1st viewing the specimens. Again, the pathologist was unaware of which animal was under consideration. RESULTS

Gel filtration on AcA 34 resulted in 3 peaks. The 1st (I) was eluted in the void volume, the 2nd (II) eluted with a molecular weight of approximately 70,000 and the 3rd (III) eluted at approximately the included volume. Peak II was determined to be almost entirely composed of albumin by the dye binding assay (American Monitor) as well as reactivity with antirabbit serum albumin antisera. Peak I was refractionated on an AcA 22 column and resulted in the resolution of essentially 1 peak, eluted in the void volume, with small amounts of material eluting at the included volume. These experiments would suggest an approximate molecular weight of 1 x 106 or greater. In an attempt to more rigorously define the molecular weight of GPl, gel filtration was carried out on supports of Sepharcryl 300 S and Sepharose 4B. Since in both cases it eluted in the void volume, approximation of size was attempted by analytical ultracentrifugation. Sedimentation velocity experiments indicated an S value of 4.5, contrary to what would be expected from gel filtration. Ultraviolet absorption spectra revealed no absorption peaks in the 250-260 nm. range, suggesting the absence of nucleic acids. Peak absorption occurred at approximately 280 nm. and was used to calculate an E, 1 per cent cm. of 11. Carbohydrate content, after hydrolysis, ranged from 27 to 30 per cent (as glucose). Table I contains ELISA data obtained from a pool of several mouse sera. Since RSA was the major contaminant of our immunogen preparations (peak II), the sera were routinely absorbed with RSA-linked Sepharose immunoabsorbents, and also tested against RSA coated plates. As can be seen, the

TABLE 1.

Binding of anti-GPI serum to GPI in ELISA* A.oo

in Plates Coated With:

Dilution of Anti-GPl 1:50 1:250 1:500 1:1000

* See text for ELISA assay.

RSA

GPl

0.073 0.032 0.031 0.008

1.324 1.110 0.880 0.653

TABLE 2.

Inhibition of binding of anti-GP] serum to GPI coated plates by soluble GPI as measured by ELISA

µg. of Soluble GPl per ml. of

% .Inhibition

Serum* 0.169 0.242 0.405 0.515 0.754 0.774 0.783 0.879 0.880

50 25 10 5

0;5 0.25 0.10 0.05 0

81 73 54 42 14 12 11 0

* Soluble GPl was preincubated with a 1:500 dilution of antiserum before being added to the plate (see text).

TABLE

3. Binding of mouse antiglycoprotein antisera to bladders from normal and oophorectomized rabbits

Control Animals #1 #2 #3 #4 #5 #6 90Day Oophorectomized Animals #1 #2 #3 #4 #5 #6

Dilution of Mouse Antiglycoprotein

Dilution of Intensity GAMig* of Staining

Extent of Staining

Observers' Scores

1:25 1:50 PBS 1:25 1:50 1:25 1:50 1:25 1:25 1:25

1:40 1:20 1:20 1:40 1:20 1:40 1:20 1:40 1:40 1:40

+3 +3 0 +3 +3 +3 +3 +3 +3 +4

100% 100% 0 70% 60% 100% 80% 80% 60% 80%

7,7 6,4 0,0 7,7 7,7 7,8 7,8 7,7 7,6 7,7

1:25 1:50 1:25 1:50 1:25 1:25 1:50 1:25 1:20 1:25 1:50

1:40 1:20 1:40 1:20 1:40 1:40 1:20 1:40 1:20 1:40 1:20

+1 +o +1 +1 +1 +1 +o +1 +1 +o +1

15% 15% 30% 30% 40% 25% 25% 30% 40% 15% 15%

0,1 1,1 0,1 0,0 4,5· 0,0 0,1 1,1 4,5 0,1 1,1

• Fluoresceinated goat antimouse gammaglobulin.

antisera gave a significant reaction with GPl even at dilutions of 1:1000. In every case, the reaction with RSA is less than 6 per cent of that with GPL Table 2 shows that the specific binding can be inhibited by preincubating the sera with GPL As can be seen, 50 per cent inhibition can be effected with 5 to 10 µg. soluble GPL In addition, RSA could not induce any si~ifi~ant inhibition of binding to GPl (data not shown), which 1s to be expected from the lack of reactivity of this pool with RSA coated plates (table I). As shown in table 3, oophorectomization resulted in a significant loss of bladder epithelial glycoprotein. Epithelium of control animals stained brightly, demonstrating a broad continuous band of fluorescence indicating large concentrations of glycoprotein. In both dilutions of reactants, all control animals stained with an intensity of either +3 or +4 on a scale of O to +4. Staining took place over 60 per cent to 100 per cent of the epithelial surface. On the other hand, the bladders from oophorectomized animals stained weakly, 0 to + 1, and over much less of the epithelial surface, 15 per cent to 40 per cent. The ratings of Kodachromes by blind observers correlated well with the pathologists' impressions upon. viewing the original slides. Results were subjected to the Wilcoxan Rank Sum Test and found to be statistically significant; p < .01 for dilutions of 1:25 of primary antibody, and of 1:40 of the secondary antibody and <.05 for dilutions of 1:50 and 1:20. '

382

WEISMAN AND ASSOCIATES

Both groups showed speckled staining of the cytoplasm of capillary endothelial cells, as well as staining of erythrocytes in capillaries close to the urothelium. Slides treated with sera ofunimmunized mice or with PBS rather than antiglycoprotein antibody showed no fluorescence. DISCUSSION

The incidence of urinary tract infection during the life cycle of women is well documented. Kass et al. have shown that the incidence of urinary tract infections falls in the prepubertal age group and rises after menopause. 13 while Marshall and Lindfoot have found a correlation between a low estrogen/progesterone ratio and incidence of urinary tract infections. 14 In addition, Schaeffer et al. have demonstrated cyclic patterns in bacterial adherence to transitional epithelial cells obtained from midstream urines of healthy females which correlated with the menstrual cycle in that adherence increased early in the cycle and decreased at ovulation. 15 Since bacterial adherence is an important step in the establishment of urinary tract infection, these clinical observations would certainly support the laboratory findings linking a thick glycoprotein layer to antiadherence properties of the bladder. In order to more clearly define such a relationship we have begun biochemical studies of the mucosal layer of rabbit bladders. The results presented here indicate that it is possible to

FIG. 1. Control rabbit bladder, 250x, stained using indirect immunofluorescence technique specific for bladder GAG. Mucosa! GAG and capillaries stain brightly.

isolate a glycoprotein from the mucosa of female rabbit bladders which is immunogenic in mice. This material, which contains 27 to 30 per cent carbohydrate, behaves in an anomalous fashion on molecular sieve columns,. that is, it is not included in gels of very large pore size. However, sedimentation velocity ultracentrifugation studies reveal a single low molecular weight species of approximately 4.5s. Although complete physicochemical studies are not yet completed, it is possible that this material is a charged polyelectrolyte, and behaves as such. 16 We have found GPl to possess an unusual stability between pH3 and 5 at temperatures of lOOC (unpublished observations). The glycoprotein GPl has proven to be immunogenic in mice, producing a humoral response measurement by ELISA. At the present time we do not know in which portion of the molecule the specificity resides, although acid and enzymatic degradation studies are underway. Our fluorescent staining studies (figs. 1, 2 and table 3) would suggest that the glycoprotein which we have isolated is most likely a surface component of the bladder epithelium, since the fluorescent label is localized mainly in that area. In addition to this, we have observed the almost complete loss of PAS positive material from the mucosal layer after our saline extraction procedure (data not shown). Diffuse areas of staining in other areas of the cell could be attributed to nonspecific interaction with some tissue components or the specific interaction of antibody with newly synthesized glycoprotein. The decrease in extent and intensity of fluorescent labelling observed after oophorectomy is in agreement with our earlier results obtained by nonimmunological techniques. 9 For the first time we have shown that the loss of ovarian control mechanisms leads to the loss of a specific glycoprotein from the bladder surface. This loss correlates to the inability of the bladder to effectively clear bacteria. This should facilitate our future studies aimed at elucidating the mechanisms involved in bacterial colonization of this organ. In our experiments it was evident that erythrocytes in capillaries in all sections of both control and oophorectomized animals fluoresced. This indicates that some antigenetic determinants were shared between the bladder urothelium MPS and RBC surface. The extent ofthis.crossreacting is currently being examined. The immunological staining method described here has several advantages over the use of the more traditional PAS technique 17 employed in the localization of glycoprotein. This technique is extremely sensitive and, observing the intensity of the fluorescence obtained, a semi-quantitative judgment of the amount of glycoprotein may be made. ·Also this material is specific and visualizes only those substances in the bladder mucosa which are extracted by the gentle saline wash designed to simulate the normal shedding of bladder glycoprotein. REFERENCES

1. Norden, C. W., Green, G. M. and Kass, E. H.: Antibacterial

2. 3. 4. 5. 6. 7. FIG. 2. Oophorectomized rabbit bladder, 25X. Mucosa! staining seen in fig. 1 is absent. Capillaries still stain brightly.

8.

mechanisms of the urinary bladder. J. Clin. Invest., 47: 2684, 1968. Mulholland, S. G.: Lower urinary tract antibacterial defense mechanism. Invest. Urol., 17: 93, 1979. Kaye, D.: Antibacterial activity of human urine. J. Clin. Invest., 46: 1078, 1967. Hinman, F., Jr. and Cox, C. E.: The voiding vesical defense mechanism: the mathematical effect of residual urine, voiding interval and volume of bacteriuria. J. Urol., 96: 491, 1966. Drach, G. W. and Cox, C. E.: Immunofluorescence studies of the bladder defense mechanism. I. Localization of escherichia coli in the human bladder. J. Am. Geriat. Soc., 15: 1114, 1967. Parsons, C. L., Greenspan, C. and Mulholland, S. G.: The primary antibacterial defense mechanism of the bladder. Invest. Urol., 13: 72, 1975. Shrom, S. G., Parsons, C. L. and Mulholland, S. G.: Vesical defense: further evidence for a charge related mucosa! anti-adherence mechanism. Surg. Forum, 29: 632, 1978. Hanno, P. M., Parsons, C. L., Shrom, S. H., Fritz, R. and Mulhol-

BLADDER GLYCOPROTEIN

90 100

lL 120 130

land, So Go: The effect of in experimental bladder infectiono Surgo Reso, 25: 324, Mulholland, So Go, Qureshi, So Mo, Fritz, R WO and Silverman, K: Effect of hormonal deprivation on the bladder defense mechanismo Jo UroL, 127: 1010, 19820 Mastroianni, L, Beer, Fo, Shah, Uo and Cleve, To R: Endocrine regulation of oviductal secretions in the rabbito Endocrinology, 68: 92, 196L Liberti, Po Ao: Incremental combining site filling of anti-polypeptide antibodieso Immunochemistry, 12: 303, 19750 Park, Jo To and Johnson, L: A submicrodetermination of glucoseo Jo BioL Chemo, 181: 149, 19490 Kass, K R, Savage, No Do and Santamarina, R Ao Go: The

140 150

160 170

383

significance of bacteriuria in preventive medicineo In: Progress in Pyelonephritiso Edited by Do R Kass, Philadelphia, Davis, 19650 Marshall, So and Lindfoot, Jo: Influence of hormones on urinary tract infectionso Urology, 9: 675, 19770 Schaefferm, Ao Jo, Amundsen, So K and Schmidt, Lo No: Adherence of Escherichia coli to human urinary tract epithelial celk InfecL ImmmL, 24: 753, 19790 Gibbons, R Ao, in Glycoproteinso Edited by Ao Gottschalk, Elsevier, New York, ppo 31-140, 19720 Shrom, So Ho, Parsons, Co L and Mulholland, So Go: Role of urothelial surface mucoprotein in intrinsic bladder defenseo Urology, 9: 526, 19770