Main

Hypercalciuria, nephrocalcinosis, and osteopenia occur in some patients with NBS(1,2). Both excessive absorption of dietary calcium and inappropriate bone resorption appear to be contributory, the former presumably resulting from elevated blood levels of calcitriol [1,25 (OH)2D] caused by prostaglandin E2 stimulation of 1-hydroxylase(2,3). Indomethacin treatment reduces hyperprostaglandinuria and partially corrects hypercalciuria, but such patients continue to excrete excessive urinary calcium derived from bone(3). Although the cause of inappropriate bone resorption is unknown, both serum and urine from patients with NBS contain a factor that decreases bone calcium uptake in a bioassay system and increases serum calcium levels in rats after parenteral injection(4). This activity is bound by an anion-exchanger (ECTEOLA-cellulose) from which it is eluted by 0.5 M NaCl. Eluates are inactivated by heparitinase and by protamine but not by trypsin. Size-exclusion filtration suggests an Mr between 10 and 30 kD(4).

These findings suggest that the calciotropic effect is related to heparan or to a structurally similar GAG acting alone or in conjunction with a cytokine. Heparin is known to accelerate bone resorption(5–7), probably by enhancing the stimulatory effect of bFGF on osteoclasts(7–9). In the bioassay used here, bFGF antibody blocks the calciotropic effect of heparin(10). In the present study, we used chromatographic techniques to investigate three possible reasons for the bone-resorbing activity of NBS urine: 1) an excess of a GAG normally present, 2) a unique GAG, or 3) a cytokine dependent on a GAG for its activity.

METHODS

Urine samples. Three subjects provided urine samples. One was a 12-y-old boy with NBS (patient 2 inRef. 3). He was receiving treatment with indomethacin, 6 mg · Kg-1 · d-1, when the urine was collected. Two were normal males, ages 10 and 70 y. Twenty-four-hour refrigerated urine samples were collected without preservative and stored at -20°C until used. Informed consent was obtained from subjects and/or their caretakers in accord with a protocol approved by the SUNY Health Science Center Institutional Review Board for the Protection of Human Subjects.

Materials. Dimethylmethylene blue was purchased from Aldrich, Milwaukee, WI; heparin from Elkins-Sinn, Cherry Hill, NJ and from Sigma Chemical Co., St. Louis, MO; recombinant human bFGF from R&D Systems, Minneapolis, MN; and bFGF fragment 120-125 from Peninsula Laboratories, Belmont, CA. DEAE-Sephacel, Sepharose CL-6B, and heparin-Sepharose were obtained from Pharmacia, Piscataway, NJ. All other reagents were purchased from Sigma Chemical Co., including heparitinase (Heparinase III, Heparitinase I, EC 4.2.2.8) from Flavobacterium heparinum and monoclonal human anti-bFGF.

Bioassay. The bone disk bioassay method has been described in detail elsewhere(4,11,12). Four to six 6-mm disks from young rat clavaria were incubated for 10 min at 38°C in 0.25 mL of PBS or in the solution to be assayed. Individual disks were removed from the solution, blotted dry on absorbent paper, and incubated for 15 min at 38°C in 0.2 mL of PBS containing 1.25 mM calcium and 2.0 mM phosphate. The disks were then removed, and the calcium concentration in the incubation medium was measured by fluorometric titration(13). Uptake was calculated from the change in calcium concentration, and the result was expressed as a ratio: E/C, E being uptake by disks preincubated in sample and C the uptake by littermate disks preincubated in PBS. Substances that increase bone resorption, such as PTH, give E/C values of <1.0. Those that enhance calcium uptake, such as calcitonin and estradiol, give values >1.0(4,11).

Concentration of calciotropic activity from urine. ECTEOLA-cellulose was washed and equilibrated with 0.05 M Tris-HCl (pH 7.5) containing 0.1 M NaCl(14). Forty-milliliter aliquots of urine were mixed with 4.0 g of ECTEOLA-cellulose that was then collected by centrifugation, washed twice with PBS, and suspended in 4 mL of 0.5 M NaCl. The suspension was shaken for 2 h at room temperature, and the ECTEOLA-cellulose was collected by centrifugation. The supernate was used as "ECTEOLA eluate."

Purification of the calciotropic activity from urine

Ion-exchange chromatography. The calciotropic activity was concentrated and purified by ion-exchange chromatography on 0.7 × 5-cm columns of DEAE-Sephacel equilibrated with 8 M urea, 0.05 M sodium acetate, pH 6.0, containing 0.5% Triton X-100 (urea buffer)(15). Up to 600 mL of urine was applied to a column at a rate of approximately 30 mL/h. The chromatogram was developed with a 200-mL linear gradient from 0.0 to 0.7 M NaCl in urea buffer. Fractions of 2-3 mL were collected and analyzed for GAG content with dimethylmethylene blue(16) and for calciotropic activity with the bone disk assay as described above.

The fractions with calciotropic activity were combined and diluted with urea buffer to give an NaCl concentration of <0.1 M. The diluted fractions were then applied to a 0.7 × 5-cm column of DEAE-Sephacel. The GAG was eluted with 1.0 M NaCl in urea buffer. Fractions of approximately 1 mL were collected and assayed for GAG. The GAG was usually eluted in one or two fractions. These were combined and stored at 4°C or at -20°C until used.

Gel-filtration chromatography. Fractions concentrated as just described were then chromatographed on 1.5 × 125-cm columns of Sepharose CL-6B equilibrated with 0.15 M NaCl in urea buffer. Fractions of 2.5 mL were collected and analyzed for GAG and calciotropic activity as before. The active fractions were combined and concentrated by passage through columns of DEAE-Sephacel as described above. Some concentrated fractions were incubated at room temperature for 18 h in 0.5 N NaOH to determine whether or not the GAG was attached to a protein core(17).

Heparitinase digestion and heparin-Sepharose chromatography. Active fractions from the Sepharose CL-6B columns were pooled and dialyzed against two changes of 80 vol of 0.1 M NaCl in 0.01 M HEPES, pH 7.5. Two 1-mL aliquots of the dialyzed solution were then incubated for 16 h at 43°C. Heparitinase, 4 U, was added to one tube. The reaction mixtures were then applied to 0.7 × 5-cm columns of heparin-Sepharose equilibrated with 0.6 M NaCl, 0.01 M Tris-HCl, pH 7.0(18). The heparin-Sepharose was washed with 25 mL of 0.6 M NaCl, 0.01 M Tris-HCl, pH 7.0, and 25 mL of 1.0 M NaCl, 0.01 M Tris-HCl, pH 7.0, and then eluted with a 30-mL linear gradient from 1.0 to 2.0 M NaCl in 0.01 M Tris-HCl, pH 7.0. Fractions were collected and analyzed for calciotropic activity with and without added heparin.

Effect of purified urine fractions on serum calcium concentration in rats. The hypercalcemic effect of a purified urine fraction containing calciotropic activity shown by disk assay was tested in 20-26-g rats by intraperitoneal injection of 0.1 mL 90 min before serum sampling. Serum calcium levels were estimated by fluorometric titration(13), and the differences between means were assessed by t test.

RESULTS

Functional characterization of the calciotropic activity from a subject with NBS. Table 1 shows that heparin reduces the calcium uptake of bone disks. This effect, which is maximal at approximately 0.6 U/mL, is prevented by bFGF receptor blockade with fragment 120-125(19) bFGF antibody and heparitinase. bFGF alone, at 10 ng/mL, also reduces uptake and its effects are prevented by receptor blockade and by antibody. The calciotropic effect of an ECTEOLA eluate of NBS urine is eliminated by heparitinase, bFGF receptor blockade, and bFGF antibody, supporting the hypothesis that its activity is due to a combination of GAG and bFGF. In two other subjects previously reported(4), the calciotropic activity unmodified urine samples was inhibited by FGF antibody and by FGF receptor blockade (data not shown).

Table 1 Effects of heparitinase, bFGF antibody, and bFGF receptor blockade on bone disk* responses to calciotropic activity
Full size table

The finding that bFGF antibody prevents the calciotropic effect of heparin suggested that heparin could be acting in concert with intrinsic bFGF in the disk matrix. To test this possibility, disks were preincubated with bFGF antibody before use in the assay. This did not affect calcium uptake in the disks incubated with buffer (Table 2) but eliminated the calciotropic effect of heparin. Antibody-treated disks did not respond to bFGF at 1 ng/mL, but their uptake was reduced by a combination of 1 ng/mL of bFGF with heparin at 0.25 or 0.5 U/mL. These results indicate that the calciotropic effect of heparin in this assay is due to a combination of GAG and intrinsic bFGF.

Table 2 Responses of bone disks preincubated with bFGF antibody*
Full size table

Purification of the calciotropic activity from urine

Ion-exchange chromatography. Figure 1 shows the results of DEAE-Sephacel chromatography of urine from a 12-y-old boy with NBS, a normal 10-y-old boy, and a normal man. As expected(20), the children excreted more GAG than the adult. The concentration was higher in the normal boy's sample, but polyuria in the NBS subject made GAG excretion nearly equal in the two juveniles (data not shown). In all three subjects, GAG eluted in a broad complex peak. Individual plots differed somewhat in their initial slopes but were closely similar in peak locations and declining slopes. In the NBS subject, bioassay showed calciotropic activity in fractions eluted between 0.27 and 0.32 M NaCl (Fig. 1 and Table 3). Neither normal subject showed activity. However, fractions eluted between 0.2 and 0.36 M NaCl in the boy and 0.22 and >0.3 M in the man became active after addition of 1 ng/mL bFGF, a concentration inactive alone. Because not all fractions were tested, it is possible that an even wider range of bFGF-reactive GAG is present in normal urine. The degree of calciotropic activity was roughly proportional to the amount of GAG present, but the data are insufficient for quantitation.

Figure 1
figure 1

DEAE-Sephacel chromatography of urine from a subject with NBS (•), a normal 10-year-old boy (○), and a normal adult male (x). The chromatograms were performed as described in "Methods." The left ordinate is the OD at 525 nm measured by the dimethylmethylene blue method. A 100-µL aliquot of each fraction was used for the NBS and adult male subjects. A 25-µL aliquot of each fraction was used for the normal boy. The right ordinate and the broken line indicate the NaCl concentration in the eluant. The shaded area shows where the calciotropic activity was found in the effluent fractions from the NBS subject.

Full size image
Table 3 Calciotropic activity of chromatographic fractions
Full size table

Gel-filtration chromatography. When active fractions from the NBS subject were combined and chromatographed on Sepharose CL-6B, a single peak of metachromatic material was eluted (Fig. 2). This contained calciotropic activity (Table 3). Treatment of the active NBS fraction with 0.5 N NaOH for 16 h at room temperature did not alter its Sepharose CL-6B elution pattern, suggesting that the compound under study is a GAG rather than a proteoglycan(17).

Figure 2
figure 2

Gel filtration on Sepharose CL-6B of the calciotropic fractions from ion-exchange chromatography. The gel filtration was carried out as described in "Methods." The ordinate shows the OD at 525 nm measured by the dimethylmethylene blue method, using 100-µL aliquots of the effluent. The shaded area indicates the fractions with calciotropic activity. Phenol red was eluted from the column as indicated by the bar at the upper right.

Full size image

Effect of bFGF antibody on the active fractions. Experiments were performed to determine the effect of bFGF antibody on the active fractions obtained by DEAE-Sephacel chromatography (Table 3). Aliquots of 1.0 mL were treated with 800 mg of ECTEOLA-cellulose and bFGF antibody was added to the mixture. The ECTEOLA-cellulose and bFGF antibody was by centrifugation, washed twice with 8 vol of PBS, and then eluted with 1 vol of 0.5 M NaCl. Control experiments showed that this procedure removed excess antibody. The calciotropic activity of the ECTEOLA eluates was then determined. bFGF antibody reduced the calciotropic activity of the fractions so treated.

Effect of heparitinase on the active fractions. Concentrated active fractions prepared by DEAE-Sephacel chromatography followed by gel filtration on Sepharose CL-6B were treated with heparitinase as described in "Methods." Heparitinase treatment eliminated the calciotropic activity, but this activity could be restored by addition of heparin to the assay system (Table 3).

To further characterize the active component in the fractionated urine, larger aliquots of the purified fractions were incubated with heparitinase in an attempt to release bound bFGF, because it is known that heparitinase digestion will release biologically active bFGF from extracellular matrix(21). The reaction mixtures were then subjected to chromatography on heparin-Sepharose columns as described under "Methods." The results are presented in Table 4. The calciotropic activity of fractions not incubated with heparitinase was recovered when the column was washed with 0.6 M NaCl in 0.01 M Tris-HCl buffer at pH 7.0, and no activity was eluted with subsequent increases in NaCl concentration of the eluant whether tested without added heparin (data not shown) or with added heparin (Table 4). In contrast, no calciotropic activity was found in the 0.6 M NaCl washes from the column to which the heparitinase-digested aliquot was added, either when tested alone (data not shown) or with added heparin (Table 4). When these columns were eluted with the NaCl gradient, fractions were obtained in which high levels of calciotropic activity could by shown if heparin was added to the assay system, but no activity was detected in fractions assayed without added heparin. Treatment of the fractions with antibody to bFGF completely eliminated the calciotropic activity (data not shown). These results show that the calciotropic activity isolated from NBS is due to a combination of bFGF and a GAG.

Table 4 Heparin-Sepharose chromatography of concentrated urinary calciotropic activity from an NBS subject
Full size table

Potential calciotropic activity of inactive GAG fractions from NBS urine

In the preceding experiments, the heparitinase-sensitive GAG component of active fractions was eluted from DEAE-Sephacel in a relatively narrow band near the middle of the peak of metachromatic material. Experiments were carried out to determine whether or not the metachromatic material in nearby inactive fractions could interact with the bFGF from active fractions to manifest calciotropic activity.

Accordingly, an active fraction was treated with heparitinase to destroy the GAG component. It was then mixed with a GAG-containing inactive fraction and the mixture was tested for calciotropic activity (Table 5). The calciotropic activity of the active fraction was destroyed by heparitinase but was restored by the addition of an inactive GAG-containing fraction eluted later from the DEAE-Sephacel column. This experiment indicates that GAG fractions not associated with bFGF but capable of interacting with bFGF are excreted by NBS patients.

Table 5 Potential calciotropic activity of inactive GAG fractions from NBS urine
Full size table

Effects of addition of bFGF to normal urine before chromatography. The data from the experiments on the normal subjects showed that normal urine contains GAG capable of interacting with bFGF to generate calciotropic activity (Fig 1). To determine the chromatographic behavior of the resulting complex, sufficient bFGF was added to urine from the normal subjects to achieve a concentration of 10 ng/mL. The urines were then chromatographed on columns of DEAE-Sephacel as described under "Methods." The elution pattern of metachromatic material was compared with that of urine with no added bFGF chromatographed on the same columns. Fractions were also tested for calciotropic activity. In both normal subjects, the elution pattern of metachromatic material from untreated urine was similar to that illustrated in Figure 1 and was unchanged by the addition of bFGF. Calciotropic activity was detected in fractions from the urine of both subjects to which bFGF had been added (Table 6). In the urine from the boy, activity was present in fractions eluted from the column with between 0.30 and 0.32 M NaCl, and, in the urine from the man, activity was eluted between 0.26 and 0.32 M NaCl. These elution patterns correspond to those obtained on chromatography of urine from the NBS subject in which the activity was eluted with 0.27-0.32 M NaCl. These experiments show that normal urine contains GAG that develop calciotropic activity in combination with bFGF and that the active complex has chromatographic properties closely similar to the calciotropic activity from NBS urine.

Table 6 Calciotropic activity of chromatographic fractions from normal urine to which bFGF was added
Full size table

Effect of purified urine fraction on the serum calcium concentration in rats. An experiment was performed to determine the in vivo effects on serum calcium levels of chromatographically purified calciotropic activity from NBS urine. A fraction of NBS urine purified by ion-exchange chromatography and subsequently concentrated gave an E/C ratio of 0.7 at a 1:250 dilution. One-tenth milliliter of undiluted fraction or of PBS was injected intraperitoneally into 20-26-g rats. After 90 min, the serum calcium level was 2.45 mmol/L (SEM 0.02) in 10 uninjected rats, 2.44 mmol/L (SEM 0.05) in eight rats injected with PBS, and 2.64 mmol/L (SEM 0.05, p versus control group <0.002) in nine rats injected with the active fraction. These results show that the calciotropic activity can be shown in intact animals as well as with the bone disks.

DISCUSSION

The data presented are consistent with the hypothesis that the calciotropic effect of urine from the NBS subject is the result of bFGF bound to a GAG susceptible to heparitinase digestion. The experiments summarized in Table 1 show that the calciotropic effect of heparin in the disk assay is removed by heparitinase digestion. The calciotropic effect of heparin is also prevented by a bFGF fragment that blocks the bFGF receptor or by antibodies to bFGF itself, supporting the concept of essential interaction of heparin and bFGF in the disk assay. Heparitinase treatment, bFGF receptor blockade, and antibody neutralization of bFGF also eliminate the calciotropic activity concentrated from NBS urine by treatment with ECTEOLA-cellulose.

The data in Table 2 show that the calciotropic effect of heparin in the disk assay can be eliminated by preincubation of the disks with bFGF antibody to neutralize intrinsic bFGF. Antibody-treated disks remain responsive to a combination of bFGF and heparin and to the calciotropic factor isolated from NBS urine, providing further support for the hypothesis that bFGF and a GAG act together to generate calciotropic activity.

The calciotropic activity of urine from the NBS subject was concentrated and partially purified by ion-exchange chromatography and gel filtration. As expected, calciotropic activity of this partially purified product was also removed by heparitinase and by bFGF antibody (Table 3). The active material obtained by the gel-filtration step was treated with heparitinase to release bound bFGF and then subjected to chromatography on heparin-Sepharose columns(18) (Table 4). The calciotropic activity of the preparation not treated with heparitinase did not bind to the heparin-Sepharose and was recovered in the initial wash. After heparitinase treatment, calciotropic activity could only be detected if heparin was added to the fractions. The GAG-dependent activity was retained on the heparin-Sepharose but could be eluted with concentrations of NaCl between 1.0 and 2.0 M, as expected for bFGF(18). It was neutralized by bFGF antibody, further confirming the identity of this component as bFGF. The in vivo hypercalcemic effect of concentrated NBS urine fractions with high levels of activity in the bone disk assay suggests that the bioassay results are clinically relevant.

The similarity of metachromatic patterns in chromatographic fractions of normal and NBS urines (Fig. 1) suggests that there are not large qualitative or quantitative differences in GAG content. The similar elution patterns for calciotropic activity in NBS and bFGF-supplemented control urines also suggest that bFGF excess rather than an aberrant GAG may be responsible for the inappropriate mobilization of bone calcium, which in some subjects leads to hypercalciuria, nephrocalcinosis, and osteopenia.

The precise chemical nature of the urinary GAG that interacts with bFGF has not been determined. As shown here (Tables 5 and 6), urinary GAG from normal subjects interact with bFGF to inhibit calcium uptake in the bone disk assay. Further, as shown in Table 5, fractions from NBS urine that are inactive in the bone disk assay will restore the calciotropic activity of active fractions made inactive by treatment with heparitinase. A number of studies have identified some of the structural features necessary for GAG to interact with bFGF and bFGF receptors(22–27). Our experiments show that GAG, which have the ability to interact with bFGF to develop calciotropic activity, are excreted in normal urine. These are eluted from DEAE-Sephacel columns over a broad range of NaCl concentrations, suggesting that significant structural variation is consistent with this property, as has been found by others for GAG supporting different bFGF activities(22–27). Similarly, in urine from the NBS subject, GAG are excreted that have no calciotropic activity in the disk assay and, therefore, are assumed to be associated with little or no bFGF but are capable of interacting with exogenous bFGF to develop calciotropic activity. The observation that GAG capable of interacting with bFGF to develop calciotropic activity are eluted outside the narrow band of calciotropic activity found in eluates of both NBS urine and normal urine containing bFGF is not explained in the present study. This finding could be due to variations in GAG affinity for bFGF because of specific differences in primary structure or to differences in size and, therefore, charge of GAG molecules. These points are currently under investigation.

Some patients with Bartter syndrome have loss-of-function mutations in one of three different genes: the Na-K-2Cl transporter NKCC2(28), the ATP-sensitive K channel ROMK(29,30), or the chloride channel CLCNKB(31). Patients with NBS have an NKCC2 or ROMK mutation(30,32). Further research will be necessary to establish the relationship of these mutations to the hypercalciuria of NBS.