Increased glomerular permeability to negatively charged Ficoll relative to neutral Ficoll in rats

【摘要】  It is established that the glomerular filter sieves macromolecules based on their size, shape, and charge. Anionic proteins are thus retarded compared with their neutral or cationic counterparts. However, recent studies have indicated that charge effects are small, or even "anomalous," for polysaccharides. We therefore investigated the impact of charge on the glomerular permeability to polysaccharides by comparing sieving coefficients (; primary urine-to-plasma concentration ratio) for negatively charged, carboxymethylated (CM) FITC-Ficoll and FITC-dextran with their neutral counterparts. For these probes, were determined in anesthetized Wistar rats [269 ± 2.7 g (±SE; n = 36)], whose ureters were cannulated for urine sampling. The glomerular filtration rate was assessed using FITC-inulin. Polysaccharides were constantly infused, and after equilibration, urine was collected and a midpoint plasma sample was taken. Size and concentration determinations of the FITC-labeled polysaccharides were achieved by size-exclusion HPLC (HPSEC). For CM-Ficoll, was significantly increased (32 times at 55 Å) compared with that of uncharged Ficoll. A small increase in for CM-dextran compared with neutral dextran was also observed (1.8 times at 55 Å). In conclusion, negatively charged Ficoll relative to neutral Ficoll was found to be markedly hyperpermeable across the glomerular filter. Furthermore, negatively charged Ficoll was observed to be larger on HPSEC compared with its neutral counterpart of the same molecular weight. It is proposed that the introduction of negative charges in the "dendrimeric," cross-linked Ficoll molecule may alter its configuration, so as to make it more extended, and conceivably, more flexible, thereby increasing its glomerular permeability.

【关键词】  charge barrier capillary permeability macromolecules fractional clearance reflection coefficients

IT IS GENERALLY ACCEPTED THAT the glomerular filter discriminates among macromolecules based on their size, shape, and net charge ( 6, 8 ). With respect to charge, the permeability of anionic dextran sulfate was found to be reduced and that of cationic, diethylaminoethyl (DEAE) dextran to be increased compared with that of neutral dextran ( 6 ). However, more recent studies have indicated that sulfated dextran may be processed in the kidney ( 28 ) and desulfated during its renal passage ( 10 ), and furthermore, that it may bind to plasma proteins ( 17 ), and to membrane phospholipids ( 25 ), causing an artifactual reduction in the sieving coefficients (; i.e., the primary urine-to-plasma concentration ratios) of dextran sulfate. In addition, isolated glomerular basement membranes (GBM) have generally failed to show charge selectivity when probed with neutral and negatively charged Ficoll ( 7 ) or native (anionic) or cationized albumin ( 4 ). In line with these findings, Schaeffer et al. ( 26 ) were unable to find (in rats in vivo) any difference between glomerular to carboxymethylated (non-sulfated) dextran or to hydroxymethyl starch (HES), both negatively charged, and their neutral counterparts. Furthermore, the HES molecules showed lower for any given Stokes-Einstein (SE) radius (cf. Ficoll) than did dextran. It was concluded that the glomerular filtration barrier restricts the transport of polysaccharide macromolecules as a function of size and configuration whereas the presence or absence of negative charge does not play any role. Further supporting these results, Guimarães et al. ( 18 ) did not find a decrease in glomerular permeability to negatively charged, carboxymethylated (CM) Ficoll compared with uncharged Ficoll, confirming a previous observation by Greive et al. ( 16 ). Instead, they found a markedly increased glomerular permeability to CM-Ficoll.

In contrast to the apparent inability of the glomerular filter to discriminate between polysaccharides of different charge, there is ample evidence that, indeed, the glomerular filter selects globular proteins based on their charge. Thus anionic proteins are retarded compared with neutral and cationic proteins, as extensively reviewed by Comper and Glasgow ( 9 ) and Venturoli and Rippe ( 29 ). The reason the glomerular capillary wall exhibits low discrimination ability with respect to differently charged polysaccharides, while being able to separate proteins of different molecular charge, is obscure. However, one clue to this enigma could be the fact that carbohydrates exhibit an extended molecular configuration, with a larger SE radius, compared with that for globular proteins, for any given molecular mass ( 19, 29 ). Such an extended configuration, conceivably, generates a more flexible (compressible) structure and hence increases the molecule's permeability through the glomerular filtration barrier ( 29 ). Charge modification of a polysaccharide may lead to a further increase in molecular extension, favoring an increased flexibility and, thereby, an increased solute permeability.

Could the process of charge modification of the highly cross linked and "ellipsoid" molecules of Ficoll ( 19 ) lead to conformational alterations, with increased molecular extension, increasing their permeability compared with their uncharged counterparts? If so, would the linear, "random coil," structure of dextran make it less affected by conformational changes, and thereby less hyperpermeable, when negatively charged? The present study was performed to test this hypothesis by comparing glomerular sieving coefficients to negatively charged, CM-Ficoll and -dextran vs. their uncharged molecular equivalents.

MATERIALS AND METHODS

Animals. Experiments were performed in 36 male Wistar rats (Møllegaard, Lille Stensved, Denmark) divided into experimental groups as shown in Table 1. The GFR for each group is reported in Table 1. The rats had free access to standard chow and water until the day of the experiment. The Animal Ethics Committee at Lund University approved the studies.

Table 1. Body weight and glomerular filtration rate for each experimental group

Surgery. The rats were anesthetized intraperitoneally with 60 mg/kg pentobarbital sodium and placed on a heating pad to maintain body temperature at 37°C. The tail artery was cannulated (PE-50 cannula) for arterial pressure recordings on a polygraph (model 7B, Grass Instruments, Quincy, MA) and for the administration of drugs. A tracheotomy was performed using a PE-240 tube. The left carotid artery and left jugular vein were cannulated (PE-50) for blood sampling and infusions, respectively. After a laparotomy, catheters (PE-10 coupled to a PE-50) were placed in the left and the right ureter and used for urine collection. After surgery, the animal was allowed to recover for at least 30 min.

Molecular probes and standards. The fluorescently (FITC) labeled polysaccharides Ficoll, dextran, and inulin where obtained from and made CM by TdB Consultancy (Uppsala, Sweden). A mixture of Ficoll 70 and Ficoll 400 was used to achieve a broad spectrum of molecular sizes. The probes were CM so as to obtain two different degrees of substitution. The lower degree of CM substitution [CM-Ficoll (5.8%), CM-dextran (3.8%)] corresponds to a net negative charge of approximately -40 and -22, respectively, for a molecule with a SE radius of 36 Å (cf. albumin with an SE radius of 36 Å and a net negative charge of approximately -20). The higher degree of substitution [CM-Ficoll (13%), CM-dextran (17%)] corresponds to a negative charge of -95 and -92, respectively. The quantity of CM substitution was determined by titration (Mikro Kemi, Uppsala, Sweden). Briefly, the sample and a blank were dissolved in NaCl and titrated to pH 2.0, whereby carboxylic groups are supposed to become protonated. The carboxylic content was calculated as the difference in titer between sample and blank.

Narrow Ficoll standards were kindly provided by Dr. Torvald Andersson (Pharmacia, Uppsala, Sweden), whereas dextran standards were purchased from Fluka (Buchs, Switzerland). The standards were labeled with FITC (Sigma, St. Louis, MO) as shown elsewhere ( 23 ). Then, 0.1 g polysaccharide was dissolved in 2 ml DMSO (Sigma), and 0.1 g NaHCO 3 and 0.5 g FITC were added. The mixture was heated for 15 min (100°C) and then slowly added to 20 ml ethanol. After an overnight precipitation, the FITC-labeled polysaccharide was centrifuged for 15 min at 1,500 g and the pellet was dissolved in 2 ml phosphate buffer. The pH was titrated to 7.0, and the sample was applied on a PD-10 desalting column (Pharmacia) for removal of free (unbound) FITC.

Experimental protocol. The investigated FITC-labeled polysaccharide was given as a bolus dose together with 100 µg of FITC-inulin. The bolus dose was immediately followed by a constant infusion (3 ml/h) containing both the FITC-polysaccharide and FITC-inulin (6.25 µg/min) to maintain a reasonably constant plasma concentration of each substance. Analyzing the plasma concentration of the differently sized polysaccharide molecules over time demonstrated a very small change over a 20-min interval. Beyond this time 50 Å) accumulated, but the concentration of molecules with SE radii <50 Å were almost constant for up to 100 min. After an equilibration period (5-20 min), urine was collected for 5 min and a plasma sample was aspirated at 2.5 min. The sieving curves obtained were validated using longer equilibration and urine collection periods and were still found to be unaltered.

A 20-min equilibration period was permitted before starting the Ficoll and CM-Ficoll experiments. Neutral FITC-Ficoll 70 (12 µg) was given together with FITC-Ficoll 400 (288 µg) to achieve a broad spectrum of molecular sizes, followed by an infusion containing FITC-Ficoll 70 (0.30 µg/min) and FITC-Ficoll 400 (7.2 µg/min). In the experiments with negatively charged probes, a bolus of 0.13 mg 5.8% CM-Ficoll 70 and 2.9 mg 5.6% CM-Ficoll 400 was given followed by a constant infusion (3.2 and 72 µg/min, respectively). The bolus dose of the 13.3% CM-Ficoll 70 contained 7.4 mg and was followed by an infusion of 0.19 mg/min.

To avoid toxic effects of dextran in the rat (similar to the studies for Ficoll in humans) ( 2, 5 ), we gave very low dextran doses, 20 times lower than those for which we could detect a drop in arterial blood pressure. The dextran was allowed to equilibrate in the animal for only 5 min before the start of the experiment. A total amount of 540 µg neutral FITC-dextran (an equal mixture of 5, 25, and 150 kDa) was injected as a bolus dose. This was immediately followed by an infusion of 9 µg/min of 5-kDa, 6 µg/min of 25-kDa, and 4.5 µg/min of 150-kDa dextrans. The negatively charged 3.9% FITC-CM-dextran 70 was given as a bolus dose (1.2 mg) followed by an infusion of 30 µg/min. The bolus of the 16% FITC-CM-dextran contained 2.8 mg and was followed by an infusion of 70 µg/min.

Effects of CM-Ficoll on glomerular barrier characteristics and systemic capillary permeability to CM-Ficoll vs. Ficoll. To address the question of whether CM-Ficoll per se alters the permeability characteristics of the glomerular barrier, we analyzed for albumin in three rats given 5.8% CM-Ficoll by intravenous infusion for 30 min. The for albumin was measured using a tissue uptake technique, to account for the (proximal) tubular uptake of albumin, plus its urinary excretion, as described at some length previously elsewhere ( 22 ).

To test whether any change in permeability to CM vs. uncharged polysaccharides would be exclusive to the glomerular capillary barrier, we analyzed the transcapillary escape rate (TER; in %/h) of eight neutral Ficoll fractions ( n = 3) vs. the corresponding fractions of 5.8% CM-Ficoll ( n = 3) in rats. The renal peduncle of both kidneys was ligated 30 min before the start of the experiment to ensure that only "nonglomerular" capillaries were included.

High-performance size-exclusion chromatography. HPLC was performed with devices from Waters (Milford, MA). Size exclusion was achieved using an Ultrahydrogel-500 column (Waters). The mobile phase (phosphate buffer with 0.15 M NaCl, pH 7.4) was driven by a pump (Waters 1525), and fluorescence was detected with a fluorescence detector with excitation at 492 nm and emission at 518 nm (Waters 2475). The system was controlled using Breeze software 3.2 (Waters). The column was calibrated with five narrow Ficoll fractions and seven dextran standards labeled with FITC as described above. The polydispersity-corrected SE radii ( 15 ) of the Ficoll standards were 70.2, 57.2, 45.1, 36.4, and 28.4 Å, and the dextran standards were 164, 125.1, 105.6, 82.8, 63.7, 38.2, and 18.7 Å, respectively. The proteins IgM, alcohol dehydrogenase, apoferritin, and aprotinin were also used to calibrate the column and detected with an absorbance detector (Waters 2487). The void volume (V 0 = 5.53 ml) and total volume (V T = 10.89 ml) of the column were determined with blue dextran (2 x 10 6 Da) and glycine (75 Da), respectively. The distribution coefficient ( K av) was plotted against the log of the SE radii of the respective standard molecules and fitted to a third-order polynomial ( Fig. 1 ). Representative molecular size distribution curves from serum and urine (urine was recalculated to obtain Bowman's space concentration) after separation on high-performance size-exclusion chromatography (HPSEC) are shown for (neutral) Ficoll and 5.8 % CM-Ficoll in Fig. 2.

Fig. 1. Calibration curve for the ultrahydrogel-500 column was obtained using FITC-labeled Ficoll and dextran standards. Albumin, alcohol dehydrogenase, apoferritin, IgM, and inulin were also used as standards. The relationship for the calibration curve is y = -475.45 x 3 + 1,059.40 x 2 - 878.78 x + 294.41 ( R 2 = 0.9981).

Fig. 2. The Stokes-Einstein (SE) radius distribution of neutral Ficoll (solid line) and 5.8% carboxymethylated (CM)-Ficoll (dotted line) in plasma and in primary urine. For clarity, the inset shows an enlarged depiction of the concentrations in primary urine vs. the SE radius. Both polysaccharides are a mixture of MW = 70 and 400, as described. The urine concentration was corrected for the loss of water during tubular passage to obtain the concentration in primary urine.

Calculations. Sieving coefficients were determined by dividing the concentrations of probe molecules in the Bowman's space by their average plasma concentrations (P x ). The former were determined by dividing the concentration of probe in the urine (U x ) by the urine-to-plasma concentration ratio of inulin (U in /P in ). The equation used was

The excretion of inulin per minute [U in x V u (urine flow)] normalized to plasma inulin was used to measure GFR according to the standard formula

The transcapillary escape rate (TER) of Ficoll was obtained by fitting the plasma disappearance of Ficoll vs. time to an exponential decay function, C t = C o e - bt, where C t and C o are the concentrations of substance in plasma at time t and time 0, respectively, b is the decay coefficient (min -1 ), and t is time. The decay coefficient was multiplied by 60 ( x 100) to give the values (in %/h).

Statistics. Values are expressed as means ± SE. Statistical significances were tested using unpaired Student's t -tests (significance levels are shown in Table 2 ). The ratios of the of neutral to that of negatively charged polysaccharides are given with their SEs.

Table 2. Sieving coefficients for neutral and negatively charged Ficoll and dextran at defined solute radii (SE radius)

RESULTS

Characterization of neutral and CM polysaccharides. Size separation of the charged and uncharged species of Ficoll on HPSEC shows that the negatively charged CM-Ficoll 70 eluted earlier than neutral Ficoll 70 molecules of similar radius ( a e; see Fig. 3 A ). This indicates that CM per se increases the molecular radius of Ficoll. To exclude the possibility that this increase in size may be an artifact, due to the presence of a minor amount of negative charges on the otherwise electroneutral column material, we performed size exclusion in a buffer with high ionic strength (0.6 M NaCl in phosphate buffer), which is expected to markedly reduce (eliminate) any charge effects.

Fig. 3. HPLC elution curves of neutral Ficoll and negatively charged CM-Ficoll at 2 different ionic strengths, I = 0.3 M ( A ) and I = 0.6 M ( B ). Both Ficoll samples have an average MW of 70.

Figure 3 B shows that the negatively charged CM-Ficoll 70 still appeared as enlarged under conditions of high ionic strength as under conditions of normal ionic strength. The K av values, at normal and high ionic strengths, for all the polysaccharide probes used in this study are summarized in Table 3. As seen from the table, 5.8% carboxymethylation of Ficoll caused an increase in its SE radius from 35 to 45 Å, and the higher CM degree of 13% caused a further enlargement to 52 Å. This increment in a e was semiquantitatively confirmed using single-angle, quasi-elastic, light-scattering (QELS; measurements kindly performed by Prof. A. Passi). On the other hand, with carboxymethylation of dextran 70 to 3.8%, the increase in the SE radius (from 61 to 62 Å) was negligible. However, with a higher degree of carboxymethylation (17%), the molecule expanded to 94 Å.

Table 3. Distribution coefficient for the molecular probes used at normal and high ionic strength

Glomerular sieving curves for carboxymethylated and neutral Ficoll and dextran. The glomerular permeability to negatively charged CM-Ficoll was markedly enhanced compared with that of its neutral counterpart, as seen by the shift in the sieving curve for CM-Ficoll to the right compared with that of neutral Ficoll ( Fig. 4 ). A slightly increased permeability was also observed for CM-dextran compared with neutral dextran ( Fig. 5 ). The difference in between the neutral polysaccharides and their negatively charged counterparts is illustrated in Fig. 6 as the ratio between the two species plotted as a function of a e. The ratio of CM-Ficoll to neutral Ficoll forms a bell-shaped curve peaking at 54 Å, where the sieving coefficient of CM-Ficoll is 33-fold higher than that of neutral Ficoll. The difference is much less pronounced for dextran, with the maximal ratio being only 2.5. Numerical values for sieving coefficients at defined SE radii are given in Table 2, and the SE radius distribution of neutral Ficoll and 5.8% CM-Ficoll in plasma and primary urine is shown in Fig. 2.

Fig. 4. Glomerular sieving coefficients ( ) vs. molecular radius ( a e; SE radius) of neutral Ficoll and negatively charged CM-Ficoll.

Fig. 5. Vs. a e of neutral and negatively charged dextran.

Fig. 6. Ratio of CM-Ficoll to neutral Ficoll and CM-dextran to neutral dextran as a function of a e.

Ficoll and dextran with high carboxymethylation. We also investigated the effect of higher degree of carboxymethylation on the for both Ficoll and dextran (with an approximate charge of -95 and -92, respectively). This resulted in an even further enhancement of transport through the glomerular filter, the magnitude of Ficoll transport being somewhat larger than that of dextran having a similar degree of carboxymethylation ( Fig. 7 ).

Fig. 7. Of neutral and 2 different CM preparations of Ficoll ( A ) and dextran ( B ).

Effects of CM-Ficoll on glomerular barrier characteristics and systemic capillary permeability to CM-Ficoll vs. Ficoll. In three rats there was no significant increase in for albumin following CM-Ficoll infusion, the being 7.94 x 10 -4 ± 1.9 x 10 -4 compared with 6.6 x 10 -4 ± 0.45 x 10 -4 ( n = 19) in historical controls ( 22 ).

A comparison of systemic vascular permeability of CM-Ficoll vs. Ficoll, in terms of systemic plasma disappearance of polysaccharide tracer vs. time in three experiments, showed that 5.8% CM-Ficoll had a lower TER ( 70% of that of neutral Ficoll) for small ( a e <55 Å) molecules (conceivably traversing through small pores). However, there was no significant difference for large ( a e 55 Å) molecules (conceivably passing through large pores). For neutral Ficoll, the TERs (in %/h, ±SE) were 81.3 ± 9.1 for 25 Å, 71.2 ± 8.3 for 35.5 Å, 23.2 ± 5.2 for 55 Å, and 15.3 ± 4.5 for 75 Å. For 5.8% CM-Ficoll, the TERs were 31.2 ± 8.1 for 25 Å, 47.2 ± 2.5 for 35.5 Å, 17.0 ± 4.3 for 55 Å, and 14.9 ± 4.1 for 75 Å.

DISCUSSION

The main finding of this study is that negatively charged CM-Ficoll, and to some extent CM-dextran, showed significantly higher than their neutral counterparts. Although these results may appear to contrast with the generally accepted view of a high charge selectivity of the glomerular filter, there is ample experimental evidence clearly indicating that the glomerular barrier does retard anionic proteins compared with neutral or cationic proteins ( 9, 13, 22, 29 ). What could then be the reason for this discrepancy among results on for differently charged proteins vs. those for polysaccharides?

Our findings for Ficoll corroborate the findings of Guimarães et al. ( 18 ), who also found an increased glomerular permeability to CM-Ficoll. In their study, the CM-Ficoll was carboxymethylated to a degree of 13.8%, whereas in our study a 5.6% as well as a 13.3% CM substitution of Ficoll were tested. In contrast to the present results, Guimarães et al. observed augmented 45 Å in radius, whereas we observed a clear-cut difference affecting a wider molecular size spectrum, i.e., down to an a e 20 Å. The reason for this discrepancy is not clear, but it could depend on a signal detection problem because of very low plasma concentrations of Ficoll molecules of <45 Å in radius in the study of Guimarães et al. (cf. Fig. 3 A in Ref. 18 ). In our experimental set-up, however, the infused polysaccharides covered a broad molecular range, from 15 to 80 Å, potentially making the signal detection more accurate in the a e range 20-45 Å. The present data also seem to be consistent with those of Schaeffer et al. ( 26 ), who did not detect any differences between the of CM-dextran vs. neutral dextran. In the present study, we found a slight increase in of CM-dextran when the charge was modified to a moderate extent, but this increase (1.8 times at 55 Å) is very small compared with the increase seen with CM-Ficoll (32 times at 55 Å).

The Ficoll and dextran primarily tested were substituted with carboxymethyl groups to an extent of 5.8 and 3.8%, respectively, corresponding to a net negative charge of approximately -40 and -22 for a molecule with a SE radius of 36 Å (cf. albumin with an a e of 36 Å and a net negative charge of 20). For comparison, we also performed experiments using Ficoll and dextran carboxymethylated to a degree of 13.3 and 16% (-95 and -92 negative charge), respectively. These "unphysiologically" high degrees of substitution resulted in even greater increases in ( Fig. 7 ). For 13.3% CM-Ficoll, the sieving coefficients were of similar magnitude (for molecules 45 Å) as those obtained by Guimarães et al. ( 18 ) using a CM-Ficoll having a 13.8% substitution.

The common view on charge selectivity of the glomerular barrier is based on studies of sieving curves for charge modified dextrans ( 6, 8 ) and on differentially charged proteins such as albumin vs. neutralized albumin, horseradish peroxidase (HRP) vs. charge modified HRP, and different isoforms of lactate dehydrogenase (LDH) with identical sizes but different net molecular charges ( 9, 21, 27, 29 ). In the classic glomerular permeability studies utilizing polysaccharides, the permeability of negatively charged, sulfated dextrans was decreased, whereas that for positively charged, DEAE-modified dextrans was increased compared with that for neutral dextrans ( 6, 8 ). On the basis of these results, it was postulated that the glomerular barrier has a pore radius of 50-55 Å, but due to its negative charge, the passage of albumin ( a e = 36 Å) would be almost completely hindered. From such data, Deen et al. ( 14 ) originally calculated a glomerular barrier fixed charge density of 120-170 meq/l.

A number of studies during the past few years have challenged this view. Comper et al. ( 10 ) showed that sulfated dextrans were actually processed by the renal cells and desulfated during glomerular filtration ( 10, 28 ). Guasch et al. ( 17 ) and Vyas and Comper ( 30 ) showed that sulfated dextran can bind to plasma proteins. Furthermore, it has been shown that dextran sulfate can easily bind to lipid membranes ( 25 ). Hence, glomerular filtration experiments using dextran sulfate result in a marked underestimation of true dextran. Thus the charge density of the glomerular barrier was probably originally overestimated by Deen et al. 1980 ( 14 ). Furthermore, an increased of positively charged, DEAE-modified dextrans ( 6 ) could not be reproduced by Adal et al. ( 1 ). They postulated that the increased sieving coefficients of DEAE dextran may be due to binding of the positively charged probes to anionic sites in the glomerulus, thereby, conceivably through disruption of the glycocalyx, making the barrier leakier.

Using extracted GBM in vitro, Bolton et al. ( 7 ) showed that the GBM permeability to sulfated Ficoll molecules was not reduced compared with neutral Ficoll. They concluded that the charge barrier of the glomerular capillary wall is either located in the endothelial cell layer or in the slit diaphragm, but not in the glomerular basement membrane. Ohlson et al. ( 23 ) compared the sieving of neutral Ficoll to that of negatively charged, native albumin, and on this basis they calculated the fixed negative charge of the glomerular barrier to be 30 meq/l. However, conclusions drawn from comparisons of such disparate molecular probes (i.e., proteins vs. polysaccharides) may be problematic, because other factors, such as discrepant configuration and deformability, could have influenced the permeability differences seen. Thus, as extensively reviewed by Venturoli and Rippe ( 29 ), polysaccharides such as Ficoll and dextran should not be compared with proteins based on their SE radii alone.

Plotting the log SE radius vs. log molecular weight (MW) reveals mutually discrepant relationships for globular proteins, Ficoll and dextran, respectively. Globular proteins appear to behave close to "hard spheres," whereas polysaccharides appear much less dense than hard spheres, dextran being less dense than Ficoll. Applying pore theory to the glomerular for polysaccharides, which show marked deviation from hard sphere behavior, will tend to overestimate the equivalent small-pore radius. It thus seems that molecular deformability, besides molecular size, shape, and charge, determines the glomerular sieving coefficients for molecules of various species (polysaccharides being more flexible than proteins). The less the density of a molecule, the more permeable it appears across the glomerular filter. If charge modification makes the molecule more "extended" (less dense) compared with its neutral counterpart, then a further increase in radius (at an unchanged MW) may be expected, which would make the molecule even more hyperpermeable.

Ficoll is a polysaccharide obtained by polymerization of sucrose with epichlorohydrine. The result is a highly cross-linked molecule with a "dendrimeric" structure. Although a detailed characterization of the effect of carboxymethylation on the structure of the Ficoll molecule has not been done, except by HPSEC and in preliminary assessments using QELS, as in the present study, conceivably, when this dendrimeric ( 3, 12 ) molecule is charge modified, its side chains will repel each other, causing them to stick out from the surface of the molecule. This would make the charge-modified molecule larger, and hence, less compact compared with its neutral counterpart. Indeed, in our study, CM-Ficoll showed an increased SE radius compared with uncharged Ficoll ( Table 3 ), the increase being larger than predicted from the mere addition of carboxymethyl groups. Dextran, on the other hand, is a linear polysaccharide of glucose molecules linked together at their -1,6 position. According to the HPSEC analysis, there was an insignificant increase in the SE radius for dextran when carboxymethylated to 3.9%. However, with the high degree of carboxymethylation (17%), the SE radius of dextran increased dramatically, indicative of a reduced molecular density or an increased asymmetry of the molecule when highly charged. However, there was only a moderate increase in the in vivo glomerular permeability of the charge-modified dextran compared with the increase seen with charge-modified Ficoll. Dextran is already markedly asymmetric and hyperpermeable across the glomerular filter as an uncharged species. It is thus speculated that a further increment in the asymmetry of the molecule (by carboxymethylation) would not markedly enhance its permeability through the highly selective glomerular membrane, as it is already extremely asymmetric. On the HPSEC column, however, the role of deformability and asymmetry of the enlarged (negatively charged) polysaccharides is of little significance due to interactions with very large pores, being one order of magnitude larger ( 500 Å) than those of the glomerular filter (40 Å). Asymmetry or flexibility facilitates transport only when the molecule is interacting with structures similar in size to that of the molecule itself ( 11 ). Analogous with this reasoning, bikunin (MW 25), which is highly asymmetric and negatively charged, has a large SE radius on the HPSEC, similar to that of albumin (36 Å), but it exhibits a very low "effective" molecular radius in vivo and a high glomerular permeability ( 20, 24 ) compared with albumin. In contrast to the "abnormal" behavior of CM-Ficoll across glomerular capillary walls, preliminary data indicate that negatively charged Ficoll molecules show lower permeability across systemic microvascular beds. The reason for this discrepancy is obscure, but it could be related to the extremely large diffusion surface area ( A 0 / x ) relative to fluid flow (cf. low Peclet number) across the glomeruli vs. systemic capillaries. However, we will return to this intriguing issue in a forthcoming paper.

Ficoll and CM-Ficoll, as well as dextran and CM-dextran, have been shown to be stable during glomerular filtration and renal passage, neither taken up by renal cells nor binding to plasma proteins ( 18, 26 ). This should omit the possibility that artifacts due to instability or renal processing of the polysaccharide molecules would have affected our results. Furthermore, even though the size-exclusion column is considered uncharged, some stationary (negative) charge could still exist on the column material, which would make the carboxymethylated polysaccharides appear larger due to electrostatic interactions with the column. To exclude this possibility, we raised the ionic strength of the mobile phase to markedly reduce any charge effects. This slightly increased the elution volumes for both the negative and neutral polysaccharide, but the change (decrease) in K av by charge modification of the molecules was maintained unchanged for a high ionic strength of the buffer. This indicates that any charge effects exerted by the column were negligible. Further evidence indicating a negligible charge effect of the column material is that there was no significant difference in elution volume for native (negatively charged) vs. neutral albumin on the column in previous studies from this laboratory ( 22 ). Furthermore, semiquantitative light scattering confirmed the increment in a e of CM-Ficoll 70 compared with Ficoll 70. Thus it seems safe to conclude that the a e increase in the negatively charged polysaccharides as observed with HPSEC is not spurious.

In conclusion, we found an increased permeability of the glomerular capillary wall to negatively charged carboxymethylated Ficoll relative to neutral Ficoll. It is speculated that this is due to an aberrant behavior of Ficoll when it is negatively charged, in that the molecule, due to its dendrimeric nature, becomes larger, less dense and, thereby, conceivably more flexible (compressible), making it hyperpermeable across the glomerular barrier. By contrast, the "random coil" dextran did not show this marked hyperpermeability when moderately carboxymethylated. The data apparently contradict previous glomerular sieving results for globular proteins of differing charge, thus indicating that polysaccharides may not by suitable as probes for studying the charge selectivity of the glomerular filter.

GRANTS

This study was supported by Swedish Medical Research Council Grant 08285, the Lundberg Medical Foundation, the Swedish Association for Kidney Patients, the University Hospital of Lund (donation funds for nephrology), and the Lars Hierta Memorial Foundation.

ACKNOWLEDGMENTS

Anna Rippe is gratefully appreciated for skillful technical assistance. We are grateful to Professor Alberto Passi at the Università dell'Insubria, Varese, Italy, for performing the light-scattering experiments on Ficoll and CM-Ficoll.

【参考文献】
  Adal Y, Pratt L, and Comper WD. Transglomerular transport of DEAE dextran in the isolated perfused kidney. Microcirculation 1: 169-174, 1994.

Andersen S, Blouch K, Bialek J, Deckert M, Parving HH, and Myers BD. Glomerular permselectivity in early stages of overt diabetic nephropathy. Kidney Int 58: 2129-2137, 2000.

Ballauff M and Likos CN. Dendrimers in solution: insight from theory and simulation. Angew Chem Int Ed Engl 43: 2998-3020, 2004.

Bertolatus JA and Klinzman D. Macromolecular sieving by glomerular basement membrane in vitro: effect of polycation or biochemical modifications. Microvasc Res 41: 311-327, 1991.

Blouch K, Deen WM, Fauvel JP, Bialek J, Derby G, and Myers BD. Molecular configuration and glomerular size selectivity in healthy and nephrotic humans. Am J Physiol Renal Physiol 273: F430-F437, 1997.

Bohrer MP, Baylis C, Humes HD, Glassock RJ, Robertson CR, and Brenner BM. Permselectivity of the glomerular capillary wall. Facilitated filtration of circulating polycations. J Clin Invest 61: 72-78, 1978.

Bolton GR, Deen WM, and Daniels BS. Assessment of the charge selectivity of glomerular basement membrane using Ficoll sulfate. Am J Physiol Renal Physiol 274: F889-F896, 1998.

Chang RL, Deen WM, Robertson CR, and Brenner BM. Permselectivity of the glomerular capillary wall. III. Restricted transport of polyanions. Kidney Int 8: 212-218, 1975.

Comper WD and Glasgow EF. Charge selectivity in kidney ultrafiltration. Kidney Int 47: 1242-1251, 1995.

Comper WD, Tay M, Wells X, and Dawes J. Desulphation of dextran sulphate during kidney ultrafiltration. Biochem J 297: 31-34, 1994.

Davidson MG and Deen WM. Hindered diffusion of water-soluble macromolecules in membranes. Macromolecules 21: 3474-3481, 1988.

Deen WM, Bridges CR, and Brenner BM. Biophysical basis of glomerular permselectivity. J Membr Biol 71: 1-10, 1983.

Deen WM, Lazzara MJ, and Myers BD. Structural determinants of glomerular permeability. Am J Physiol Renal Physiol 281: F579-F596, 2001.

Deen WM, Satvat B, and Jamieson JM. Theoretical model for glomerular filtration of charged solutes. Am J Physiol Renal Fluid Electrolyte Physiol 238: F126-F139, 1980.

Frigon RP, Leypoldt JK, Uyeji S, and Henderson LW. Disparity between Stokes radii of dextrans and proteins as determined by retention volume in gel permeation chromatography. Anal Chem 55: 1349-1354, 1983.

Greive KA, Nikolic-Paterson DJ, Guimaraes MA, Nikolovski J, Pratt LM, Mu W, Atkins RC, and Comper WD. Glomerular permselectivity factors are not responsible for the increase in fractional clearance of albumin in rat glomerulonephritis. Am J Pathol 159: 1159-1170, 2001.

Guasch A, Deen WM, and Myers BD. Charge selectivity of the glomerular filtration barrier in healthy and nephrotic humans. J Clin Invest 92: 2274-2282, 1993.

Guimãraes MA, Nikolovski J, Pratt LM, Greive K, and Comper WD. Anomalous fractional clearance of negatively charged Ficoll relative to uncharged Ficoll. Am J Physiol Renal Physiol 285: F1118-F1124, 2003.

Lavrenko PN. Hydrodynamic properties and the shape of the molecules of the polysaccharide Ficoll-400 in solution. Polymer Sci USSR 28: 576-584, 1986.

Lindstrom KE, Blom A, Johnsson E, Haraldsson B, and Fries E. High glomerular permeability of bikunin despite similarity in charge and hydrodynamic size to serum albumin. Kidney Int 51: 1053-1058, 1997.

Lindstrom KE, Johnsson E, and Haraldsson B. Glomerular charge selectivity for proteins larger than serum albumin as revealed by lactate dehydrogenase isoforms. Acta Physiol Scand 162: 481-488, 1998.

Lund U, Rippe A, Venturoli D, Tenstad O, Grubb A, and Rippe B. Glomerular filtration rate dependence of sieving of albumin and some neutral proteins in rat kidneys. Am J Physiol Renal Physiol 284: F1226-F1234, 2003.

Ohlson M, Sorensson J, and Haraldsson B. Glomerular size and charge selectivity in the rat as revealed by FITC-Ficoll and albumin. Am J Physiol Renal Physiol 279: F84-F91, 2000.

Ohlson M, Sorensson J, Lindstrom K, Blom AM, Fries E, and Haraldsson B. Effects of filtration rate on the glomerular barrier and clearance of four differently shaped molecules. Am J Physiol Renal Physiol 281: F103-F113, 2001.

Santos HA, Garcia-Morales V, Roozeman RJ, Manzanares JA, and Kontturi K. Interfacial interaction between dextran sulfate and lipid monolayers: an electrochemical study. Langmuir 21: 5475-5484, 2005.

Schaeffer RC Jr, Gratrix ML, Mucha DR, and Carbajal JM. The rat glomerular filtration barrier does not show negative charge selectivity. Microcirculation 9: 329-342, 2002.

Sorensson J, Ohlson M, Lindstrom K, and Haraldsson B. Glomerular charge selectivity for horseradish peroxidase and albumin at low and normal ionic strengths. Acta Physiol Scand 163: 83-91, 1998.

Tay M, Comper WD, and Singh AK. Charge selectivity in kidney ultrafiltration is associated with glomerular uptake of transport probes. Am J Physiol Renal Fluid Electrolyte Physiol 260: F549-F554, 1991.

Venturoli D and Rippe B. Ficoll and dextran vs. globular proteins as probes for testing glomerular permselectivity: effects of molecular size, shape, charge, and deformability. Am J Physiol Renal Physiol 288: F605-F613, 2005.

Vyas SV and Comper WD. Dextran sulfate binding to isolated rat glomeruli and glomerular basement membrane. Biochim Biophys Acta 1201: 367-372, 1994.

作者单位：Department of Nephrology, University Hospital of Lund, Lund, Sweden