|
- 积分
- 0
- 威望
- 0
- 包包
- 0
|
作者：Giuliano Ciarimboli, Clara Hjalmarsson, Arend Bökenkamp, Hans-Joachim Schurek, Börje Haraldsson作者单位：Division of Nephrology, Children‘s Hospital, Hannover Medical School, D-30623Hannover; Children‘s Hospital, Bonn University,D-53113 Bonn, Germany; and Departments of Physiologyand Nephrology, Göteborg University, SE-405 30Göteborg, Sweden & P$ \9 U! b" n- D
- X2 B- S p/ H3 o0 K- P. S! d
7 y* k3 l8 h% P+ K
7 W5 y* {8 S6 {7 m
, k4 k% E, X* r" ]/ D! s
3 I# s" S3 \" T* u" E
2 T. |4 C/ D! F; E1 e. G( g 3 l- D: R' g1 {8 M/ l1 t
9 ~, H. w0 z. f
% }) x1 Y/ o6 a7 o
- H) k; F2 {& Z: h* v
5 m4 G6 e7 R+ `4 F1 O: Q% K) ~
! j: c; D z9 T( x 【摘要】+ G9 ~' I9 ?% H7 j8 Q
In a previous paper, we found that low ionic strength (I) reversibly reduced the glomerular charge density, suggesting increased volume of thecharge-selective barrier. Because glutaraldehyde makes most structures rigid,we considered the isolated, perfusion-fixed rat kidney to be an ideal modelfor further analysis. The fixed kidneys were perfused with albumin solutionscontaining FITC-Ficoll at two different Is (I = 151 and 34 mM). At normal I,the fractional clearance ( ) for albumin was 0.0049 (SE -0.0017, 0.0027, n = 6), whereas for neutralFicoll 35.5Å of similar size was significantly higher 0.104(SE 0.010, n = 5, P At low I, foralbumin was 0.0030 (SE -0.0011, 0.0018, n = 6, not significantfrom albumin at normal I) and for Ficoll 35.5Å was identical to that at normal I, 0.104 (SE 0.015, n = 6, P albumin at low I). According to a heterogeneous chargedfiber model, low I reduced the fiber density from 0.056 to 0.0315, suggestinga 78% gel volume expansion. We conclude that 1 ) there is asignificant glomerular charge barrier. 2 ) Solutions with low Iincrease the volume of the charge barrier even in kidneys fixed with glutaraldehyde. Our findings suggest that polysaccharide-rich structures, suchas the endothelial cell coat, are key components in the glomerularbarrier. $ e+ r4 b3 G7 F; k- D6 F, G
【关键词】 glomerular permeability charge selectivity
3 P& p/ d# \7 L0 {0 f TO ACHIEVE THE rapid rate of filtration required to regulate thecomposition and volume of body fluids, glomerular capillaries possess uniquefunctional and structural characteristics. A striking example is anextraordinarily high hydraulic permeability. Glomerular capillaries are one totwo orders of magnitude more permeable to water than are capillaries fromvarious other microvascular beds( 9 ). Nevertheless, the samestructure normally imposes an extremely efficient barrier to the passage ofplasma proteins so that the concentrations of albumin and larger proteins areminute.
2 T8 t( c1 \6 G5 i
& g- z1 u3 v1 Y7 @4 s( h% `8 x+ I( BThe transport of solutes across microvascular walls can be described by atwo-pore theory of capillary permeability( 30 ). Measurements ofsteady-state sieving coefficients ( ) of proteins from plasma tointerstitium or lymph are used to predict the pore radius and distribution. Inglomerular capillaries, the presence of tubular reabsorption and secretionprocesses that modify final urinary composition is a formidable obstacle tothe determination of sieving coefficients for proteins and, consequently, to the study of glomerular permselectivity. Experimental in vitro systems such asthe isolated nephron or the isolated perfused glomerulus are free from theinfluence of tubular transport processes( 26, 37 ). These systems, however,are less suitable for studies of macromolecular transport due to the smallamounts of solute filtrated in minute volumes. With a normal albumin fractionof only a few tenths of a percent, the measurements are less accurate than forsmaller solutes.) }& d) x: L6 P+ u, z
) m5 O6 t& n1 j& ^. ]
Oliver et al. ( 27 ) foundthat Ficolls (globular uncharged cross-linked copolymer of sucrose andepichlorohydrin that is neither secreted nor reabsorbed by the renal tubules)of various radii had a lower fractional clearance than dextrans of equalStokes-Einstein radius ( a SE ). This implies that Ficoll maybe a reliable transport probe for the measurement of small and large poreradii.; }- f% [' ~1 ?
6 G1 y! \* U# p8 j i, }Fractional clearance experiments of charged dextrans substantiated thehypothesis of a charge-dependent glomerular filtration of macromolecules( 4 ). On the basis of fractionalclearance data of dextran in the rat, in a now classic work Deen et al. ( 10 ) calculated an apparentfixed charge concentration on the glomerular capillary wall (GCW) of120-170 meq/l. However, when using charged dextran in permselectivitystudies, uptake and desulphation of dextran sulfate by the glomerular andtubular cells ( 6, 42, 43 ) and the ability of certaindextran sulfates to bind to plasma proteins( 13 ) complicate theinterpretation of the results.+ \" V9 K+ H8 L3 O- p$ u$ e5 Z
9 r/ i6 S: ^! o9 \7 [) JWe developed a modified isolated rat kidney model in which the tubularreabsorption processes were eliminated by glutaraldehyde fixation( 5 ). Gluteraldehyde is afixative that acts rapidly and offers accurate tissue preservation. There isevidence that in the isolated kidneys fixed by perfusion with glutaraldehyde, no ultrastructural alteration of the GCW is detectable; an intact organizationof the glomerular cells and an unaltered distribution of glomerular polyanionswere reported ( 36, 37 ). In one of the mostdetailed descriptions of the fine structure of isolated kidneys afterperfusion fixation with glutaraldehyde, Kriz et al.( 19 ) showed that the integrityof the barrier is remarkably preserved. No cell lysis is noticed, and thestructure of the glomerular capillaries is undistorted( 11, 12 ), whereas the metabolicprocesses are eliminated. With the use of this model, the glomerularpermeability properties can be directly studied, without interferences of thetubular apparatus and without influence of hemodynamic factors and bloodconstituents such as hormones. The charge of the GCW was determined( 5 ) using albumin solutionsbuffered at different pHs spanning the isoelectric points of albumin and ofthe glomerular basal membrane.9 B2 `( _9 x. k7 }: C, U/ m
$ J {6 o8 M6 V# y, X k1 Q
In the present work, the fractional clearance of FITC-Ficoll was measuredto determine the glomerular size and charge selectivity in theperfusion-fixed, isolated rat kidneys. Previous studies suggest low ionicstrength (I) to reversibly reduce the glomerular charge density, most likelydue to volume expansion of the compartment responsible for chargeselectivity.3 Z2 S" T% u$ O
3 O4 @; Q2 S# g
Therefore, it was of particular interest to estimate the charge density atdifferent Is of the perfusate. Our hypothesis was that lowering I would notinduce dynamic alterations of the estimated charge density in the fixedkidneys if the glomerular charge selectivity were to reside in the basementmembrane and/or in the podocyte slit membrane. On the other hand, marked changes in charge density could be expected in the fixed kidneys if glomerularcharge selectivity were dependent on the endothelial cell coat barrier, whichis more resistant to glutaraldehyde ( 31, 35 ).! r' W8 p S, O" B( c
% Y" E5 w) Z+ w0 W8 M( ~MATERIALS AND METHODS
! R) {% m% f+ R
& D4 a, v$ D3 r. vExperimental Animals
`4 f7 k/ E3 O" x) \, Z6 m$ N: ?9 h$ q1 N! c
Experiments were performed on male Sprague-Dawley rats weighing 200-300 g. The animals had free access to food (standardized pellets,Altromin, Altromin Gesellschaft für Tierernährung mbH Lage, Germany)and tap water until the experiment. The rats were anesthetizedintraperitoneally with 100 mg/kg body wt thiopental-sodium (Trapanal,Byk-Gulden, Konstanz, Germany). The local ethics committee approved theexperiments.$ i/ m( P& c9 o0 m
/ e2 X9 i/ D; q! U+ V- \8 |" Z+ } U
Kidney Isolation5 E' W& I0 q6 K6 B$ E. @
. N& Y8 T+ F( _3 k) x5 l0 N
Surgery. The rats were placed on a temperature-regulated table. The surgical procedure was a modification of that reported by Weiss et al.( 40 ) and by Nishiitsutsuji-Uwoet al. ( 22 ). The right kidneywas always used for perfusion. As ureter catheters, we used short (10 mm)polypropylene catheters (PP-10, Portex, Hythe, UK) connected to largerpolyethylene catheters (PE-50, Portex), thereby preventing a buildup ofureteral backpressure. After heparin injection (Liquemin, Hoffmann-LaRocheGrenzach-Wyhlen, Germany), the kidney was placed in a temperature-controlled metal chamber. Before the perfusion was started, the aorta was clamped distalto the right renal artery, and a double-barreled cannula was inserted into theabdominal aorta distal to the clamp. Perfusion was started in situ by openingthe clamp and tying the proximal aortic ligature. Thus zero perfusion of the experimental kidney never occurred.2 e5 a7 t9 y( ?1 _) x7 f
/ Y& U( ^0 d) b1 @Perfusion apparatus and technique. The apparatus was designed as arecirculation system with dialysis because of a higher stability. Theperfusion technique and apparatus have been previously described in detail( 33, 34 ). Experiments were performed using a substrate-enriched Krebs-Henseleit bicarbonate solutioncontaining 50 g/l BSA (Fraction V, Sigma, Deisenhofen, Germany)( 34 ). Verapamil (Isoptin,Knoll, Minden, Germany) at a dose of 4.4 µmol/l was added to the perfusionmedium. The effective perfusion pressure was 100 mmHg.
) {" M& I; m9 T
+ E3 B6 n9 q s5 ]Kidney Fixation1 m& t0 J4 S! H( l! J: t
& \, Y' T0 s6 q* x) ^After isolation, the kidney was perfusion-fixed with a 1.25% monomericglutaraldehyde (Polyscience, Warrington, PA) solution in 0.1 M phosphatebuffers (final pH 7.2). The fixation solution was made isooncotic to plasma byaddition of hydroxyethyl starch (Plasmasteril, Fresenius, Bad Homburg,Germany) to a final concentration of 60 g/l. In previous experiments, it hadbeen observed that the perfusion resistance was increasing dramatically whenthe perfusate was colloid free. For fixation, the kidney was perfused for 6 to8 min at a pressure of 150 mmHg.
; O8 z7 A' ]9 u0 X- q; F, d5 y/ Z' _3 d/ ~/ t" D
Reperfusion of the Fixed Kidney
: m! ^2 P, M7 K7 e& Q: ^9 e8 F/ u; B1 i
Before reperfusion experiments were started, the fixed kidney was washedfree from glutaraldehyde by a 60-min single-pass perfusion with 0.9% saline ata pressure of 100 mmHg. This step was necessary to avoid the formation ofprotein-glutaraldehyde aggregates, which can significantly reduce theperfusion flow rate in protein perfusion experiments (not shown).
: x4 _+ [$ B3 E/ I7 \" a0 G( ~5 W! ]2 K) j. v5 S- B
Every solution used in perfusion experiments of the fixed kidney contained100 mg/l polyfructosan. Experiments were performed at a perfusion pressure of100 mmHg.
5 w/ u" k" r: h+ q
# B8 h8 ?+ t6 u9 zStudy Design
# H( a& M. Z/ z5 V- b& O
* L0 p/ h$ t+ I7 ?The fixed kidneys were perfused successively with phosphate-buffered (136.9mM NaCl, 2.7 mM KCl, 0.5 mM MgCl 2, 0.9 mM CaCl 2, 8.1 mMNa 2 HPO 4, and 1.5 mM KH 2 PO 4, pH7.4) solutions containing 10 g/l BSA and 70 mg/l FITC-Ficoll (Ficoll-70,Bioflor, Uppsala, Sweden) containing Ficoll molecules of different size. The I of the "normal" perfusate was 151 mM. The low I perfusates (I = 34mM) contained the same concentrations of BSA and Ficoll, respectively, butotherwise had the following composition: 26 mM Na, 4.3 mM K, 2.5 mM Ca, 8.4 mMCl, 0.8 mM Mg, 25 mM HCO 3, 0.5 mM H 2 PO 4, 5.6mM glucose, and 241 mM mannitol.) w) {8 l# o- \8 V4 T5 R P( V
* ]$ R2 p5 {# }: l4 g& C
Electrolytes, `) z |7 D& A4 c2 [0 s
0 F4 H0 O- @% l1 G
Na and K concentrations in perfusate and urinarysamples were determined with an ion-selective electrode analyzer (System E2Aelectrolyte analyzer, Beckman, Brea, CA).
, h! q$ Q; `+ F8 O& s5 X2 ]4 `! S" c& z( h1 o- Z1 w; n, K
Analysis of Ficoll Sieving2 W3 L1 R ?/ f; e8 X1 s* A
6 b* E/ J2 l% T7 V4 V% H/ x& E, V
For calculating the sieving coefficients for FITC-Ficoll, all perfusate andurine samples were subjected to gel filtration (BioSep-SEC-S3000, Phenomenex,Torrance, CA) and detection of fluorescence (RF 1002 Fluorescence HPLCMonitor, Gynkotek, Germering, Germany) using Chromeleon (Gynkotek) software.As eluent, we used a 0.05 M phosphate buffer with 0.15 M NaCl with pH 7.0.From each sample, a volume of 5-10 µl was analyzed at an emissionwavelength of 520 nm and an excitation wavelength of 492 nm; during analysisflow rate (1 ml/min), sampling frequency (1 per second), pressure (4 MPa), andtemperature (8°C) were maintained constant. We estimated the error in theC U /C P ratios for Ficoll to be sizes.2 |8 Z7 C) m, _, _ u
/ V1 ]. O: T2 Z' s6 _+ c
Other Analytic Methods6 z6 K* K. h# S% y" J$ }% \+ P
: ~5 L+ [6 O1 e9 Q* A. b. @Total protein was determined using the Bradford method( 3 ). Inulin was measured afteracid hydrolysis by the hexokinase/glucose-6-phosphate dehydrogenase method( 32 ) by including aphosphohexose isomerase reaction into the assay.$ [' `7 h" L! z$ k
% `: i2 i8 Z" M" L, D3 C0 r
Calculations( K# m( R0 U. U; Q0 u7 e
) B: S+ ~3 U' S. k8 I+ PGlomerular filtration rate. The glomerular filtration rate (GFR) of the isolated and of the fixed kidney was determined by measuring inulin(polyfructosan) clearance. For calculating the GFR, we used the followingformula
6 w+ F/ K0 s4 l( R9 R$ `$ z! t
x2 ~# I$ J# V( m: c( 1 )
l9 G' ^0 c- t( S( g
" k( f: y( S" D4 c" Y. A% ]C p is the concentration of inulin in plasma, and C u represents its concentration in urine; Q u is urine flow rate.
+ z$ i& A; q1 O* K F- D* w( D W0 c* ~( K K( r% z$ {) z. T1 y8 H9 w
Fractional clearances of albumin and Ficoll,. Thefractional clearance for solute X was calculated as
) U; \$ C3 R' X3 p @+ e* h/ T9 c% N# Q2 A+ D8 ^/ f9 l" I4 r
( 2 )
# B7 B w" `2 c) ~6 a8 Z4 _
7 A1 ?- u/ q/ N& d: LModels of Glomerular Size and Charge Selectivity" D' k/ W1 p! P; x6 S
5 R& ~7 j0 \0 W+ B
We used two different theoretical models for analysis of glomerular sizeand charge selectivity, namely the gel-membrane model ( 24 ) and a charged fiber model( 16 ) with smalldiscontinuities with extremely low concentrations of fibers (large pores).
" s! s) c, D6 L7 T! r
' P1 a) ?: \) U' T8 [% W+ uGel-Membrane Model
- K7 W& W" Z8 V! y) F8 n: C3 ^$ h- K: O2 ^8 ]4 L- D
The gel-membrane model ( 24 )assumes the glomerular barrier to be composed of two separate compartments inseries: one charge selective (gel) and one size selective (membrane). The gelcontains fixed negative charges, and the concentration of an anionic moleculesuch as albumin will be lower in the gel than in the plasma. The secondcompartment of the barrier behaves as a membrane exerting size discriminationbut no charge selectivity. Thus, in this model, the effects of size and charge are treated in two different compartments, which greatly facilitate thecalculations but naturally represent a gross oversimplification, because thesieving coefficient for a certain solute is given by the product of for each individual component of a serial barrier( 8 ). Furthermore, it can beargued that the limitations of the model affect the results leading toerroneous conclusions. However, if we consider the "gel" to be apart of the plasma compartment rather than the barrier, the model may still be valid.
/ { y5 ?! D5 J! F% H+ j+ _4 f' ~( C" V
Charge selectivity is estimated from the for albumin and its neutral counterpart of similar size, Ficoll 35.5Å, giving adensity of fixed charges, (see Ref. 24 ). The fractional clearancefor Ficolls of Stokes-Einstein radii between 30 and 70 Å (180 datapairs) allows estimates of size selectivity using a two-pore model, which hasthe following four parameters: the functional small pore radius(r S ), the large pore radius (r L ), the large porefraction of the hydraulic conductance (f L ), and the unrestrictedexchange area over diffusion distance (A 0 / x). For moredetails, please consult Ref. 24.
+ l: O: s) K/ N7 t$ u4 C S5 D) H3 p2 _/ R* k
Charged Fiber Model with Discontinuities of Low Fiber Density/ \( m9 ~ ]! V Z$ U
& G/ I' C' p" G* H" U& k* PTo combine size and charge selectivity in one model is highly complicated,but Johnson and Deen ( 16 )extended the partitioning theory of Ogston( 23 ) to develop a chargedfiber model to predict the concentration ratio of a solute at equilibrium in and outside a gel. The endothelial surface layer (glycocalyx) and theglomerular basement membrane are examples of such more or less charged gels.We previously used the model in a quantitative analysis of charge selectivity( 38 ), but the present analysis differs in two important aspects. First, the present analysis takes intoaccount that there may be heterogeneous fiber densities with regions with lowfiber concentrations, i.e., large pores. Second, due corrections are made forthe diffusivity in a gel ( 29 ).
* C/ f' f1 t7 j( ~5 j6 C: C8 [
* i2 {3 A' b) \9 G% l5 \0 R2 CThe gel/plasma concentration ratio at equilibrium in a fiber matrix isdescribed by the partition coefficient ( )+ \& T$ r; {6 [9 A$ _" A
& c' {; T" l& [ o7 |. \9 J6 M( 3 )+ E) d7 F) ~' Y" F- y1 I, N
" c! h8 K# S! V" y6 X
where g(h) is the probability of finding the closest fiber at a distance, h,from a spherical solute in a dilute solution
. _& J# V8 D( K; `, i* Q. L9 A( u) C& ?% J. I
( 4 )
5 z3 B: B' K! ^# b) W$ O
3 o% o0 T& ~6 _! \# l4 `where is the volume fraction of fibers, r s is the soluteradius, and r f is the fiber radius. By integrating Eq. 4,Ogston ( 23 ) reached the following expression for
: k" B2 D+ m6 `) }0 K" E: C
" h" _/ Q- F( A( 5 )0 E8 i) t* v, I7 r
: ^% |! d( k6 |Johnson and Deen ( 16 )introduced a Boltzmann factor to describe the relative probability atdifferent energy states in charged gels. Multiplying g(h) by this factor gives
0 B( ^7 i4 H; a" |. X; T- Q
! f( ~1 H2 L8 F2 m9 `( 6 )4 k+ q4 n6 W! R* _5 x8 W: f" p: s
; z; d- @; w( G9 ]/ U( ~
where E(h) is the electrostatic free energy of the interactions between thesolute and nearest fiber divided by k T ( k is Boltzmann's constant and T is the absolute temperature). E is dependent of one positionvariable, h, only. In a true system, the solute would interact with multiplefibers (and other solutes). Johnson and Deen( 16 ) solved Eq. 6 using a linearized Poisson-Boltzmann equation with dimensionless parametersscaled by the electrical potential R T/ F (R is the gasconstant and F is Faraday's constant).1 X9 @: U# |3 i
/ | H& H5 x4 T3 P' u
Interactions between solute and fiber cause changes in the free electrochemical energy0 e$ l$ i5 T4 Y6 I- }/ ~
3 @: G# S; ~9 g. u
( 7 )3 m d+ L2 f% O' [9 b
: y) _' z+ w) K, O6 Q8 @4 S: G
where the subscripts s, f, and sf refer to the isolated solute, isolated fiber, and combined solute-fiber system. Note that to obtain G, nestedpolynomial equations are required making the calculations much more complex( 16 ).
; D' h! S" C) K: ^9 n
. P) D0 v; z8 L2 uThe energy (E) needed in Eq. 6 is given by
( I$ I; |, _$ n, j1 L. F" K0 J3 l9 {, R0 c. ]: k
( 8 )
( W0 p7 E! H8 V& x2 C( l+ m7 v9 {- _8 ]1 U
where is the dielectric permittivity for the solvent. R, T, and F were previously explained. The dielectric permittivity is therelative dielectric constant multiplied by 0, the constant for vacuum ( 0 = 8.8542 · 10-12 C ·V - 1 · m - 1 ). Inthe case of uncharged solutes or fibers, the low dielectric constant willchange the potential field of the charged solute or fiber surrounding, thusincreasing the electrostatic free energy. For further details of theequations, please consult Johnson and Deen( 16 ).
& o2 o. X" w$ f( j, K. T, g5 }7 l2 x* N/ `5 ~0 {4 |$ A/ b
To apply the partition coefficients to experimental data, one mustcalculate fractional clearances ( ). In the concept of "fibermatrix," Curry and Michel( 7 ) used the expression of Anderson and Malone ( 2 ) tocalculate the reflection coefficient ( ) from the partition coefficient( ): `. n! T7 Y e) t) q5 W/ H+ w+ S
- `/ d7 ]. _+ P* ?, e/ J( 9 ) M3 t, `5 K$ J8 t* I. [+ t
: V7 r% I7 ]# G* M* w6 C" T
Note that experimental observations in agarose gels give reflection coefficients that differ somewhat from those of Eq. 9 ( 17 ). There are, however,limited experimental studies of reflection coefficients in biological gels,and, to our knowledge, this is the best equation available.8 u K2 u+ }2 p) p, e# q
3 T8 r) x. o7 P* m' _
Finally, the diffusion capacity (PS) is given by# f n* w& V) N7 ?1 \
+ i# ^; u- J) M$ P
( 10 )
" K. y. s) E0 g! C' n# q7 n3 |0 k- K1 m3 b" v
where A 0 / x is the unrestricted exchange area over diffusiondistance and D is the free diffusion constant. D/D 0 is the relativediffusivity in a gel as presented by Phillips ( 29 )" S2 d6 G: i* c# [
. G3 Y; ]( O3 y7 [- _& h. e; ~* F
( 11 )4 j! P7 }1 _: W
1 y+ h9 I, z; Z! y6 H& |
The fractional clearance ( ) is obtained using a nonlinear flux equation ( 30 )
' w: q/ ?: H% q
, x' S" b( B* e2 Y5 p( 12 )
' z; R% H* Y* n/ F9 y5 N; U; x( b8 `6 p. R
where GFR was 0.1 ml · min - 1 · gkidney - 1 (wet wt) in this study., E c0 U. ~6 R1 B
1 `2 {$ I( ]! j' ^8 c2 D$ i; R1 LIn a previous study ( 38 ),we noted that the charged fiber model did not adequately describe the effectsof changing I. However, introducing large pores improves the precisionsignificantly. Moreover, introducing a heterogeneous fiber network with small regions with low fiber concentrations (1/20th of the average) further improvesthe agreement between theory and experimental data. Thus the total fractionalclearance for a solute is the sum of the through the main gel( main gel ) and that occurring through the large porediscontinuities ( L ). The large pores represent a smallfraction (f L ) of the total hydraulic conductance and an evensmaller fraction (f 2 L ) of the exchange area(A 0 / x). Hence, L is calculated as for main gel except for the fact that
* w7 X& L: D0 S/ X" D e2 e; ]4 w3 q) G" \9 G5 x
which will affect and PS and the resulting fractional clearance can bewritten as
( ~: r, F9 t. Q1 R( }6 |
3 C2 b) j: k) t3 H) j5 a, u( 13 )/ |4 L* l# N5 Q5 N) u
0 [, k$ V8 D, K% }3 `" u3 s1 K- {The important parameters in the model are: the fiber radius (r f ), the relative concentration of fibers in the gel ( ), the surface charge densities of solute (q s ) and fiber (q f ),the unrestricted exchange area over diffusion distance(A 0 / x), the large pore fraction of the hydraulic conductance(f L ), and the dilution factor for the fiber density in the large pores. Some of these parameters were constant (r f, q s for albumin and Ficoll, the large pore dilution factor), whereas others weremodified (A 0 / x,, q f, f L ) toachieve a good fit between experimental and theoretical data.
" H+ e1 F4 [1 M# a3 ~
0 m+ d3 _4 p! P/ h! h% e4 c Q' dCurve-Fitting Procedures
0 f2 D+ P$ d; o5 r* t2 |
# W8 o, D/ o+ V. C9 {6 U) JIn the present study, the fractional clearances for Ficolls of differentmolecular sizes were modeled using Mathcad 2001i (MathSoft Engineering &Education, Cambridge, MA). First, different values for A 0 / x,, q f, f L were tested to achieve acceptable fittingbetween modeled and experimental data at normal I (151 mM). Second, the sameparameter values were used to calculate the fractional clearance for Ficoll atlow I, 34 mM. This resulted, however, in poor fitting between experimental andmodeled data particularly for the smaller solutes. Finally, the concentrationof fibers was gradually reduced by low I until acceptable agreement wasobtained between the experimentally determined fractional clearance for Ficolland the values obtained by the heterogeneous charged fiber model. Details ofthe calculations are given in a separate PDF file available at the Journalwebsite ( http://ajprenal.physiology.org/cgi/content/full/00227.2001/DC1 ).; S% d$ R6 Z8 m( T/ U% J' z6 R
, H9 ?1 F; I8 q; v, v3 c. Z* @8 ]% }Statistics
% H& E* k" m* A/ s4 g* K, W
2 ]% K8 m" w/ A) z3 BData are presented as means ± SE or with 95% confidence intervals(CIs). For the two-pore model parameters and for albumin,the statistical analysis was based on the logarithmic values, due to theskewed distribution of data. Differences were tested using Student's t -test paired design./ b0 {/ ?! W$ d* a; L* I2 y
3 ~; J" G, I! p9 L& O4 g8 R# f: fRESULTS8 R5 _& z/ G0 } I8 u( [% D
* [ G: P3 E( ~: V1 ?# n, o/ p- l
Reperfusion Experiments of the Fixed Kidney' c6 A7 s, _1 g( x9 W, V; u- U
& r4 f* K8 P1 _" v6 c( cThe urine concentrations of sodium and potassium were equal to those inperfusate (data not shown). The inulin concentration ratio between perfusateand urine was 1.01 ± 0.02 ( n = 10), i.e., not significantlydifferent from unity.
1 L/ c4 p- K. f- N3 G( ^' K8 k9 w8 s, {8 Q* J" T0 ^
GFR and Renal Perfusate Flow
# [9 a% i2 R* m. d' Z& @! t7 p9 a1 p w: _. ]: g& j# G
The values for the GFRs and the renal perfusate flow (average ± SE)are reported in Table 1.
. P/ j1 u/ `1 K: ?. p
% U; i1 _% l4 ~, t kTable 1. GFR and RPF0 P: p" H# c3 _/ \
( \6 H! G$ ~7 Q rFractional Clearance of BSA and Ficoll 35.5Å7 r, t z7 _$ P1 r+ Z* e
M9 @- K% x) V
At normal I, the sieving coefficient,, for albumin was 0.0049 (SE-0.0017, 0.0027, n = 6), i.e., about 1/20th of that forneutral Ficoll of similar size ( a SE = 35.5 Å) 0. 104(SE = 0.010, n = 5, P for albumin was 0.0030 (SE -0.0011, 0.0018, n = 6), not significant compared with that at normal I. ForFicoll 35.5Å was 0.104 (SE 0.015, n = 6, notsignificant compared with normal I). Thus forFicoll 35.5Å was significantly higher than for albuminat low I as well ( P Fig. 1. R1 v" m" b' a( `' l+ m
( O. v) P5 f. G' [ WFig. 1. Fractional clearance ( ) ± SE for Ficoll 35.5Å ( right ) and albumin ( left ) with normal (151 mM, open bars)and low (34 mM, closed bars) ionic strength perfusates. The for Ficollwas higher than for albumin at both low (** P P
6 {, \/ O \6 F3 n0 y; G% L7 c# v$ U$ M! l
Figure 2 shows the sievingcoefficients obtained in individual reperfusion experiments of the fixedkidney for BSA compared with that of Ficoll of a SE 35.5Å. All data fall to the right of the line of identity indicatingrestriction of the anionic albumin compared with the neutral Ficoll of similar hydrodynamic size, i.e., a significant glomerular charge barrier is evident. Figure 3 illustrates the U/Pconcentration ratios for Ficoll of various molecular radii.
% o |- c% @' ?
- i# j7 Q K. u7 O0 DFig. 2. Fractional clearance of Ficoll 35.5Å plotted against thefractional clearance for albumin in the perfusion-fixed rat kidney. Data areshown for kidney perfusion with low ( ) and normal ( ) ionic strengthsolutions.. _5 s; |# G& [! M. h1 M
~/ f" `) K. Y, o* `Fig. 3. Urine over plasma concentration ratios for Ficoll plotted against theStokes-Einstein radius. Data are shown for kidney perfusion with low ( )and normal ( ) ionic strength solutions.
, M1 F L7 C9 O( V' ^3 G& E( f* k% C2 d! n* e3 `8 \# s
Gel-Membrane Model Analysis7 l6 U3 e7 a8 i) u3 l
) v/ W2 ], x2 A" ]4 ^ ?, ]The functional small pore radius was 36 Å (27-44 meq/l, 95% CI)at normal I and 33 Å (23-43 meq/l, 95% CI) at low I. The largepore radius was 137 ± 8 and 197 ± 30 Å for normal and lowI, respectively. The large pore fraction of the total hydraulic conductance was 2% for normal and 3% for low I. The glomerular charge density,,was estimated to be 38 meq/l (28-71 meq/l, 95% CI) for normal I and 13meq/l (11-16 meq/l, 95% CI) during perfusion with low I perfusate,suggesting a threefold increase in volume of the "charged gel"( Fig. 4 ).# G ~6 r' R6 ?7 Q& q, n
0 J1 i) @4 z8 F/ \% a6 rFig. 4. Estimated charge density using the gel-membrane model at normal and lowionic strength perfusion (** P9 x P! ?9 V4 ^8 K$ \3 q/ S$ w! w
1 r, A& E. `" R$ aHeterogeneous Charged Fiber Model0 c2 m. P8 V' ]) Y2 _
4 j2 }* Q F/ q1 m( T/ iIn this model, the unrestricted exchange area over diffusion distance,A 0 / x, was 100,000 cm. The fiber radius was 4.5, the relative fiber volume,, was 5.6%and the fiber surface charge density was -0.3 C/m compared with -0.022 C/m for albumin ( Table2 ). The gel was heterogeneous with discontinuities with 1/20th ofthe fiber density accounting for 8.5% of the hydraulic conductivity or 0.72%of the total area. With these parameters, there was an acceptable fit between the 180 Ficoll data pairs (U/P ratio vs. Stokes-Einstein radius) obtained atnormal I and the modeled values. As the I was reduced, however, the modeledvalues deviated from the measured data and more so for smaller solutes, i.e.,higher U/P ratios. Figure 5 shows a Blandt-Altman plot demonstrating the deviations between measured andmodeled U/P ratios. To achieve a better fit at low I, a twofold ( 78%)expansion of the gel was assumed, reducing the fiber density to 3.15%( Fig. 5 ).
* P! l0 ~, J3 b: ~, }9 \+ s. x7 \& v
Table 2. Parameters of the hetergenous charged fiber model
$ x0 r5 c! A- l( }' B7 [6 l
0 N0 c O2 A1 o" |4 S4 jFig. 5. Blandt-Altman plot describing the agreement between measured and modeledU/P concentration ratios for Ficolls of different sizes. At normal ionicstrength (I), there is an acceptable agreement. For low ionic strength,however, the fitting is poor unless the fiber density is markedly reduced asfor the diluted model.
+ l! Q2 Z H! C0 S4 K& E* Z3 Z7 K
2 M7 O) b/ [. A0 ^* Y2 NIn contrast to the Ficoll data, the U/P ratios for albumin were notadequately described by the heterogeneous charged fiber model, whichoverestimated for albumin six to eight times. Possible explanationsare presented in DISCUSSION.
3 Q& n# C; m& v' [; N3 h \, X
5 o% g+ v/ G' b2 f" CDISCUSSION, j3 v5 m5 p5 b! t" i0 T, P
' A Y& b* \, p6 ?; c0 H! ~) ?In this study, the functional properties of the glutaraldehyde-fixed glomerular barrier were evaluated using a broad fraction of neutral sphericalFicoll molecules together with albumin using perfusate solutions of variableI. This experimental model may be considered highly artificial, but it isactually ideal for the purpose of this study, as previous studies suggested marked volume changes to occur in the charge-selective compartment in responseto alterations of perfusate I( 38 ).0 X( U0 q9 v. V% D3 g8 o
. R3 B5 ?0 N. s" v1 D( I3 ]2 x/ GOur hypothesis was that if cellular structures and/or the collagen IV-richglomerular basement membrane were responsible for charge selectivity, thenfixation would abolish dynamic changes of the charge density in response toalterations of I. On the other hand, if fixation does not affect the dynamicsof charge density, then mucous structures such as the endothelial cell coatare likely to be involved because they are more resistant to glutaraldehydefixation ( 1, 35 ).
3 U9 k: B+ K! }8 \( m
5 H7 l) I4 F, ` Y: L- WOur main findings were that the fractional clearance,, for albuminwas one order of magnitude less than that for a neutral Ficoll of similarhydrodynamic size (35.5 ). From this chargeselectivity, a charge density of 38 meq/l could be calculated using thegelmembrane model. The value is surprisingly similar to those estimated invivo ( 41 ) and in vitro( 15, 21, 25, 38 ), as noted in a previousstudy on fixed kidneys ( 5 ).Data could also be interpreted in terms of charged fiber densities assuming acertain degree of heterogeneity. Reducing the I of the perfusate did notaffect the relationship between albumin and Ficoll35.5Å as much as expected based on the increasedcharge-charge interactions. Consequently, both theoretical models predict thatthe volume of the gel did increase during low I perfusion. The more accurateheterogeneous charge fiber model suggests a volume expansion of 78%, whereas the gel-membrane model suggests a threefold volume expansion. These dynamicalterations of the charge density in a fixed kidney suggest that the structureresponsible for glomerular charge selectivity is a polysaccharide-rich layerresistant to fixation such as the endothelial cell coat (or possibly the glomerular basement membrane).) X) M/ X. K x% l* R
6 A' P U; ~8 l, X! U" B: z, w! JPermeability Characteristics of the Fixed Kidney7 E, Q9 l& \3 i% x n: H7 M
2 t! K! J# ?) k/ d& s
The sieving coefficient for albumin obtained at neutral pH in the fixedkidney is higher than in vivo but similar to that reported for the unfixedisolated, perfused rat kidney( 37 ). The glomerularpermeability is, however, heterogeneous( 5 ).
0 I. k4 o9 p: ~1 Z) w2 {) c1 C' T; p
% N1 r' S/ j0 @( |) XThe sieving coefficient for BSA at a concentration of 50 g/l does notsignificantly differ from that obtained at a concentration of 10 g/l( 44 ). Similar results havebeen obtained in micropuncture experiments of the isolated rat kidney( 37 ). Therefore, perfusion experiments of the fixed kidney were performed at an albumin concentration of10 g/l to maintain the costs of the experiments low. Recent experimentsshowed, however, that the albumin concentration indeed may affect the sievingof tracer macromolecules ( 20 ). This deviation from the normal physiological protein concentration, albeitdisturbing, will, however, not affect the conclusions of the study.
( j% }* R! a9 v" ?+ w; Y
4 J6 W' p/ Z8 R2 q) A* |Glomerular Barrier4 k/ G' A }0 e
- O8 x. M) P3 C' }; wThe finding that the sieving coefficient of albumin is much lower than thatof Ficoll of equivalent size (35.5 Å) supports the classic notion of acharge barrier ( 4 ). Recently, this notion has been challenged due to some limitations of the dextran used asa tracer ( 28 ). The calculationof the GCW charge distribution according to a simplified model ofcharge-charge interactions( 24 ) gave a charge density of38 meq/l. Thus glomerular charge selectivity was overestimated in theclassical studies ( 9 ) due tothe use of sulfated dextrans. The gel-membrane model has the virtues of beingable to describe glomerular permeability in a variety of situations, and thecalculations are rather straightforward. It is, however, an oversimplified view of the reality because charge and size interactions cannot really beseparated. Thus barriers in series will contribute to the overall sievingcoefficient of a tracer as products (i.e., tot = 1 · 2 · 3 · 4... n ), which suggests that theremust be some degree of size restriction in the gel compartment as well( 20 ). The charged fiber model is theoretically more correct, but it suffers from being highly complex.Indeed, some of the fiber matrix equations have not yet been fully developed.The equations required to estimate the partition coefficients aresophisticated but do nevertheless have certain limitations. They do, forexample, assume random interactions between a solid sphere and one fiber. Inreality, the glomerular barrier is composed of multiple fibers and plasma proteins in an orderly fashion. Moreover, the equations to estimate thereflection coefficient and the diffusion capacity in the gel are crude atpresent. Still, we consider the charged fiber model to be the most accuratetheory for analysis of glomerular permeability.
0 N8 _; i% h; R1 O) v: I8 ]# H, h9 Q7 Q8 F
The gel-membrane model adequately describes both albumin and Ficoll data.The heterogeneous charged fiber model grossly overestimated the fractionalclearance for albumin. This probably indicates that the latter model, despiteits complexity, has limitations. Alternatively, it may suggest that albuminbinds to tubular structures in the fixed kidney causing underestimations of for albumin. There are, however, no indications of such bindingproblems in these kidneys that have been extensively prewashed.: q; W$ X; n9 Y! M
/ ^0 H9 A S& \In the present study, we introduce heterogeneity into the charged fibermodel. Hereby, the adaptation to the experimentally determined Ficoll dataimproved dramatically. It is important to note that similar conclusions weredrawn using the two different models of glomerular charge and size selectivitynamely that perfusion of rat kidneys with low I seems to induce a volume expansion of the gel (by 78% or more).& Z; y% Z; I8 w' v" I5 f
1 o2 V' `8 t2 V' S. vFixation with Glutaraldehyde
6 s6 y% E, S2 w# `5 Q5 X: P
) V Z% \: C3 I9 IThe urine-to-perfusate ratio of one for both inulin and the electrolytesdemonstrates that the urine collected represents glomerular ultrafiltrate. Thetubules are therefore part of an inert system in which the metabolic processeshave been eliminated. The fixed kidney can thus be regarded as a pure "membrane." Histological studies( 36 ) have shown thatglomerular structures of isolated, perfusion-fixed kidneys were similar tothose in vivo, including the distribution of anionic sites in the glomerularbasement membrane as characterized with Ruthenium red. Moreover, in rathindquarter preparations, fixation with glutaraldehyde reduced surface areafor capillary exchange but had no effect on capillary permeability( 14 ).
S! T8 F, E$ [& H/ d, l1 \0 V w
) ?, v/ W7 }- K/ P C' TMucous structures rich in polysaccharides, however, are more resistant toregular fixation techniques and studies on the microanatomy of such structuresmust employ special, nonconventional fixation regimes( 1, 18, 31, 35 ). For this reason, it would be expected that the endothelial cell coat, a polysaccharide-rich layercreating an interface between plasma and endothelial cells, should allowvolume changes even in a fixed kidney. In our study, the low I perfusion didindeed reduce the estimated charge fiber density in the isolated,perfusion-fixed kidneys. As all fixed structures, except the endothelial cellcoat, are rigid and incapable of undergoing the large volume changes requiredto alter the charge density observed in our experiment, this supports thehypothesis that glomerular charge selectivity is related to the cell coatcovering the endothelial cells." U5 R% @; Q, h2 `7 R8 p8 h6 |) P' s
/ K) b8 O/ c% {: K, I$ O
Finally, we have to consider some alternative interpretations of ourresults. Could, for example, the observed alteration in estimated chargedensity be due to something else than volume changes of the glomerular chargebarrier? Indeed, the biophysical models for transport of charged solutesacross charged membranes or gels are far less precise than the theoriesdealing with transport of neutral solutes. However, in a recent study, we compared different models including the most advanced charged fiber-matrixanalysis ( 38 ) and the resultsare more or less the same. All current theories predict that the foralbumin should be reduced by more than one order of magnitude when the I isreduced from 151 to 34 mM. The experimental observations suggest a modest, butstatistically significant, reduction of albumin at low I. Atpresent, the only plausible explanation is that the glomerular charged fiberdensity is reduced. The reversibility of this process demonstrated bySörensson et al. ( 39 )seems to rule out other possibilities than volume changes of the chargebarrier with a constant number of fixed charges. Indeed, dramatic fluid shiftsare to be expected because the electroosmotic pressure of the gel increasesdrastically (reaching 160 mmHg) as the I is reduced (for more details, see Ref. 38 ).2 v0 K" V o* r3 ?$ A
. z/ E& o5 W7 u* l3 C: P: M/ sIn conclusion, the glomerular barrier is size and charge selective. Perfusion with solutions of low I reduced the estimated charged fiber densityby at least 78%, probably due to volume expansion of gel. Because almost allconstituents of the glomerular barrier, except the polysaccharide-richendothelial cell coat, are rigid in the fixed kidney, our findings support theview that the endothelial cell coat can be an important component of the glomerular barrier.1 x* `* Y1 N4 R
& W3 H% m( Q$ w- n: I3 f
DISCLOSURES
1 q! e3 x6 {& j3 X; X& U! k% ]* Y2 A+ g9 o/ {. n& K
This study was supported by Swedish Medical Research Council Grants 9898,the Knut and Alice Wallenberg Research Foundation, the IngaBritt and ArneLundbergs Research Foundation, and Sahlgrenska University Hospital GrantLUA-S71133.
" z7 N& m+ J$ D& H7 A5 @6 S 【参考文献】9 Y' `. { A& H/ Q! B
Allan-Wojtas P,Farnworth E, Modler H, and Carbyn S. A solvent-based fixative for electronmicroscopy to improve retention and visualization of the intestinal mucusblanket for probiotics studies. Microsc Res Tech 36: 390-399,1997./ C7 s7 Z0 O/ t- ?1 Y7 o. {- T
/ h: c* K$ V# M/ H) M, q5 p$ z# J) C8 Z$ v+ q
3 E# o4 p$ g" d+ i" {
Anderson JL andMalone DM. Mechanism of osmotic flow in porous membranes. Biophys J 14:957-982, 1974.
9 M+ E( G1 j* T1 ]1 d* \8 {
z' Z# v3 }; K8 A
' m1 z% X( A$ N/ m5 Y% h
- j9 `7 `4 c( QBradford MM. A rapid and sensitive method forquantification of microgram quantities of protein utilizing the principle ofprotein dye binding. Anal Biochem 72: 248-254,1976.
0 ~ w" y1 h! Y2 V! F' a" s$ V5 f% X8 J: F8 r" B
* U) {2 g9 u! @+ c( ?3 {8 g" K2 b' E
9 Z: h# U0 [2 q: {% I; h' nChang RKS, DeenWM, Robertson CR, and Brenner BM. Permselectivity of the glomerularcapillary wall. III. Restricted transport of polyanion. KidneyInt 8:212-218, 1975.( J C" o% \2 @
" g) Y+ H9 U2 o$ t$ j
4 L+ ^; F S) F3 ^$ k: O+ H+ |
$ u1 g6 T' L* d# pCiarimboli G,Schurek HJ, Zeh M, Flohr H, Bökenkamp A, Fels LM, Kilian I, and StolteH. Role of albumin and glomerular capillary wall charge distribution onglomerular permselectivity: studies in the perfused-fixed rat kidney model. Pflügers Arch 438:883-891, 1999.
! }3 n. d6 W2 ~4 f+ @
2 r. Y7 K' W' D; a% A2 k; i3 T+ N6 _# ?# P% g1 ?. J- L( }
7 l# n" x' U% D- b
Comper WD, TayM, Wells X, and Dawes J. Desulphation of dextran sulphate during kidneyultrafiltration. Biochem J 297:31-34, 1994.: \( [3 J5 o& P2 e- _
; e" g1 u5 k1 S+ }4 R. M% V7 U8 ^1 v, p$ W: S
0 e) T: a1 Z6 l$ q( k# k
Curry FE andMichel CC. A fiber matrix model of capillary permeability. Microvasc Res 20:96-99, 1980.0 `9 M- M' ]6 D) I4 ^& V! [
I4 l3 W8 V9 c
5 x- l" E6 r1 j9 \# V5 P+ c/ Z
4 }6 o ?/ H1 s' ?$ G) vDeen WM,Lazzara MJ, and Myers BD. Structural determinants of glomerularpermeability. Am J Physiol Renal Physiol 281: F579-F596,2001.
; {4 [& T! V# m& n$ `3 I2 @. F# {- f% v2 _
0 ~9 K) x( i8 j) Y3 H) W' y S+ M( Q3 S! @* L" R
Deen WM,Robertson CR, and Brenner BM. Transcapillary fluid exchange in the renalcortex. Circ Res 33:1-8, 1973.( Q/ ~ Z9 ^0 j6 g6 Q# \+ w2 i
/ M7 b; Y% F- r/ d1 _0 A
1 T( ]* I- H$ n. W! T0 {+ q5 I a) D% F+ O# ]
Deen WM, SatvatB, and Jamieson JM. Theoretical model for glomerular filtration of chargedsolutes. Am J Physiol Renal Fluid Electrolyte Physiol 238: F126-F139,1980.
: r" I: D* o2 `( |! _! S) L3 B
3 d, @+ R t+ U& K/ s a
3 i/ D: a, b/ @- f1 A- y# v
# d( ~7 C6 ?7 X& AElling F,Hasselager E, and Friis C. Perfusion fixation of kidneys in adult pigs forelectron microscopy. Acta Anat (Basel) 98: 340-342,1977.' z+ F/ k' Q+ L5 a
* L' `! w& S1 Q; p" y5 j
7 P, B% M8 v/ n- J8 F
% G7 P% t$ Y5 x( i0 z4 V; `7 [) f( c
Gade R. Microangiography in the perfusion-fixed rabbit kidney: a single injectiontechnique for the study of the nephron. Invest Radiol 12: 348-352,1977.
0 R% v. M6 X1 X3 Q1 D- H
2 c& A7 ^/ {. a- ~1 Z! p5 A5 b' ?; A+ L
, s; Z6 m1 c W
Guash A, DeenWM, and Myers BD. Charge of the glomerular filtration barrier in healthyand nephrotic humans. J Clin Invest 92: 2274-2282,1993.
1 Q2 V+ S8 f2 q! w& [/ ^. i
4 Y- T b1 f {1 N( N- R, S+ e/ _" k: l% Y% r# ^- f) I2 d0 |
% f+ [9 T, X* F: S& e: o
Haraldsson B and Johansson BR. Changes in transcapillary exchange induced by perfusionfixation with glutaraldehyde, followed by measurements of capillary filtrationcoefficients, diffusion capacity and albumin clearance. ActaPhysiol Scand 124:99-106, 1985.; ] M+ s1 w9 T! k1 g
. D% R3 H. y/ q; [7 D m7 _( N
9 C! Y+ f' f' x3 {& e% {: v" k
! s8 N4 \7 i' D, _0 P
Hjalmarsson C,Ohlson M, and Haraldsson B. Puromycin aminonucleoside damages theglomerular size barrier with minimal effects on charge density. AmJ Physiol Renal Physiol 281:F503-F512, 2001.
; Q& e- a" T0 O: n/ b# Q7 g9 V7 T6 s2 o- l, j8 T
5 K; v+ g# y. D( Y0 ]; l+ W
9 X. d k- P8 e' A, o5 |# C
Johnson EM andDeen WM. Electrostatic effects on the equilibrium partitioning ofspherical colloids in random fibrous media. J Colloid InterfaceSci 178:749-756, 1996.5 }6 ]5 N! I# r0 t4 k' V1 H
% e$ s0 n2 ~" u" B* e) N8 q6 U" m
! ^2 s" F8 l' T$ ?. b9 Y0 g# i
( E1 ^7 E" ]2 t0 _# { l
Johnston ST andDeen WM. Hindered convection of proteins in agarose gels. JMembr Sci 153:271-279, 1999.: R# q. P* ?/ Y3 g: E
/ I5 p1 k1 o5 T y: ^. k9 L1 S
" k1 j& I- o* H6 q+ n1 Q2 O4 b1 ?2 @5 S4 U6 h
Jongebloed W,Stokroos I, Van der Want J, and Kalicharan D. Noncoating fixationtechniques or redundancy of conductive coating, low kV FE-SEM operation andcombined SEM/TEM of biological tissues. J Microsc 193: 158-170,1999.
; l5 X5 X/ j9 e: X- d
+ _* g- @& {. e
8 G+ z- F0 q5 q4 h1 M* |- a( A& J# O: I8 {4 z7 b% q
Kriz W,Hackenthal E, Nobiling R, Sakai T, Elger M, and Hahnel B. A role forpodocytes to counteract capillary wall distension. KidneyInt 45:369-376, 1994.. M5 h, Y: p( } Q. ~6 Z
( }' g+ {' }0 j3 X: _
3 Q. {9 S) x; V% `0 W. M
. y+ c' i0 V6 T
Lazzara MJ andDeen WM. Effects of plasma proteins on sieving of tracer macromolecules inglomerular basement membrane. Am J Physiol RenalPhysiol 281:F860-F868, 2001.
2 S( `# w. f- l3 U2 C1 j/ Q2 O" K, q" W2 `
6 ~$ S b. m$ A7 K; K! K& ~, a/ T$ z
# F: e8 k, [; t- A6 D2 Z
Lindström KE, Johnsson E, and Haraldsson B. Glomerularcharge selectivity for proteins larger than serum albumin as revealed bylactate dehydrogenase. Acta Physiol Scand 162: 481-488,1998.
: s2 v, U6 q' Q& T, C o; a, {1 {+ d* s# ]/ y8 F2 Z5 F0 @1 T& v. H7 ^
0 [: r7 X" R& ?( v8 S. [: {* V
5 w- M: z5 r y8 x6 I3 y2 |6 cNishiitsutsuji-Uwo JM, Ross BD, and Krebs HA. Metabolic activitiesof the isolated perfused rat kidney. Biochem J 103: 852-862,1967.+ ^' ~$ d) D {
0 n. E0 b7 B; X! F
; R7 J8 V$ N Q6 y }7 |2 _0 L2 Q; m) x2 T, F& ]$ X6 _
Ogston AG. The spaces in a uniform random suspension of fibres. Trans FaradaySoc 54:1754-1757, 1958.
$ K, R/ o9 s3 [2 Y
6 P5 s7 ^& p9 z f3 A4 Y* |. F8 D( x% d
2 z2 q/ L' g$ {; YOhlson M,Sorensson J, and Haraldsson B. A gel-membrane model of glomerular chargeand size selectivity in series. Am J Physiol RenalPhysiol 280:F396-F405, 2001.1 N" F4 n$ l: x
/ `$ R! s5 U+ }1 H! ?1 l B
: n0 l+ }" }& d: s4 [
. v x. B I% ?- W5 X+ \Ohlson M,Sorensson J, and Haraldsson B. Glomerular size and charge selectivity inthe rat as revealed by FITC-ficoll and albumin. Am J Physiol RenalPhysiol 279:F84-F91, 2000.. Q9 i8 ] Y0 K: o0 m" g4 n$ ]
% r. e# A* h: o8 O, r: ?0 w) d
) S0 B7 v* l6 ]. w" y4 b& A
4 F/ H8 P: E/ s, j1 xOken D andFlamenbaum W. Micropuncture studies of proximal tubule concentrations innormal and nephrotic rats. J Clin Invest 50: 1498-1505,1970.9 k: z9 T G, [; Q% I ?
6 L5 z: g$ A& ?1 d( F7 j
3 X+ s# ]' o, X8 x s' e* t
" g8 y p$ O5 U
Oliver JP,Anderson S, Troy JL, Brenner BM, and Deen WM. Determination of glomerularsize selectivity in the normal rat with Ficoll. J Am SocNephrol 3:214-228, 1992.
5 S3 I) m5 A$ \4 j' m( m% U: ?7 O% d( [1 u5 V9 m
' s6 D4 [ s4 X3 F8 I1 Z' e) g9 E
* x& A2 V& X. N" POsicka TM,Pratt LM, and Comper WD. Glomerular capillary wall permeability to albuminand horseradish peroxidase. Nephrology 2: 199-212,1996.1 Y! @' w4 J! }5 m
- |4 a9 g) ?! N+ |, i
& n" p! ^# H) \, o' \! H
: T' q2 m; l. F' L4 U2 RPhillips RJ. A hydrodynamic model for hindered diffusionof proteins and micelles in hydrogels. Biophys J 79: 3350-3353,2000.
5 ^3 d6 r, V4 C+ Y4 l1 k
; m' c* T! K) @4 H
6 Y7 z! ^& \4 r' E5 i( G' M$ e. @7 w* U m# N# E* {2 Y9 B6 F
Rippe B andHaraldsson B. Transport of macromolecules across microvascular walls: thetwo-pore theory. Physiol Rev 74: 163-219,1994.
9 C+ |7 c3 \/ ~6 E( m+ J8 w7 ^
! m( Z$ I9 y) C* W# J9 P! C/ r8 `$ C2 }5 D7 b6 u
0 l( O: t- ?6 s! \5 o* S
Sanchez J,Speare D, Sims D, and Johnson G. Adaptation of a fluorocarbon-basednon-aqueous fixation regimen for the ultrastructural study of the teleostepithelial mucous coat. J Comp Pathol 117: 165-170,1997." J( @/ r2 T) H8 a
$ R1 I* C8 ~% Y) w. A( b
, J2 e* U" {$ R+ E5 |+ t. E" v( ]
6 H; _' d; L7 E7 d; CSchmidt F. Die enzymatische Bestimmung von Glucose und Fructose nebeneinander. Klin Wochenschr 39:1244-1247, 1961.
: F! D& O @! _2 I& i* T4 X# L8 P6 }9 _
% A( P ]# B" S. h# j
7 h( N9 v- B% u$ \; ]$ o
Schurek H. Application of the isolated perfused rat kidney in nephrology. Contrib Nephrol 19:176-190, 1980.9 ?/ f- g3 H, E
0 ]6 p q: m" \8 V9 X9 J% N7 d$ E' h" s5 M' A4 c8 U
+ X# w& O3 T) Z1 PSchurek H andAlt J. Effect of albumin on the function of perfused rat kidney. Am J Physiol Renal Fluid Electrolyte Physiol 240: F569-F576,1981.
% ^8 q" ]7 f, Q: k" M: V5 ~( j
3 _, V( x" R, M+ S7 p6 F" s4 I& G* d/ N" z8 r0 O' t$ V7 `
+ n5 i, U' G% O: B$ C, H: I
Sims D andHorne M. Non-aqueous fixative preserves macromolecules on the endothelialcell surface: an in situ study. Eur J Morphol 31: 251-256,1993.
1 R$ q; \! N6 n& z; N! J6 s: c9 a9 P
+ q Z9 B5 r3 L
/ P7 k' t4 k9 S Y) e& M2 L# eSonnenburg-Hatzopoulos C, Assel E, Schurek H, and Stolte H. Glomerularalbumin leakage and morphology after neutralization of polyanions. II.Discrepancy of protamine induced albuminuria and fine structure of theglomerular filtration barrier. J Submicrosc CytolPathol 16:741-751, 1984.* v. `" n* s/ l8 |# A
% B% f, f/ g4 ]3 C0 k/ X: k9 Y2 W
; C2 R3 F1 Q: H8 p, Q' m; l
# c4 B0 H7 H/ `* t7 w% yStolte H,Schurek H, and Alt J. Glomerular albumin filtration: a comparison ofmicropuncture studies in the isolated rat kidney with in vivo experimentalconditions. Kidney Int 16:377-384, 1979.
4 Q8 n7 Y: W) `, ^7 L# T6 J# p% ^
% }/ e/ `% X5 N4 S. n
c9 F$ U" W x4 E9 U# W$ _$ ISörensson J, Ohlson M, and Haraldsson B. A quantitativeanalysis of the glomerular charge barrier in the rat. Am J PhysiolRenal Physiol 280:F646-F656, 2001.
# x& w2 x% k; M' P( y3 b
9 E1 H+ q1 a" c+ }# x1 {" {9 D# I% s/ m9 z
9 ]" n; A4 }/ p: I9 kSörensson J, Ohlson M, Lindström K, and Haraldsson B. Glomerular charge selectivity for horseradish peroxidase and albumin at lowand normal ionic strenghts. Acta Physiol Scand 163: 83-91,1998.# V' {! e% r8 {
5 h1 V' g/ k8 o6 V0 T3 u
r. l7 Z& b* z
+ U8 d5 [( a/ {Weiss C, PassowH, and Rothstein A. Autoregulation of flow in isolated rat kidney in theabsence of red cells. Am J Physiol 196: 1115-1118,1959.5 ^& Y* n3 @" l( s, }
8 l7 `3 x! z. P5 I/ i
/ J+ g/ P/ \+ e w5 R) P% G4 X
; l) V G2 z0 R% b9 GWolgast M,Källskog, and WahlströmH. Characteristics of the glomerular capillary membrane of the rat kidney as ahydrated gel. II. On the validity of the model. Acta PhysiolScand 158:225-232, 1996.2 O% F/ C: z& ^# b
% [- r" S; K# ~
+ p/ Q( |* j6 m/ D0 a, Y; F. c* p, R: F, e
Vyas S andComper W. Dextran sulphate binding to isolated rat glomeruli andglomerular basement membrane. Biochim Biophys Acta 1201: 367-372,1994.
, J; K b u( [" O! F8 F Q$ U L9 o3 T! s8 G( k
' H$ p) d- k( U: l, _( @3 p' d Y1 F& d# H3 H# X& F
Vyas S, ParkerJ, and Comper W. Uptake of dextran sulphate by glomerular intracellularvesicles during kidney ultrafiltration. Kidney Int 47: 945-950,1995.; T& k9 j3 U( R$ J7 }3 ]6 Y
$ q& i5 x c$ R2 j( M7 G5 N& d) ~% M& w) B# V8 S8 @$ x
; |: j/ z: {2 _2 u% VZeh M. Zur glomerulären permeabilität geladener und ungeladenermakromolekule (Doctoral Thesis). Hannover, Germany: Hannover MedicalSchool, 1994. |
|
发表于 2009-4-21 13:45
