Determinants of GFR depression in early membranousnephropathy - 《American Journal of Physiology》美国生理学杂志干细胞之家



免疫细胞治疗专区	欢迎关注干细胞微信公众号

12 3 4 5 6 7 8 9 10 ... 23 下一页

返回列表

查看: 362012\|回复: 221	go Determinants of GFR depression in early membranousnephropathy [复制链接]

轻羽

新手上路

Rank: 1

积分: 0
威望: 0
包包: 0

楼主

发表于 2009-4-21 13:37 |只看该作者 |倒序浏览 |打印

作者：M. A.Hladunewich, K. V.Lemley, K. L.Blouch,  B. D.Myers作者单位：Divisions of Nephrology, Departments of Medicine andPediatrics, Stanford University School of Medicine, Stanford,California 94305 7 g8 i8 t+ H6 r: q- C6 H# ?( `





         ; ?* @0 T6 b0 K" ^, P
         ( l! S0 i3 f$ P
         7 y7 O4 c) e) ~, K- c* `

      3 ~1 b& s" V- ^& o

      0 A& N- H! y& ~# `
      【摘要】
   Weevaluated the glomerular filtration rate (GFR) in 34 subjects withmembranous nephropathy (MN) of new onset. We used physiological techniques to measure GFR, renal plasma flow, and oncotic pressure andcomputed a value for the two-kidney ultrafiltration coefficient ( K f ). A morphometric analysis of glomeruli inthe diagnostic biopsy permitted computation of the single-nephronultrafiltration coefficient (SN K f ). MN subjectswere divided into two groups: moderate or severe, according to whetherGFR was depressed by less or more than 50%.SN K f was subnormal but similar in moderate andsevere MN. In contrast, two-kidney K f wassignificantly more depressed in severe than in moderate MN. Weestimated the total number of functioning glomeruli( N g ) by dividing two-kidney K f by SN K f. Whereas mean N g was similar in controls and moderate MN (1.5 and 1.4-1.7 × 10 6, respectively), it wassignificantly lower in severe MN (0.5 × 10 6 ). Thisdegree of glomerulopenia was not reflected in the rate of globalsclerosis. We conclude that a combination of depressed SN K f (due to foot process broadening) andprofound glomerulopenia 50% early inthe course of MN. The cause of the glomerulopenia remains to be elucidated. 5 V% U: E% v. d; o$ o: }
      【关键词】 glomerular filtration rate glomerular hemodynamics ultrafiltration coefficient glomerular morphometry glomerular number" c4 O' p# N& Y. J% D7 c, L# k$ X
   INTRODUCTION
8 k& B  n, ~9 \+ |) E
THE ONSET AND EARLY STAGES of membranous nephropathy (MN) are often associatedwith depression of the glomerular filtration rate (GFR) ( 11, 25-27 ). This phenomenon has been shown by micropuncture study of early experimental (Heymann's) MN in the rat to be a consequence of depression of the ultrafiltration coefficient( K f ), a measure of the intrinsic ultrafiltrationcapacity of glomerular capillary walls ( 1, 13, 29 ). Thenet pressure for ultrafiltration, the remaining determinant of GFR, hasinvariably been found to be elevated in experimental MN, indicatingthat K f depression is the sole factor leading tohypofiltration under these circumstances ( 1, 13, 29 ).) E7 E( }1 {7 M0 e- q0 c0 X

K f is the product of the hydraulic permeabilityof glomerular capillary walls and the surface area available forfiltration ( 9 ). We studied these determinants of K f by using a stereological approach to quantifythe structural changes in glomeruli obtained by percutaneous renalbiopsy from patients with active MN ( 11, 27 ). Such studiesreveal the autoimmune injury to glomerular epithelial cells (podocytes)that underlies MN to lead to gross deformation of their foot processes.An ensuing decline in the number of filtration slits, through whichfiltrate gains access to Bowman's space, impairs the hydraulicpermeability of the affected glomerular capillary walls( 11 ). Our morphometric analyses have previously pointed toimpaired hydraulic permeability as the only identifiable GFR-loweringfactor early in the course of MN ( 11, 26, 27 ). Incontrast, a recent analysis of serial biopsies revealed that bothimpaired hydraulic permeability and a loss of filtration surface areacontributed to chronic and persistent depression of K f and GFR after 2-5 yr of MN( 26 ).
# s7 f" v7 J) Z; W1 i
GFR depression, presenting as azotemia, at the onset of MN has beenidentified as an early predictor of eventual progression to end-stagerenal failure ( 6, 20, 22, 30 ). We thus designed thepresent study to further elucidate the mechanism of hypofiltration inearly MN. We once again used morphometric techniques to examine glomerular structure in the diagnostic biopsies of a large number ofpatients with MN of new onset. We combined the structural findings witha physiological evaluation of GFR and its hemodynamic determinants. Wethen used mathematical modeling to estimate ultrafiltration capacity,both at the level of individual glomeruli [single-nephron K f (SN K f )] and of theaggregate of all functioning glomeruli in the two human kidneys(2-kidney K f ). The subject of this report is therelationship among these two quantities and the extent to which GFR wasdepressed in two groups of subjects with MN of graded severity.9 Z6 X! j- i9 u5 o8 |  A, K
9 l6 k; R% k/ C* T0 n
METHODS1 Q" c' D+ _3 K

Patient Population# ^4 u) g2 G  B, ~
  W9 N% y" C+ {, D# u
The subjects of our study were 34 adult patients who presentedto our clinic with a nephrotic syndrome and a histopathological diagnosis of MN. The patients varied in age from 17 to 71 yr, and 22 were men. All physiological (clearance) studies were performed within ayear of the diagnostic biopsy (median interval = 2 mo). Two groupsof healthy individuals were studied to provide control values for theglomerular functional and structural parameters. Control group1 was composed of 130 healthy volunteers. Their ages variedbetween 18 and 80 yr, and 89 were men. They underwent renal clearancestudies comparable to those performed in the patient population.Control group 2 was composed of 19 living kidney transplant donors. Their ages varied between 23 and 48 yr, and 11 were men. Eachunderwent a renal biopsy at the time of transplantation. All denied ahistory of renal disease, hypertension, or diabetes mellitus. At thetime of evaluation, each was found to be normotensive andnormoglycemic, to have a normal serum creatinine level, and to have aurinary protein excretion rate in the normal range.
+ g* I4 H9 _# X; W5 E" k* }
Physiological Evaluation

All patients and volunteers underwent a determination of GFR,renal plasma flow, and preglomerular vascular pressures according to aprotocol approved by the Institutional Review Board at the StanfordUniversity School of Medicine. Initially, blood was sampled fordetermination of plasma oncotic pressure ( A ). Urine wasvoided spontaneously after diuresis had been established with an oral water load (10 to 15 ml/kg). A priming dose of inulin (50 mg/kg) andpara-aminohippuric acid (PAH; 12 mg/kg) was then administered. Thereafter, inulin and PAH were given by continuous infusion to maintain plasma levels constant at ~20 and 1.5 mg/dl, respectively.* N  t! l7 I5 O5 V# f0 E/ [
7 {; [& p" i: S3 v# b( z4 x
Sixty minutes after the priming infusion, arterial blood pressure wasdetermined. Four timed urine collections were then made, each of whichwas bracketed by a blood sample drawn from a peripheral vein. The GFRwas expressed as the average value for the four timed inulinclearances. The rate of renal plasma flow was estimated by dividing thecorresponding clearance of PAH by an estimate of the prevailing renalarteriovenous extraction ratio for PAH. We showed previously thatreductions of GFR and peritubular capillary protein concentration exertan additive effect to lower the PAH extraction ratio in patients withglomerular disease ( 3 ). From the relationship observed inthat study between the PAH extraction ratio and GFR, we assigned thefollowing values to the subjects of the present study: 0.9 for healthycontrols, 0.8 for patients 80ml · min 1 · 1.73 m 2 ), and 0.7 for patients with MN and a depressed GFR.Inulin and PAH were determined with colorimetric methods using aTechnicon Auto Analyzer II ( 3 ). Plasma oncotic pressurewas measured directly using a Wescor 4400 membrane osmometer (Wescor,Logan, UT) and serum creatinine levels by a rate-dependent modification of the Jaffe reaction, employing a Beckman Creatinine Analyzer (model2, Fullerton, CA).* n& f& C$ f& w  p' ~$ k
8 m4 I: s8 S7 K
Morphometric Evaluation
+ u: ]+ Y; k' }. i( ?3 w9 ]; [
Light micrososcopy. All glomeruli in a single, 1-µm-thick section stained with periodicacid-Schiff reagent were analyzed at the light microscopic level. Onaverage, 14 (range 4-49) glomeruli were examined in each diagnosticbiopsy in the patients with MN. The average number of glomeruli amongthe 19 control biopsies was 19 (range 5-58). A dedicated computersystem (Southern Micro Instruments, Atlanta, GA), consisting of a videocamera and monitor, microscope, and digitizing tablet, was used toperform the measurements. The outline of each glomerular tuft in thesection was traced onto the digitizing tablet and the mean tuftcross-sectional area was determined using computerized planimetry. Themeasured tuft area included any parts with segmental sclerosis. We nextcounted the number of patent ( N p ) and globallysclerotic ( N s ) glomeruli in a single section ofcortical tissue. Serial sections were examined to verify the assignmentof N s in the single section. The percentageof globally sclerotic glomeruli (G l ) was calculated by 1 = N s N s +N p ( D s /D p ) ×
. `1 v, Z, D  r9 i4 m
( 1 )

where D s and D p are the mean diameters of globallysclerotic and patent glomeruli, respectively, derived from the tuftcross-sectional areas. The ratio accounts for the difference in theprobability of encountering a glomerulus of either type in a randomcross section due to their different sizes. Glomerular volume(V G ) was calculated from the average tuft cross-sectionalarea ( A G ) as follows G = &bgr; d ( A G ) 3 / 2 ( f s9 x" K; x. F, }! N

( 2 )3 F, ?( U) `$ W1 g! k1 P
/ q5 M9 J# n5 M
where is a dimensionless shape coefficient( = 1.38 for spheres), d is a sizedistribution coefficient ( d = 1.1), which is used toadjust for variations in glomerular size ( 28 ), and f s is a correction factor for the tissueshrinkage associated with paraffin embedding( f s = 1.64) ( 17 ). Thefractional interstitial area was examined at ×600 magnification. A10 × 10-point grid was superimposed over each field in the entirecross section, and the fraction of total area occupied by interstitiumwas determined by point counting. Interstitial area was defined as thatoutside of tubular and vascular structures, other than peritubular capillaries.
' P! @9 ~; P: z; w- l* u% }4 H" }
Electron Microscopy9 h1 v, V' b( P  W4 b
$ D7 D$ f+ m0 {2 F. ~
For transmission electron microscopy, tissue was fixed in 2.5%glutaraldehyde and 2.0% paraformaldehyde in 0.1 M cacodylate bufferand then embedded in Epon. Toluidine blue-stained sections weresurveyed to locate blocks with patent glomeruli present entirely withinthe block. An ultrastructural analysis was performed on two glomerularprofiles in each patient. Ultrathin sections (~90 nm) of the selectedglomeruli were stained with lead citrate and uranyl acetate. A completemontage of each glomerulus (×2,850 magnification) was prepared andline-intercept counting was used to calculate the fractional surfacedensity of the peripheral capillary wall by standard stereologicmethods ( 28 ). Six to eight images of peripheral capillaryloops in each of the glomerular profiles were then photographed at×11,280 to evaluate the frequency of epithelial filtration slits andthe thickness of the peripheral glomerular basement membrane (GBM).Filtration slit frequency was determined by counting the total numberof epithelial filtration slits and dividing that number by the totallength of the peripheral capillary wall at the epithelial interface( 11 ). The harmonic mean basement membrane thickness( bm ) was calculated for each individual usingthe method of orthogonal intercepts ( 14 ) bm = 8 3 &pgr; ×&dgr;′ bm′

( 3 )
  S' ^- R3 ~$ W& p8 ]% j" h
where ' bm' is theapparent harmonic mean basement membrane thickness. Measured thicknessincluded both normal GBM material and intramembranous deposits. Thenumber of intercepts per individual was between 142 and 192 on average.
9 i8 q7 K0 {0 X" X4 t
Calculations
2 L+ P' o8 T7 L( d. O. s1 N; T
Glomerular oncotic pressure. We showed that oncotic pressure in nephrotic humans rises linearly asplasma flows axially along the glomerular capillaries and water isremoved by ultrafiltration ( 5 ). We first calculated efferent (postglomerular) oncotic pressure( E ) as follows: E =&pgr; A (1 −

( 4 )+ _  X% U. R- \* o
. L7 @, f2 Y3 x: z: C2 y9 c7 Z
where A is the afferent (systemic)oncotic pressure and the FF is the filtration fraction. We thenestimated mean glomerular oncotic pressure( GC ) as the arithmetic mean of A and E.
) R, `4 @' q; k8 c! N
Two-kidney K f. A mathematical model for the glomerular filtration of water ( 9, 27 ) was used to calculate the two-kidney K f, which is defined in this study as theproduct of glomerular hydraulic permeability and the total filtrationsurface area of all glomerular capillaries in the two human kidneys.The input values for the model included the measured values of GFR,renal plasma flow, and A, as well as anassumed value for the glomerular transcapillary hydraulic pressuredifference ( P). The latter quantity cannot be directly measured inhumans. However, using an indirect curve-fitting technique, weestimated that P approximates 40 mmHg in the healthy human kidneyand assigned this value to both the control and MN groups in thepresent study ( 18, 27 ). Micropuncture determinations inHeymann nephritis, a rodent model of MN, indicate that P is invariably elevated in this form of glomerular injury ( 1, 13, 29 ). Moreover, human MN is accompanied by arterial hypertension (see below). Given that a fraction of the increment in arterial pressure is likely transmitted into glomerular capillaries, it isprobable that P in human MN is also elevated. Thus, an assumption that P in MN is the same as in healthy controls is a conservative one and should provide an upper bound for the average K f in this disorder ( 27 ). To allowfor the effect of possible variations in P on computed membraneparameters in patients with MN, we performed a sensitivity analysis,repeating all calculations over a hypothetical P range (35 to 45 mmHg) that brackets the assumed control value of 40 mmHg.
3 f6 O& e1 x) I8 m% w2 T
Single-nephron K f. The total filtration surface area in a single glomerular tuft wascalculated from v × V G

( 5 )! x: ^. H6 a' ?# L' s

where S v and V G are,respectively, the filtration surface density and glomerular tuft volume.) @5 F5 E6 g5 |: R1 ?3 o/ S* i

The intrinsic hydraulic permeability of the glomerular capillary wall( k ) was estimated from the filtration slit frequency (FSF)and basement membrane thickness by using a hydrodynamic model ofviscous flow that has been described in detail elsewhere ( 8, 11 ). In this model, the capillary wall consists of a largenumber of repeating structural units, each of which is based on asingle filtration slit. The width of such a structural unit ( W ) is calculated from the FSF by 2 &pgr; × 1 FSF

( 6 )
- j' W# t# _4 B9 M! r: G' ?
where 2/ is a stereologic factor that accounts forthe random angle of sectioning.# U: P0 y) s4 z3 D* L& b

Considering the capillary wall as a system of resistances in series,the overall hydraulic permeability is calculated from thepermeabilities of each component layer by 1 k en + 1 k bm + 1 k ep −13 r* A# Y  A! x: m/ R
6 Y/ Q- |2 B/ x5 L4 U0 u8 E4 V
( 7 )6 ~5 r/ `: n1 q- i6 ]
" n: J" m1 H5 X) ?, s. F
where k en, k bm, and k ep are, respectively, the hydraulicpermeabilities of the endothelium, basement membrane, and epithelium. Many of the needed structural parameters have not been measured for thehuman glomerular capillary wall, necessitating substitution ofcorresponding values derived from rats, as described in detail by uspreviously ( 11, 26 ). The values derived from previous studies in the rat and used in the model calculation include the permeabilities of the endothelium ( k en, 2.0 × 10 7 m · s 1 · Pa 1 )and of the slit diaphragm ( k s, 7.9 × 10 8 m · s 1 · Pa 1 ),the filtration slit diaphragm width ( W s, 41 nm),and the Darcy permeability of the glomerular basement membrane( k D, 2.7 nm 2 ) ( 7, 10 ).A preliminary study in our laboratory showed that the value for W s in humans is probably quite similar (36 ± 4 nm, n = 4) and does not appear to differ betweenpatients with MN and healthy controls ( n = 2 each).

The permeability of the epithelial layer was calculated using ep =&egr; s k s = W s W  k s3 r9 b  c7 X7 E* @2 n# R

( 8 ). ]# k) m$ u' K' B: V
7 c, B4 u% t! w- |
where s is the fraction of thebasement membrane area occupied by filtration slits and W s is the slit width( s = W s / W ). The permeability of thebasement membrane ( k bm ) was calculated using Eq. 21 of Drumond and Deen ( 10 ).  \9 ?0 Q, c5 c4 e8 `3 z, [- [
" L, L4 X! ?8 \7 K
The single-nephron ultrafiltration coefficient(SN K f ) was calculated from the product offiltration surface area ( S ) and the local hydraulicpermeability of the walls of patent glomerular capillaries( k ) in the glomeruli that were examined ultrastructurally. In making this calculation, we corrected for the effect of immersion fixation to decrease glomerular dimensions relative to in situ perfusedglomeruli ( 17 ).
. g4 h1 l; n& H6 j* S- y9 d1 I, b" A
Number of glomeruli. We estimated the total number of functioning glomeruli( N g ) in the two kidneys as the quotient g  = 2-kidney  K f  /single-nephron  K f
+ t8 U2 T* D) S  C. P6 k
( 9 )

Statistical Analysis
$ h8 y9 S2 S5 K* |! [' q2 T' _
Initially, Student's t -test was used to assess thedifference in the GFR between the control group and all patients withMN. Linear regression analysis was used to elicit possiblerelationships between the GFR and a number of morphometric measurementsin patients with MN. For the remainder of the analysis, we divided thepatients with MN into two grades of injury. Values of GFR above orbelow 50% of the average normal (control) level were used tocategorize the MN as either moderate or severe. The degree of nephrosisin the two grades of injury as well as the number of patent glomeruli ( N g ) was compared using theWilcoxon-Mann-Whitney test. Either an analysis of variance combinedwith the Newman-Keuls test for post hoc comparisons or theKruskal-Wallis test with the Dunn procedure was used to assess thesignificance of differences among the groups of patients with moderateMN, severe MN, and controls. Results are reported as means ± SDor the median (range).  G7 B5 o! D: P2 _6 _- r. h
) C+ ~" c4 ~! \4 R! |
RESULTS  T" E- Y+ _2 S9 H/ D1 ]
& E* ~$ ^% c0 }# H. k
Physiological Assessment
: d' v  ]. @1 f4 N' s9 i4 e+ o& F
The mean GFR in healthy controls was 101 ± 17 ml · min 1 · 1.73 m 2. By contrast, the finding in MN that GFR = 62 ± 30 ml · min 1 · 1.73 m 2 ( P depressed, the GFR varied widely (range10-119ml · min 1 · 1.73 m 2 ). As stated above, we used a GFR above or below 50% ofthe average value in healthy individuals (i.e., 50 ml · min 1 · 1.73 m 2 ) to categorize the MN as moderate ( n = 21) or severe ( n = 13), respectively. Judged by themedian levels of proteinuria of 6.2 g/24 h (2.2-10.5) and 13.6 g/24 h (4.8-25.9) in moderate and severe MN, respectively, thesevere MN group had significantly worse nephrosis ( P = 0.001).

Results of our physiological assessment of GFR and its determinants aresummarized in Table 1. Whereas GFR inmoderate MN tended to be slightly depressed (83 ± 15 ml/min), thecorresponding rate of renal plasma flow tended to be elevated, 806 ± 188 vs. 566 ± 128 ml/min in controls ( P depression of GFR (29 ± 11 ml/min)according to which subjects were assigned to the severe MN group wasnot associated with a significant depression of the rate of renalplasma flow (504 ± 382 ml/min; Table 1 ). Thus, a markeddepression of the filtration fraction in each category of MN, 0.11 ± 0.03 in moderate and 0.07 ± 0.03 in severe (vs. 0.18 ± 0.03 in controls), indicates that changes in determinants of GFR otherthan renal plasma flow must explain the observed level ofhypofiltration.

Table 1. Physiological data
+ R. r# z3 J* _6 P% \
Reflecting the marked hypoproteinemia, afferent oncotic pressure( A ) was markedly depressed in moderate MN,15.2 ± 3.9 vs. 24.4 ± 2.4 mmHg in the control group( P value for mean oncoticpressure along the glomerular capillaries ( GC ) was proportionately more depressedcompared with the control group, mean 16.2 ± 4.2 vs. 27.1 ± 2.6 mmHg, respectively ( P 0.001). A (11.2 ± 3.3) and GC (11.7 ± 3.6 mmHg) were significantly more depressed in the severe MN group (Table 1 ). The depression in GC in MN can be inferred to elevate the netultrafiltration pressure by ~10 and 14 mmHg in the moderate andsevere groups, respectively. Because GC isthe force opposing the formation of filtrate, depression of either Pand/or K f must be invoked to explain theobserved hypofiltration.( S, L7 d9 h- l5 m2 X
! g* D) U2 o9 U9 D8 o
In an effort to estimate the magnitude of the effect attributable to K f depression, we first assumed a normal P of40 mmHg, a value similar to that observed by micropuncture in thenormal euvolemic rat ( 24 ). With measured values ofGFR, renal plasma flow, and this value for P, the ultrafiltrationmodel of Deen et al. ( 9 ) yielded a value for two-kidney K f of 11.0 ± 5.7 ml · min 1 · mmHg 1 in healthy controls (Table 1 and Fig. 1 ).We next used a sensitivity analysis to estimate the influence of Pon K f in each grade of MN. We examined theeffects of a P that was the same (40 mmHg), higher (45 mmHg), orlower (35 mmHg) than normal. The computed values indicate that K f is depressed in MN regardless of the actual value of P within this range. There is negligible overlap among controls, moderate and severe MN under any combination of P values (Fig. 1 ). Because arterial pressure was elevated in MN (Table 1 ), weinfer that P is in fact likely to be elevated. For purposes of theanalysis that follows, however, we made the conservative assumptionthat none of the increment in arterial pressure was transmitted intoglomerular capillaries and that P was equivalent to the controlvalue (i.e., 40 mmHg). Because P and K f arereciprocally related, this should provide an upper bound for two-kidney K f in MN relative to the control. This quantity,which we will refer to as K f40, was only 34 and10% of control K f40 in moderate and severe MN,respectively (Table 1 ).( F7 g! f  I# @) e
6 K  H4 ?) Q) k+ R
Fig. 1. Box plots of calculated 2-kidney ultrafiltrationcoefficient ( K f ) in controls ( column1 ), moderate membranous nephropathy (MN; columns2 - 4 ), and severe MN ( columns5 - 7 ). Box plots display the 10th, 25th, 50th, 75th,and 90th percentiles on the vertical axis. Outliers among the 130 control subjects are indicated by. The effects on the K f of variations of P in MN are displayed onthe horizontal axis.) j1 w# s: T/ ^) f7 M# I; m
+ k+ V2 f) W: c( w: s
Morphological Assessment
1 B1 _3 _- M& c3 c
Our morphometric analysis is summarized in Table 2. The first finding that is remarkableis that despite the striking differences in GFR and two-kidneyfiltration capacity ( K f40 ) between moderate andsevere MN, quantitative glomerular morphology was similar in the twogrades of injury. The only histopathological finding in the diagnosticbiopsy that distinguished severe from moderate injury was a substantialexpansion in the former of the interstitial compartment (Table 2 ). Themean percent global sclerosis was similar in each injury grade (Table 2 ) as was the prevalence of patients with global sclerosis (8/21 and5/13 in moderate and severe, respectively; Fig. 2 ). We determined filtration surface areafrom the product of glomerular volume and filtration surface density inthe patent glomeruli (Table 2 ). Reflecting a near doubling ofglomerular volume (Table 2 ), filtration surface area was increased ineach MN subset (Fig. 3 A ).Almost all of the resistance to transcapillary water flow is exerted bythe glomerular basement membrane and the diaphragms at the bases of theepithelial filtration slits ( 8, 10, 11 ). Basement membranethickness was increased twofold, a phenomenon that is predicted tolower hydraulic permeability (Table 2 ). Also, broadening of footprocesses lowered the frequency of intervening filtration slits toapproximately one-third of normal in both injury grades of MN (Table 2 ), further limiting water flux into Bowman's space./ @# _) K5 e; R2 H" F7 Z8 [' W

Table 2. Morphometric analysis+ v: Y& L5 L$ P" F- q  D- p0 I1 L
% [3 s, c$ p1 v: X3 L
Fig. 2. Prevalence and extent of glomerulosclerosis in moderate( left ) and severe MN ( right ).; A, o* N( ?" P# J5 u- V, a3 k& K
% K. J# |9 r+ f. S# K
Fig. 3. Box plots comparing filtration surface area (S; A ), hydraulic permeability ( k; B ), andsingle-nephron K f (SN K f; C ) in controls, moderate MN, and severe MN. Represent outliers. * P P
! K) d8 b9 H% Q6 B: o, }
We applied the foregoing findings to the mathematical model of viscousflow of Drumond and Deen ( 10 ) to calculate a value forlocal hydraulic permeability. It was profoundly impaired in eachcategory of MN. Surprisingly, however, the impairment of hydraulicpermeability was similar in the two injury grades, 8 ± 3 and6 ± 3, respectively, in moderate and severe MN vs. 22 ± 3 m · s 1 · Pa 1 × 10 10 in controls (Fig. 3 B ). We nextcalculated single-nephron K f for each individualfrom the product of filtration surface and hydraulic permeability. Thedetermination of single-nephron K f frommorphometric data is completely independent of the physiologicaldetermination of K f40. It is thus of interestthat a striking disparity emerged between computed single-nephron K f and two-kidney K f inthe two injury grades. Whereas the value for two-kidney K f was severe $ @2 X4 e& ]- [/ t/ ?7 P. z

Glomerular Density
2 O9 b. E, }# m9 P4 s5 b
Computation of functional glomerular number( N g ) from Eq. 9 suggests that moresevere glomerulopenia is the reason for the disproportionately low GFRand two-kidney K f in severe vs. moderate MN(Table 1 ). Because the numerator (2-kidney K f )and denominator (single-nephron K f ) in Eq. 9 were determined in two separate control groups (see METHODS ), only a group mean value for N g in controls could be calculated. Thisquotient yields a value for N g of 1.5 × 10 6 for healthy controls (Table 3 ), which is close to the value of1.2 × 10 6 found by direct morphometric analysis ofnormal kidneys at autopsy ( 19 ). The corresponding value of N g in moderate MN (i.e., assuming P = 40 mmHg) is computed to be 1.7 ± 1.2 × 10 6.Allowing for an elevation of P in the hypertensive patients with MNto 45 mmHg, the corresponding value in moderate MN for N g would be 1.4 ± 1.0 × 10 6 (Table 3 ). Corresponding values for the N g in those with severe MN are 0.54 ± 0.5 and 0.45 ± 0.4 × 10 6, respectively (Table 3 ),suggesting that N g in severe MN was considerablylower than would be suggested by the low frequency of global sclerosis.The possibility that resorption of sclerotic glomeruli masked the trueextent of glomerular loss is suggested by the finding that severe, butnot moderate, MN was accompanied by marked collagenization andexpansion of the interstitial compartment (Table 2 ).: q. y. w( {- G# A5 }$ C: a- p

Table 3. Estimation of number of functioning glomeruli

DISCUSSION

As might be expected, the extent to which GFR declines early inthe course of MN is proportional to the magnitude of depression oftwo-kidney K f, a measure of the total capacityfor ultrafiltration of all functional glomeruli in the two humankidneys. In contrast, no such relationship is apparent between GFR and K f determined by morphometric analysis ofindividual glomeruli. The reason for the decline inSN K f, in early MN, is a fall in hydraulicpermeability ( k ). We cannot exclude the possibility thatmolecular rather than structural alterations in the filtration slitdiaphragm lower k more in severe than in moderate MN( 15 ). It seems to us, however, that the most plausibleexplanation for the disparity is that the disproportionate reduction oftwo-kidney K f in severe MN is due to a steepreduction in the number of glomeruli.
- |9 S% L2 C; A8 C9 h' O
Glomerular number has been estimated in the dog kidney. Glomerulardensity in a kidney biopsy core of known volume was extrapolated tocortical volume, as assessed by MRI. Subsequently, a fractionator method used for validation after nephrectomy demonstrated good agreement ( 2 ). However, this technique has not yet beenapplied to estimate glomerular number in humans. Because there is notechnique available that is sufficiently sensitive to directly imagehuman glomeruli in vivo at present, we estimated the number offunctional glomeruli in our experimental subjects from the quotient,two-kidney K f /single-nephron K f. Our estimate in control subjects is in good agreement with the number of glomeruli estimated by morphometric analyses of kidneys of subjects coming to autopsy with no evidence ofrenal disease ( 19 ). The number was similar in patientswith moderate MN. Our estimate in subjects with severe MN, however, isfar lower, averaging only 500,000 glomeruli. Although we cannot excludethe possibility that the latter subjects may have been endowed withonly a small number of glomeruli at birth, this seems unlikely to usfor the following reasons. First, the estimated glomerular number isover two standard deviations below the mean value for normalindividuals in the aforementioned autopsy study (mean = 1,234,000; coefficient of variation = 0.25). Second,glomerulopenia of similar magnitude has been demonstrated at autopsy insubjects with severe diabetic nephropathy ( 4 ).Finally, marked expansion and collagenization of the interstitialcompartment in severe but not moderate MN point to an advanced stage ofchronic renal injury in the former. Taken together, these observationssuggest that severe MN predisposes to glomerulopenia, perhaps as aresult of sclerosis and resorption of heavily damaged glomeruli.$ I, y- h" V! M; ]1 K
) w/ l; g/ _! s* ?7 Z5 S0 ]' B. t
We acknowledge that our estimate of N g haslimitations as neither of the values needed for the estimate, namelytwo-kidney K f and single-nephron K f, is precisely known. The most notable errorin calculating two-kidney K f is likely to arisefrom discrepancies between our assumed P value of 40 mmHg and theactual value of P, which cannot be determined in humans. Severalfactors could lead to errors in calculation ofSN K f. For example, the use of fixed correctionfactors for the glomerular shrinkage associated with paraffin embeddingand immersion fixation could compromise the accuracy of our estimationof glomerular volume and hence filtration surface area( 17 ). Similarly, the need to use data from rats couldcompromise the accuracy with which we estimated hydraulic permeability( 8 ).
3 O. u8 ]' y: {+ ^! I% @
Whereas we are able to determine the dimensions of major glomerularstructures in humans morphometrically, the characteristics of several"nanostructures" have to be extrapolated from reported data for therat. These latter include the width of the filtration slit( s ) and the fractional area of fenestrae( f ), both of which we showed to be similar inthe human glomerulus. As stated in METHODS, we find s in both normal human subjects and those with MN toaverage 36 ± 4 nm, a value quite similar to the 41 nm reported for rats ( 10 ). Similarly, using scanning electronmicroscopy, we showed that f in humansaverages 0.16 vs. 0.20 in the rat ( 16 ). Because of theirlarge dimensions, the resistance imposed by endothelial fenestraeaccounts for only 1-2% of total resistance to water flow. Thus,the small aforementioned difference between humans and the rat shouldhave a negligible influence on computed k ( 8, 10 ).5 N/ N2 {$ n. g7 O0 X
& E9 |- z& z$ Q3 M+ a$ }
A key example of a nanostructure that has not been validated in humansare the dimensions of the apertures in the filtration slit diaphragms,as determined in the normal rat by Rodewald and Karnovsky( 23 ). Given that MN results primarily from an injury topodocytes, it is possible that changes in their foot processes couldalter the dimensions of the apertures. That the latter do not influencemodel predictions strongly, however, has been shown in minimal changenephropathy, a glomerular injury characterized by essentially identicalchanges in foot processes to those seen in MN. Drumond and Deen( 10 ) used micropuncture determinations of K f and a morphometric determination offiltration surface area in rats with adriamycin nephrosis, an analog ofminimal change nephropathy, to compute an experimental value of k ( k exp ) for this disorder( 10 ). They showed that model predictions for k were within the same range as k exp.
. J* Q3 t. _' u) |8 R" q) A
We also provided similar evidence to validate the model in humans withminimal change nephropathy ( 11 ). We computed k exp from the above-described physiologicaldetermination of two-kidney K f, an assumed valueof 1.2 × 10 6 glomeruli and morphometricallydetermined filtration surface area. Once again, there was remarkablygood agreement between k exp and k predicted by the model ( r = 0.71, P alterations in foot processes do not seem to cause largeenough changes in epithelial permeability ( k epi )to influence the value of k computed by the model using thenormal rodent dimensions of the apertures in the filtration slitdiaphragm. It appears that a reduction in fractional area of filtrationslits, in turn a function of reduced filtration slit frequency, ratherthan changes in intrinsic slit diaphragm structure, is responsible forlower k epi and hence k under thesecircumstances ( 11 ). We accordingly submit that ourestimate of k should yield a reasonable approximation ofSN K f and thus a reasonable estimate ofglomerular number. That this is indeed the case is suggested by therelatively good agreement between the mean number of glomeruliestimated in our control subjects from the quotient two-kidney K f /SN K f and valuesdetermined directly in nonnephropathic individuals by using unbiasedstereologic techniques at autopsy ( 19 ). The rather normalvalue for estimated N g in moderate MN isconsistent with a relatively low frequency of globalglomerulosclerosis. We infer that Eq. 9 should thus be noless successful in estimating N g in severe MNand that the marked reduction that we calculate in this setting islikely to be real, if not absolutely precise.; H, u, u4 H1 \& ]( Y5 ?

Our computation of greater depression of two-kidney K f in severe compared with moderate MN isinfluenced by the assumption of a value of P of 40-45 mmHg ineach grade of injury. An alternative explanation for the greaterdepression of GFR in severe MN is that there was marked reduction in P in that group. Given known values for GFR, A, renalplasma flow, and single-nephron K f, one can thenuse the ultrafiltration model of Deen et al. ( 8, 9 ) toestimate the extent to which P would have to be depressed to explainthe observed hypofiltration in severe MN, assuming that the premorbidnumber of glomeruli in both moderate and severe MN was the same as incontrols, i.e., 1.5 × 10 6 ( 21 ).This calculation revealed that a reduction of P to 18 mmHg in severeMN vs. 34 mmHg in moderate MN would be required to account for thegreater depression of GFR observed in those with severe injury atbaseline. There are two reasons that make this possibility unlikely,however. As stated previously, micropuncture studies in rat analogs ofMN have invariably revealed afferent arteriolar dilatation with anensuing elevation of P ( 1, 13, 29 ). Even if segmentalrenovascular resistance in human MN differs from that in the rat, it ishard to conceive how P could have been depressed by over 20 mmHg inour subjects with severe injury. Arterial pressure in these subjectsexceeded normal by 21 mmHg, (Table 1 ). Transmission of even a minorfraction of this increment into glomerular capillaries should haveelevated and not reduced P. By exclusion, this points to a reductionin glomerular number as the most likely explanation for thedisproportionate depression of GFR in severe MN./ b1 b5 i8 d2 n& Y5 f$ o5 W) C

Another potential alteration of glomerular hemodynamics that couldpotentially contribute to greater GFR depression in severe thanmoderate MN is the significantly lower rate of renal plasma flow in theformer, 504 ± 382 vs. 806 ± 188 ml · min 1 · 1.73 m 2, respectively ( P is suggested by two findings, however. Thefirst is that renal plasma flow in severe MN is not significantlydifferent from the normal control value (566 ± 128), despite thefinding that GFR is depressed by ~70% on average in the former(Table 1 ). Also, the significantly lower filtration fraction in severethan in moderate MN, 7 ± 3 vs. 11 ± 3%, respectively( P 1 ), points to a GFR-lowering effectby a determinant other than renal plasma flow. Greater K f depression owing to glomerulopenia in severeMN could be such a determinant of the lower GFR than in moderate MN.Dividing the observed rate of total renal plasma flow by the calculatednumber of glomeruli in Table 3 yields a rate of renal plasma flow pernephron (assuming P = 40 mmHg). Whereas the latterquantity is 816 ± 558 nl · min 1 · nephron 1 in moderate MN, it is almost twofold higher in severe MN at 1,679 ± 1,989 nl · min 1 · nephron 1.Thus, if as we propose, glomerulopenia indeed contributes to the lowerGFR in severe MN, the relative depression of total renal plasma flow inthis circumstance simply represents a loss of capacity by the corticalmicrovascular bed and not a reduction in the actual glomerularperfusion rate.
2 ^! ^" t: Q1 I5 \3 Z
We conclude that the onset of MN is accompanied by a severe depressionin hydraulic permeability of the glomerular capillary wall (Fig. 3 B ). This is partially offset by enhancement of filtration surface area (Fig. 3 A ) and by profound depression ofglomerular oncotic pressure (Table 1 ). As a result, GFRinitially remains in the normal range or is depressed by severe MN, by contrast, we propose that equivalentdepression of hydraulic permeability in patent glomeruli is compoundedby a marked reduction in functional glomerular number( N g ). Together, these two phenomena lowertwo-kidney K f to a level where increases infiltration surface area in remnant glomeruli and depression of oncoticpressure can no longer adequately compensate and the 50%. We showed previously that a progressive reduction of GFR in MNover the medium term is a consequence of declining K f ( 26 ). The latter isattributable, in part, to an increasing prevalence of globalglomerulosclerosis, and in part to a progressive loss of filtrationsurface area in remnant glomeruli, with an ensuing decline insingle-nephron K f. We propose thatsuperimposition of these medium-term changes on a markedly reducednumber of glomeruli likely accounts for the subset of patients with MN,who progress rapidly to end-stage renal failure. Advances in imagingthat will permit human glomeruli to be counted in vivo will be required to validate our proposal and to confirm that glomerular number isindeed depressed early in the course of severe MN.8 _3 W7 U% u' r, z" x& d

ACKNOWLEDGEMENTS
2 k1 O6 P9 m. l  X0 p9 f7 v! V
We acknowledge L. Anderson, electron microscopy laboratorysupervisor, for assistance with morphometry.& N1 \% A# F6 ~" D( y5 k# E
      【参考文献】
1. Allison, MEM,Wilson CB,andGottschalk CW. Pathophysiology of experimental glomerulonephritis in rats. J Clin Invest 53:1402-1423,1974  .
7 d2 [' J( ], o3 Y$ H! w% @

4 u* N0 B+ T" l6 j" Y
2. Basgen, JM,Steffes MW,Stillman AE,andMauer SM. Estimating glomerular number in situ using magnetic resonance imaging and biopsy. Kidney Int 45:1668-1672,1994  .& c7 u3 l! L& I2 ^/ T

5 a3 W' b$ Y+ L( z% ?

3. Battilana, C,Zhang H,Olshen R,Wexler L,andMyers BD. PAH extraction and the estimation of plasma flow in the diseased human kidney. Am J Physiol Renal Fluid Electrolyte Physiol 261:F726-F733,1991 .

4. Bendsten, TJ,andNyengaard JR. Number of glomeruli in diabetic kidneys. Diabetologia 35:844-850,1992  .

9 i2 [" |8 a& j1 N
5. Canaan-Kühl, S,Venkatraman ES,Ernst SIB,Olshen RA,andMyers BD. Relationship among protein and albumin concentrations and oncotic pressure in nephrotic plasma. Am J Physiol Renal Fluid Electrolyte Physiol 264:F1052-F1059,1993 .3 f" X2 G$ v& }0 c% x

& {) y+ ?0 h* \; N! r
6. D'Amico, G. Influence of clinical and histological features on actuarial renal survival in adult patients with idiopathic IgA nephropathy, membranous nephropathy, and membranoprofilerative glomerulonephritis: survey of the recent literature. Am J Kidney Dis 20:325-333,1992./ d9 w& j' n6 K! L0 S- q0 V  O

( H7 Q4 x* a4 i2 W; g% {5 J
7. Daniels, BS,Hauser EB,Deen WM,andHostetter T. Glomerular basement membrane: in vitro studies of water and protein permeability. Am J Physiol Renal Fluid Electrolyte Physiol 262:F919-F926,1992 .6 A9 V" L/ y# p# }
2 T* I& s! `' p: `! ]1 o
: |, ~( w( p, ~+ w

8. Deen, WM,Lazzari MJ,andMyers BD. Structural determinants of glomerular permeability. Am J Physiol Renal Physiol 281:F579-F596,2001 .

8 U: v5 t' X5 e
7 m' E# c+ h' j& [; I+ n
9. Deen, WM,Robertson CR,andBrenner BM. A model of glomerular ultrafiltration in the rat. Am J Physiol 223:1178-1183,1972 .
& D' M- v8 S0 @+ r* J

" Y' R. a0 ]5 u+ D9 [
10. Drumond, MC,andDeen WM. Structural determinants of glomerular capillary hydraulic permeability. Am J Physiol Renal Fluid Electrolyte Physiol 266:F1-F12,1994 .
6 {4 O# A& B1 o: v
% H1 m1 S" }' k

11. Drumond, MC,Kristal B,Myers BD,andDeen WM. Structural basis for reduced glomerular filtration capacity in nephrotic humans. J Clin Invest 94:1187-1195,1994  .

12. Guasch, A,Sibley RK,Huie P,andMyers BD. Extent and course of glomerular injury in human membranous glomerulopathy. Am J Physiol Renal Fluid Electrolyte Physiol 263:F1034-F1043,1992 .

13. Ichikawa, I,Hoyer JR,Seiler WM,andBrenner BM. Mechanisms of glomerulotubular balance in the setting of heterogeneous glomerular injury. J Clin Invest 69:185-198,1982  .

; Y2 r5 N2 R/ d7 i8 S: K) ^5 t% Q

14. Jensen, EB,Gundersen HJB,andOsterby R. Determination of membrane of thickness distribution from orthogonal intercepts. J Microsc 115:19-33,1979  .
; Y7 G$ Z8 I4 n" C0 o2 e7 Y

: z# f5 `  t- m6 u0 ^( S7 q# X
15. Kerjaschki, D. Caught flat-footed: podocyte damage and the molecular bases of focal glomerulosclerosis. J Clin Invest 108:1583-1587,2001  .

# ~, t# C: B1 J
; h& f  Y1 Z: u& d7 K
16. Lafayette, RA,Druzin M,Sibley R,Derby G,Malik T,Huie P,Polhemus C,Deen WM,andMyers BD. Nature of glomerular dysfunction in pre-eclampsia. Kidney Int 54:1240-1249,1998  .
. Y9 T& c: ?! v' L2 Y
- A& g0 n3 a4 |* Q9 z7 \$ t
1 I! N, D1 Z! \8 _+ |
17. Miller, PL,andMeyer TW. Effects of tissue preparation on glomerular volume and capillary structure in the rat. Lab Invest 63:862-866,1990  .+ h1 r4 Z0 ?2 Z- x9 K& g

0 ~. Z# @! H# D! `5 Z
" E& {* N0 R1 p/ G1 X% B& u
18. Myers, BD,Peterson C,Molina CR,Tomlavonich SJ,Newton LD,Nitkin R,Sandler H,andMurad F. Role of cardiac atria in the human renal response to changing plasma volume. Am J Physiol Renal Fluid Electrolyte Physiol 254:F562-F573,1988 .* W; z: E/ B( i( Z! W
$ {! M% @; I* Y% M; z( S# p
1 h* C9 G, b  X2 `$ _; V. A
' _# c( b0 a! X% c1 T- l% F
19. Nyengaard, JR,andBendsten TF. Glomerular number and size in relation to age, kidney weight, and body surface in normal man. Anat Rec 232:194-201,1992 .  r+ `+ A4 `- d  Q

; h* n  f) O7 u8 S# M9 |) ?
1 j: `( K  v8 w- V( Q# n- n- O3 ^
20. Pei, Y,Cattran D,andGreenwood C. Predicting chronic renal insufficiency in chronic membranous glomerulonephritis. Kidney Int 42:960-966,1992  .

6 ^! E8 x9 R" U0 H; Q% @4 o
21. Ramaswamy, D,Corrigan G,Polhemus K,Boothroyd D,Scandling J,Sommer FG,Alfrey E,Higgins J,Deen WM,Olshen R,andMyers BD. The maintenance and recovery stages of postischemic acute renal failure in humans: a study of cadaveric renal allografts. Am J Physiol Renal Physiol 282:F271-F280,2002 .- i( Y8 n, E) S- P' ^3 e

22. Ramzy, MH,Cameron JS,Turner DR,Neild GH,Ogg CS,andHicks J. The long-term outcome of idiopathic membranous nephropathy. Clin Nephrol 16:13-19,1981  .. o4 C2 I. R; g+ |# ~
. a4 Z, j" ~/ V& G
6 u/ {! B: S2 ^

23. Rodewald, R,andKarnovsky MJ. Porous substructure of the glomerular slit diaphragm in the rat and mouse. J Cell Biol 60:423-433,1974 .1 T/ b9 f  j( t7 Q, h. K0 Z- C
- C+ Z9 K! i6 a2 v& I8 F, p
$ B8 q% W  ?1 y- y# O

24. Scholey, JW,andMeyer TW. Control of glomerular hypertension by insulin administration in diabetic rats. J Clin Invest 83:1384-1389,1989  .

25. Shemesh, O,Ross JC,Deen WM,Grant GW,andMyers BD. Nature of the glomerular capillary injury in human membranous glomerulopathy. J Clin Invest 77:868-877,1986  .

$ @- j+ p# `7 N7 D
26. Squarer, A,Lemley KV,Ambalavanan S,Kristal B,Deen WM,Sibley R,Anderson L,andMyers BD. Mechanisms of progressive glomerular injury in membranous nephropathy. J Am Soc Nephrol 9:1389-1398,1998 .1 b) @, y# \' d0 l) d" l

4 J: ]$ `, S& _7 c8 w
" R& G, o! X( R
27. Ting, RH,Kristal B,andMyers BD. The biophysical basis of hypofiltration in nephrotic humans with membranous nephropathy. Kidney Int 45:390-397,1994  .9 m! ^% [7 y# o$ B- u' f9 q2 h5 L! U- p
. i4 m5 y; T  |
# k, c4 E2 q; k; H
, h* R0 H& b6 a: T2 Z3 G& t# o1 _
28. Weibel, ER. Sterological Methods: Practical Methods of Biological Morphometry. London: Academic, 1979, vol. 1, p. 44-45 and 131-134.- `+ c+ s1 ~' v( T
& T( X. [3 ^; e2 z; O
4 x$ {, {$ H% @$ x" g4 Q
+ H& a7 D! r2 m5 B& s- Y6 Z* p
29. Yoshioka, T,Rennke HG,Salant DJ,Deen WM,andIchikawa I. Role of abnormally high transmural pressure in the permselectivity defect of glomerular capillary wall: a study in early passive Heymann nephritis. Circ Res 61:531-538,1987 .

) D3 Z1 D, B2 ~$ c% s
" y2 O( a- w* q3 Q% {$ v
30. Zuchelli, P,Ponticelli C,Cagnoli L,andPasserini P. Long-term outcome of idiopathic membranous nephropathy with nephrotic syndrome. Nephrol Dial Transplant 2:73-78,1987 .

收藏0 分享0 顶0 踩0