The biophysical basis of hypofiltration in nephrotic humans with membranous nephropathy

Kidney International, Vol. 45 (1994), pp. 390—397

The biophysical basis of hypofiltration in nephrotic humans with membranous nephropathy ROBERT H. TING, BATYA KRISTAL, and BRYAN D. MYERS Division of Nephrology, Department of Medicine, Stanford University School of Medicine, Stanford, California, USA

The biophysical basis of hypoffitration in nephrotic humans with membranous nephropathy. Physiologic and morphologic techniques were used to elucidate the determinants of the glomerular filtration rate in 62 nephrotic patients with membranous nephropathy and 104 healthy controls, Renal plasma flow, glomerular filtration rate, afferent oncotic pressure and dextran sieving coefficients were determined. Mathematical models of glomerular filtration were then used to compute likely upper bounds for the ultrafiltration coefficient and pore areallength ratio (a measure of pore density). These upper bounds for each measure of intrinsic ultrafiltration capacity were both depressed by 75% in mem-

branous nephropathy. A corresponding excess of ultrafiltration pres-

suggested by the consistent demonstration of restricted transglomerular sieving of uncharged dextran or polyvinyl pyrollidone macromolecules in these disorders [1—5, 13]. Whereas this finding is consistent with a reduced Kf, depression of either the glomerular perfusion rate or pressure is predicted to result in enhancement and not depression of the transglomerular passage of uncharged macromolecules [14].

The purpose of the present study was to quantitate each determinant of Kf and glomerular ultrafiltration pressure, so as

sure (versus control), attributable solely to reduced intracapillary to elucidate the mechanisms of glomerular hypofiltration in oncotic pressure was by 12 mm Hg. Glomerular morphometry revealed peripheral capillary filtration surface area to be enhanced in membranous nephropathy (3.27 x l0 vs. 2.03 X 10' sm'). However, there was

membranous nephropathy, a common cause of the nephrotic syndrome in humans. To do this we have estimated glomerular a massive reduction in filtration slit frequency due to epithelial flows, preglomenilar vascular pressures and dextran sieving podocyte broadening (336 vs. 1204 slits/mm capillary length) as well as coefficients in 62 patients with membranous nephropathy. We marked thickening of the glomerular basement membrane (1130 vs. 388 nm). We conclude that the latter two findings lower the ultrafiltration

then analyzed the findings with theoretical models of glomerular filtration in an effort to distinguish between possible contribu-

coefficient in membranous nephropathy by depressing the hydraulic permeability of the glomerular capillary wall. We propose that this tions by reduced ultrafiltration pressure and diminished ultraabnormality is solely responsible for the hypofiltration observed in a ifitration capacity to the prevailing GFR. To define the strucmajority of patients with this disease. tural correlates of impaired ultrafiltration capacity we also

subjected glomeruli obtained by needle biopsy of the kidney to a morphometric analysis.

The glomerular filtration rate (GFR) is the most widely used laboratory measure of the extent and course of renal injury. It

Methods

represents the net flux rate of water across the walls of all

Patient population

patent glomerular capillaries and is equal to the product of the net pressure for ultrafiltration and the ultrafiltration coefficient, (Kr).

Studies of patients with nephrotic glomerular injuries suggest

that ultrafiltration pressure is likely to be elevated in this

circumstance [1—5]. The invariable presence of hypertension in

such diseases is likely associated with a high transcapillary hydraulic pressure difference (P). The ubiquitous hypoproteinemia that attends the nephrotic syndrome lowers the opposing oncotic pressure. Together, these lead to an increase in the ultrafiltration pressure. By exclusion, it seems likely that hypoifitration in this setting is caused uniquely by a depression in K. Depression of Kf accompanied by an offsetting increase in the

net pressure for ultrafiltration has been demonstrated repeatedly by micropuncture of rats with experimental analogues of the primary glomerular diseases most often seen in humans [6—12]. That this is also likely true of the human counterparts is

The subjects of our study were 62 adult patients who presented consecutively to our clinic with a nephrotic syndrome and a histopathological diagnosis of membranous nephropathy (MN). The patients varied widely in age (17 to 76 years) and 40 of 62 were male. We selected MN for study for two primary reasons. Firstly, the prevalence of glomerulosclerosis is low so that the effects of obliteration of glomeruli by this process can be circumvented [4]. Secondly, the thickening of the glomerular capillary wall due to subepithelial immune deposits and associated basement membrane thickening is rather uniform among glomeruli in a given patient. This rather homogeneous appearance suggests that all glomeruli could contribute uniformly to any abnormality of glomerular filtration. The control population was comprised of 104 healthy volunteers who spanned an age range from 18 to 73 and who were also predominantly male. They provide a normal range for the glomerular functional quantities of interest. All denied a history

of renal disease, hypertension or diabetes. At the time of © 1994 by the International Society of Nephrology

evaluation each was found to be normotensive and normoglycemic, and to have a negative dipstick test for urinary protein.

390

391

Ting et a!: Glomerular hypofihtration

filtration surface area of all glomerular capillaries in the two All patients and volunteers underwent a determination of human kidneys. The input values for the model included the GFR, renal plasma flow (RPF) and preglomerular vascular measured values of GFR, RPF and 11A and an assumed value pressures according to a protocol which had been approved for the glomerular transcapillary hydraulic pressure difference previously by the Institutional Review Board at the Stanford (1P). The latter quantity cannot be directly measured in huUniversity School of Medicine. A subset of each group, (con- mans. However, using an indirect curve-fitting technique we trols N = 19, membranous N = 28), also underwent an have estimated that P approximates 35 mm Hg in the healthy examination of dextran sieving. Urine was voided spontane- human kidney, and have assigned this value to both the control Physiologic evaluation

ously after diuresis had been established with an oral water load and MN groups in the present study [18]. Micropuncture

(10 to 15 ml/kg). A priming dose of inulin (50 mg/kg) and determinations in Heymann nephritis, a rodent analogue of para-aminohippuric acid (PAH, 12 mg/kg) was then adminis- MN, indicate that P is invariably elevated in this form of tered. In the subsets receiving dextran, this was followed by a priming infusion of dextran 40 (130 mg/kg/lO mm). Thereafter, inulin and PAH were given by continuous infusion to maintain plasma levels constant at 20 and 1.5 mg/dl, respectively. Dextran 40 was infused constantly at half the rate calculated for inulin. Sixty minutes after the priming infusion, arterial blood pressure was determined, and blood was sampled for examination of plasma oncotic pressure (HA) and plasma protein concentra-

tion. Four timed urine collections were then made, each of

which was bracketed by a blood sample drawn from a peripheral vein. The GFR was expressed as the average value for the four timed inulin clearances. The rate of RPF was estimated by dividing the corresponding clearance of PAH by an estimate of the prevailing renal arteriovenous extraction ratio for PAH (E). We have shown previously that reductions of GFR and peritubular capillary protein concentration exert an additive effect to lower EPAR in patients with glomerular disease [15]. From the relationship observed in that study between EPAH and GFR, we have assigned the following values for EPAH: 0.9 for healthy controls, 0.8 for patients with membranous nephropathy and a normal age-adjusted GFR and 0.7 for patients with membranous nephropathy and a depressed age-adjusted GFR. Fractional dextran clearances (OD) were calculated using the equation:

= (U/P)/(U/P)1

(1)

where (U/P)D and (UfP)1 refer to the urine-to-midpoint plasma concentration ratio of dextran and inulin, respectively. The concentrations of inulin and PAR were determined with

an automated assay [1]. Concentrations of dextran were assayed with anthrone after separation of the component molecules of dextran 40 in urine and deproteinized plasma into 2 A fractions by gel permeation chromatography using precalibrated Ultragel AcA44 columns (LKB, Pleasant Hill, California, USA). Concentrations of albumin and IgG in serum and urine were determined by immunochemical methods as de-

glomerular injury [6—8]. Given that human MN is accompanied

by arterial hypertension (vide infra), it is probable that a fraction of the increment in arterial pressure is transmitted into glomerular capillaries, and that P is likely, if anything, to be 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 prevailing level of K1 in this disorder [17]. We have accordingly designated the latter computation in MN as "maximum" K1 (Kfmax). We also used sieving profiles of uncharged and non-reabsorbable dextrans of broad size distribution to compute additional membrane parameters for the 28 individuals with MN and the 19 normal controls who received an infusion of dextran. For this purpose we used a mathematical model which represents the glomerular capillary wall as a heteroporous membrane, one containing two populations of pores of widely varying size in parallel [19]. According to this model the membrane is dominated by a population of small and restrictive cylindrical pores of identical radius (r0). The second and parallel population of

pores are large and non-discriminatory, serving as a "shunt pathway" through the membrane. The shunt pathway is characterized by a parameter con, which governs the fraction of the total filtrate volume passing through the non-restrictive portion

of the membrane. The membrane barrier to filtration of water and uncharged macromolecules in this "isoporous plus shunt" membrane model is characterized fully by the values of r0, 0,,, and Kf. An additional membrane parameter that can be derived from K1 and r,, is the ratio of effective pore area-to-pore length

(S'Il), which is closely related to the apparent number of restrictive pores in the two kidneys (N). More precisely,

N S'/11r02

(2)

Thus, if 1 and r0 are nearly constant, changes in S'/l are very nearly proportional to changes in the total number of restrictive pores [14].

The approach used for modeling the intrinsic membrane parameters separates their effects on dextran sieving coeffi-

scribed previously [16]. Plasma oncotic pressure was measured cients from those due to purely hemodynamic changes [19]. To directly using a Wescor 4400 membrane osmometer (Wescor allow for the effect of possible variations in P on computed

Inc., Logan, Utah, USA) and serum creatinine levels by a membrane parameters in patients with MN, we performed a

rate-dependent modification of the Jaffe reaction, employing a sensitivity analysis, repeating all calculations over a hypothetical range of P values (30 to 45 mm Hg) that brackets the Beckman Creatinine Analyzer (Model 2, Fullerton, California, assumed control value of 35 mm Hg. USA).

Morphometric evaluation Theoretical analysis of filtration data Tissue blocks suitable for morphometric analysis were availA mathematical model for the glomerular filtration of water [17] was used to calculate K1, which is defined in this study as able for 27 patients in whom a kidney biopsy had been perthe product of glomerular hydraulic permeability, and the total formed at the time of the clearance study. Biopsies performed

392

Ting et a!: Glomerular hypofiltration

Table 1. Clinical features and proteinuria Controls Scr mg/dl

AIb excretion rate jsg/min

IgG excretion rate isg/min SMb gluIer SISG giliter

200 Membranous

09b (0.4—1.3)

(0.5—16,6) 5321b

(2.2—32.1) 1 3b (0.2—3.9)

(624-26300) 276b

6.2'

47 2b

(7.9—5950) 19 7b

(5.8—14.5)

(1.1—21.6)

95

P< 0.0001

150

43

E

Values are median (range). ap <0.001 b P < 0.0001

100

E U-

0 on 16 living kidney donors at the time of renal transplantation

were used to provide control values for the morphometric quantities of interest. Each of these transplant donors had been shown during routine pre-operative evaluation to have normal renal function and anatomy. All glomeruli up to a maximum of 30 in a single 1 pm section stained with periodic-acid Schiff reagent were analyzed at the light microscopic level. On average, 18 glomeruli per biopsy were examined in each patient with MN (range 4 to 30). The average number of glomeruli among the sixteen control biopsies was 19 (range 13 to 30). The numbers of open and occluded glomerular tufts were recorded, where occluded glomeruli were defined as those exhibiting global sclerosis. A dedicated computer system (Southern Micro Instruments, Inc., Atlanta, Georgia, USA), consisting of a video camera, screen, microscope and digitizing tablet, was used to perform measurements [3].

50

0 Controls

Membranous

Fig. 1. GFR in controls and patients with MN. The data are illustrated as quartile box plots where the lower, middle and upper horizontal lines

of the box represent the 25th, 50th and 75th percentiles, respectively. The vertical lines represent the remaining two quartiles.

The outline of each glomerular tuft in the cross section was traced onto the digitizing tablet at 900 X and the tuft cross sectional area (A0) computed using area perimeter analysis. Glomerular volume (V0) was calculated from A0 as described magnification electron photomicrograph was overlaid by a grid previously [3]. An equation which takes the reduced diameter of intersecting lines. The GBM thickness was then measured by observed in globally sclerotic glomeruli into account was used to calculate their actual prevalence (G1).

the orthogonal, intercept method [20].

Statistical analysis For transmission electron microscopy tissue was fixed in 2.5% glutaraldehyde and embedded in epon. Toluidine blueDifferences between the MN and control subjects were stained sections were then surveyed to locate the two open evaluated by either the Wilcoxon rank sum test or an unpaired glomeruli closest to the center of each section. Ultra thin Student's t-test depending on the distribution of the findings. sections (60 to 70 nm) of the selected glomeruli were next Tabulated results are expressed as either the mean SEM stained with lead citrate and photographed. A complete mon- where the distribution of values is Gaussian or as the median tage of each glomerulus was used to calculate the peripheral (range) where the distribution of values is skewed. Most illuscapillary filtering surface density (Sw) by point and intercept trated results are presented as quartile box plots, where the counting at low magnification (2820 x). The filtration surface lower, middle and upper horizontal lines of the box represent area, S (in /Lm2/glom), was defined as the interface between the the 25th, 50th and 75th percentiles, respectively. The vertical peripheral capillary wall and epithelium, and was calculated as lines represent the remaining two quartiles. the product of S,,, and the corresponding value of V0 [4]. Six to eight high-power electron photomicrographs were then obResults tained from each of the two glomerular proffles and printed at a Clinical features and proteinuria magnification of 11,280 x to evaluate the frequency of epithelial The median serum creatinine level was significantly elevated filtration slits and the thickness (harmonic mean) of the peripheral glomerular basement membrane (GBM). Filtration slit in MN, 1.2 versus 0.9 mgldl in controls. The excretion rates of frequency was determined by dividing the total number of albumin and IgG and the corresponding circulating levels of epithelial filtration slits by the total length of the GBM that was each protein in patients with MN are typical of the nephrotic captured on the electron photomicrographs [3, 4]. Each high syndrome, and are summarized in Table 1.

393

Ting et a!: Glomerular hypofiltration

A

B

1500

0.5

P< 0.0001 0.4 C

0 C) a 0.3

1000

C

0 a 0.2

E LL

0

500

LL

0.1

0

I

__

1

___ I

0 Controls Membranous

50

I

Afferent P< 0.001

Controls Membranous

Fig. 2. Renal plasma flow (A) and filtration fraction (B) in controls and patients with MN.

140

Efferent

P <0.001

P.c 0.0001

40 30 E E

20

Ll _

I

120

FT'1

10

E E 100

0 0

Controls Membranous Controls Membranous

Fig. 3. Afferent (A) and efferent oncotic pressure (B) in controls and patients with MN.

80

Glomerular filtration of water and macromolecules The GFR was significantly depressed below control values in the patients with MN. The respective median values were 100 and 63 mIIminil.73 m2 (Fig. 1). Only 15 of the 62 subjects with

60 Controls

Membranous

Fig. 4. Mean arterial pressure in controls and patients with MN.

MN achieved a GFR value above the 25th percentile of the normal range. Although values for RPF were more widely distributed in MN than in controls, there was no significant difference between the must contribute to hypofiltration in the minority of subjects in groups. In fact, the median RPF in MN was numerically higher whom RPF is depressed. than the corresponding control value, 628 vs. 548 mI/mini! .73 Reflecting the marked hypoproteinemia (Table 1), the median m2, respectively (Fig. 2A). Even in the minority of cases in 11A in MN was markedly depressed below the corresponding

whom RPF was depressed below the 25th percentile of the control values, 15 versus 24 mmHg (Fig. 3A). Conversely, normal range, the depression was less than in proportion to the median arterial pressure (MAP) was significantly elevated GFR, with the result that the filtration fraction was significantly depressed, (0.12 vs. 0.19, respectively; Fig. 2B). These findings suggest that glomerular hypofiltration cannot be attributed to a decline in RPF in most cases of MN, and that additional factors

above the control value, 105 versus 90 mm Hg (Fig. 4). Given that some fraction of the 15 mm Hg increment in MAP is likely to have been transmitted into glomerular capillaries, a parallel elevation in P is likely to have eventuated, increasing the net

394

Ting et a!: Glomerular hypofiltration

30

100

P< 0.0001

E

a)

0C

20

10-'

(a) 0 C

C

0 E

10

(5

1 02

U-

0

Controls

Membranous

10 22

28

34

Fig. 5. Computed maximum K in controls and patients with MN.

ultrafiltration pressure at the afferent end of the glomerular

46

52

64

58

Dextran radius, A Fig. 6. Dextran sieving profiles in the control (0) and MN groups (nj. For MN, all sieving coefficients in the 26 to 48 A interval are significantly lower and all sieving coefficients in the 52 to 60 A interval are significantly higher than in controls.

capillary tufts. Based on the conservative estimate of equivalent zP (35 mm Hg) in the MN and control groups, we computed K1 to have been profoundly depressed in the former. The median value for this likely upper bound for Kf in MN (Kimax) was only 3.6

Table 2. Membrane parameters Controls

mlJ(min mm Hg)/1.73 m2, versus a corresponding control Restrictive pore radius r0, A value of 15 mll(min mm Hg)/1.73 m2 (Fig. 5). There was Pore arealpore length virtually no overlap between the two groups, suggesting that K1 S'/l, Km Shunt parameter is invariably depressed in nephrotic subjects with MN. x l0 The configuration of the sieving profile for dextran macro-

molecules in the 26 to 60 A radius interval was altered from control in patients with MN (Fig. 6). Sieving coefficients at the low radius end of the MN profile were markedly depressed, while sieving coefficients for the largest dextran molecules examined were elevated. Analysis of sieving data with the heteroporous membrane

40

Membranous

57.7"

54.2"

(55.9—60.6)

(49.5—57.5)

305.7a

75.6"

(155.6—818)

(16.0—175.5)

0.4a

6.3"

(0—1.6)

(2.6—31.2)

Values are median (range).

ap< 0.0001

along with a sh in a small fraction of such pores from the restrictive to the shunt-like portion of the membrane. A sensitivity analysis of the effects of variations of i.P on L(f more prominent in MN than in controls (Table 2). Linear and pore density, the two most sensitive measures of intrinsic regression analysis revealed o to be correlated with the ultrafiltration capacity are illustrated in Figures 7 and 8. Even in fractional clearances of both albumin (r = 0.90, P <0.0001) and the unlikely event that P was lowered by S mm Hg (to only 30 IgG (r 0.89, P < 0.0001), suggesting that impairment of mm Hg) in our hypertensive subjects with MN, the Kf and pore barrier size-selectivity contributed to the nephrotic-range pro- density would still be significantly lower than in healthy conteinuna. Further, w0 and S'/l were inversely related (r —0.64, trols (P < 0.00 1). Thus while potential contributions to the low p < 0.001). Thus, the membrane dysfunction seen in membra- filtration fraction and the low GFR in MN by depression of P nous nephropathy appears to reflect a reduction of pore density, cannot be excluded, impairment of the intrinsic ultrafiltration model (Table 2) revealed both pore density (S'/l) and restrictive pore radius to be markedly reduced in MN. In contrast, judged shunt-like pores were 16-fold by substantial enhancement of

395

Ting et a!: Glomeru!ar hypofiltration

5OO 400

- 300

20 -

200 10 -

100

0"

0—

-'P

Membranous

Controls

Fig. 7. Effects of variations in .P (301045 mm Hg; right hand bars) on values for Kf in MN compared to Kf at S.P = 35 in controls, Over the

entire range of assumed P the values of K in MN are significantly lower than Kf in controls (P < 0.001).

Controls

30 35 40 45 Membranous

Fig. 8. Effects of variations in t.P (30 to 45mm Hg; right hand bars) on values for S'/l in MN compared to S'/l at .P = 35 in controls. For all values of assumed P, S'/l in MN is significantly lower than at P = 35 in controls (P < 0.001).

capacity of glomerular capillary walls appears to be predomi- actually enhanced (Fig. 9). By exclusion, we infer that either nant, and an invariable feature of the glomerular injury in this reduced hydraulic permeability of glomerular capillary walls or diminished ultrafiltration pressure (or some combination of the disorder. two) must provide the basis for the observed hypofiltration [17]. Glomerular morphometry The net pressure for ultrafiltration represents the imbalance The percent of glomeruli exhibiting global glomerulosclerosis between the glomerular transcapillary hydraulic pressure difwas only slightly higher in MN than in controls (median = 5 vs. ference (P) and the opposing oncotic pressure exerted by 0). The median volume of the remaining 95% of glomeruli that retained proteins as plasma transits axially along the glomerular remained patent was greater than twofold larger in MN than in capillaries (HGC). Whereas we were unable to determine the controls, 3.1 x 106 versus 1.4 x 106 p.m3 (Figure 9A). Thus, value for zP, our findings indicate that 11GC must have been despite a trend to lower filtration surface density (S = 0.11 in profoundly lowered in MN. Not only was the oncotic pressure MN vs. 0.15 in controls), the actual surface area of the interface of plasma entering the glomerular tuft (HA) severely depressed, between peripheral capillary walls and capillary lumen (S) was but the extent to which intraluminal protein concentration, and significantly higher in MN than in the control group, 347 versus hence oncotic pressure, could rise during axial plasma flow was 183 p.m2 x io (Fig. 9B). Given that the overwhelming majority severely limited by the low filtration fraction observed in this of glomeruli in patients with MN were patent, it appears that disorder (Fig. 2B). diminished hydraulic permeability, rather than a loss of filtraWe have shown [221 that Hc in nephrotic humans rises in a tion surface area, must be implicated in the depression of linear fashion as water is removed by ultrafiltration; it follows computed Kf in this disorder. that: The properties of two structures along the filtration pathway — FF) (3) that can influence hydraulic permeability are the thickness of the glomerular basement membrane (GBM) and the density of where HE is the efferent oncotic pressure (Fig. 3). From the interpodocytic diaphragms at the base of the epithelial calculated 11E and observed A, we estimate that 11GC (the filtration slits [21]. The median GBM thickness in MN was 889 arithmetic mean of HA and IIE) was depressed in MN to a nm, a value which is more than twofold larger than control median value of only 15.9 mm Hg, considerably below the value of 380 nm, P < 0.0001 (Fig. lOB). Just as striking was a corresponding value of 27.7 mm Hg in healthy controls (Fig. marked reduction in filtration slit frequency consequent upon 11). Thus, at the low filtration fraction observed in humans with broadening of the epithelial podocytes (Fig. bA). Of note, MN, depression of 11GC alone can be inferred to elevate the ifitration slit frequency was reduced in MN to only 308 slits/mm pressure for ultrafiltration by approximately 12 mm Hg. GBM compared with 1152 slits/mm GBM in controls, P < Analyses of segmental vascular resistance in experimental 0.0001. Clearly, the number of diaphragms at the base of these models of the nephrotic syndrome in the rat using the mislits must have declined in parallel, limiting water flux into cropuncture technique have revealed resistance in afferent arterioles to be lowered in proportion to, or proportionately Bowman's space. Discussion

more than that in efferent arterioles [6—8]. A comparable change

in segmental resistance in human MN would facilitate the In the current study, patients with biopsy-proven membra- transmission into glomerular capillaries of arterial pressure, nous nephropathy were analyzed consecutively. Among pa- which we have found to be elevated on average by 15 mm Hg in tients bearing this diagnosis, the majority were found to have a the patients of the present study (Fig. 4). The glomerular depressed GFR. A morphometric analysis confirmed that the capillary hypertension in turn, would serve to elevate and not overwhelming majority of glomeruli were patent and that the depress zP. Although we cannot exclude a species difference in glomerular capillary surface area available for filtration was segmental vascular resistance in humans with MN, it is difilcult

396

Ting et al: Glomerular hypofiltration

A

B

x

P < 0.0001

'

I

a)

ca

?c5,,

E0 can

2. a)

0E

]

1

>)< 2 T

1

I

400

-.

ca\< 300 1: cflt: 200

.

100 Fig. 9. Glomerular volume (A) and filtration

0

0 Controls Membranous

Controls Membranous

A

B

1500 U)

2000

U)

a)

>'

C.)

1000

surface area per glomerulus (B) in the control and MN groups,

P < 0.0001

1500

f

a)

C

1000 .Q

S

500

500 0

0 Controls Membranous

Controls Membranous

Fig. 10. Epithelial filtration slit frequency (A) and glomerular basement membrane thickness (B) in the control and MN groups.

to conceive of a decline in P of sufficient magnitude to offset

in patients with MN, explaining our computation that Kf is depressed by approximately 75% in these patients. accordingly led to the conclusion that net ultrafiltration presWe conclude that hydraulic permeability of glomerular capsure is likely to have been markedly elevated in MN and that illary walls is likely depressed in all nephrotic patients with MN profound depression of Kf was the proximate cause of the because basement membrane thickening and foot process hypofiltration observed in these subjects. broadening are prominent in virtually all these patients. An We have already defined K1 as the product of hydraulic offsetting increase in the surface area available for filtration, permeability and the surface area available for ifitration. As and an increase in the net pressure for ultrafiltration are usually stated previously, the latter quantity was significantly larger in insufficient to compensate for the massive depression of hyMN than in controls, a consequence of the glomerular hyper- draulic permeability of glomerular capillary walls. As a result, trophy observed in this disorder [4] (Fig. 9). By exclusion, a the vast majority of nephrotic patients with MN exhibit glomer-

the 12 mm Hg depression of the opposing

We are

reduction in hydraulic permeability due to alterations along the ular hypofiltration. extracellular pathway for transcapillary water flux appears to be

implicated as the predominant if not exclusive cause of the

Acknowledgments

computed depression of Kf. The GBM and interpodocytic diaphragms have recently been

This study was supported by grants #DK29985-09, DK40800-03 and GCRC grant MO1-RR0007O from the National Institutes of Health. Drs.

estimated to each account for approximately 50% of the hydraulic permeability of normal glomerular capillary walls [21].

Robert Ting and Batya Kristal are research fellows of the Juvenile Diabetes Foundation International.

The striking increase in thickness of the GBM as well as decreased frequency of filtration slits (and by extension the interpodocytic diaphragms) are likely to have acted in concert to lower hydraulic permeability of the glomerular capillary wall

Reprint requests to Bryan D. Myers, M.D., Stanford University School of Medicine, Division of Nephrology, Room S-069, Stanford, California 94305, USA.

Ting et al: Glomerular hypofiltration

50

397

MYERS BD: Slope of serial GFR and the progression of diabetic glomerular disease. JAm Soc Nephrol 3:1358—1370, 1993 6. ALLISON MEM, WILSON CB, GOTTSCHALK CW: Pathophysiology

P< 0.001

of experimental glomerulonephritis in rats. J C/in Invest 53:1402— 1423, 1974 7. ICHIKAWA I, H0YER JR, SElLER WM, BRENNER BM: Mechanisms

of glomerulotubular balance in the setting of heterogeneous glomerular injury. J C/in Invest 69:185—198, 1982

40

8. YOSHIOKA T, RENNKE HG, SALANT DJ, DEEN WM, IcHIKAwA I:

Role of abnormally high transmural pressure in the permselectivity

defect of glomerular capillary wall: A study in early passive Heymann nephntis. Circ Res 61:53 1—538, 1987 9. BOHRER MP, BAYLIS PC, ROBERTSON CR, BRENNER BM: Mecha-

30

nisms of the puromycin-induced defects in the transglomerular passage of water and macromolecules. J Cliii lnvest 60:152—161, 1977 10. ANDERSON S, DIAMOND JR, KARNOVSKY Mi, BRENNER BM:

E E

Mechanisms underlying transition from acute glomerular injury to late sclerosis in a rat model of nephrotic syndrome. J C/in Invest

U (I,

82:1757—1768, 1988

20

11. SCHOLEY JW, MILLER PL, RENNKE HG, MEYER TW: Effect of converting enzyme inhibition on the course of adriamycin-induced nephropathy, Kidney Int 36:816—822, 1989 12. ICHIKAWA I, RENNKE HG, HOYER JR: Role of intrarenal mecha-

nisms in the impaired salt excretion of experimental nephrotic syndrome. J Cliii Invest 71:91—103, 1983

10

13. ROBSON AM, GIANGICOMO J, KIENSTRA RA, NAQVI ST, INGEL-

FINGER JR: Normal glomerular permeability and its modification by minimal change nephrotic syndrome. J C/in Invest 54:1190—1199, 1974 14. CHANG RLS, ROBERTSON CR, DEEN WM, BRENNER BM: Permse-

0

lectivity of the glomerular capillary wall to macromolecules: I.

Controls

Membranous

Fig. 11. Integrated glomerular intracapillary oncotic pressure in the

control and MN groups.

Theoretical considerations. Biophys J 15:861—886, 1975 15. BATTILANA C, ZHANG H, OLSHEN R, WEXLER L, MYERS BD: PAH

extraction and the estimation of plasma flow in the diseased human kidney. Am J Physiol 30:F726—F733, 1991 16. NAKAMURA Y, MYERS BD: Charge-selectivity of proteinuria in diabetic glomerulopathy. Diabetes 37:1202—1211, 1988 17. DEEN WM, ROBERTSON CR, BRENNER BM: A model of glomerular

ultrafiltration in the rat. Am J Physiol (Renal Fluid Electrol Physiol 10)223:1178-1183, 1972 18. MYERS BD, PETERSON C, MOLINA CR, TOMLANOVICH SJ, NEW-

References 1. SHEMESH 0, Ross JC, DEEN WM, GRANT GW, MYERS BD: Nature

of the glomerular capillary injury in human membranous glomerulopathy. J Clin Invest 77:868—877, 1986 2. CHAGNAC A, KIBERD BA, FARINAS MC, STROBER S, SIBLEY RK,

HOPPE R, MYERS BD: Outcome of the acute glomerular injury in proliferative lupus nephritis, J Cl/n Invest 84:922—930, 1989 3. GUASCH A, HASHIMOTO H, SIBLEY RK, DEEN WM, MYERS BD:

TON LD, NITKIN R, SANDLER H, MURAD F: Role of cardiac atria

in the human renal response to changing plasma volume. Am J Physiol 23:F562—F573, 1988 19. DEEN WM, BRIDGES CR, BRENNER BM, MYERS BD: A Hetero-

porous model of size-selectivity: Application to normal and nephrotic humans. Am J Physiol 249:F374—F389, 1985 20. HIROSE K, OSTERBY R, NOZAWA M, GUNDERSEN HJG: Develop-

ment of glomerular lesions in experimental long-term diabetes in

Glomerular dysfunction in nephrotic humans with minimal changes

the rat. Kidney liii 21:688—695, 1982 21. DRUMOND MC, DEEN WM: Structural determinants of glomerular

or focal glomerulosclerosis. Am J Physiol (Renal Fluid Electrol

capillary hydraulic permeability. (abstract) J Am Soc Nephrol

Physiol) 29:F728—F737, 1991 4. GUASCH A, SIBLEY RK, HUIE P, MYERS BD: Extent and course of

3:560, 1992 22. CANAAN-KUHL S, VENKATRAMAN ES, ERNST SIB, OLSHEN RA,

glomerular injury in human membranous glomerulopathy. Am J

MYERS BD: Relationships among protein and albumin concentra-

Physiol (Renal Fluid Electrol Physiol) 32:F1034-F1043, 1992

tions and oncotic pressure in nephrotic plasma. Am J Physiol

5. AUSTIN SM, LIEBERMAN IS, NEWTON LD, MEJIA M, PETERS WA,

(Renal Fluid Electrol Physiol) (in press)