Effect of flow and stretch on the [Ca2 ]i response of principal and intercalated

轻羽

作者：Wen Liu, Shiyun Xu, Craig Woda, Paul Kim, Sheldon Weinbaum,  Lisa M. Satlin作者单位：1 Department of Pediatrics, Mount Sinai School of Medicine, New York 10029-6574; and Department of Mechanical Engineering and Center for Biomedical Engineering, The City College of New York, New York, New York 10031 ! I$ k# M1 H8 S) f. D  T
                  
) ?+ b4 b$ w. {                  / X( ~2 `! s4 D5 C3 q3 {
         
) B/ e) k: K- a                        
1 E  h- P6 V- M* F0 V9 z            
  ^! @+ ]- z3 V            ! x3 t# W/ P8 I+ s7 v
            
/ n3 h1 V  P/ m7 M. E- V            
3 \' S8 ~7 Y, R4 z3 `: Z8 K% K                     
7 N3 ?; i$ a2 G+ o        
- p+ S! r9 Y7 T8 j5 O- Y+ {5 c- `        2 J4 F# w$ n6 m" S* b# l
        ' {! I+ M$ |' h  [0 v! e) y
          【摘要】: E. C! n  V4 m5 {
      An acute increase in tubular fluid flow rate in the microperfused cortical collecting duct (CCD), associated with a 20% increase in tubular diameter, leads to an increase in intracellular Ca 2  concentration ([Ca 2  ] i )in both principal and intercalated cells (Woda CB, Leite M Jr, Rohatgi R, and Satlin LM. Am J Physiol Renal Physiol 283: F437-F446, 2002). The apical cilium present in principal but not intercalated cells has been proposed to be a flow sensor. To determine whether flow across the cilium and/or epithelial stretch mediates the [Ca 2  ] i response, CCDs from New Zealand White rabbits were microperfused in vitro, split-open (to isolate the effect of flow across cilia), or occluded (to examine the effect of stretch and duration/magnitude of the flow impulse), and [Ca 2  ] i was measured using fura 2. In perfused and occluded CCDs, a rapid ( 3 min) increase in luminal flow rate and/or circumferential stretch led to an approximately threefold increase in [Ca 2  ] i in both principal and intercalated cells within 10 s. This response was mediated by external Ca 2  entry and inositol 1,4,5-trisphosphate-mediated release of cell Ca 2  stores. In split-open CCDs, an increase in superfusate flow led to an approximately twofold increase in [Ca 2  ] i in both cell types within 30 s. These experimental findings are interpreted using mathematical models to predict the fluid stress on the apical membranes of the CCD and the forces and torques on and deformation of the cilia. We conclude that rapid increases in luminal flow rate and circumferential stretch, leading to shear or hydrodynamic impulses at the cilium or apical membrane, lead to increases in [Ca 2  ] i in both principal and intercalated cells.
7 _4 V1 M8 ^5 W$ j# [  `          【关键词】 cilium cytoskeletal deformation fluid shear stress mechanotransduction fura intracellular calcium concentration# u$ X8 ^8 C; I7 x4 X. W0 l
                  THE CORTICAL COLLECTING DUCT (CCD) of the mammalian nephron contributes to the final renal regulation of Na  , K  , acid-base, and water homeostasis. The CCD is a heterogeneous epithelium comprised of two morphologically and functionally distinct cell types. Principal cells reabsorb Na   and water (in the presence of vasopressin) and secrete K  , whereas intercalated cells transport acid-base and can, under certain conditions, absorb K   ( 5, 17, 29, 35, 37 ). Although these cells reside side by side in the CCD, they are considered not to be coupled, maintaining different resting intracellular pH ( 32 ) and showing no apparent functional intercellular communication under resting conditions ( 44 ).
" ?4 D7 G& B$ P+ z) ]
3 E2 a5 P- l% [' b2 tWe have recently shown that an acute increase in tubular fluid flow rate in the microperfused CCD, associated with an 20% increase 150 nM increase in intracellular Ca 2  concentration ([Ca 2  ] i ) in both principal and intercalated cells comprising this segment ( 44 ). 1 Whether the increase in [Ca 2  ] i reflects increased Ca 2  influx from the external solutions, Ca 2  mobilization from internal stores, and/or decreased Ca 2  efflux from the cell remains uncertain. Also unknown is the identity of the stimulus for the flow and/or stretch-induced [Ca 2  ] i transient. Like endothelial cells lining the vasculature ( 6, 38 ), renal tubular epithelial cells likely experience at least three types of mechanical forces in response to variations in urinary flow rate: hydrostatic pressure, circumferential stretch, and fluid flow-induced shear or drag forces.
1 H) ~& U+ ^* m+ I3 Q4 k0 `, ^" h9 v. G1 D3 [( @' P3 r6 h' ]5 }* B
A likely candidate for the proximate flow sensor in the tubule is the apical primary cilium, a nonmotile structure projecting from the centriole of all renal tubular cells except intercalated cells of the collecting duct ( 18, 24, 34 ). Madin-Darby canine kidney (MDCK) principal cells express a primary cilium that is 8 µm in length ( 26 ). Praetorius and Spring ( 26 ) recently reported that bending of the cilium of MDCK cells either directly with a micropipette or by increasing the rate of flow superfusing the apical surface of monolayers resulted in an increase in [Ca 2  ] i, a response attributed to external Ca 2  entry through mechanosensitive channels followed by Ca 2  release from inositol 1,4,5-trisphosphate (IP 3 )-sensitive internal stores. Direct mechanical stimulation of the apical membrane of MDCK cells also led to a transient increase in [Ca 2  ] i that differed from the response induced by bending of the cilium in its larger amplitude and shorter time delay between stimulation and peak [Ca 2  ] i ( 10 vs. 40 s for bending of the cilium). Furthermore, the response was not affected by removal of extracellular Ca 2 , suggesting that direct mechanical manipulation of the apical membrane stimulates release from internal stores and does not require external Ca 2  entry. The physiological importance of structurally and functionally intact cilia in renal epithelial cells is underscored by the growing body of evidence that disruption of proteins localized to the cilia, such as polaris ( 45 ), cystin ( 15 ), and polycystin 1 ( 22 ) in orpk, cpk, and Pkd1 del34/del34 mice, respectively, is associated with a renal cystic phenotype.
! t9 L! s; i% v) U' ^0 F6 E8 m' h4 _
The flow channel used by Praetorius and Spring ( 26 ) in the studies described above had a cross section 3-mm wide and 3-mm deep and thus generated an average linear velocity of 110 µm/s at a flow rate of 1 µl/s. The authors attempted to simulate in their flow chamber the average velocity that would be encountered in an intact perfused tubule. An average velocity of 110 µm/s corresponds to a tubular fluid flow rate of 7 nl/min when the internal diameter of the tubule is assumed to be 37 µm, the value used by Praetorius and Spring ( 26 ). Model calculations performed in the present study, however, assume an internal tubular diameter of 25 µm, a dimension we ( 30, 44 ) and others ( 42 ) have previously reported for the rabbit CCD. Yet, as will be demonstrated in the present study, it is not the average velocity that one should attempt to reproduce in simulating mechanotransduction in a flow chamber, but rather a dynamic similarity of the fluid shear stress acting on the apical surface of the cells and, more importantly, the hydrodynamic forces and torques required to bend the cilia, at least in the case of principal cells.; s2 x8 Q; w: h2 z5 J
8 G7 U# i; H& Y: R/ w2 u1 ~
The internal structure of the principal cell cilium is characterized by a 9   0 organization of microtubules ( 34, 40 ). The deformation of this type of cilium in a cultured kidney epithelial cell line was studied by Schwartz et al. ( 34 ) using nonlinear beam theory ("heavy elastica" model). These authors were primarily interested in determining the flexural rigidity of the cilium. Because a detailed hydrodynamic model for determining the drag forces on the cilia and their hydrodynamic interaction with one another was not available until recently ( 12 ), Schwartz et al. ( 34 ) used an empirical equation for the flow past a single cylinder to calculate the drag on the cilium. Using this approach, the authors were able to demonstrate good agreement between theoretical predictions and experimental observations of the bending response of the cilium subject to a physiologically appropriate range of flow rates. These authors also raised a question relevant to the studies performed by Praetorius and Spring ( 26 ) and that posed in the present study; specifically, does stretch-activation of channels (e.g., Ca 2  channels) reflect bending deformation of the cilium with resultant tension in its plasma membrane or deformation of the cortical cytoskeleton in the terminal web at the base of the cilium arising from the resisting moment of the linkages that attach the cilium to the cortical network of microtubules and actin filaments at its base? Both of these possibilities will be explored in the present study using a mathematical model.
4 W5 I: w1 j* P+ q: b3 W' P
; c5 m1 W! x; A, a* V6 J2 y# ~Although the data summarized above provide compelling evidence that the cilium is a flow sensor in collecting duct epithelia, two important differences are evident on comparing the response of the native collecting duct ( 44 ) with that of the cultured cell line. First, in the perfused tubule, both principal and intercalated cells, the latter devoid of cilia, respond to an increase in flow with an increase in [Ca 2  ] i ( 44 ). In addition, there is no evidence of gap junctional communication in native tubules ( 44 ), whereas in MDCK cells the Ca 2  signal spreads laterally by the diffusion of IP 3 through gap junctions ( 26 ).
0 j& n% q$ L/ v, f0 }4 a2 C. w: X2 O9 H$ @% H- x% F
The purpose of the present study was thus to identify the cellular mechanisms underlying the flow and/or stretch-induced [Ca 2  ] i transient in the native CCD. We specifically sought to determine whether 1 ) flow across the apical membrane of the CCD, comprised of principal cells with apical cilia and intercalated cells decorated with apical microvilli/microplicae, and/or circumferential stretch mediates the flow-induced increase in [Ca 2  ] i in this segment and 2 ) the [Ca 2  ] i transient induced by flow/stretch reflects external Ca 2  entry and/or mobilization of internal stores. To examine these questions, we used the Ca 2  -sensitive fluorescent dye fura 2 to measure changes in principal and intercalated cell [Ca 2  ] i in isolated CCDs microperfused in vitro in their native geometry (and thus subject to both flow and stretch), split-open to expose the luminal surface to superfusate flow in the absence of stretch (simulating an epithelial monolayer preparation) or occluded to isolate the effect of circumferential stretch in response to transient fluid flow loading of increasing magnitude and duration. We also performed a rigorous hydrodynamic analysis of the flow in our perfusion chamber and then developed a detailed model to predict the fluid shear stress at the apical surface of the split-open tubule and the hydrodynamic forces and torques acting on the cilia. Finally, these predictions were quantitatively compared with the equivalent forces and torques in the intact perfused tubule by suitably modifying the theoretical model developed for brush-border microvilli by Guo et al. ( 12 ). We show that the local velocities, forces, and torques reproduced in existing flow chambers are significantly smaller than those encountered in vivo because the cilia lie within a hydrodynamic boundary layer at the base of the flow chamber where the local velocity at the cilia tips in current experiments is far less than encountered in vivo.) m, b3 o: h3 c2 D9 X7 K! B

6 X' \, j1 ]! _% ?MATERIALS AND METHODS+ W$ U! g( H* X  w& N5 \% Q  b
+ V: e% {  |% |# z, U% k
Animals% X2 d" J5 }  e* N9 Q& v" H4 V

# p" B0 _4 a; J7 FAdult female New Zealand White rabbits were obtained from Covance (Denver, PA) and housed in the Mount Sinai School of Medicine animal care facility. The animals were fed standard rabbit chow and given free access to food and water. Animals were killed by intraperitoneal injection of a lethal dose of pentobarbital sodium (100 mg/kg). All experiments were conducted in accordance with the National Institutes of Health guidelines for the care and use of laboratory animals." z* Z% e  @+ ~6 Z

- f3 V. U5 s, z  }The kidneys were removed, and single tubules were dissected freehand in cold (4°C) dissection solution containing (in mM): 145 NaCl, 2.5 K 2 HPO 4, 2.0 CaCl 2, 1.2 MgSO 4, 4.0 sodium lactate, 1.0 sodium citrate, 6.0 L -alanine, and 5.5 D -glucose, pH 7.4, 290 ± 2 mosmol/kgH 2 O ( 44 ). A single tubule was studied from each animal.
+ `1 L# b* T% z9 ^0 u+ {. s8 Y3 ^+ h! f/ N, i
Isolated Tubules& D9 K8 n& t- F3 h

( E  e8 Q1 j: n' B. |Isolated tubules were either microperfused in vitro (native geometry) or split open to expose the luminal surfaces of all cells to superfusate flow, as previously described ( 44 ). Tubules were perfused and bathed at 37°C with Burg's solution, which resembled the dissection solution except that 25 mM NaCl was replaced by NaHCO 3, and the solution was gassed with 95% O 2 -5% CO 2 at room temperature to reach a pH of 7.4 ( 44 ). In some experiments, CCDs were perfused with Burg's solution prepared without Ca 2  (Ca 2  -free perfusate).* I" Y# d% |5 z7 l0 [

# V; h! [* d/ \To distinguish between the contribution of circumferential stretch and a flow-induced response in tubules perfused in their native geometry, the distal ends of some perfused CCDs were occluded manually, using a blunt pipette, after baseline [Ca 2  ] i was measured, and the pressure in the tubule was increased by raising the height of the perfusate reservoir. The dilation of the tubule creates a transient fluid flow whose magnitude and duration depends on the rate at which the reservoir is elevated, the final height of the reservoir, and the distance [length ( L ), µm] of the measurement location from the occlusion site. The instantaneous average velocity u av (in µm/s) at the measurement site is given by u av = (2 L / R )d R /d t, where R ( t ) is the instantaneous tubule radius (in µm) and d R /d t is the rate at which it is changing in response to the time-dependent increase in pressure. When the pressure reservoir is raised slowly, d R /d t 0, leading to a purely elastic response; rapid increases in height generate a transient flow impulse in addition to the circumferential stretch response. Thus the rate at which the reservoir height is raised enables the elastic response to be distinguished from a combined response that involves both fluid flow and circumferential stretch.# Z% R/ e" Q# r' ^8 z; @/ s6 n4 i. U

: R3 p0 {: Z: AMeasurement of [Ca 2  ] i( q& }$ J2 F% y

, N& Q9 c- R& D6 ?" p, A7 l4 rAfter equilibration, tubules were loaded with 20 µM of the acetoxymethyl ester of fura 2 (Molecular Probes, Eugene, OR) added to the bath for 20 min. In several experiments, rhodamine-labeled peanut lectin (PNA; Vector Laboratories, Burlingame, CA) was added to the luminal perfusate for 5 min to identify intercalated cells; rabbit principal cells do not bind PNA ( 31 ). [Ca 2  ] i was measured in individually identified fura 2-loaded cells, as previously described ( 44 ). Two to four principal and intercalated cells residing in the lateral wall of each perfused CCD or localized to the center of each split-open segment were analyzed. For each cell in each tubule, three (for peak) to five (for baseline) measurements of [Ca 2  ] i were averaged to generate a mean value for baseline (obtained immediately before flow was increased) and peak (maximal) concentrations. The time to peak was defined as the time interval between the onset of high flow rate and the detection of the maximal [Ca 2  ] i value. The mean responses of principal and intercalated cells in each tubule were used for further cell-specific data analysis.4 _0 E! n) I* I- }; f

0 c1 q! M$ U7 o, MTo assess the source of Ca 2  contributing the [Ca 2  ] i transient, CCDs were studied in the absence of luminal and/or basolateral Ca 2  or pretreated with basolateral 2-aminoethoxydiphenyl borate (2-APB; 10 µM), a cell-permeant inhibitor of the IP 3 receptor ( 10, 19 ), or basolateral thapsigargin (100 nM), an irreversible inhibitor of endoplasmic reticulum (ER) Ca 2  -ATPase that prevents refilling of intracellular Ca 2  pools and leads to depletion of internal stores. Several additional studies were performed in microperfused CCDs bathed in Burg's solution containing apyrase (10 U/ml; Sigma-Aldrich), an enzyme that rapidly hydrolyzes 5'-nucleotide triphosphates to monophosphates, to examine whether autocrine/paracrine signaling by 5'-nucleotides mediates the response of the CCD to flow/stretch.. [% O$ A- y( T. ^# [' W
7 V/ Q7 s8 C$ G. ^8 Q* c! Q0 x- y
Effect of Flow on Gap Junctional Intercellular Communication
7 j3 a. e: c, M8 W3 O" I$ }' I% t+ l% U" O2 m
Randomly identified individual cells in split-open CCDs were injected with Lucifer Yellow for monitoring of intercellular coupling, as previously described ( 44 ), before and after initiation of high superfusate flow (25 µl/s).
: I0 v- ~) u( f4 Y4 u$ _6 M$ T) f
Measurement of Cilia Length6 _) A: J# s8 C! ~2 d

5 ?9 c' ~' T4 u( y% C# aCilia length was measured by indirect immunofluorescence microscopy using a monoclonal antibody specific for acetylated -tubulin (6-11B-1; generous gift from G. Piperno; see Ref. 25 ) and a commercially available anti-bovine -tubulin mouse monoclonal antibody (Molecular Probes). Split-open CCDs were fixed for 15 min in 2.5% paraformaldehyde at room temperature, permeabilized with Triton X-100 (0.1% for 30 min for 6-11B-1 and 0.3% for 15 min for anti-tubulin antibody) in PBS containing BSA (15 mg/ml), and incubated for 2 h at room temperature with 6-11B-1 or overnight at 4°C with -tubulin antibody (1 µg/ml). After a thorough washing, the tubules were incubated for 2 h with a 1:1,000 dilution of the secondary antibody, a FITC-conjugated anti-mouse IgG (Sigma, St. Louis, MO), at room temperature. After being washed with PBS three times, an 8-µl droplet of Prolong anti-fade solution was deposited on the tubule, which was then covered with a glass coverslip applied with slight pressure to flatten and bend the apical cilia in a uniform direction. Confocal microscopic analysis of six tubules from three different rabbits ( n = 2 cilia/tubule) revealed that the cilia length was 2.2 ± 0.4 µm with cilia possessing bulbous distal ends ( Fig. 1 ).9 c* e3 Q! H6 u6 ?+ W
( z" V5 }* T7 ^! X0 N8 H% m; r1 e
Fig. 1. Indirect immunofluorescence microscopy of cilia in fixed and permeabilized split-open cortical collecting ducts (CCDs). CCDs were labeled with antibodies directed against Tg737 ( A ) or -tubulin ( B ) and visualized by confocal microscopy (0.098 µm/pixel) after application of an FITC-conjugated secondary antibody.! N( x8 |4 X5 b+ ^& _% R; D/ g

, k3 a- y/ n5 \8 N3 ?' _6 S0 MMathematical Models for Shear Stresses, Forces, and Torques in Perfused and Split-Open CCDs& z- p. t- U. _  I# q' F6 p) [" v

6 n0 |: U% O/ y3 V1 E/ B/ o' |1 s, SAs indicated above, the microperfused CCD subject to an increase in tubular fluid flow rate experiences a shear stress on the apical membranes of the cells comprising the segment, a drag and torque on their primary cilium, and circumferential stretch ( 44 ). To examine the isolated response to apical shear stress, we studied split-open tubule monolayers. To compare the shear forces acting on the apical surface of the tubular epithelial cells and the forces and torques acting on the cilia, mathematical models were developed to describe the fluid flow past the cilia in the experimental chamber used for measurement of [Ca 2  ] i in split-open CCDs. These predictions were then compared with those of a newly developed mathematical model for the flow past the cilia in a perfused tubule.
# r+ E' w+ Q, i, ?& ?9 q- Q6 B$ `) o: K( K6 m( d) D
Flow in the channel. A sketch of the geometry (transverse and longitudinal sections) of the flow chamber is shown in Fig. 2. The channel has a free water surface that is maintained at a constant height of 1.5 mm by a vacuum siphon placed just at the top of the bevel in the diagram. The tip of the flow inlet (a bent 18-gauge needle, with sharp tip removed) was positioned just above the floor of the specimen chamber. The split-open tubule is positioned in the center of the bottom surface of the chamber with the tubule axis aligned parallel to the longitudinal section. An en face view of the specimen with representative cellular dimensions is shown in Fig. 3 A. Because the dimensions of the cross section of the chamber are much greater than the width or height of the split-open tubule, the basic flow in the tubule can be calculated as if the tubule was absent. This flow can be closely approximated as a unidirectional viscous flow in a beveled trough that satisfies no-slip boundary conditions on its bottom and slanted side walls and a zero fluid shear stress at the free surface. Because of the complex cross-sectional geometry, the governing Navier-Stokes equation was solved numerically. The relationship between the shear stress on the center line at the bottom wall of the channel, w (0) in dyn/cm 2, and the flow rate Q, is
1 \+ b+ Q. I+ m. q/ q4 d
. X  Q5 n/ B0 y1 Y+ J4 Z# XFig. 2. Geometry of the flow chamber used for measurement of flow-induced changes in intracellular Ca 2  concentration ([Ca 2  ] i )in split-open CCDs shown in transverse ( A ) and longitudinal ( B ) sections.
6 I0 S9 [) t/ X% l) V& q* \. z( `* m8 W. B6 g. U
Fig. 3. Theoretical models for split-open CCD showing, in an en face view, the hexagonal cell geometry and cell dimensions ( A ) and a perfused tubule showing the arrangement of primary cilia at the tubule wall ( B ). Solid and dotted lines represent cilia in different planes.
# u1 S' M; R/ y# c3 s" F" |/ _' Z3 O, `6 C1 X8 i
Velocity in cilia layer of split-open tubule. It is a difficult and subtle problem to accurately calculate the flow in the disturbed region at the apical surface of the split-open tubule because of the presence of the protruding cilia. Our idealized mathematical model is shown in Fig. 3 A. The cells are arranged in a hexagonal array. Given that intercalated cells account for only 25% of the total number of cells comprising the rabbit CCD, this cell population is neglected in the model. At the center of each hexagonal cell is a vertical cilium that is approximated as a circular cylinder of 2.5 µm height and 0.2 µm diameter ( 1 ). Because the height of the cilia is small compared with the width of the open tubule sheet (78.5 µm), the flow in the central region of the sheet can be approximated as a two-dimensional flow in which edge effects from the lateral boundaries of the sheet can be neglected. The Reynold number based on the cilia tip velocity and height is of the order of 10 -4 and, thus, in the Stokes slow flow regime. The viscous dissipation resulting from the protruding cilia can be described by a distributed body force, or Darcy resistance. An approximate governing equation for the flow in the disturbed layer is thus given by a modified Stokes equation with a Darcy term, as described by Bird et al. ( 3 ); m" t* s) z# F0 v6 K* H. I
  f$ F& s5 O3 ?# u
Here, z (µm) is the normal coordinate measured from the apical surface, U ( z ) is the local average velocity (µm/s) at height z surrounding each cilium, P is the pressure (dyn/cm 2 ), µ is the fluid viscosity (0.01 dyn · s/cm 2 ), and k is the Darcy permeability (30 µm 2 ). At the apical membrane of the tubule epithelium ( z = 0 µm), the flow velocity vanishes, whereas at the cilia tips ( z = height in µm), a shear stress imposed by the outer flow is essentially the same as that given by w (0) in Eq. 1.2 C/ k' M2 m1 y! y+ C5 N, _! C

- T: B4 v5 g: UThe use of Eq. 2 to determine the average velocity field and the forces and torques on the cilia has been examined in a hydrodynamically equivalent problem, the Stokes flow past a periodic array of slender vertical fibers in a parallel walled channel ( 39 ). Comparison with exact Stokes solutions for the average flow past each fiber and the resulting drag show excellent agreement for fiber solid fractions up to 0.7." F) K0 ]9 A5 w

' q/ U- L  |1 A" `: W1 q* l. gThe solution of Eq. 2 with the foregoing boundary conditions is given in APPENDIX 1. Representative solutions for the disturbed average velocity profile are shown in Fig. 4 for cilia of different lengths ranging in height from 2.5 µm in the adult rabbit principal cell to 8 µm in confluent monolayers of MDCK cells examined by Praetorius and Spring ( 26 ). The nearly straight line labeled "without cilia" in Fig. 4 is the velocity profile that would exist at the bottom of the channel in the absence of a split-open tubule. The distortion of the velocity profiles is significantly greater for the MDCK cilia because of their greater protrusion length.
; O# i" Q9 {6 e, p7 ~- L0 P
/ |- z: @% U2 L) ^  U% oFig. 4. Model predictions for velocity profiles in cilia layer of a split-open tubule showing distortion of the average velocity profile near the tubule wall resulting from cilia of different lengths. Chamber flow rate, 25 µl/s.2 Z+ S2 i. m. S: ~1 f  q4 V
$ `2 a5 k  W7 p
At a high superfusate flow rate of 25 µl/s, the maximum velocity in the channel is 6.56 mm/s, but the local velocity at a distance 2.5 µm (height of the cilia) off the bottom wall on the channel centerline is only 17.9 µm/s, a value 360 times smaller ( Table 1 ). At a low flow rate of 3.2 µl/s, the velocities, which scale linearly, are 12.7% of those just cited. In contrast, we shall show that, for a tubule perfusion rate of 5 nl/min, the velocity at the tip of a 2.5-µm cilia ( Fig. 1 ) in a typical tubule of 25-µm internal diameter ( 42 ) is 119 µm/s, whereas the peak velocity at the tubule center line is 350 µm/s ( Table 1 ). It is clear from these theoretical predictions that the maximum or average channel velocity has little bearing when trying to achieve dynamic similitude with the forces and torques experienced by cilia in intact tubules.% H% T( S  o: L% R: i, b+ R

& Y" t! E* _+ nTable 1. Calculated results of theoretical models for the split-open and perfused CCDs
2 ?" W/ z7 D0 z- T( e  `' p
7 t* Z0 A. ?6 X( l( D$ c+ ]Forces and torques on cilia in split-open CCDs. With the use of the theoretical approach developed by Guo et al. ( 12 ), local solutions of the Stokes equations for creeping flow can be used to calculate the local drag on the cilia per unit length. This drag is proportional to the local velocity U ( z ) given by Eq. A5 in APPENDIX 1. The local drag can be integrated along the length of the cilia to provide the total drag on each cilium. If this local drag is multiplied by the local moment arm from the base of the cilia at z = 0, and this local moment is integrated along the length of the cilium, the total torque can be obtained. The expressions for the total drag and torque are given by Eqs. A8 and A9 in APPENDIX 1.
# U5 _' X$ v. x+ ]# ?3 W3 S8 ]& Q1 x8 C4 R2 r6 `
Velocity profile forces and torques on cilia in perfused tubules. Guo et al. ( 12 ) developed a theoretical model to describe the flow in the proximal tubule and the forces and torques acting on the brush-border microvilli. The flow in the brush-border layer is described by the axisymmetric equivalent of Eq. 2. The flow geometry for the proximal tubule is conceptually equivalent to our model for the CCD shown in Fig. 3 B except that, in the case of the proximal tubule, the layer of microvilli at the tubule wall forms a relatively dense ordered hexagonal array in which there is virtually no axial flow except in the immediate vicinity of the microvilli tips. The same basic model can be applied to the flow in the intact CCD, although the flow now can easily pass through the border layer of much more widely separated cilia that line the wall of the CCD. The disturbed velocity profiles predicted by the model are shown in Fig. 5 where one observes that the presence of the cilia causes only a modest distortion of the parabolic profile for Poiseuille flow in a tube without cilia and a modest change in the wall shear stress. These calculations assume a flow rate of 5 nl/min, an inner tubule diameter of 25 µm, and a value for k based on the separation distances of the cilia tips (see Eq. A7 in APPENDIX 1). The forces and torques acting on the cilia are determined in the same manner as described above for the split-open tubule using the velocity profiles shown in Fig. 5.
. o: \; ]9 V# i8 V8 K) i# R$ L
0 g, U  l3 U4 I5 c4 Z: sFig. 5. Model predictions for velocity profiles in perfused tubule for cilia of different lengths. Tubule diameter, 25 µm and flow rate, 5 nl/min. v = Velocity (µm/s).
7 [2 A) L" b$ s* J4 A7 }
" c% @* H" P' b8 JA central question that must be addressed in understanding mechanotransduction in the CCD is the relative magnitude of the drag force on cilia compared with the total drag force resulting from fluid shear on the apical membrane of an individual cell. The latter is the product of the wall shear stress w and the area of a hexagonal unit cell in Fig. 3 A.. c$ E6 ?8 v6 a; i+ _

) d; G: Y3 |6 n& F: d. ~" ]Statistical Analysis
9 D% n* h# M8 A' u1 O! [
& {3 _3 R: F( SResults are expressed as means ± SE; n equals the number of animals. Significant differences were determined by paired or unpaired t -tests, as appropriate, using the software program SigmaStat (SPSS, Chicago, IL). Significance was asserted at P
" j$ t% V! t  B  i& z4 y
# Y/ {6 F6 m: E' y% aRESULTS
6 k! G! k- H: T6 C3 h7 \6 N, o0 \; _3 ?5 T: _* l7 ^$ W
Effect of High Luminal Flow Rate on [Ca 2  ] i in Perfused CCDs- O  E# ~7 U: v, r/ o

4 t5 ^: r% `" o. d9 @In perfused tubules, a rapid increase in tubular fluid flow rate, sufficient to increase diameter by 14.4 ± 1.7% ( n = 7) within 1 s, led to a prompt increase in [Ca 2  ] i in both principal (106.3 ± 15.7 to 338.9 ± 67.2 nM) and intercalated (124.1 ± 20.9 to 346.3 ± 48.4 nM; Figs. 6 A and 7 ) cells within 10.6 ± 1.9 s in both cell types ( Fig. 8 ), followed by a gradual decay to a plateau value. [Ca 2  ] i remained significantly elevated above baseline for at least 20 min during a period of sustained high flow (Figs. 6 A and 7 ).
& m. x# @5 @: H$ o# K8 |! {( [2 G3 N- n, Q4 G
Fig. 6. Representative tracings of the effects on [Ca 2  ] i in a principal (PC; black) and intercalated (IC; gray) cell of a sustained increase in luminal flow rate in a microperfused CCD ( A ), high rate of superfusate flow in a split-open CCD ( B ), and circumferential stretch in an occluded CCD ( C ). [Ca 2  ] i increased in both cells in response to each maneuver. The bar indicates the period of sustained high flow.
! d: f1 i( p: K# j3 N$ B. s( {
1 O5 h7 N  L8 I* d0 O' o. H& JFig. 7. Summary of [Ca 2  ] i measurements in principal and intercalated cells in perfused ( n = 7), split-open ( n = 4), and occluded ( n = 5) CCDs in response to flow/stretch. Baseline and peak [Ca 2  ] i measurements are given, as are the values detected at 20 min of sustained high flow/stretch and 10 min after the flow rate was reduced to baseline in perfused and split-open CCDs. In all cells studied, an increase in flow/stretch led to an increase in [Ca 2  ] i. Principal and intercalated cell [Ca 2  ] i did not differ from baseline when measured 10 min after luminal or superfusate flow was reduced ("recovery"). Data are shown as mean values ± SE. * P
1 P3 L7 z/ T* \  [' U0 n
7 \* N! F* {7 w% ?1 vFig. 8. Time-to-peak [Ca 2  ] i in principal and intercalated cells in perfused ( n = 7), split-open ( n = 4), and occluded ( n = 5) CCDs in response to flow/stretch. Data are shown as mean values ± SE. * P
% [( f9 s1 ^; x3 |% \
2 I% e) Q. q# p) b2 l1 HEffect of High Superfusate Flow on [Ca 2  ] i in Split-Open CCDs9 Y3 B" a' V3 X  \) L

2 a0 w& f, t) s6 P6 s/ r# M( RIn split-open CCDs, an increase in superfusate flow from 0 to 3.2 µl/s failed to generate a [Ca 2  ] i response. However, an increase in superfusate flow from 3.2 to 25 µl/s led to an increase in [Ca 2  ] i in both principal (103.9 ± 9.2 to 213.9 ± 38.8 nM; P (101.1 ± 15.8 to 200.9 ± 9.4 nM; P cells in four CCDs (Figs. 6 B and 7 ). The time to peak [Ca 2  ] i averaged 36.3 ± 4.7 s in these preparations ( P compared with perfused CCDs; Fig. 8 ). At 20 min of sustained high flow, [Ca 2  ] i was not significantly different [ P = not significant (NS)] from the peak value attained and tended to be higher than baseline ( P = 0.08). A reduction in superfusate flow from 25 to 3.2 µl/s led to a fall in principal (239.2 ± 41.5 to 149.2 ± 37.2 nM; P (199.8 ± 13.2 to 127.4 ± 29.5 nM; P = 0.05) cell [Ca 2  ] i from peak values in three out of the four tubules studied; of note was that the flow-induced increase in principal [Ca 2  ] i in the fourth CCD was modest (35 nM) compared with the increase of 135.2 ± 35.8 nM detected in the other three CCDs. The recovery [Ca 2  ] i indicated above was not significantly different from those measured during the initial superfusion period at 3.2 µl/s.
6 }' R4 T* g4 e$ G, L' O) j. ~0 v+ N) p4 }
Effect of Circumferential Stretch and Flow Rise Time on [Ca 2  ] i in Occluded CCDs
$ s9 K/ l2 z0 n! @
' }& R+ {" d) }1 rIn the occluded CCD, a rapid increase in luminal volume, sufficient to increase tubular diameter by 19.5 ± 2.4% ( n = 5) within 0.6 s, led to an increase in [Ca 2  ] i in both principal (83.9 ± 9.5 to 364.9 ± 16.9 nM; P (84.2 ± 18.8 to 370.7 ± 79.1 nM; P cells (Figs. 6 C and 7 ). Peak [Ca 2  ] i was reached within 7.6 ± 0.8 s in both cell types ( Fig. 8 ). The magnitude of the change in [Ca 2  ] i in principal (281.0 ± 19.0 nM) and intercalated (286.6 ± 66.9 nM) cells in occluded CCDs exceeded that measured in split-open tubules (110.0 ± 35.7 and 99.8 ± 12.1 nM, respectively; P Similarly, a rapid increase in luminal volume, sufficient to increase tubular diameter by 5.4 ± 0.9% ( n = 3), led to a significant increase in [Ca 2  ] i in both principal (133.3 ± 5.3 to 233.0 ± 37.5 nM; P (155.0 ± 8.8 to 264.4 ± 17.3 nM; P cells. However, a slow increase (over 3-5 min) in tubular volume, to expand tubular diameter by 20%, failed to elicit a [Ca 2  ] i transient (data not shown; n = 3).2 d, ~+ }- A5 T1 W
) F0 ?# z! I! R/ M" }7 k
Source of Ca 2  Leading to Flow-Induced Rise in [Ca 2  ] i
" r" k6 _4 ~- A& F0 F3 h  G- x+ `, t
We considered it likely that the flow-induced increase in [Ca 2  ] i in the perfused CCD was the result of release of internal Ca 2  stores and/or external Ca 2  influx. These possibilities were tested using an array of inhibitors. Pretreatment of CCDs with thapsigargin, in the presence of external Ca 2 , led to a modest increase in [Ca 2  ] i as internal stores were emptied, followed by a slow decay ( Fig. 9 A ). A subsequent increase in tubular fluid flow/stretch failed to elicit an increase in [Ca 2  ] i in principal (183.7 ± 26.2 to 207.5 ± 21.2 nM; P = NS) and intercalated (196.0 ± 36.4 to 208.8 ± 30.7; P = NS) cells in seven CCDs. Similarly, exposure of tubules ( n = 4) to 2-APB, an IP 3 receptor antagonist that had no effect on baseline [Ca 2  ] i (97.1 ± 22.5 nM in principal cells and 99.1 ± 18.4 nM in intercalated cells), completely abolished the [Ca 2  ] i response to an increase in flow rate/stretch ( Fig. 9 B ). These data indicate that the flow-induced increase in [Ca 2  ] i requires release of IP 3 -sensitive internal Ca 2  stores.
( W" v1 n. \  y# c5 @- [4 i
2 F# k$ K+ z6 O% j; u/ |, }5 b4 lFig. 9. Representative tracings of flow and/or stretch-induced changes in [Ca 2  ] i in principal (black) and intercalated (gray) cells in CCDs pretreated with inhibitors of internal Ca 2  mobilization or in the absence of extracellular Ca 2 . Inhibition of internal Ca 2  release, either by pretreatment with thapsigargin ( A ) or the inositol 1,4,5,-trisphosphate (IP 3 ) receptor antagonist 2-aminoethoxydiphenyl borate (2-APB; B ), blocked the expected response to an increase in luminal flow/stretch. C : in CCDs perfused with a nominally Ca 2  -free perfusate, an increase in flow/stretch led to an increase in [Ca 2  ] i that rapidly returned to baseline. D : no response was detected in CCDs perfused and briefly bathed ( 2  -free perfusate. E : basolateral apyrase failed to inhibit the rapid, high-amplitude flow and/or stretch-stimulated increase in [Ca 2  ] i in CCDs perfused in the nominal absence of Ca 2 ." }& @' J* t, V. ~& I. E5 O9 M

5 R! i& B. a1 w" r) S  {0 VTo discern whether external Ca 2  entry participates in the response, CCDs were perfused with a Ca 2  -free perfusate (with or without 1 mM EGTA), and the response to flow/stretch was monitored. As shown in Fig. 9 C, in the absence of luminal Ca 2  and presumably Ca 2  influx across the apical membrane, an increase in flow/stretch led to a rapid rise in [Ca 2  ] i in principal (113.9 ± 10.2 to 398.7 ± 53.6 nM; P (138.3 ± 8.7 to 408.3 ± 63.9 nM; P = 0.005) cells in six CCDs, a response that could reflect either release of Ca 2  from internal stores or Ca 2  influx across the basolateral membrane. Unlike our observation of sustained plateau elevations of [Ca 2  ] i for at least 20 min in CCDs perfused at high flow rates with standard Ca 2  -containing perfusate (Figs. 6 and 7 ), [Ca 2  ] i in CCDs perfused in the nominal absence of luminal Ca 2  fell to baseline levels (108.6 ± 18.0 and 113.3 ± 17.2 nM in principal and intercalated cells, respectively; P = NS for each value compared with paired baseline [Ca 2  ] i ) by 10 min in the continued presence of fast flow. These data suggest that the flow-induced plateau elevation in [Ca 2  ] i in perfused CCDs is mediated by luminal Ca 2  entry. In the nominal absence of luminal ( x 10 min) and basolateral ( x 3 min) Ca 2 , no [Ca 2  ] i increase was detected in CCDs ( n = 4) subject to an increase in luminal flow rate ( Fig. 9 D ). The latter observation provides compelling evidence for coupling between extracellular Ca 2  entry at the basolateral membrane and internal Ca 2  release. Of note are reports by others that interruption of store-operated channels (SOCs), either by genetic disruption ( 21 ) or by pharmacological inhibition, significantly attenuates IP 3 -mediated ER Ca 2  release. Apyrase added selectively to the bathing solution ( n = 4) failed to inhibit the rapid high-amplitude flow and/or stretch-stimulated increase in [Ca 2  ] i in either principal ( [Ca 2  ] i = 190 ± 65.9 nM from a baseline of 127.7 ± 16.7 nM within 7.8 ± 1.2 s) or intercalated ( [Ca 2  ] i = 193.5 ± 51.9 nM from a baseline of 115.3 ± 11.3 nM within 8.8 ± 1.3 s) cells in CCDs perfused in the nominal absence of Ca 2  ( Fig. 9 E ); in these tubules, [Ca 2  ] i returned to baseline within 10 min (i.e., no sustained plateau elevation of [Ca 2  ] i was detected). These results suggest that 5'-nucleotides, acting as autocrine/paracrine mediators, potentially released at the basolateral membrane in response to epithelial stretch, do not mediate the flow-induced high-amplitude [Ca 2  ] i response.
, d; Y) m! L, B6 h; ]8 f( n
/ K& B' V. b/ E, y/ P3 A# }8 OEffect of Flow on Gap Junctional Intercellular Communication
, a* P8 P2 Z6 _% h- D7 w+ f9 j7 N5 Z4 S' F  ]5 L
Intercalated cells do not possess apical cilia, yet they respond to an increase in flow/stretch with an increase in [Ca 2  ] i. This could be explained if principal and intercalated cells are directly coupled and/or gap junctions are opened in response to flow. To examine these possibilities, individual cells in split-open CCDs were microinjected with the cell-impermeable fluorescent dye Lucifer Yellow. Superfusion for 5 min did not lead to intercellular spread or significant leakage of the dye (data not shown). Exposure of split-open CCDs to 0.4% trypan blue at the conclusion of the experiment ( n = 3) revealed occasional trypan blue-positive cells along the edges of the split tubule; microinjected cells, however, excluded the viability marker.
- K& S$ D0 Y4 d4 q0 H
3 n* z* o. o% `Forces and Torques on Cilia in Split-Open and Perfused CCDs! {0 V% k( ^' K( d
8 S% K# u$ ^9 n2 f: f' V( I! |$ b
Table 1 summarizes the major predictions of our theoretical models for the split-open and perfused CCDs and presents results that would have been obtained for a confluent monolayer of MDCK cells if they were exposed to shear in our fluid flow chamber. Although the results cannot be compared quantitatively with the experiments reported by Praetorius and Spring ( 26 ), since the geometry of their flow chamber differed substantially from our own, instructive qualitative comparisons can be made. The primary difference from a hydrodynamic viewpoint between the cells analyzed in this study and cultured MDCK cells is the large difference in length of the cilia (2.5 vs. 8 µm, respectively).
0 \' [# u$ I- n7 n, C
8 B2 M& A7 R( AIn split-open CCDs, an increase in superfusate flow from 0 to 3.2 µl/s failed to generate a [Ca 2  ] i response, whereas a further increase in flow rate to 25 µl/s led to a moderate response ( Fig. 6 B ). 2 The average velocity in our flow chamber at the high flow rate is about four times the maximum average velocity in the experiments reported by Praetorius and Spring ( 26 ), but, as indicated earlier, it is the local velocity at the tips of the cilia and not the average velocity of the flow in the chamber that determines the dynamic behavior of the cilia in response to flow.9 Z# ^- {( l  G

% u' s4 W& a, o+ s5 _+ XThe first important observation in Table 1 (also see Figs. 4 and 5 ) is that the tip velocity at the high flow rate for a cilium of 2.5-µm height and 0.2-µm diameter is 17.9 µm/s, approximately one-sixth of that anticipated in an intact tubule of 25-µm internal diameter ( 42 ) perfused at a rate of 5 nl/min. Both the wall shear stress on the apical membrane and the total drag on the cilia at the high flow rate are nearly one order of magnitude smaller than predicted for the perfused tubule, as is also true for the total torque ( Table 1 ). The predictions for the 8-µm cilium provide new insight into the experiments with MDCK cells reported by Praetorius and Spring ( 26 ). Cilium tip velocities do not scale linearly with the length of the cilia because of the increasing hydrodynamic interactions between cilia as their length increases (Figs. 4 and 5 ). Thus the tip velocity on an 8-µm cilium is only about two times that on 2.5-µm cilium ( Table 1 ). The total drag force, in that it is proportional to the product of tip velocity and cilia length, would scale as the square of the cilium length were it not for this cilium-cilium interaction. For the split-open tubule, the drag on the 8-µm cilium is less than six times greater than that on the 2.5-µm cilium at the flow rate of 25 µl/s ( Table 1 ). Similarly, the torque, as a product of drag and moment arm, would scale as the cube of the tip length without the ciliumcilium interaction. Instead, the torque experienced by the 8-µm cilia is 22-fold greater than that predicted for the 2.5-µm cilium ( Table 1 ). The torque on an 8-µm cilium at the low flow rate of 3.2 µl/s (0.37 pN · µm; results not shown in Table 1 ) significantly exceeds that on the 2.5-µm cilium at the high flow rate of 25 µl/s (0.13 pN · µm).: c/ S8 F) N0 O' p1 D: n- c9 p# H

; k/ H3 p$ G- iAnother major result is the prediction of the relative magnitude of the total shear force on the apical membrane of the principal cells compared with the drag force on the primary cilium. The total drag resulting from fluid shear on the plasma membrane in our high-flow experiment, assuming hexagonal cells whose surface area is 111 µm 2, is 0.62 pN. In contrast, the drag on a 2.5-µm cilium itself at a flow of 25 µl/s is only 0.078 pN, or nearly one order of magnitude smaller than the drag resulting from fluid shear. The predicted tip deflection ( Table 1 ) resulting from this drag, while small for the 2.5-µm cilium (only 1.9 nm or 0.08% of cilium length), increases to 227 nm for the 8-µm cilium at the same flow rate. The predicted shapes of the deformed cilia for the split-open tubule for a flow of 25 µl/s are shown in Fig. 10. The flexural rigidity of the cilia used in these calculations is 1.4 x 10 -23 N · m 2 ( 34 ). Unfortunately, the calculation for the tip deflection could not be performed for an 8-µm cilium in a perfused tubule since the predicted tip deflection exceeded the limit of validity of the small deflection theory applied herein. For these large deflections, an "elastica" model would need to be employed similar to that used by Schwartz et al. ( 34 ). For comparison, the tip deflection of a 2.5-µm cilium in a perfused tubule subject to a flow of 5 nl/min is 13 nm or 0.52% of the cilium length. As a rough guide, the tip deflection increases as the fourth power of the cilia length and the deflection of an 8-µm cilium is 100 fold greater than the 2.5-µm cilium just described.9 ?! A# }1 w- l  p, O( p
2 |& n) R- S, h1 U4 O; V/ N
Fig. 10. Model predictions for cilia deflections in a split-open tubule with a chamber flow rate of 25 µl/s. Note that the tip deflection increases approximately as the fourth power of the cilia length. y = deflection distance.
2 x7 t1 {( W# g
3 j% ]$ K0 J# _DISCUSSION
$ e% u# D& A" B4 R
8 Y3 q% r# ~5 J8 ^" Z  ]- fUrine flow has been proposed to be pulsatile in the collecting duct, leading to cyclical expansion and collapse of the tubules ( 33 ). Cumulative evidence from studies in other tissues possessing a lumen, such as microvessels ( 6, 38 ), predicts that increases in luminal flow rate within the CCD should alter the mechanical forces, including fluid shear stress, hydrostatic pressure, and membrane stretch and deformation, to which principal and intercalated cells comprising the epithelium are exposed. The purpose of the present study was to begin to explore the response of CCD epithelial cells to hydrodynamic forces. Specifically, we sought to examine the effects of luminal flow on [Ca 2  ] i in microperfused CCDs studied in their native geometry (subject to flow/circumferential stretch) and split-open tubules (subject to apical shear).0 g9 x- C6 i5 Z' u' I
5 y" y! I, ]( ^  @
Microperfused CCDs/ Q9 F" V- i8 y2 n+ C# M- y
* X, f' a' l9 @) S. O
Increases in tubular fluid flow rate in the CCD, isolated from its surrounding interstitium and microperfused in vitro, were associated with circumferential stretch and increases of up to 20% in tubular diameter. Note that, although circumferential stretch of this magnitude would not likely occur in the kidney with an intact capsule under conditions of health, comparable increases in collecting duct diameter have been reported after ureteric obstruction ( 16, 43 ). Both principal and intercalated cells within the CCD responded to increases in flow/circumferential stretch ( 5% increase in tubular diameter) with an increase in [Ca 2  ] i (Figs. 6 A and 7 ).
! h* c4 f3 F3 y: R+ ~  Y( q5 o4 o* [) ]' z
The source of the Ca 2  giving rise to the flow and/or stretch-induced rise of [Ca 2  ] i includes release from internal stores and influx from the extracellular space. The complete inhibition of the response by thapsigargin and 2-APB ( Fig. 9 ) points to a requirement of IP 3 -mediated release of ER Ca 2  stores in the response. Typically, depletion of Ca 2  from internal stores triggers the capacitative influx of extracellular Ca 2  across the plasma membrane through SOC or Ca 2  release-activated Ca 2  currents ( 2, 23 ). Praetorius and Spring ( 26 ) proposed that the Ca 2  signal generated by bending of the cilium in MDCK cells was the result of Ca 2  influx through mechanosensitive channels residing in the cilium or its base. We thus considered it likely that the flow and/or stretch-induced Ca 2  influx in the CCD studied in its native geometry was also localized to the apical membrane. Indeed, our finding that removal of Ca 2  from the luminal perfusate led to a fall in [Ca 2  ] i to baseline levels within 10 min in the continued presence of high flow rate, an observation that contrasted with the sustained 20 min; Figs. 6 and 7 ) elevation in [Ca 2  ] i in CCDs perfused with our standard Ca 2  -containing perfusate, suggests that the flow-induced plateau elevation in [Ca 2  ] i in perfused CCDs is mediated by luminal Ca 2  entry in the epithelial cells. The complete inhibition of the flow and/or stretch-induced increase in [Ca 2  ] i in tubules perfused and briefly bathed in the absence of Ca 2  suggests that basolateral Ca 2  entry pathways exist in the CCD and may be coupled to internal Ca 2  release.. o$ u6 H# [3 s+ \! Q% @+ l
; D; a1 P2 w1 A  X- v' k
Compatible with a basolateral Ca 2  entry pathway linked to release of internal Ca 2  stores are recent reports of cell-specific and polarized expression of IP 3 receptor isoforms in the kidney in patterns that suggest compartmentalization of distinct IP 3 -sensitive Ca 2  pools ( 20 ). Molecular cloning studies have demonstrated three types of IP 3 receptor (types 1-3) derived from different genes ( 42 ). Type 2 is expressed exclusively in collecting duct intercalated cells in a diffuse cytoplasmic distribution, whereas principal cells express the type 3 receptor mainly in the basolateral portion of the cytoplasm. Of note are recent reports that the IP 3 receptor in the ER and SOCs in the plasma membrane may directly interact ( 2, 19 ).
/ O( P# L$ C/ j4 ~+ B! t9 k5 n: x) ^2 j) I0 P4 v2 |
Split-Open CCD
* G) X) X. J1 P4 C! F
* S. E, `% ]2 ]" JIn split-open CCDs, an increase in superfusate flow from 0 to 3.2 µl/s failed to elicit a response. However, an increase in flow rate to 25 µl/s led to a modest sustained increase in [Ca 2  ] i with an 30-s time delay between stimulus and response (Figs. 6 B and 7 ). Note that, even at the high flow rate, the forces and torques experienced by the cilium are approximately sixfold lower than those predicted in the microperfused CCD ( Table 1 ). Thus the flow-induced approximately twofold increase in [Ca 2  ] i in principal and intercalated cells in split-open CCDs may be below the potential maximal response, which, for the MDCK cells used by Praetorius and Spring ( 26 ), was also a twofold increase. Although the highest average velocity in the flow chamber in the latter study was less than in the present experiments, Table 1 shows that the torque on an 8-µm cilium in MDCK cells is 19 times that on a 2.5-µm cilium for the same flow conditions. This observation suggests that, if the torque is the triggering mechanism for the response, a saturation behavior could have been achieved for the 8-µm cilium studied by Praetorius and Spring ( 26 ), even for a lower tip velocity. A different type of flow chamber is needed for future studies where local velocities, drag forces, and torques can be increased by an order of magnitude to quantify the maximum response. To test for the maximum cilium-generated response in split-open CCDs, the bending response of micropipette suction should be compared with that resulting from fluid flow.( }: q' l0 c9 @
! k+ D5 M) y+ ~( a! F
Our theoretical model allowed us to predict the relative magnitude of the total shear force on the apical membrane of the principal cells compared with the drag force on the primary cilium. The response to fluid shear of endothelial cells has been studied extensively. A [Ca 2  ] i response has been observed for fluid shear stresses as low as 1 dyn/cm 2 ( 6 ). An important question that arises from the data in the CCD is whether it is the high total drag resulting from fluid shear on the plasma membrane or the lower drag on the cilium itself that is the primary mechanism involved in the flow-induced [Ca 2  ] i response for the split-open CCD. The observation that the fluid shear stress on the apical membrane at the high flow rate is only 0.056 dyn/cm 2, a value substantially below the shear stress threshold for a [Ca 2  ] i response in endothelial cells, and the surface area of an endothelial cell is at least three times larger than a CCD principal cell suggests that it is not the drag force or shear stress per se that is the activating mechanical stimulus but either the torque on the anchoring microtubules and actin filaments at the base of the cilium where it attaches to the cytoskeleton or, alternatively, the bending deformation of the cilium and the resulting opening of stretch-activated Ca 2  channels.  z& ]0 X7 I! _9 C6 C2 U9 b
  X8 _* W8 ?! m! z7 q( B6 u8 Y
Of note are recent studies that show that polycystin-1 ( PC1 ) and -2 ( PC2 ), two genes implicated in autosomal dominant polycystic kidney disease, participate in fluid-flow sensation by the primary cilium in renal epithelial cells. Cells isolated from embryonic transgenic mice lacking functional PC1, and grown as monolayers on coverslips, formed cilia but did not exhibit a superfusate flow-induced Ca 2  influx as did control cells ( 22 ). Blocking antibodies directed against PC2, a Ca 2  -permeable cation channel ( 9 ) that interacts with PC1 ( 13 ), also abolished the flow response in wild-type cells as did inhibitors of the ryanodine receptor ( 22 ). Based on these data, Nauli et al. ( 22 ) proposed that conformational changes of ciliary PC1 transduce a mechanical signal in a chemical response by activating associated PC2 Ca 2  channels; the local Ca 2  influx in the cilium subsequently triggers internal Ca 2  release. This paradigm is similar to that proposed by Praetorius and Spring ( 26 ), i.e., an initial Ca 2  influx in the primary cilium is important for the flow-induced Ca 2  response. In contrast, the data in the present study suggest that the flow-induced Ca 2  increase is generated by the cilium transmitting a mechanical force to the cytoskeleton, which is associated, at least in part, with basolateral Ca 2  entry that appears to be associated with Ca 2  -induced Ca 2  release from the intracellular IP 3 -dependent Ca 2  stores. This hypothesis is based on a model for cilium torque. The discrepancy between the findings in perfused tubules and cell monolayers ( 22, 26 ) raises the possibility of two distinct mechanisms underlying the Ca 2  response observed in the CCD studied in its native geometry: flow and stretch activation.; R* U9 ^" Z0 m; B0 Q1 m6 ?' d
- B" b) f% P4 G/ t. ]! ]" m: ?$ F
The differences (temporal delay, amplitude of response) in [Ca 2  ] i response noted between the split-open CCD and the microperfused tubules also suggest that different Ca 2  signaling pathways mediate the two responses. Praetorius and Spring ( 26 ) proposed that the Ca 2  signal generated by bending of the cilium in MDCK cells was the result of Ca 2  influx through mechanosensitive channels residing in the cilium or its base. They demonstrated that the [Ca 2  ] i response resulting from the bending of the cilium by either flow or micropipette suction was abolished in confluent MDCK monolayers when extracellular Ca 2  was removed. However, under the latter conditions, the response to direct mechanical stimulation of the apical membrane was retained.2 E% o* f) N$ \. V" B) {0 g
& E# ^3 B* f. Z, k& [) P" r
Because our split-open tubule preparation does not allow for the selective manipulation of apical and basolateral superfusate composition, we could not examine the effects of apical vs. basolateral Ca 2  -free solution or Ca 2  channel inhibitors to discern whether the flow response in the native CCD, like that reported in MDCK cell monolayers ( 26 ), is absolutely dependent on external Ca 2  through stretch-activated Ca 2  channels in the cilium. However, the very small tip deflections for a 2.5-µm cilium (1.9 nm) elicited by a flow rate sufficient to generate a Ca 2  response, as predicted by our model in Table 1, would not be expected to be a sufficient stimulus to open mechanosensitive Ca 2  channels in the plasmalemma of the cilium. This suggests that the primary site of activation may be at the base of the cilium because of the transmission of the torque on the cilium to the actin cytoskeleton in the terminal web.
. W+ j/ k7 ~* I3 {" W' S* ?
8 ^0 z) k1 j$ p# QWeinbaum et al. ( 41 ) have reported that there is a large amplification of the force on the axial structural elements for brush-border microvilli, because of the resisting moment at the base. In the case of the microvilli, this resisting moment led to a 38-fold increase in the force on the axial actin filaments, since the moment arm at the base is small compared with the length of the microvilli. The strong wind blowing across a tall tree will seldom split its trunk but instead topple the tree at its roots. The 9   0 axial structure can be simplified to consider eight symmetrical microtubule pairs at the base of the axoneme part of the cilium where it enters the transition zone at the top of the basal body ( Fig. 11 ). If the bending axis is in the center plane of the cross section passing through two of the microtubule pairs 1 and 5, the other six must provide the resisting moment applied by the torque on the protruding cilium. Assuming that the microtubule pairs lie on a circle of 200-nm diameter, it is relatively straightforward to compute the tensile and compressive forces on microtubule pairs 6, 7, and 8 and 2, 3, and 4, respectively. The symmetry of the arrangement predicts that the tensile and compressive forces are equal and opposite. The compressive forces on individual microtubule pairs are listed in Table 1. For a 2.5-µm cilium, the compressive force on the outermost microtubule pair 3 (0.86 pN) is 11-fold greater than the drag force on the cilium. For an 8-µm cilium, this compressive force increases to 16 pN at the same flow rate, whereas, for a 2.5-µm cilium in a perfused tubule, this force is 6 pN for a flow rate of 5 nl/min. Thus the forces exerted on the supporting structures in the basal body greatly exceed the typical 0.1-pN forces measured for the deformation of the cortical actin filament network beneath the plasma membrane ( 25 ) when membrane proteins are dragged in the plane of the membrane by optical traps and their cytoplasmic tails collide with microdomain boundaries. This suggests that the basal body must be anchored by more rigid supporting structures such as the microtubule complex shown in Fig. 11.
1 W  ]3 ]. K8 |2 H% X0 Y. g  T1 g/ f1 g$ U: b' g6 X
Fig. 11. A : simplified mathematical model of microtubule distribution inside a primary cilium. Microtubules, which terminate in the transition zone at the top of the basal body, transfer the torque on the axoneme to the centriole in the basal body. Because of the stiffness of the microtubules in the primary cilium and its supporting structure, a 2.5-µm cilium bends only slightly because of the drag forces exerted by the fluid flow. U, average velocity. B : cross section of the cilium. The "9 0" microtubule structure of primary cilium is reduced to 8 symmetrically arranged microtubule doublets. The model is used to predict the tensile and compressive forces at the base of each microtubule pair when a torque is applied about an axis passing through microtubules 1 and 5.; y7 l) P( c, t- F

! ]" K6 r( t, ^6 n5 @Occluded Tubules
8 K! p, b" P3 P0 p  N" j
, G. D7 m4 e9 f. I0 Z) k, X6 BAnalysis of the results generated in occluded tubules provides valuable insight into the nature of the mechanical signal that triggers the flow-induced [Ca 2  ] i response. Measurements of [Ca 2  ] i were obtained at a distance approximating two tubular diameters proximal to the occlusion site. Thus the average velocity for the flow transient in the tubule elicited during the "filling" period, u av = (2 L / R )d R /d t, can be approximated by 8 d R /d t. For the rapid increase in luminal volume leading to a 20% increase in tubular diameter, d t was at most 600 ms (i.e., the increase in diameter was achieved between two image acquisition frames using our digital camera), d R was 7 µm, and thus u av was at least 93 µm/s. This value is fivefold greater than the 17.9-µm/s tip velocity predicted for the split-open tubule at the high flow rate ( Table 1 ). If the [Ca 2  ] i response in the occluded tubule is triggered by a shortduration deformation of the cytoskeleton in the terminal web in which the cilium is anchored, then it is no surprise that this increase in [Ca 2  ] i actually exceeds the response elicited by the high flow rate in the split-open tubule. The notion that it is not the magnitude of epithelial stretch but the initial impulse loading resulting from the flow transient is supported by the important observation that, when a 20% stretch is achieved slowly over 3 min, where u av is 2  ] i response was observed. Finally, a rapid small 5% circumferential stretch produced a peak flow velocity of 23 µm/s ( 25% of the 93 µm/s response for the 20% increase in diameter), a rate comparable to the 17.9-µm/s tip velocity in the split-open tubule subject to high flow ( Table 1 ). Of note is that the magnitude of the increase in [Ca 2  ] i in the latter series of tubules was similar to that detected in the split-open tubule superfused at high flow. In summary, these observations tentatively suggest that circumferential stretch may not be the triggering mechanism for the [Ca 2  ] i response but that the magnitude of the tip velocity and hence the torque on the cilium is the important mechanical signal. To confirm this conjecture, one needs to rule out rate-dependent stretch, not just stretch magnitude, and establish a threshold for fluid shear.
% q/ p# _% A" y& r9 A! c) |$ w0 q  i* P  Q
Intercalated Cell Response
& T- q# h) p9 Q1 r  ^' @6 f
7 b- i% T. L8 b4 _* r  Q% VOur detection of comparable flow-stimulated increases in [Ca 2  ] i in principal and intercalated cells, the latter devoid of apical cilia, indicates that the cilium is not the sole mechanotransducer of the flow response in the CCD. Given that principal and intercalated cells appear not to be coupled in the native epithelium ( 44 ), we speculate that flow over the microvilli and microplicae that decorate the apical surfaces of intercalated cells ( 8 ) generates a bending moment equivalent to that created by flow-induced bending of the apical cilium of principal cells. The flow-induced deformations of individual microvilli in the brush border of the proximal tubule ( 12 ) are predicted to be substantially smaller than the flow-induced bending response of the primary cilium ( 26, 34 ). Although the bending moment on each microvillus may be very small ( 0.01 pN · µm at a flow rate of 30 nl/min), these microvilli are quite numerous (4,000/cell) and collectively generate a combined torque (4 pN · µm; see Ref. 12 ) that exceeds that predicted for a central cilium (0.91 pN · µm) in a perfused CCD. A similar model must be developed for the microvilli/microplicae at the surface of the intercalated cell to provide a theoretical basis for understanding their response to flow./ g' f* c5 t; u$ U! I
% Z: c' M4 @1 d
Alternatively, the response to flow/stretch in principal and intercalated cells may be mediated by paracrine/autocrine signaling secondary to stretch-induced release of an extracellular factor. It is well documented that extracellular nucleotides, including ATP and UTP, can induce a variety of cell responses, including increases in [Ca 2  ] i. Burnstock ( 4 ) proposed that deformation of cells in distended epithelial "tubes" leads to release of ATP by the epithelium. In fact, mechanical stress has been shown to lead to release of ATP and UTP across both apical and basolateral membranes in polarized airway epithelia ( 14 ). ATP functions as an extracellular signaling molecule through activation of members of the P 2X and P 2Y receptor families. P 2X receptors are Ca 2  -permeable, nonselective cation channels identified, at the mRNA level, in both principal and intercalated cells ( 36 ). Binding of ATP to G protein-coupled P 2 purinergic receptors activates phospholipase C, leading to hydrolysis of phosphatidylinositol 4,5-bisphosphate to IP 3 and release of internal Ca 2  stores ( 7, 27 ). We have recently identified functional P 2Y2 but not P 2X receptors on the apical surfaces of both principal and intercalated cells of the CCD ( 44 ). Furthermore, we reported that [Ca 2  ] i transients induced by an acute increase in tubular fluid flow in the CCD were not mediated by apical P 2 purinergic receptor signaling ( 44 ). Our present observation that basolateral apyrase fails to prevent the flow and/or stretch-induced high-amplitude increase in [Ca 2  ] i suggests that 5'-purine and -pyrimidine nucleotides, possibly released at the basolateral membrane in response to epithelial stretch, do not mediate this response.! B8 m' C, U3 Z, d5 b

$ N6 X# _" U3 b  {6 q* FPhysiological Significance# p; a/ M  I6 i# f, b
- ?' S6 \9 s# x* v! j
Nearly all epithelial and endothelial cells that are subjected to mechanical strains that exceed a few percent will exhibit [Ca 2  ] i transients. Thus the flow and/or stretch-induced increase in [Ca 2  ] i observed in the microperfused or occluded tubule is not unexpected. The central question is whether the force experienced by the isolated CCD (flow/circumferential stretch) or split-open tubule (no circumferential stretch) is more representative of the physiological condition. As stated above, it is unlikely 15% under physiological conditions, since this would cause a substantial change in volume of the kidney and produce large compressive stresses on the tissue opposing expansion. We believe the more reasonable hypothesis to be that individual tubules in vivo experience only minimal circumferential stretch in response to increases in urinary flow rate and that the primary cilium in principal cells and microvilli/microplicae in intercalated cells serve as flow sensors. If this is indeed the case, the response observed in the split-open tubule is the more relevant, and the models developed herein for the forces and torques on these structures and the consequent deformation of the cytoskeleton provide the key to understanding the mechanical behavior of the afferent sensor.
6 Q8 `; B  J8 T5 t# m, V' j+ m( f2 q! {) v2 W5 f
APPENDIX I
; C; `0 ]8 B3 W
) Q6 K0 s: q8 N4 Q  DThis appendix describes the solution to Eq. 2 in MATERIALS AND METHODS for the velocity profile U ( z ) in the cilia layer in the split-open tubule and also provides the expression for the drag ( D ) and torque ( T ) on each primary cilium. Equation 2 is solved subject to boundary conditions
1 `) L; s2 g3 ]! |7 }6 y! y
9 i0 @  _& S* u& c7 [* p(A1a)/ [: E3 n" N0 T, z% ~

' y' |- A/ u$ S6 B7 H(A1b)
3 A: n/ B8 x( n4 C# H
; v4 N! Q; N, Pwhere h is the tip height of the cilia and ( h ) is the shear stress at the top of the cilia layer, defined by: q8 r8 X$ t7 W# K% Z; h1 b

# C! I! H( Y- B+ B: T' X(A1c)
( |) b6 S, x2 f+ R/ z# ~9 j( k# L0 G$ h
Note that w (0) is given by Eq. 1, since this is the shear stress near the wall imposed by the outer flow in the channel.: M. S$ V& O6 [, Y% V' w
* H  n8 |1 [2 @0 S) t
The general solution of Eq. 2 is the sum of particular and homogeneous solutions
2 }. `: ]& N0 M5 N* `! Y) G/ m$ j( d1 @
(A2)
* ^: d5 R7 j# g6 D0 v% \( l0 R1 @' x+ u3 q% n( J6 n  p
where A and B are unknown constants.
% w1 z4 [1 Q8 `' ~. \1 X3 ?# L
3 v+ a; @* ]/ r" O* NThe solution for U ( z ) that satisfies Eqs. A1a and A1b is
: H) F" b- p# F: H, c& k# |% L+ J9 \+ S: R/ M
(A3)
! E1 |4 u# i  s4 e3 c2 G/ B- w8 C8 }" A5 z
The local drag force (drag force/unit fiber length) on each cilium, F ( z ), is related to the local velocity distribution U ( z ) and Darcy permeability coefficient k by
' R0 o2 v/ R0 h, K! c( X+ _; \( j' v% K" j; ^% v
(A4)
. a3 H# L/ L$ H2 ?! J# M' p3 L  {! {) @* _8 [/ r  p
Here s, the area of each periodic unit shown in Fig. 12, is given by
) i, Q. T* M5 N
. R, M. ]7 q* J7 w5 s$ Ewhere l is the side length of each hexagonal cell.
: A. s1 o3 i+ ~. z4 V! O
1 M- B; n% Y$ ]  m1 _4 v5 U0 f6 O( oFig. 12. A : idealized model of the apical surface of a number of principal cells with cilia located at the center of each hexagonal cell. l, Length. B : repetitive periodic unit in the cilia array used to calculate the force and torque on the central cilium. Note that this unit contains the equivalent of one cilium, and the area equals the cell surface area formed by the region 2A   B in A.2 P! o0 a& @1 f$ t, r/ ^7 M) w
/ h$ }1 ?8 @- D. d
Sangani and Acrivos ( 28 ) have obtained a numerical solution for the Stokes flow past a periodic fiber array shown in Fig. 12. These authors showed that the dimensionless drag, F /µ U, can be approximated by
2 w3 ^! t. a% X$ [- s- f
  C, P! k: q, s; t2 R- m- }; ^6 {(A5)
3 Q/ l2 g1 {) p3 o0 z. G6 O" @9 i/ Y9 X: o$ H* e1 m
Here c, the solid fraction is defined by
5 J4 v8 ^1 U( s$ P/ S
2 k6 R* w0 B$ K7 k(A6)
0 n, j9 O3 X5 K& ?) ^" n$ _9 r' |) n, Y  f4 L2 H0 }
where
( N4 b8 O% s9 R- y. U          【参考文献】* t, C$ p$ D; q9 M. U3 G' K# e( U
Andrews PM and Porter KR. A scanning electron microscopic study of the nephron. Am J Anat 140: 81-115, 1974.
! v; i9 G9 y& G2 y/ x# k6 V$ U/ G1 o3 D( M0 D; a0 O

: R9 [& o1 h$ q# g- O
: o* B% e9 T" P9 }4 n* fBerridge MJ. Capacitative calcium entry. Biochem J 312: 1-11, 1995.9 i$ B) `+ M/ A: s  f9 L

2 z* L( [+ g$ d1 N( h# a
* _$ G- }8 l7 B. O  w# M1 c& R4 }8 d, H( C
Bird R, Stewart W, and Lightfoot E. Transport Phenomena. New York: Wiley, 1960.
' h  f. i. c; |/ }7 A3 |6 Z
1 z) J8 A; ~. U  m
5 Z5 \- [; ]: V4 v
, w6 S% \# [. t) F1 BBurnstock G. Release of vasoactive substances from endothelial cells by shear stress and purinergic mechanosensory transduction. J Anat 194: 335-342, 1999.
  o0 M! m' m1 h" {% }7 W+ o- a
8 n  V+ H& G6 \2 t7 [4 ^! J
' V2 k& L" @5 X) A6 }' S( d3 ^0 _$ X3 E( R* C( P) c6 p
Constantinescu A, Silver RB, and Satlin LM. H-K-ATPase activity in PNA-binding intercalated cells of newborn rabbit cortical collecting duct. Am J Physiol Renal Physiol 272: F167-F177, 1997.
3 w6 I7 u0 S7 ?0 o& u% I4 e7 Y
9 [% _8 k8 i7 ?' `  t% k5 A5 s. }3 x0 j) b4 T/ ], i

* O6 m+ @9 Y1 r% lDavies PF. Flow-mediated endothelial mechanotransduction. Physiol Rev 75: 519-560, 1995.
0 T/ D, _/ w% U& ]; s* D. v, v5 }: z+ W2 P3 ?; U

9 @6 o. `9 e$ u: K! T8 h0 E% W4 B
, G- V7 v$ F2 O/ ~Dubyak GR and el-Moatassim C. Signal transduction via P2-purinergic receptors for extracellular ATP and other nucleotides. Am J Physiol Cell Physiol 265: C577-C606, 1993.
! q  I8 B. \) N& S6 F, r+ _+ j
; X; p1 M' u' I& J" s
. m! p0 E" e7 ]0 B8 V1 F
" b% Y  r* ^  Z$ z8 jEvan AP, Satlin LM, Gattone VH II, Connors B, and Schwartz GJ. Postnatal maturation of rabbit renal collecting duct. II. Morphological observations. Am J Physiol Renal Fluid Electrolyte Physiol 261: F91-F107, 1991.$ p! N4 `% k$ J4 ?
, P3 o$ W. k% I* ~9 c  o$ s
$ i' @" u; G0 q

+ K  v* s/ e. t7 v! y: ^Gonzalez-Perret S, Kim K, Ibarra C, Damiano AE, Zotta E, Batelli M, Harris PC, Reisin IL, Arnaout MA, and Cantiello HF. Polycystin-2, the protein mutated in autosomal dominant polycystic kidney disease (ADPKD), is a Ca 2  -permeable nonselective cation channel. Proc Natl Acad Sci USA 98: 1182-1187, 2001.
7 A9 R1 U9 x' C, l8 m4 }# b- p! c
# c# ~8 f6 e0 A
5 f: |  ?4 M/ V* V! B7 Q# M  c0 o
! B7 y5 ^8 ?, g1 D$ MGregory RB, Rychkov G, and Barritt GJ. Evidence that 2-aminoethyl diphenylborate is a novel inhibitor of store-operated Ca 2  channels in liver cells, and acts through a mechanism which does not involve inositol trisphosphate receptors. Biochem J 354: 285-290, 2001.* a/ g1 J+ F3 K8 X' l

2 p: p7 i0 Z0 M# `
8 X7 p. ]3 f; [. |  T8 c: \/ Z; T5 y+ n) w; \: @
Grynkiewicz G, Poenie M, and Tsien RY. A new generation of Ca 2  indicators with greatly improved fluorescence properties. J Biol Chem 260: 3440-3450, 1985.
0 K7 p5 V  O- Z8 \! q
7 a4 Z4 X, v  ~8 A$ D% }% q
* D% C5 x4 u$ N7 Y
3 ~' _8 P- z' q/ S4 gGuo P, Weinstein AM, and Weinbaum S. A hydrodynamic mechanosensory hypothesis for brush border microvilli. Am J Physiol Renal Physiol 279: F698-F712, 2000.
) R+ S2 D( U4 K2 k' R3 Q2 v( o7 [1 W# |$ u7 _. s& \. s- D/ V

2 i& j1 m, U. `# {3 S4 b$ v) V+ @9 ?/ w1 y3 {" z: J2 f
Hanaoka K, Qian F, Boletta A, Bhunia AK, Piontek K, Tsiokas L, Sukhatme VP, Guggino WB, and Germino GG. Co-assembly of polycystin-1 and -2 produces unique cation-permeable currents. Nature 408: 990-994, 2000.
) |3 M  E$ y* w3 c. p9 @. N5 {( q0 x' W
" t+ o5 f" Z: {' B. F

; \$ C% `6 n/ ^8 ]' m1 eHomolya L, Steinberg TH, and Boucher RC. Cell to cell communication in response to mechanical stress via bilateral release of ATP and UTP in polarized epithelia. J Cell Biol 150: 1349-1360, 2000.
, n' h' n. I" \$ P4 Q2 _1 G% E
* X( w2 K; t; M/ M8 o
3 U/ E8 _1 Z$ T$ A1 E1 X6 s. K3 h8 _0 p' P1 G0 f7 M9 @
Hou X, Mrug M, Yoder BK, Lefkowitz EJ, Kremmidiotis G, D'Eustachio P, Beier DR, and Guay-Woodford LM. Cystin, a novel cilia-associated protein, is disrupted in the cpk mouse model of polycystic kidney disease. J Clin Invest 109: 533-540, 2002.4 C# U) L* C* X& ?. G7 X
  c: i; _# E+ q3 z- M( N4 v
2 }7 x3 u' ~$ s5 k, e; b7 w
+ q  \/ B* N) [& ~
Kimura H and Mujais SK. Cortical collecting duct Na-K pump in obstructive nephropathy. Am J Physiol Renal Fluid Electrolyte Physiol 258: F1320-F1327, 1990.' \7 u% J: V3 C  B4 E
  _7 x3 r; h" `- U# ^) M! s

3 P# U" N* l- V7 O$ n6 ]3 B( f
6 }: R& c! @8 Q$ f1 K! z$ EKnepper MA, Verbalis JG, and Nielsen S. Role of aquaporins in water balance disorders. Curr Opin Nephrol Hypertens 6: 367-371, 1997.6 I4 i. N& {+ l! |3 q' D' Z$ Z

7 b0 G6 e8 R# Y9 w2 L
. i0 j1 i: F/ Z& ^0 u9 u1 B* [  J2 Z9 x5 W
Latta H, Maunsbach AB, and Madden SC. Cilia in different segments of the rat nephron. J Biophys Biochem Cytol 11: 248-252, 1961.0 H* g. d3 _6 {& ?2 {% _; E$ T6 k
" z/ y( n& ?5 S

# ~. F: o* n& j1 c  ~5 v/ \9 w3 V# R8 Y& K4 E: ^2 d% B; y
Ma HT, Patterson RL, van Rossum DB, Birnbaumer L, Mikoshiba K, and Gill DL. Requirement of the inositol trisphosphate receptor for activation of store-operated Ca 2  channels. Science 287: 1647-1651, 2000.
  b5 V+ F, p2 k( \6 O! Y
) [  F3 G9 y0 j( d; h( ?
) A: x  Z* r/ P2 z: Z' p! X2 Z1 z) _8 A
Monkawa T, Hayashi M, Miyawaki A, Sugiyama T, Yamamoto-Hino M, Hasegawa M, Furuichi T, Mikoshiba K, and Saruta T. Localization of inositol 1,4,5-trisphosphate receptors in the rat kidney. Kidney Int 53: 296-301, 1998.
7 W5 f; a9 j4 L1 n$ P. E  R% M3 P) ~. a

! B% l* l( d7 ]1 s! }) ~$ s1 v( P: u9 q/ |" ~9 l6 g! |+ F8 Q
Mori Y, Wakamori M, Miyakawa T, Hermosura M, Hara Y, Nishida M, Hirose K, Mizushima A, Kurosaki M, Mori E, Gotoh K, Okada T, Fleig A, Penner R, Iino M, and Kurosaki T. Transient receptor potential 1 regulates capacitative Ca 2  entry and Ca 2  release from endoplasmic reticulum in B lymphocytes. J Exp Med 195: 673-681, 2002.
) ]) u1 W/ i+ R3 w+ {% t5 @1 \8 H# P
+ N  X3 U0 J/ n' a! a

1 h, i( L4 ?( ^4 V4 `! f" o4 RNauli SM, Alenghat FJ, Luo Y, Williams E, Vassilev P, Li X, Elia AE, Lu W, Brown EM, Quinn SJ, Ingber DE, and Zhou J. Polycystins 1 and 2 mediate mechanosensation in the primary cilium of kidney cells. Nat Genet 33: 129-137, 2003.
9 ~; ^) L; i- m$ L) K& _5 R% K
8 y+ b8 e! D) h3 W2 F
0 O% j- a# c& O+ H& a2 c' ^& J
1 N1 c) ~3 P1 u0 T, qParekh AB and Penner R. Store depletion and calcium influx. Physiol Rev 77: 901-930, 1997./ B' k5 S3 R, F9 H, D

9 w  B5 W6 x6 m
3 X4 o, `! E( S% X' N) c4 [7 |  `9 v* C" A  l0 ^
Pfaller W and Klima J. A critical reevaluation of the structure of the rat uriniferous tubule as revealed by scanning electron microscopy. Cell Tissue Res 166: 91-100, 1976.
1 A1 v- E& R% Z* q- y0 O, o* y1 m/ S! C! Y" A, H

) |9 v8 O% a' C3 Q! C5 d# l
! U5 z, E& z+ ~7 C" p/ }Piperno G, LeDizet M, and Chang XJ. Microtubules containing acetylated alpha-tubulin in mammalian cells in culture. J Cell Biol 104: 289-302, 1987.
* ?: |$ F1 S. s% {# f
* @& H6 |# ?" }, ?. u# d0 f' z" S+ y
* i' z% j  q* }$ D# V& r3 d
Praetorius HA and Spring KR. Bending the MDCK cell primary cilium increases intracellular calcium. J Membr Biol 184: 71-79, 2001." p$ Z- u9 J8 Q' S: Q: v/ d
7 b$ d' @$ |9 d7 q
1 q- @8 x3 O  E: |' q- v9 X

& R- D$ j# M# [8 p# eRalevic V and Burnstock G. Receptors for purines and pyrimidines. Pharmacol Rev 50: 413-492, 1998.5 p  i6 q7 O7 U7 y
+ x) H5 W; J$ S9 V* D2 I  t

- c. N! X( N; i8 z, e0 u2 w, Z! L/ h# t7 n
Sangani AS and Acrivos A. Slow flow past a periodic array of cylinders with application to heat trasnfer. Int J Multiphase Flow 8: 193-206, 1982.+ U' ]# U0 D" B3 ^. r

3 \: F. n+ {# x/ ~
, l' x4 Z+ ^: ^
4 {1 g. x0 g4 s) I" M, dSatlin LM. Regulation of potassium transport in the maturing kidney. Semin Nephrol 19: 155-165, 1999.
3 R) L7 C8 o9 A3 I4 N
: M, D8 p6 {! q9 Y1 \2 G; u- V3 M9 g# u& i# a" u

+ d/ t6 M$ q, l! F2 S+ pSatlin LM, Evan AP, Gattone VH III, and Schwartz GJ. Postnatal maturation of the rabbit cortical collecting duct. Pediatr Nephrol 2: 135-145, 1988.
: ^& O9 w8 m, x0 s2 k
0 A5 t) q5 N6 O' M; B$ c9 v  O4 N! y6 ?; ~8 K6 z2 \4 u

+ P: m$ \' s# lSatlin LM, Matsumoto T, and Schwartz GJ. Postnatal maturation of rabbit renal collecting duct. III. Peanut lectin-binding intercalated cells. Am J Physiol Renal Fluid Electrolyte Physiol 262: F199-F208, 1992.
2 \4 A1 R6 J# u. f7 k9 |7 N! _
3 h8 ^: C* _. `% B, y% z, p* Z. s( \% c/ Q7 F
; t# [  _- O4 W$ C2 S! E: m
Satlin LM and Schwartz GJ. Postnatal maturation of rabbit renal collecting duct: intercalated cell function. Am J Physiol Renal Fluid Electrolyte Physiol 253: F622-F635, 1987.: O$ `; s+ T$ z. c) O7 w& W8 p* o/ q
6 n/ N- x) n" Y; S$ g

0 w. ^- h5 L" P, h( A0 f9 ^. ^$ P3 I9 _) b, D2 ~$ Y" c
Schmidt-Nielsen B and Graves B. Changes in fluid compartments in hamster renal papilla due to peristalsis in the pelvic wall. Kidney Int 22: 613-625, 1982.
  f  ?5 N9 d& a7 @# Z% x0 U
( M  A4 D7 k" U
) b( M  g, @& Z7 M/ j4 d, c3 ^" R
/ k6 a' V. _* Z: y1 ]& VSchwartz EA, Leonard ML, Bizios R, and Bowser SS. Analysis and modeling of the primary cilium bending response to fluid shear. Am J Physiol Renal Physiol 272: F132-F138, 1997.
& H6 Q# v0 t" p$ g% p- N8 \& |; F* z, Q$ g% d# C
, Y8 s. ?0 V% Y0 Z" ^& `
4 L" V7 W& w$ B' Y0 O
Schwartz GJ, Barasch J, and Al-Awqati Q. Plasticity of functional epithelial polarity. Nature 318: 368-371, 1985.
0 _0 j8 q% A7 o* U7 G( \; c8 T( _2 v# ?

, F+ K7 }% D% i
6 E$ H$ L: N, q8 e7 O3 ~2 iSchwiebert EM and Kishore BK. Extracellular nucleotide signaling along the renal epithelium. Am J Physiol Renal Physiol 280: F945-F963, 2001.
, ^, B5 E/ g% q/ x/ f% s' c2 a; n1 h1 h
  w. B* w' Q1 T5 q# p
" @- c  b9 r: d9 T6 V* Z
Stokes JB. Ion transport by the collecting duct. Semin Nephrol 13: 202-212, 1993.. Q# B8 }% u0 H4 N6 Z6 u

: |) ~3 y$ e- r2 a& Y$ B  T/ P$ d) S0 m
  m% H- k; M. w* D
Traub O and Berk BC. Laminar shear stress: mechanisms by which endothelial cells transduce an atheroprotective force. Arterioscler Thromb Vasc Biol 18: 677-685, 1998.
$ b! j$ S/ H7 S! X0 W  q
: b+ t5 X7 I- i5 q1 f% n7 _7 `' Q( z$ N5 w
* G  s, P  s3 N" z4 B* o; }4 q
Tsay RY and Weinbaum S. Viscous flow in a channel with periodic cross-bridging fibres: exact solutions and Brinkman approximation. J Fluid Mech 226: 125-148, 1991.
0 H# P( C) ?# q  p& T4 R4 n6 u' |9 \
# u6 _: b: y7 {% n
* y- N9 `! n) h9 M* w
) y* c8 s& I0 I6 Z9 rWebber WA and Lee J. Fine structure of mammalian renal cilia. Anat Rec 182: 339-343, 1975.( t- q* o, \& ~- E9 r

! o" }% i6 j% x( ]* x5 y# E8 r
. h4 d, l/ _/ n0 Y
) R; }$ v) X' S5 k" w/ I; z- WWeinbaum S, Guo P, and You L. A new view of mechanotransduction and strain amplification in cells with microvilli and cell processes. Biorheology 38: 119-142, 2001.( q3 A- W: y; ^0 ^. s* N3 S

$ h% P. W" e( {6 X
  Y' J! |  ~# ~9 c/ L4 X9 p% ?3 Q% ]5 c" x- q& k& b
Welling LW, Evan AP, and Welling DJ. Shape of cells and extracellular channels in rabbit cortical collecting ducts. Kidney Int 20: 211-222, 1981.+ U9 b3 \3 l8 N+ f# ~# g( X

, K9 _2 _9 c! Z  w* S6 v& g$ v8 A" w' `  }/ C0 X% p1 |9 s
7 m% ]6 B4 ~) ]( ^/ W' o8 I
Wilson DR. Micropuncture study of chronic obstructive nephropathy before and after release of obstruction. Kidney Int 2: 119-130, 1972.! ?3 T8 ^+ u* Y  {
% y0 a  H! G* m. Q# m3 i- `* M

# E5 o+ ?9 e+ W7 f# a" H# s4 d2 W4 F6 _$ h$ {$ e0 E# s$ G$ B
Woda CB, Leite M Jr, Rohatgi R, and Satlin LM. Effects of luminal flow and nucleotides on [Ca 2  ] i in rabbit cortical collecting duct. Am J Physiol Renal Physiol 283: F437-F446, 2002.. t6 l! y# ~/ P& l9 y
( F6 X2 j- \. l: D
& d1 R* I# E8 @9 J. a" M8 W
' M& F' A; X/ m/ L7 v5 e( Y( O
Yoder BK, Tousson A, Millican L, Wu JH, Bugg CE Jr, Schafer JA, and Balkovetz DF. Polaris, a protein disrupted in orpk mutant mice, is required for assembly of renal cilium. Am J Physiol Renal Physiol 282: F541-F552, 2002.

biobio

顶一个先  

tuanzi

天啊. 很好的资源

剑啸寒

拿把椅子看表演

awen

其实回帖算是一种没德德，所以我快成圣人了  

tempo

今天临床的资料更新很多呀

yukun

写得好啊  

泡泡鱼

快毕业了 希望有个好工作 干细胞还是不错的方向

橙味绿茶

家财万贯还得回很多贴哦  

biobio

牛牛牛牛