Calculations of capillary hydraulic conductivity (L_p ) have been made over the last thirty years by measurement of filtration rate in cannulated, perfused capillaries during an imposed hydrostatic or oncotic pressure gradient (Michel et al. 1974). To make these measurements it has been necessary to cannulate the capillary and perfuse it under a constant pressure head. Permeability was measured by occluding the vessel with a glass rod or other instrument to equilibrate the pressure in the vessel to the driving pressure. Any movement of marker cells in the vessel will therefore be due to filtration or absorption of fluid across the vessel wall, which could be calculated from the surface area of the vessel and the velocity of the cells. The L_p (filtration rate per unit area per unit pressure) could therefore be calculated from measured parameters. This technique could not be used to measure L_p between periods of controlled shear rate because the shear rate is dependent on the flow in the vessel, which is in turn dependent on the difference between the inflow and outflow pressures. The outflow pressure in these experiments is dependent on the outflow resistance of the microcirculation of the animal. This resistance is highly variable and depends on the pressure and diameter of the vessels downstream of the cannulated vessel.

We have measured the velocity of the red cells in the pipette before they enter the capillary, and calculated the shear rate based on the ratio of the cross-sectional areas of the vessel and the point in the pipette where red cell velocity was measured. This simple calculation showed that the shear rates (and velocities) in the vessel varied by an order of magnitude during a short perfusion time (15 min), presumably due to alterations in outflow pressure. In order to overcome this difficulty we have designed a novel method for measurement of L_p during controlled outflow pressures. The mesentery of a pithed frog, Rana temporaria, was visualised under a Wild M8 microscope. At the end of the experiments frogs were killed by cranial destruction. A glass rod was used to restrain the tissue close to one end of a capillary (15-25 µm wide and > 800 µm long). The vessel was cannulated with a bevelled micropipette (tip 15-20 µm diameter) containing 1 % bovine serum albumin, frog Ringer solution and rat red blood cells (collected by cardiac puncture of halothane-anaesthetised rats, later killed by cervical dislocation). A second restrainer was placed at the other end of the capillary, and the vessel cannulated anterogradely. The double cannulated section of the vessel was perfused with the upstream pipette set to 35 cmH₂O and the downstream pipette at a variable pressure, lower than 35 cmH₂O. L_p was measured by altering the downstream pressure until there was no flow into either the downstream or upstream pipette, by switching the pressure to a third manometer set to the pressure at which balance is maintained. This was usually within 0.5 cmH₂O of the upstream pressure. The differential velocity was measured by recording alternately from the vessel segments close to each cannulation site on a video tape. These two sections were recorded by moving the microscope independently of the stage by 0.2-2 mm between two metal bossheads. Velocity was calculated as the distance moved between a marker red cell and the pipette tip at various times over the course of 3-4 s for cells near both pipette tips.

The product of the differential velocity and the cross-sectional surface areas of the vessel at each marker red cell site was the flow rate across the vessel wall. The surface area was calculated from the mean radius sampled at a number of points along the vessel. Since the applied pressure was known (the average of the balance pressure and the upstream pipette pressure) then the filtration rate per unit area per unit pressure – L_p – could be calculated. Using this technique, combined with measurement of red cell velocity between balance pressures, L_p can be calculated between periods of controlled shear rate.

This work was funded by The Wellcome Trust (58083) and the British Heart Foundation (FS20057).

Michel, C.C., Mason, J.C., Curry, F.E., Tooke, J.E. & Hunter, P.J. (1974). Q. J. Exp. Med. Cog. Sci. 59, 283-309.