Bert-Inge Rosengren

Transendothelial transport: the vesicle controversy.

The relative contribution of transcytosis vs. large pore transport to the passage of macromolecules across microvascular endothelia has been a controversial issue for nearly half a century. To separate transcytosis from ‘porous’ transport, the transcytosis inhibitors N-ethylmaleimide (NEM) and filipin have been tested in in situ or ex vivo perfused organs with highly conflicting results. In continually weighed isolated perfused organs, where measurements of pre- and post-capillary resistances, capillary pressure and capillary filtration coefficients can be repeatedly performed, high doses of NEM and filipin increased the bulk transport of macromolecules from blood to tissue, despite producing vasoconstriction. By contrast, in in situ perfused organs, marked reductions in the tissue uptake of albumin tracer have been observed after NEM and filipin. When tissue cooling has been employed as a means of inhibiting (active) transcytosis, results have invariably shown a low cooling sensitivity of albumin transport, compatible with passive transendothelial passage of albumin. This observation is further strengthened by the commonly observed dependence of albumin transport upon the capillary pressure and the rate of transcapillary convection. For low-density lipoprotein (LDL), a cooling-sensitive, non-selective transport component has been discovered, which may be represented by filtration through paracellular gaps, lateral diffusion through transendothelial channels formed by fused vesicles, or by transcytosis. From a physiological standpoint there is little evidence supporting active transendothelial transport of most plasma macromolecules. This seems to be supported by studies on caveolin-1-deficient mice lacking plasmalemmal vesicles (caveolae), in which there are no obvious abnormalities in the transendothelial transport of albumin, immunoglobulins or lipoproteins. Nevertheless, specific transport in peripheral capillaries of several hormones and other specific substances, similar to that existing across the blood-brain barrier, still remains as a possibility.

The Peritoneal Microcirculation in Peritoneal Dialysis

This paper deals with the peritoneal microcirculation and with peritoneal exchange occurring in peritoneal dialysis (PD). The capillary wall is a major barrier to solute and water exchange across the peritoneal membrane. There is a bimodal size‐selectivity of solute transport between blood and the peritoneal cavity, through pores of radius ∼40–50 Å as well as through a very low number of large pores of radius ∼250 Å. Furthermore, during glucose‐induced osmosis during PD, nearly 40% of the total osmotic water flow occurs through molecular water channels, termed “aquaporin‐1.” This causes an inequality between 1−σ and the sieving coefficient for small solutes, which is a key feature of the “threepore model” of peritoneal transport. The peritoneal interstitium, coupled in series with the capillary walls, markedly modifies small‐solute transport and makes large‐solute transport asymmetric. Thus, although severely restricted in the blood‐to‐peritoneal direction, the absorption of large solutes from the peritoneal cavity occurs at a high clearance rate (∼1 mL/min), largely independent of molecular radius. True absorption of macromolecules to the blood via lymphatics, however, seems to be occurring at a rate of ∼0.2 mL/min. Several controversial issues regarding transcapillary and transperitoneal exchange mechanisms are discussed in this paper.

Blood Flow Limitation In Vivo of Small Solute Transfer during Peritoneal Dialysis in Rats

The aim of this study was to determine whether or to what extent transperitoneal flux of small solutes is reduced at low blood flows during peritoneal dialysis (PD) in rats. Peritoneal blood flow reductions were achieved by bleeding anesthetized (300 g) rats by 25% of their blood volume. After bleeding, a 2 h PD dwell was started using standard PD fluid. The permeability-surface area product (PS) for (51)Cr-EDTA and glucose were assessed, as well as the transperitoneal clearance (Cl) of albumin. Control animals were not bled. After bleeding, peritoneal blood flow declined from 145 +/- 17 perfusion units (PU) to 59 +/- 12 PU (P = 0.001). Concomitant with this reduction, PS for (51)Cr-EDTA fell from 0.284 +/- 0.01 ml/min to 0.216 +/- 0.01 ml/min (P = 0.006) and PS for glucose from 0.338 +/- 0.02 ml/min to 0.294 +/- 0.01 ml/min (P = 0.046). Mean arterial BP (MAP) dropped from 133 +/- 4 mmHg to 61 +/- 5 mmHg (P = 0.008). Cl of albumin fell largely in proportion to the estimated capillary hydrostatic pressure drop, i.e., from 6.1 +/- 0.7 microl/min to 2.3 +/- 0.3 microl/min (P = 0.001). The results demonstrate that the transperitoneal clearances of small solutes are blood flow limited during PD, when peritoneal perfusion is markedly reduced. The level of flow limitation was, however, much lower than expected and observed in other tissues. Albumin transport, which is not blood flow limited, was reduced largely in proportion to the calculated capillary hydrostatic pressure decrease.

Transvascular Passage of Macromolecules into the Peritoneal Cavity of Normo- and Hypothermic Rats in vivo: Active or Passive Transport?

During the last decades there has been a debate regarding whether transvascular protein transport is an active (transcytosis) or a passive (porous) process. To separate cooling-sensitive transcytosis from passive transport processes between blood and peritoneal fluid, we induced hypothermia in rats in vivo, reducing their body temperature to 19°C. Control rats were kept at 37°C. Either human albumin, or IgG, or IgM, or LDL, radiolabeled with 125I, was given intra-arterially together with 51Cr-EDTA. During tracer administration, a 2-hour peritoneal dialysis dwell was performed. Clearance of the tracers to dialysate, and the permeability-surface area coefficient (PS) for 51Cr-EDTA and glucose were assessed. During cooling, mean arterial blood pressure (MAP) was reduced to 40% of control and plasma viscosity increased by 48.5%, while peritoneal blood flow was reduced to 10%. At 19°C, clearance of albumin to dialysate fell from 9.30 ± 1.62 (SEM) to 3.13 ± 0.28 µl/min (p < 0.05), clearance of IgG from 6.33 ± 0.42 to 2.54 ± 0.12 µl/min (p < 0.05), clearance of IgM from 3.65 ± 0.33 to 1.10 ± 0.12 µl/min (p < 0.05), and clearance of LDL from 3.54 ± 0.20 to 0.73 ± 0.06 µl/min (p < 0.05). The fall in PS for 51Cr-EDTA was from 0.320 ± 0.01 to 0.075 ± 0.003 ml/min (p < 0.05), and that for glucose from 0.438 ± 0.02 to 0.105 ± 0.01 ml/min (p < 0.05). Tissue cooling reduced large solute transport largely in proportion to the cooling-induced reductions of MAP (to 40%), and the concomitant increase in viscosity (to 67%), i.e. to ≈20–30% (0.40 × 0.67) of control, though LDL clearance was reduced further. The fall in small solute PS, in excess of the viscosity effect, mirrored the fall in peritoneal blood flow occurring during hypothermia. In conclusion, the good correlation of predicted to calculated changes suggests that the overall transendothelial macromolecular passage in vivo occurs passively, and not due to active processes.

Transendothelial transport of low-density lipoprotein and albumin across the rat peritoneum in vivo: Effects of the transcytosis inhibitors NEM and filipin

This study was performed to investigate the mechanisms responsible for the transport of albumin and low-density lipoprotein (LDL) across capillary walls in vivo. To separate transcytosis from passive, ‘porous’ transport, we tested the effects of the transcytosis inhibitors N-ethylmaleimide (NEM) and filipin given intraperitoneally on the peritoneal capillary clearance of LDL and albumin in anesthetized rats undergoing peritoneal dialysis. Radiolabeled human albumin or LDL was given intra-arterially, and 51Cr-EDTA was infused intravenously. A 2-hour peritoneal dialysis dwell was performed using 16 ml of conventional 1.36% glucose-based dialysis fluid. The clearance of LDL and albumin to the dialysate and the peritoneal mass transfer coefficient for 51Cr-EDTA were assessed. Following intraperitoneal NEM incubations (0.5–5 mM), there were marked increases in the peritoneal transport of albumin and LDL for NEM doses exceeding 1 mM. For lower NEM doses, there were no reductions in clearance. Filipin incubations (0.2–4 µg/ml) did not affect the clearance of either macromolecule. In conclusion, neither NEM nor filipin caused reductions in albumin or LDL clearance across the peritoneal capillaries. The present data clearly show that NEM and filipin are unsuitable as transcytosis inhibitors in vivo.

Acute Peritoneal Dialysis in Rats Results in a Marked Reduction of Interstitial Colloid Osmotic Pressure

The aim of this study was to investigate whether the interstitial colloid osmotic pressure (COP(i)) of peritoneum is of importance for peritoneal fluid reabsorption during peritoneal dialysis (PD). For testing this hypothesis, a method to isolate interstitial fluid from the peritoneal membrane and to measure the interstitial COP(i) in the normal peritoneum and directly after the initiation of PD needed to be developed and validated. Eighteen female rats were anesthetized subcutaneously in the neck region with fentanyl-midazolam (1:1). Nylon wicks were implanted postmortem by means of a plastic catheter in the tissue just underneath the peritoneal surface. The characteristics of this fluid were compared with that isolated from wicks that were implanted in intermuscular spaces in hindlimb muscle and back subcutis. All wicks were removed after 20 min and centrifuged. The wick fluid was collected and analyzed in a colloid osmometer constructed for submicroliter samples, and interstitial fluid COP was compared with that of plasma. PD was initiated by injecting 20 ml of 3.86% glucose-containing PD fluid (Dianeal) into the peritoneal cavity, over a dwell time of 4 h. Control rats received no PD. The ratio of COP(i) to that of plasma (COP(p)) during control was 0.65 +/- 0.05 in peritoneum, 0.53 +/- 0.04 in muscle, and 0.59 +/- 0.05 in skin. After a PD dwell, the ratio was 0.29 +/- 0.03 in peritoneum, 0.54 +/- 0.08 in muscle, and 0.41 +/- 0.06 in skin. Thus, the COP ratio in peritoneum fell by 55% (P = 0.014) and in skin by 30% (P = 0.03), whereas the COP ratio in muscle was unchanged. The results suggest that acute PD results in a marked fall of the COP(i) in the peritoneal membrane, shifting the Starling equilibrium toward an absorptive state. The effect was restricted to the peritoneal membrane and, to some extent, the skin. It is speculated that the increase in peritoneal hydrostatic pressure after PD causes an increase in interstitial tissue volume, with primarily a dilution of interstitial colloids.