Eugene M. Landis and the physiology of the microcirculation

This essay looks at the historical significance of two APS classic papers that are freely available online:

Landis EM. The capillary pressure in frog mesentery as determined by micro-injection methods. Am J Physiol 75: 548–570, 1926 ( http://ajplegacy.physiology.org/cgi/reprint/75/3/548 ).

Landis EM. Micro-injection studies of capillary permeability. II. The relation between capillary pressure and the rate at which fluid passes through the walls of single capillaries. Am J Physiol 82: 217–238, 1927 ( http://ajplegacy.physiology.org/cgi/reprint/82/2/217 ).

Measurement of microvascular transport parameters of macromolecules in tissues and organs of intact animals

A critical analysis is presented of two widely used approaches to measurement of microvascular transport of large molecules in intact animals: (1) measurement of lymph flow and macromolecular solute concentration relative to plasma, and (2) tissue accumulation of tracer macromolecules. To demonstrate the advantages and limitations of each method, experimental results which permit direct comparison of the two methodologies are reviewed and analyzed, and sources of error in each pointed out. It is concluded that both approaches are valid under the appropriate conditions: steady state of lymph, initial transient state for tissue uptake. These conditions are mutually exclusive, but complementary. When the requirements for neither lymph collection or tissue accumulation alone can be satisfied experimentally, a combination of the two approaches can yield valid results.

Interstitial exclusion of macromolecules studied by graded centrifugation of rat tail tendon

Mechanical compression of cartilage and tendon has been shown to expel fluid both from collagen fibrils and from the extrafibrillar space. As reported previously, albumin (Alb) concentration and colloid osmotic pressure in tendon fluid (TF) expelled by repeated centrifugations fell progressively at increasing centrifugation force (G = 600, 2,400, and 13,100), suggesting either molecular sieving in compressed tendon or mobilization of protein-free (excluded) fluid. The present experiments, including analysis of 51Cr-EDTA, aprotinin (Ap), Alb, immunoglobulin G (IgG), and hyaluronan (hyaluronic acid; HA) with molecular weight (MW) ranging from 341 to 5 x 10(6), strongly favored the exclusion hypothesis; the fraction of Alb, IgG, and HA-free fluid (excluded) was already 0.23-0.36 in the first centrifugate, increasing to 0.73-0.82 in the third. The corresponding numbers were, respectively, 0.11 and 0.43 for Ap (MW 6,500), and 0 and 0.08 for 51Cr-EDTA. These data, combined with calculated exclusion by collagen fibrils, proteoglycans, and HA, indicated that the first centrifugate was mainly derived from the extrafibrillar space, with increasing addition of macromolecular free intrafibrillar fluid in the second and third centrifugates, with each space contributing about equally to the total centrifugate volume. The calculations also indicated that Alb-, IgG-, and Ap-free fluid was mobilized from extrafibrillar space by increasing overlap of excluded territories. An excess of HA in tendon compared with that estimated from centrifugate concentrations suggests a large bound or immobilized HA fraction.

Acute alveolar hypoxia increases blood-to-tissue albumin transport: role of atrial natriuretic peptide

Plasma immunoreactive atrial natriuretic peptide (irANP) and blood-to-tissue clearance of 131I-labeled rat serum albumin (CRSA) were examined in anesthetized rats during hypoxic ventilation (n = 5-7/group). Hypoxia (10 min) increased irANP from 211 +/- 29 (room air) to 229 +/- 28 (15% O2, not significant), 911 +/- 205 (10% O2), and 4,374 +/- 961 pg/ml (8% O2), respectively. Graded increases in CRSA were significant at 8% O2 in fat (3.6-fold), ileum (2.2-fold), abdominal muscles (2.0-fold), kidney (1.8-fold), and jejunum (1.4-fold). CRSA was decreased in back skin and testes; heart, brain, and lungs were unaffected. The increases in CRSA were related to irANP and not to arterial PO2. Circulating plasma volume was negatively correlated with whole body CRSA. Graded increases in extravascular water content (EVW) were found in the kidney, left heart, and cerebrum and were positively related to CRSA in the kidney. EVW decreased in gastrointestinal tissues; the magnitude was inversely related to CRSA. We conclude that ANP-induced protein extravasation contributes to plasma volume contraction during acute hypoxia.

Atrial natriuretic peptide levels and plasma volume contraction in acute alveolar hypoxia

Arterial oxygen tensions (PaO2), atrial natriuretic peptide (ANP) concentrations, and circulating plasma volumes (PV) were measured in anesthetized rats ventilated with room air or 15, 10, or 8% O2 (n = 5-7). After 10 min of ventilation, PaO2 values were 80 +/- 3, 46 +/- 1, 32 +/- 1, and 35 +/- 1 Torr and plasma immunoreactive ANP (irANP) levels were 211 +/- 29, 229 +/- 28, 911 +/- 205, and 4,374 +/- 961 pg/ml, respectively. At PaO2 < or = 40 Torr, irANP responses were more closely related to inspired O2 (P = 0.014) than to PaO2 (P = 0.168). PV was 36.3 +/- 0.5 microliters/g in controls but 8.5 and 9.9% lower (P < or = 0.05) for 10 and 8% O2, respectively. Proportional increases in hematocrit were observed in animals with reduced PV; however, plasma protein concentrations were not different from control. Between 10 and 50 min of hypoxia, small increases (+40%) in irANP occurred in 15% O2; however, there was no further change in PV, hematocrit, plasma protein, or irANP levels in the lower O2 groups. Urine output tended to fall during hypoxia but was not significantly different among groups. These findings are compatible with a role for ANP in mediating PV contraction during acute alveolar hypoxia.

Atrial Natriuretic Peptide as a Regulator of Transvascular Fluid Balance

Unlike other natriuretics, which act via the kidneys to reduce interstitial fluid volume with little change in plasma volume, atrial natriuretic peptide has important extrarenal actions that enable it to reduce plasma volume preferentially.

Cellular aspects of transvascular exchange: a 40-year perspective

Study of the microcirculation began with microscopic anatomy in the 17th century, but it was not until the mid-19th century that the concept of microcirculatory exchange was developed, in association with the great expansion in knowledge of physical chemistry and cell biology. In the first half of the 20th century, the roles of diffusion and osmosis (ultrafiltration) were put on a quantitative foundation and the physiological basis for regulation of the microcirculation was established. During the past 40 years, our understanding of transvascular exchange has been deepened and widened through two major developments: 1) the application of electron microscopy to the study of microvascular structure and 2) the use of sophisticated physical and mathematical models of passive transport processes to analyze and interpret experimental data. In the past few years, the focus of microvascular research has been returning to cell biology and the role of endothelial cells in controlling transvascular exchange. The 21st century holds promise of many exciting new developments.

Limits to steady-state lymph flow rates derived from plasma-to-tissue uptake measurements

Upper and lower limits to basal lymph flow rates per unit tissue mass calculated from measurements of initial tissue uptake rates of 131-I albumin and steady-state lymph/plasma or interstitial fluid/plasma concentration ratios of endogenous albumin are compared with published measurements of lymph flow. The upper limit corresponds to purely convective transport of albumin, the lower to purely diffusive transport. Of a total of nine feasible comparisons, five show good agreement defined by the measured value falling at or between the calculated limits. These include two direct comparisons in the same or closely similar animal species and experimental conditions, and three cross-specific comparisons. Disagreement, with measured lymph flows much higher than the estimated upper limit, was observed in three comparisons, one of which was direct and conspecific and two that were cross-specific. In one instance, also cross-specific, reported lymph flows were below our minimal estimate. We suggest that estimates of lymph flow derived from tissue uptake measurements may be useful in setting limits to lymph turnover of fluid and plasma proteins in tissues and organs from which it is not feasible to collect lymph. They may also provide a means of evaluating experimentally induced disturbances of lymph drainage and over- or underestimation of the tissue mass from which lymph is collected, these being serious experimental problems associated with lymphatic cannulation and lymph collection. If both direct and indirect measurements can be made in the same tissue or organ under the same conditions, the relative contributions of convective and dissipative (diffusion and vesicular) transport processes to protein extravasation can be evaluated from the relation of the directly measured flow to the two calculated limits. These requirements were met for only two tissues among those surveyed: rat tail skin and rabbit leg muscles. For rat tail skin, our results suggest that approximately half the albumin transport is convective and half dissipative, and for rabbit muscle, about two-thirds is convective and one-third dissipative.

Renal cortical interstitium and fluid absorption by peritubular capillaries

Every minute, the cortical peritubular capillaries in a 1-g rat kidney take up more than 0.5 ml tubular reabsorbate. Studies of renal lymph and measurements of pressure in capillaries (Pc) and interstitium (Pi) indicate that normally the protein colloid osmotic pressure of peritubular capillary plasma (COPp) provides the necessary absorptive force, keeping Pi at 2-4 mmHg, i.e., 8-10 mmHg lower than Pc. At reduced COPp, continued delivery of fluid from the tubules automatically raises Pi to maintain capillary fluid uptake. The transient Pi response to sudden exposure of the kidney to subatmospheric pressure shows that such adjustment of forces may take place in only 5 s. Most remarkable, adjustment of forces may take place in only 5 s. Most remarkable, reabsorption continues during protein-free perfusion of the isolated rat kidney, apparently effected by a Pi exceeding Pc. A relative retardation of interstitial uptake of ferritin from plasma in this case suggests fluid reabsorption through both small and large pores in the capillary wall. Collapse of the capillaries is presumably prevented by tight tethering to the capillary wall, giving the narrow interstitium a very low compliance.

Interstitial exclusion of IgG in rat tissues estimated by continuous infusion

Interstitial exclusion, defined as the fraction of interstitial fluid volume inaccessible to a solute, was evaluated for immunoglobulin G (IgG) in selected tissues of rats by a method previously applied to serum albumin (29). IgG distribution volumes were also measured for intestine. 125I-labeled rat IgG was infused for 5 or 7 days (n = 4 rats each) with an implanted osmotic pump (Alzet). At the termination of infusion, the rat was anesthetized, nephrectomized, and injected with 51Cr-labeled EDTA (4 h) to label total extracellular fluid volume and 131I-labeled bovine IgG (5 min) to label plasma volume. Samples of skin, muscle, and tendon were assayed for total and extractable tracer activity. Interstitial fluid from these tissues was sampled postmortem with nylon wicks for assay of 125I-labeled IgG and endogenous albumin and IgG. Exclusion of IgG was calculated from the difference between extravascular 125I-labeled IgG and 51Cr-labeled EDTA distribution volumes. In contrast to our previous experience with tracer albumin, 125I-labeled IgG was not fully extractable from minced skin, muscle, or tendon by isotonic saline; only 71-83% was recovered under conditions that eluted 92-96% of tracer albumin and 94-99% of tracer EDTA. We conclude that approximately 20% of extravascular 125I-labeled IgG in these tissues is sequestered or bound in the interstitium. Calculation of IgG fractional exclusion from extractable tracer yielded the following values (means +/- SE, n = 8 rats): leg muscles 0.37 +/- 0.09, leg skin 0.44 +/- 0.03, back skin 0.36 +/- 0.04, tail skin 0.40 +/- 0.08, and tail tendon 0.55 +/- 0.04.(ABSTRACT TRUNCATED AT 250 WORDS)

Blood-tissue transport of exogenous albumin and immunoglobulin G in genetically analbuminemic rats

Tracer uptake studies were carried out in adult female Nagase (NA) strain analbuminemic rats [derived from Sprague-Dawley (SD) stock] and in adult female SD controls to determine the extent to which capillary permeability to plasma proteins is altered in the absence of endogenous albumin. Accessory measurements (arterial pressure, central venous pressure, plasma and interstitial fluid protein concentrations and oncotic pressures, plasma volume, and interstitial fluid volume) confirm the report of Joles et al. [Am. J. Physiol. 257 (Renal Fluid Electrolyte Physiol. 26): F23-F28, 1989] that shows elevated plasma volumes, normal interstitial fluid volumes, nearly normal plasma oncotic pressures (due to elevated globulin concentrations), and lower interstitial fluid oncotic pressures. In skin, skeletal muscles, and heart muscle, clearances of exogenous heterologous (bovine) albumin were 20-40% higher in NA than in SD controls. In intestine, albumin clearances were 20-30% lower. In NA rats blood-to-tissue clearances of heterologous (bovine) immunoglobulin G in skin and heart were higher and in the intestine they were lower than in SD controls; however, clearances in skeletal muscles were not elevated. The differences between NA and SD are small compared with the large increases in macromolecular permeabilities reported by others for organs and single microvessels perfused with albumin-free fluids.(ABSTRACT TRUNCATED AT 250 WORDS)

Cellular and intercellular transport pathways in exchange vessels

The endothelium of lung alveolar capillaries is of the continuous type, that of airway exchange vessels (capillaries and pericytic venules) includes both continuous and fenestrated types. Water and small lipophilic solutes penetrate via the endothelial cells (cell membrane pathway) as well as through intercellular junctions. Hydrophilic solutes are limited to junctional pathways and cytoplasmic vesicles. Permeation of hydrophilic solutes is progressively restricted with increasing molecular size, as by a sieve, with many openings 8 nm and a few 40 to 60 nm wide. In response to local tissue injury or to certain chemical mediators, larger junctional pathways may be opened, greatly increasing permeability to large molecules. Both alveolar capillaries and airway exchange vessels exhibit this response, but the effective stimuli may differ (e.g., alveolar capillaries are insensitive to histamine and bradykinin). Hydrophilic solutes are transported by diffusion, convection, and vesicular exchange (transcytosis). For small ions and molecules (radii < 2 nm), diffusion is the dominant transport mode; contributions of convection and transcytosis are negligibly small. Because diffusion decreases with increasing molecular size, all three mechanisms may contribute substantially to transport of large molecules (radii > 2 nm). Fenestrated endothelia have higher hydraulic conductivities and are more permeable to small ions and molecules than are continuous endothelia. However, their permeabilities to plasma proteins are about the same. Lung alveolar capillary endothelium has lower hydraulic conductivity and lower solute permeabilities than do other continuous endothelia (heart, skeletal muscle). Airway exchange vessel endothelium has about the same permeability to serum albumin as alveolar capillary endothelium.

Interstitial exclusion of albumin in rat tissues measured by a continuous infusion method

Steady-state 125I-labeled rat serum albumin (125I-labeled RSA) concentration in plasma was maintained by intravenous infusion of tracer for 72-168 h with an implanted osmotic pump. At the end of the infusion period, the rat was anesthetized and nephrectomized, and extracellular fluid was equilibrated with intravenous 51Cr-labeled EDTA for 4 h. Five minutes before final plasma and tissue sampling, 131I-labeled bovine serum albumin (131I-labeled BSA) was injected intravenously as a plasma volume marker. Samples of skin, muscle, tendon, and intestine were assayed for all three tracers. Apparent distribution volumes were calculated as tissue tracer content/plasma tracer concentration. Interstitial fluid volume (Vi) was calculated as V51Cr-EDTA-V131I-BSA. Steady-state extravascular distribution of 125I-labeled RSA as plasma equivalent volume (Va,p) was calculated as V125I-RSA-V131I-BSA. Steady-state interstitial fluid concentrations of 125I-labeled RSA in skin, muscles, and tendon were measured with nylon wicks implanted postmortem, and steady-state interstitial albumin distribution volumes were recalculated as wick-fluid equivalent volumes (Va,w). Relative albumin exclusion fraction (Ve/Vi) was calculated as 1-Va,w/Vi. For skin and muscle, steady-state 125I-labeled RSA tissue concentrations were reached at 72 h. Ve/Vi for albumin averaged 26% in hindlimb muscle, 41% in hindlimb skin, 30% in back skin, 39% in tail skin, and 54% in tail tendon. For muscle, Ve/Vi corresponds to expectation if all tissue collagen and hyaluronan is dispersed in the interstitium. However, for skin and tendon, albumin exclusion is considerably lower than expected on this basis, suggesting that much of their collagen is organized into dense bundles of fibers containing no fluid accessible to 51Cr-labeled EDTA or 125I-labeled RSA.

Tissue-specific effects of physiological ANP infusion on blood-tissue albumin transport

Blood-tissue transport of 131I-labeled bovine serum albumin (BSA) during intravenous infusion of synthetic atrial natriuretic peptide (ANP) was examined in anesthetized male Wistar rats. Plasma volumes were maintained at pre-ANP levels by infusion of 2% BSA in lactated Ringer solution (LR) to minimize compensatory responses to ANP-induced hypovolemia. 131I-BSA clearance was measured over 30 min, and 125I-BSA was injected terminally to correct for intravascular volume. Thirty-minute infusion of 20 ng.kg-1.min-1 ANP resulted in a tissue-selective increase in 131I-BSA clearance in jejunum and colon compared with controls given LR only. Smaller but significant increases in tracer clearance also were observed in fat, kidney, left ventricle, and skeletal muscle exposed to 400 ng.kg-1.min-1 ANP. The observed elevation in tracer albumin extravasation was not associated with any measurable increase in tissue extravascular water content. Furthermore, it was shown that coupling of 131I-BSA transport to filtration induced by hindlimb venous congestion was similar in control and ANP-treated rats. In a second series of experiments, plasma ANP levels were determined after 30-min ANP infusions from 0 to 180 ng.kg-1.min-1. Significant linear associations between physiological ANP levels (62-578 pg/ml) and 131I-BSA clearance were demonstrable for small intestine, colon, fat, kidney, and skeletal muscle but not for skin, heart, diaphragm, and lung. We conclude that raising plasma ANP by infusion of the synthetic peptide results in a filtration-independent, tissue-selective increase in albumin transport. Tissue uptake of albumin is a potential mechanism for extrarenal fluid shift during circulatory volume overload.

Plasma volume expansion with colloids increases blood-tissue albumin transport

Extravasation of plasma proteins is increased after volume expansion with whole blood or plasma. To investigate the mechanisms responsible for this phenomenon, we measured extravascular accumulation of exogenous 131I-labeled bovine serum albumin in several tissues and organs of anesthetized rats. Plasma volume was increased acutely by infusion of isoncotic albumin or polyvinylpyrrolidone, with or without subsequent infusion of a 1:10 dilution of the colloid to induce blood-to-tissue fluid movement. Controls were given only a slow sustaining infusion of saline. The amounts of fluid and plasma protein lost from the circulation were followed simultaneously by two methods: 1) material balance in the whole animal, and 2) changes in 131I-labeled albumin uptake (VA) and water content (VW) in the individual tissues. Plasma volume expansion of 80-90% increased plasma protein extravasation in the whole rat by an average of 2.7-fold over a 30-min period. Of the protein extravasated, 42% entered the abdominal cavity. The rest was distributed in the interstitial compartment of various tissues and organs. Tracer albumin accumulation (averaged over 30 min) was increased 38-82% in skin and paw, 40-59% in skeletal muscles, 131% in hearts, and 167-230% in different parts of the intestine. Increased convective transport does not appear to be a major factor. There was little or no relation of albumin transport increase to the magnitude or direction of net fluid transfer. Coupling of albumin transport to volume flow was not greater than previously reported for saline infusion or venous congestion. Convective redistribution (convective transport without net fluid transfer, "volume recirculation") is estimated to increase albumin transport no more than 10% under the conditions of our experiments. The greater part of the increase is thus dissipative, i.e., attributable to increased diffusion or increased vesicular exchange. Control of dissipative transport of albumin may play an important role in regulating plasma volume.

Sampling interstitial fluid from rat skeletal muscles by intermuscular wicks

A modification of the implanted wick method (K. Aukland and H. O. Fadnes. Acta Physiol. Scand. 88: 350-358, 1973) was devised to sample interstitial fluid from rat muscles. Dry nylon wicks were inserted postmortem into intermuscular spaces between leg muscles by means of a plastic catheter, which was subsequently withdrawn. Inserting the wicks postmortem avoids contaminating wick fluid with proteins extravasated as a result of local inflammatory reactions; placing them intermuscularly avoids contamination by fluid and proteins from damaged muscle cells. Wick fluid protein concentrations (mg/ml) averaged 24.1 +/- 1.1 and 28.5 +/- 1.5 (means +/- SE) in medial and lateral hindlimbs muscles, respectively. The corresponding albumin concentrations were 13.0 +/- 0.7 and 13.9 +/- 0.7 mg/ml. Total protein and albumin concentrations in plasma were 54.1 +/- 0.8 and 22.5 +/- 0.3 mg/ml. Electrophoresis of wick fluid showed a pattern of peaks similar to that of plasma, with albumin relatively high and larger molecules relatively low. Proteins from muscle cells were not detected. Isotope studies (125I-labeled albumin, 51Cr-EDTA) showed that less than 2% of the albumin in wick fluid came directly from plasma and that wick fluid was not concentrated by cell swelling postmortem. Wick fluid from intermuscular wicks implanted in anesthetized rats in vivo had nearly the same total protein concentration as fluid from postmortem wicks, but albumin-to-globulin (A/G) ratios were slightly lower (1.22 +/- 0.07 vs. 1.53 +/- 0.21 measured by gel electrophoresis), and more significantly, nearly 50% of the albumin leaked to wick fluid from plasma as a result of wick implantation.(ABSTRACT TRUNCATED AT 250 WORDS)

Influence of saline infusion on blood-tissue albumin transport

Anesthetized rats were infused with lactated Ringer solution (LR) at constant rate for 30 or 60 min; delivered volume loads ranged from 0.03 to 0.08 ml/g body wt. Controls were given only a sustaining infusion of saline at 0.002 ml.g-1.h-1. Only 7-14% of the LR remained in the plasma at the end of the infusion; 76-88% entered the interstitial compartment, and 7-17% was excreted. The amount of plasma protein lost from the circulation with the extravasated fluid was studied simultaneously by two methods: 1) material balance in the whole animal and 2) changes in 131I-labeled albumin uptake (VA) and water content (VW) in individual tissues. The extravasation of 0.03-0.06 ml fluid/g body wt (75-160% initial plasma volume) did not significantly increase plasma protein extravasation in the whole rat. Nearly all of the sampled tissues of LR-infused rats had higher VW than controls. Tissue VA tended to increase with VW, but the regression slopes (delta VA/delta VW), a measure of the tracer albumin concentration of capillary filtrate relative to plasma, were low; skin, 0.006; paw, 0.018; skeletal muscles, 0.007; heart, 0.057; jejunum, 0.095; ileum, 0.045; cecum, 0.026; and colon, 0.027. These ratios are consistent with the very small loss of total plasma protein observed and attest to high solvent-drag reflection coefficients (sigma approximately equal to 1 - delta VA/delta VW): greater than 0.98 in capillaries of skeletal muscles, skin, and paw and 0.91-0.97 in heart and intestine.

Albumin extravasation rates in tissues of anesthetized and unanesthetized rats

Bovine serum albumin (BSA) labeled with 131I was injected intravenously in chronically prepared, unanesthetized rats and into pentobarbital-anesthetized rats that had received 2 ml 5% BSA to help sustain plasma volume. Initial uptake rates (clearances) in skin, skeletal muscles, diaphragm, and heart (left ventricle) were measured over 1 h. BSA labeled with 125I was injected terminally to correct for intravascular 131I-BSA. Observed clearances were in the following order in both groups of animals: heart much greater than diaphragm approximately equal to skin greater than resting skeletal muscles. Differences between unanesthetized and anesthetized animals were small and inconsistently directed. Our results suggest that the lower albumin clearances reported in the literature for anesthetized rats are not the result of their immobility or any direct effect of anesthesia on albumin transport in these tissues. The lower transport rates appear to result indirectly from changes produced by anesthesia and/or surgery in controllable parameters such as plasma volume and intravascular protein mass.

Coupling of albumin flux to volume flow in skin and muscles of anesthetized rats

Bovine serum albumin (BSA) labeled with 131I or 125I was injected intravenously in pentobarbital sodium-anesthetized rats, and tracer clearances into leg skin and muscles were measured over 30, 60, and 120 min. BSA labeled with the alternate tracer was used as vascular volume reference. Two minutes before injection of the tracer, a ligature was tied around one femoral vein to occlude outflow partially and raise capillary pressure in that leg. The unoccluded leg served as control. Skin and muscles of the occluded leg had variably and substantially higher water contents (delta W) than paired control tissues and slightly but consistently increased albumin clearances (CA). The delta CA/delta W, equivalent to the albumin concentration of capillary filtrate relative to plasma determined by linear regression, were as follows: leg skin 0.004 (95% confidence limits -0.001 to +0.009), muscle biceps femoris 0.005 (0.001-0.010), muscle gastrocnemius 0.011 (0.004-0.019), muscle tibialis anterior 0.016 (0.012-0.021). All these values are significantly less than 0.10, which corresponds to a reflection coefficient for serum albumin (sigma A) of 0.90. Convective coupling of albumin flux to volume flux in skin and muscles of intact, anesthetized rats is low, with sigma AS in the range 0.98 to greater than 0.99.

Wick sampling of interstitial fluid in rat skin: further analysis and modifications of the method

We compared modifications of the wick technique for analysis of interstitial fluid in rat subcutis. Nylon wicks were implanted for 60 min in back skin of rats after anesthesia with pentobarbital or after sacrifice by potassium chloride injection. Wicks were implanted dry or loaded with saline or varied dilutions of rat serum. Implantation of dry wicks and wicks loaded with diluted serum in living, anesthetized animals produced similar results; the protein concentration of wick fluid averaged about 60% that of the plasma protein concentration. The saline loaded wicks produced wick fluid with a lower protein concentration, average about 45% that of plasma protein concentration. The lower concentrations apparently resulted from simple dilution. Wick fluid sampled from dead animals had similar total protein concentrations, but in the dead animals there was a lower concentration of the large plasma proteins and a relatively higher concentration of the smaller proteins.

CONCLUSIONS

Wick implantation in living animals causes a transitory inflammatory reaction and a decrease in the size selectivity of macromolecular sieving, but local osmotic forces bring about a concentration equilibrium with undisturbed interstitium. Implantation of dry wicks in subcutis either in vivo or post mortem provides a simple, direct method for sampling the total protein concentration and colloid osmotic pressure of interstitial fluid. Implantation of dry wicks postmortem permits measurement of individual component protein concentrations and evaluation of molecular selectivity between plasma and interstitium.

Some consequences of capillary permeability to macromolecules: Starling's hypothesis reconsidered

Starling's hypothesis ascribes fluid movements across capillary walls to the interaction of hydrostatic and colloid osmotic forces. For 90 years it has been recognized as the basis of plasma-to-interstitial fluid balance. Its original statement was based on the notion of capillary impermeability to plasma proteins. However, as knowledge of transcapillary exchange of plasma proteins developed, its formulation was progressively modified to allow for protein transport and for interaction of protein transport with volume flow. The most important aspects of the conceptual evolution of Starling's hypothesis are reviewed in the text of this lecture.

B. W. Zweifach Award lecture. Regulation of the microcirculation

Microcirculatory blood flow and transport are controlled to meet local and systemic demands for material exchange and body fluid balance. Control mechanisms act through effectors (smooth muscle cells) at many sites within the microvascular bed. Responses at different sites are not uniform, resulting in a broadly heterogeneous distribution of pressures and flows which is constantly changing. Simplified, uniform models of microvascular networks have made it possible to identify the principles governing blood circulation and blood-tissue transport. However, knowledge of how these principles are integrated at the microcirculatory level requires the variability and heterogeneity to be taken into account. Indeed, there is much reason to believe that heterogeneity is an important part of microcirculatory control.

Capillary transport of macromolecules: pores and other endothelial pathways

Can a pore or pore-equivalent model account for transport of macromolecules across microvascular endothelium, or are alternate nonpore pathways necessary? Pores may be defined as aqueous channels of any shape or configuration, including those through a fiber matrix. Such pathways exhibit selective restriction to passage of macromolecules depending on their size, shape, and electrical charge. At least two pore pathways (small and large), differing in both sieve-element spacing and in hydraulic conductivity by an order of magnitude, are required to account for observed size selectivity for plasma proteins of similar shapes and charges. For the two organs examined critically in this review (cat ileum and dog paw), transport of macromolecules through small and large pore pathways is predominately convective. Total transport through small and large pores (alternatively, narrow and wide slits or fine and coarse fiber matrices) is insufficient to account for observed transport rates at low-to-moderate levels of volume flow. Either the estimated pore sizes and hydraulic conductivities derived from measurements of high volume-flow sieving are incorrect or other nonconvective transport pathways contribute substantially to macromolecular transport at low (normal) volume flow.

Flow and distribution of India ink in microvessels of the frog

Velocities of an India ink front and of RBCs moving ahead of it were studied by intravital microscopy in capillaries of frog mesentery and skeletal muscle. Measurements were made during microperfusion of single vessels or groups of connected vessels in mesentery, and following intravenous ink injection in both tissues. The presence of ink did not appear to interfere with microvascular flow or vasomotion during the period of observation. On the average, the ink spearhead moved only slightly faster than the RBCs. There was substantial variation in relative velocities of RBCs in the same vessel and in the relative velocities of ink front and RBCs. The time course of ink filling showed substantial heterogeneity of flow in mesentery, more nearly uniform flow in skeletal muscle. Comparison of the measured velocity ratios of ink to RBCs with published observations on relative velocities of RBC to blood suggest that the advancing, apparently parabolic front of ink moves at less than twice the mean blood velocity. This is due in small part to diffusive dispersion of the ink particles in the laminar flow gradient, but more largely to stochastic dispersion of the front by interaction with RBCs and by displacement at branches.

Protein concentration of lymph and interstitial fluid in the rat tail

Lymph was collected from tail lymphatics of anesthetized rats, subcutaneous interstitial fluid was obtained by implanting nylon wicks, and tendon interstitial fluid was obtained by centrifugation of pieces of tendon. Spontaneous lymph flow rates averaged 70 nl X min-1 X g skin-1. Protein concentrations and colloid osmotic pressures of sampled fluids differed significantly. Tail lymph had the highest protein concentration relative to plasma [lymph-to-plasma ratio 0.71 +/- 0.03 (SE) n = 10], followed by wick fluid (0.62 +/- 0.02, n = 9), with tendon fluid lowest (0.50 +/- 0.03, n = 10). Albumin and immunoglobulin G (IgG) concentrations in samples of tail skin and tendon were assayed by rocket immunoelectrophoresis. Comparison of their distribution volumes at lymph or tendon fluid concentrations, respectively, with interstitial fluid volumes measured as 2-h 51Cr-ethylenediaminetetraacetic acid space minus 5-min 125I-albumin space indicated that 50-60% of the interstitial volume in these tissues is not available for distribution of albumin or IgG. Low lymph flow and high interstitial protein content of rat tail indicate a slow turnover of interstitial protein. This suggests that interstitial washout of protein plays a role in limiting edema only after a sustained or chronic increase in fluid filtration.

Effects of hypoproteinemia and increased vascular pressure on lung fluid balance in sheep

We compared the effects of a sustained decrease in plasma oncotic pressure on lung fluid balance with those of an increase in vascular pressure in six unanesthetized sheep. Initial plasma protein concentration of 58.0 +/- 2.2 (SE) mg/ml was quickly reduced to 34.0 +/- 1.4 mg/ml via plasmapheresis and held at this value for 24 h. Red cells were returned with lactated Ringer solution infused at a rate adjusted to maintain central venous pressure; cardiac output and pulmonary vascular pressures also remained at base line. Steady-state lymph flows increased from a base-line value of 8.8 +/- 3.2 to 20.1 +/- 5.6 ml/h, while the lymph-to-plasma protein concentration ratio ( [L/P] ) decreased from 0.65 +/- 0.03 to 0.44 +/- 0.04. Decreased lymph protein resulted in reestablishment of base-line plasma-to-lymph oncotic gradient. The increased lymph flow was not the result of increased filtration forces, since all vascular pressures and the oncotic gradient were unchanged; nor was it due entirely to increased surface area since [L/P] was decreased. The decrease in plasma oncotic pressure, delta pi P, was twice as effective at increasing lymph flow (1.66 ml X h-1 X mmHg-1, delta pi P) as an equivalent increase in microvascular pressure, delta PC, at normal plasma protein concentration (0.82 ml X h-1 X mmHg-1, delta PC). Elevation of microvascular pressure during hypoproteinemia had a greater effect on lymph flow (1.44 ml X h-1 X mmHg-1, delta PC) than at normal plasma protein concentration.(ABSTRACT TRUNCATED AT 250 WORDS)

Mechanisms for redistribution of plasma protein following acute protein depletion

Acute plasma protein depletion is followed by a rapid and substantial replenishment of the protein deficit. We studied the effects of plasmapheresis on flow and composition of peripheral lymph in 11 unanesthetized sheep. Whole blood was replaced with red blood cells and lactated Ringer solution to reduce plasma protein concentration ([P]) 26-54%. At 24 h after plasmapheresis, [P] had returned halfway to base line. Lymph flow (L) increased immediately after plasma protein reduction, was maximal 3 h later, and remained elevated for more than 3 days. The increase in L was coupled with a decrease more than 3 days. The increase in L was coupled with a decrease in lymph-to-plasma protein concentration ratio ([L/P]). The plasma-to-lymph oncotic gradient was reestablished by 24 h due to the reduction in lymph protein and the partial return of [P]. After 24 h, L remained elevated despite base-line levels for all measured vascular pressures and plasma-to-lymph oncotic gradients. Although lymph flow was increased, the permeability-surface area product for protein was decreased below base line. The data confirm that the partial return of [P] in the first day after plasmapheresis is due largely to a shift of extravascular protein mass into the vascular compartment and show that redistribution is initiated by increased lymphatic return and maintained by a sustained increase in L and a decrease in protein permeability of the plasma-lymph barrier.

Early changes in fiber profile and capillary density in long-term stimulated muscles

Predominantly fast skeletal muscles of rabbits [tibialis anterior (TA), extensor digitorum longus (EDL)] were stimulated at a frequency naturally occurring in nerves to slow muscles (10 Hz continuously) for 8 h/day for 2--4 days. Such stimulation is known to convert all glycolytic fibers to oxidative and to increase capillary density. Our aim was to study early stages of conversion to investigate the factors responsible for the changes. Staining of quick-frozen sections for myosin ATPase, succinic dehydrogenase, and alkaline phosphatase was used to study the distribution of different fiber types and to measure fiber cross-sectional areas, capillaries per square millimeter, and capillary-to-fiber ratios in each fiber category. TA but not EDL showed conversion of fast glycolytic to fast oxidative fibers after 2 days, more after 4 days of stimulation. In both muscles, the largest fast glycolytic fibers were diminished in number after stimulation. There was significant increase in total capillaries per square millimeter after 4 days and some increase after 2 days of stimulation. The increase in capillaries per square millimeter exceeded the increase in the number of fibers per square millimeter, and since there was no change in mean fiber area, the increase is attributed to capillary growth. In EDL, there was an increase in the number of capillaries supplying both fast glycolytic and fast oxidative fibers, suggesting that capillary growth precedes fiber type conversion. In TA, the number of capillaries supplying fast oxidative fibers was increased but that to fast glycolytic fibers, was not. This is consistent with capillary growth simultaneous with or following fiber conversion. In both TA and EDL the number of capillaries perfused after contraction was higher in stimulated muscles, suggesting that increased capillary flow contributed to capillary growth.

The effects of hypoproteinemia on blood-to-lymph fluid transport in sheep lung

We studied the effects of reducing the plasma protein concentration on flow and composition of pulmonary lymph in 12 unanesthetized sheep. Whole blood was removed while red cells were returned and lactated Ringers was infused at a rate sufficient to maintain pulmonary vascular pressures at baseline values. A 44-54% reduction in plasma protein concentration resulted in a decrease in the plasma oncotic pressure from 18.6 +/- 1.1 to 7.8 +/- 0.9 mm Hg. Within an hour after plasmapheresis, lymph flows increased to a maximum of 4 times baseline. Subsequently, lymph flow gradually decreased and were close to baseline at 24 hours. The plasma-to-lymph oncotic gradient was reestablished in 5 hours due to decreased lymph protein. Maintained elevation of lymph flow with hydrostatic and oncotic gradients at baseline values suggest that the blood-to-lymph barrier offers less resistance to fluid transport. The calculated filtration coefficient increased 2- to 3-fold after plasmapheresis. Protein clearances remained normally coupled to lymph flows. Thus the enhanced fluid transport cannot be attributed to a permeability change in the large pore pathways. Hypoproteinemia may alter the interstitial gel so that there is less resistance to fluid movement. Such changes in fluid conductivity between blood capillaries and lymphatics may augment the lymphatic safety factor against pulmonary edema during hypoproteinemia.

Influence of venous congestion on blood-lymph transport of fluid and large molecules in the heat-injured dog's paw

Hind paws of anesthetized dogs were exposed to moderate continuous heat (45--52 degrees C). Lymph flow (L) was increased to 5.61 (SD 3.26) times control levels, lymph:plasma concentration ratios for total plasma protein (RT) were increased to 3.23 (SD 1.33), and for exogenous Dextran-110 (RD) to 8.41 (SD 4.00) times control levels. Selectivity of the blood-lymph barrier for individual plasma proteins was also decreased. Elevation of venous pressure during heating resulted in an increase of L in 5 out of 8 experiments, with no consistent change in RT or RD. Increased protein flux was thus closely coupled to increased volume flow. These observations are consistent with the presence and persistence of intercellular openings or gaps in the microvascular endothelium.

Filling of microcirculation in skeletal muscles during timed India ink perfusion

Lower leg muscles of anesthetized rabbits were perfused in situ with heparinized India ink at flows and pressures comparable to normal resting levels of 4 ml.min-1.100 g-1). Paired cross sections were counterstained with eosin to show ink-containing microvessels and reacted for alkaline phosphate to show all vessels. The fraction of microvessels filled with ink (Fi) increased progressively with perfusion time. At 3.5 s, mean Fi for the muscles studied fell between 0.12 and 0.19. At 60-90 s, the following levels were reached: medial gastrocnemius 0.74, lateral gastrocnemius 0.76, tibialis anterior 0.59, and soleus 0.80. The number of open capillaries and their distribution of flow velocities can be inferred from such data only if the perfusion rate is known or by recourse to a specified anatomic model. The time course of ink appearance shows best agreement with 60-80% of the vessels open and accessible to ink, with microvascular transit times ranging from less than 3 to greater than 30 s. If microvascular path lengths are assumed to be uniform, the range of velocities must be four times to one-fourth the mean, with 15-30% of the microvessels perfused at velocities equal to or greater than the mean. Alternatively, if microvascular velocity is assumed uniform, flow path lengths must vary from one-fourth to four times the average. Anatomic measurements of other suggest that less than one-half the variability in ink transit is attributable to differences in microvascular length. Thus both length and velocity must vary among alternate arteriovenous pathways.

Early kinetics of large molecule transport between plasma and lymph in dogs

Using a mathematical model in which the interstitial space resembled a gel filtration column, Watson and Grodins (Microvasc. Res. 16: 19-41, 1978) predicted that molecules larger than serum albumin should have an early approach to steady state in lymph faster than that of serum albumin. This hypothesis was tested in dogs anesthetized with pentobarbital sodium by use of radiolabeled albumin and fibrinogen and a fluorescent dextran and by collecting lymph from cannulas in the prepopliteal lymphatics of the hind paws. The steady-state lymph-plasma concentration ratios (R), measured chemically, were 0.24 +/- 0.02 (SE, n = 16) for albumin and 0.040 +/- 0.004 (n = 16) for fibrinogen. In 5 h, albumin and fibrinogen tracers had reached approximatly 25 and 50% of their steady-state levels, respectively. The dextran (mol wt 1.5 x 10(5)) behaved like fibrinogen. The data support the hypothesis that the interstitium has nonrestricting pathways, and that there are gel regions that have a low permeability to very large molecules.

Exclusion of plasma proteins in interstitium of tissues from the dog hind paw

Exclusion of plasma albumin and fibrinogen from portions of the extravascular sucrose space in tissues of the hind paw of anesthetized dogs was studied by use of radioactive traces. Thirty hours after a bolus injection of the labeled proteins, tracer concentrations were measured in plasma, prenodal popliteal lymph, and samples of hind paw skin, tendon, muscle, and toe pad. In a few experiments, constant plasma levels of tracer sucrose were maintained for 2 h before terminal sampling. In all experiments a vascular volume marker was given 15 min before termination. A linear correlation was found between the amounts of extravascular protein tracers in the tissue samples and their concentrations in lymph from the same paw, as would be expected if their actual interstitial concentrations were closely related to their lymph concentrations. For albumin, the extravascular distribution volume calculated at lymph concentration was less than the sucrose space for all tissues except toe pad. For fibrinogen, the extravascular distribution volume was less than albumin and sucrose for all tissues except skin. Differences in pattern and degree of exclusion may reflect the differences in interstitial gel-phase composition and structure.

Multiple pathways of capillary permeability

Morphological and physiological studies indicate multiple routes for transport across capillary endothelium. However, the identification of the morphological counterparts of specific transport processes (or the assignment of specific transport roles to morphologically identifiable pathways) has been only partly achieved: the contribution of endothelial cell membranes to transport of water and small, lipid-insoluble molecules needs to be evaluated. The identification of the "small pore" pathway for water and lipid-insoluble molecules with the intercellular junctions still remains questionable. The contributions to total macromolecular transport of junctions, single vesicles (pinocytosis, cytopempsis), chains of vesicles, and fenestrae are not yet known.

Influence of sympathetic nerves on oxygen uptake or resting mammalian skeletal muscle,

Gracilis muscles of anesthetized dogs were isolated and perfused with blood at constant flow. Oxygen uptake (Vo2) was measured by spectrophotometry of paired arterial and venous samples and monitored by continuous recording of HbO2 saturation. The sympathetic chain, connected to the isolated muscles, was stimulated electrically with supramaximal pulses. High frequencies (8-16/s) increased vascular resistance (R, ratio to control 1.86 +/- 0.6 SEM) and decreased Vo2 (0.61 +/- 0.02). Both effects could be maintained 10 min or longer. After stimulation there was an increase in Vo2, which could be less than equal to, or greater than the O2 deficit, depending on the stimulation frequency, the period of stimulation, and the blood flow. Low frequencies (0.5-4/s) increased R (1.39 +/- 0.04) and increased Vo2 during (1,18 +/- 0.03) and after stimulation. In several experiments there was an initial decrease in Vo2 which changed to an increase during the stimulation period while R remained elevated. These observations indicate that the decrease in Vo2 is due to closure of precapillary sphincters, whereas the increase is a direct effect on cell metabolism.