EXCHANGE OF SUBSTANCES THROUGH CAPILLARY WALLS 



983 



lated average protein content of the nitrate was about 

 0.3 g per 100 ml of filtrate, though with very wide 

 variations. At a venous pressure of 80 mm Hg the 

 average protein content of capillary filtrate was 1.5 g 

 per 100 ml, but this was attributed to leakage because 

 of marked vascular distention and slowed blood flow 

 with possible increased permeability from local 

 hypoxia. Fractionation studies of plasma proteins 

 before and after congestion indicated qualitatively 

 that capillary filtrate contained both albumin and 

 globulin. All that this approach could provide, how- 

 ever, was a rough order of magnitude for the normal 

 protein content of capillary filtrate. It indicated 

 merely that it was, on the average, considerably less 

 than the protein content of lymph. In passing, it 

 should be emphasized that this venous congestion 

 method is of doubtful utility in measuring capillary 

 permeability to protein in individual patients with 

 various diseases (e.g., 6, 32, 33) because normal 

 values vary greatly from subject to subject even 

 under carefully controlled conditions (17, 211). The 

 reasons for this variability are still not clear. Probable 

 factors are streamline flow in veins with sampling 

 errors and, in addition, analytic errors even with 

 triplicate hematocrits and protein analyses. 



Shortly after the studies of Landis et al. (211), White 

 et al. (379) approached the problem more directly 

 and found that venous congestion of the dog's leg 

 reduced the concentration of protein in lymph from 

 0.77 per cent to 0.21 per cent while the rate of lymph 

 production increased sixfold, as would be expected 

 with exclusion of absorption. In another experiment 

 lymph protein fell during venous congestion from 

 1 .24 per cent to 0.78, but lymph flow failed to increase 

 as much as in the previous experiment, and the lymph 

 contained a considerable number of erythrocytes. By 

 still another method Pappenheimer & Soto-Rivera 

 (282) found in the perfused hind limbs of cats that 

 capillary filtrate contained an average of 0.3 per cent 

 protein with values of 0.2 and 0.4 per cent falling 

 within the range of error. 



Thus, in mechanical, i.e., noninflammatory, 

 edemas and in congestion of the human forearm, the 

 capillary walls were found to be approximately 95 

 per cent effective as protein-retaining membranes 

 (207). During rapid filtration the protein concentra- 

 tion of capillary filtrates is extremely small even in the 

 more permeable capillaries of the intestine (168a). It 

 is probable that under these conditions the protein 

 concentration is reduced by molecular sieving as dis- 

 cussed in sections 7D and 10. 



Concerning the fate of the protein that has passed 

 from the blood stream with capillary filtrate into the 



interstitial fluid compartment, information is far from 

 complete though generally in favor of minimal re- 

 entry into capillary blood despite absorption of fluid. 

 Lewis (215) injected horse serum subcutaneously in 

 dogs and detected it in thoracic duct lymph within 

 40 min., but in blood only after 3.5 hours. Field & 

 Drinker (97) also injected horse serum subcutane- 

 ously into dogs and found, by a precipitin reaction, 

 that when all possible lymphatics were blocked no 

 foreign serum could be detected in the blood stream 

 in periods up to 7 hours after injection. They con- 

 cluded that the blood capillaries of the subcutaneous 

 tissues are not ordinarily concerned in the absorption 

 of protein. Courtice et al. (56, 58, 59) injected plasma 

 protein labeled with T-1824, into the peritoneal 

 cavity and recovered nearly all of it in lymph. Jepson 

 et al. (167) followed the removal from skin of protein 

 labeled with I 131 and in normal dogs found that lymph 

 exhibited far more radioactivity than blood plasma. 

 From similar studies in man, Hollander et al. (159) 

 concluded that protein is returned to the circulation 

 chiefly or exclusively by lymphatic flow. In the lung 

 of dogs, however, Drinker et al. (78) found that pro- 

 tein injected into the alveoli appeared first in blood 

 though very slowly and in small amount. 



It seems likely that most of the protein in capillary 

 filtrate normally fails to enter capillary blood, even 

 during absorption of fluid, because such direct return 

 to plasma involves the movement of protein molecules 

 against a considerable concentration gradient. More 

 information is needed because the mechanism and 

 completeness of this exclusion of protein during ab- 

 sorption of fluid is still far from clear, particularly in 

 abnormal states. Thus, Field & Drinker (98) found 

 in dogs with ligated lymphatics that acute plasma- 

 pheresis increased the absorption of foreign protein 

 from the tissue spaces. Jepson et al. (167) found this 

 true of labeled protein in lymphedema also. 



B. Interstitial Fluid; Protein Content and II, 7 



The protein content of interstitial fluid under 

 normal, resting conditions, and the protein osmotic 

 pressure of that fluid, U,/, have also been determined 

 so far only by indirect methods. The protein concen- 

 tration in capillary filtrate being 0.2 to 0.4 g per 100 

 ml, and that in lymph from a resting extremity being 

 1.3 to 3.3 g per 100 ml it follows that the protein 

 content of interstitial fluid under resting conditions 

 lies between these figures and is not uniform. Depend- 

 ing upon localized filtration or absorption, it can 

 range from protein-poor capillary filtrate, just pro- 

 duced, to an interstitial fluid which is protein rich 



