The effects of a single injection of a bacterial endotoxin on the cellular changes of a primary immune response to a standard dose of sheep red blood cells were studied in the spleens and mesenteric lymph nodes of mice. Daily histological comparisons of these organs in mice, injected with endotoxin, or with antigen, or both, showed that endotoxin given simultaneously with sheep red blood cells, as antigen, significantly enhanced all of the cellular changes that appear in the mesenteric lymph nodes and spleens of mice that form antibody when that antigen is given alone. First, in the white pulp of the spleens and cortical regions of the nodes, there appeared an early and excessive proliferation of the large pyroninophilic cells which seems to be responsible for the earliest formation of antibody, as judged by this work and that of others cited in the body of the paper. Polymorphonuclear cells invaded the spleens of these animals early after simultaneous challenge with antigen and endotoxin, and in far greater numbers than have ever been seen in mice given the same antigen without endotoxin. "Activated" germinal centers formed in the lymphoid tissue either 1 day before the appearance of antibody in the blood stream or on the same day, and they became larger than in the mice given antigen only. On the other hand, these specific and characteristic cellular changes failed to appear in mice prevented from forming any antibody at all by injections of endotoxin given 2 days before the antigenic challenge. These findings are discussed in the light provided by data from recent reports of others as well as in the light of the accompanying paper (1) which demonstrated not only the enhancement of antibody formation following simultaneous injections of antigen and endotoxin, as already known, but a totally unexpected, complete suppression of its formation when endotoxin was given 2 days before antigen.

The manner in which a single injection of S. typhosa endotoxin effects the primary hemolysin response to sheep erythrocytes in the mouse has been shown to depend on the dosage, route, and time of administration of the endotoxin, as well as on the route employed for the injection of antigen. The normal production of antibody, following an intravenous or an intraperitoneal injection of red blood cells, is suppressed if the bacterial lipopolysaccharide is given before and by the same route as the antigen. The response to an intraperitoneal injection of sheep red cells is also inhibited if preceded by an intravenous injection of endotoxin. By contrast, hemolysin formation to intravenous antigen is enhanced considerably by a previous intraperitoneal injection of endotoxin, and the response both to intravenous and to intraperitoneal injections of the antigen increases if the endotoxin is given by the same route either simultaneously or shortly after the foreign red cells. These findings are discussed in regard to the physiological action of bacterial endotoxins and the early events in antibody formation.

The findings presented in this paper indicate that 60 gm. of liver tissue, taken from adult donor rabbits 2 weeks after injecting them with large amounts of bovine γ-globulin, and transferred to the peritoneal cavities of recipient rabbits prepared for a secondary reaction, contain enough of the antigen to induce the formation of detectable amounts of precipitin in the recipient rabbits. On the other hand, liver taken from donors 21 and 28 days after injecting them with the same antigen fails to bring about a similar effect. The findings outlined in the paper suggest that the antigen is degraded after the 2nd week to such an extent that it fails to engender the formation of complete antibody in the recipient animals, although previous work has shown (2–4) that the transfer of liver tissue from similar donor rabbits to mice renders the recipient mice anaphylactically sensitive to the antigen, bovine γ-globulin. The implications of the findings are discussed.

Following single intravenous injections of a foreign protein antigen, bovine γ-globulin, into mice and rabbits, antigenic material persisting in the liver could be detected for several weeks. Ground liver tissue—taken from the mice and rabbits, just mentioned, either 4 or 6 weeks after injecting the antigen—when transferred repeatedly, at 2 or 3 day intervals, to the peritoneal cavities of normal, or unilaterally adrenalectomized, recipient mice, rendered the recipients sensitive to active anaphylaxis when they were challenged after a suitable interval by intravenous injections of the original antigen. The work throws some light on the state of the antigenic material that persists for 4 to 6 weeks in the livers of the donor animals. Obviously it is sufficiently unchanged, at least in its reactive groups, to engender in the recipient mice antibodies capable of reacting with the original antigen.

A sensitive biological test has been used to detect the persistence of minute traces of a foreign protein, bovine γ-globulin, in the blood and livers of rabbits intravenously injected with it, as an antigen. At various intervals after injecting these rabbits (donors) serum or liver tissue was transferred from them to the peritoneal cavities of normal or unilaterally adrenalectomized mice (recipients) with the aim of rendering the latter hypersensitive to the antigen that might be persisting in the transferred materials; a state of affairs detectable, 2 days later, by the appearance of signs of reversed passive anaphylaxis when the recipient mice were intravenously challenged with a strong anti-bovine γ-globulin rabbit serum. The protein persisted in the blood of the donor rabbits, in readily demonstrable amounts for 1 month, and in the blood of one animal, in minute traces, or as long as 6 weeks. It was detectable in the livers for 8 weeks. The persistence of bovine γ-globulin in rabbits, which form circulating antibodies to it well, is not as long as that in mice, which form antibodies to it poorly, since in previous work with the mouse the antigen was found (1) in the blood after 8 weeks and in the liver for 14 weeks. Nevertheless the antigen persists in the rabbit much longer than is generally supposed. Indeed it can be found in the liver all through the period in which circulating antibody is demonstrable in the blood. Explanations for the phenomenon have been suggested. Its significance in relation to the mechanisms of antibody formation is obvious.

Methods have been devised whereby the persistence of foreign antigens in mice can be detected. A highly diffusible, blue azo dye, echt-säure-blau was coupled to bovine γ-globulin and human serum albumin and injected into the animals. In this way deep blue tracer antigens were obtained. These were promptly stored in cells widely distributed throughout the body, especially in the reticulo-endothelial elements of the liver, spleen and mesenteric lymph nodes. The dye as such was not stored but rapidly excreted. A blue coloration in the organs just mentioned was still perceptible after 85 to 120 days in the case of the azoglobulin and 36 to 44 days in that of the azoalbumin. To determine whether these substances had actually persisted, as well as to learn how long uncoupled globulin and albumin remained after injection, recourse was had to the phenomenon of reversed passive anaphylaxis, which was found to be characterized by extraordinary changes in the vessels of the ears (EVR) in the mouse, plainly visible under the microscope when called forth by an antiserum specific for the antigen to which the animal had been sensitized. So sensitive is the vascular response that as little as 0.5 to 0.1 µg. of protein as antigen, previously injected into the peritoneal cavity of a 30 gm. mouse, can be detected a few days later by an intravenous injection of antiserum. By means of the EVR the globulin antigen has been detected in the blood and livers of injected mice for as long as 56 and 101 days, respectively; the albumin and azoalbumin for only 21 and 36 days. In the mesenteric lymph nodes of injected mice the albumin and azoalbumin antigens were found after 42 and 44 days, respectively. The hepatic tissue and that of the mesenteric lymph nodes of mice injected with azoalbumin, containing in consequence stored blue material, when transferred to recipient mice yielded positive tests for antigen (EVR) as long as blue color could be perceived in these tissues with the unaided eye, or at low magnification. After the color had disappeared from the tissues the transfer tests were found to be negative. This fact speaks for the antigenicity of the colored material. In summary it is plain that certain antigenic proteins, after introduction into the blood stream of mice, are stored in certain tissues and that they may persist there for weeks or even months, far longer than has generally been supposed. This persistence of antigen within the body, especially after detectable amounts of antigen have apparently disappeared from the blood, provides a reason for prolonged antibody formation, a phenomenon for which no adequate explanation has hereto been offered.

Intensely blue dye-azoproteins have been prepared by diazotization and coupling of the highly indiffusible blue dye T-1824, Evans blue, with various serum proteins and egg albumin. The products, whether purified by precipitation with alcohol or by chromatography, have a constant dye-to-protein ratio and tests have shown them to be essentially free from unlinked dye. An extremely diffusible dye, echt-säure-blau, has also been coupled to bovine γ-globulin. These materials are adapted to physiological experimentation. They seem to behave in the bodies of mice like other proteins; they fail to appear in either the bile or urine of normal animals, and they are strongly antigenic. When these soluble antigenic azoproteins are injected into the blood stream of mice for the first time they enter reticulo-endothelial cells in almost every organ of the body; the final distribution is like that of intravenously injected, finely divided particulate matter. The azoproteins appear in the cells which classical immunological studies have shown to be active in removing particulate antigenic materials or bacteria from the blood or body fluids. The Kupffer cells of the liver and sinus and reticular cells in lymph nodes, especially the great mesenteric node, are particularly active in the removal of the blue antigens from the blood, but many other R-E cells are active to a lesser degree. The storage of the antigenic material is in the cytoplasm only; it has not been seen within nuclei, nor has it been seen within cells of the brain. Serological methods disclose that the blue material seen within Kupffer cells of the liver after as long a period as 2 days is still antigenic in its reactions. The blue azoproteins, therefore, serve excellently as tracer antigens, especially since they can be seen directly in fresh and fixed tissue preparations and in the body fluids.

Pronounced vascular changes occurring in the ears and claws of mice during anaphylactic shock are described. Practically at once after a foreign serum (pig, horse, or rabbit) enters the blood stream of sensitized animals both the arterial and venous vessels undergo marked, local or generalized constriction in the organs mentioned. Usually spasm of the vessel walls occurs simultaneously in the arteries and veins, but it may appear first in the arteries, or occasionally in the veins. When venous spasm precedes arterial spasm, the true capillaries become distended with cells; if the reverse order holds, the ears appear bloodless. There is no active constriction or dilatation of capillaries; the capillary behavior follows passively the changes in the large vessels. Peripheral vascular spasm occurs while the carotid blood pressure is high, but a few minutes later, while this still holds true, the ear vessels begin to relax and the circulation is resumed. Shortly afterwards the blood pressure falls to levels far below normal, but the vessels remain open. If the circulation of one ear is obstructed while anaphylactic shock is produced, no vascular spasm occurs in it. Release of the obstruction during the animal's recovery results in belated constriction of the blood vessels of this ear although by now the vessels in the other ear are dilated and the general systolic blood pressure is very low. The vascular reactions in the ears appear to be uninfluenced by the blood pressure in the large vessels, and they are not a response to nervous stimuli. They are local in origin. The vascular changes are often not clearly perceptible in the gross but are plainly to be seen under a low power of the microscope. They occur in some sensitized mice exhibiting no manifest signs of shock, differing only in degree from the changes taking place when shock is severe or fatal.

The pressure in the cutaneous lymphatic capillaries of normal mice anesthetized with nembutal ranged between 0.0 and 2.7 cm. of water. Measurements of the interstitial pressure in the tissue immediately next the lymphatics showed that, in more than half the instances studied, there was a slight gradient of pressure from the tissues to the lymph. In nearly all the other instances the pressures inside and outside the lymphatic capillaries were approximately equal. In two cases in which lymph flow in the capillaries was rapid, the lymph pressure may have been negative. Under these circumstances there must have been a considerable gradient of pressure from the tissues to the lymph. In skin which was rapidly becoming, or had recently become, edematous as result of the application of xylol or of heat, the intralymphatic capillary pressure generally was increased, yet when compared with the pressure prevailing in the edema fluid outside of the capillaries it was usually found to be relatively much lower, at times by as much as 5.9 cm. of water. The findings indicate that a pressure gradient is an important factor in lymph formation under normal and pathological circumstances.

Means have been described for the study of pressure conditions in normal and pathological skin of living human beings and mice. The true pressure in normal skin cannot be measured directly by any of the means hitherto described, because there is insufficient free fluid to make manometric determinations. However, for practical purposes, the intracutaneous pressure has been approximately estimated by introducing into skin exceedingly small amounts of a relatively unabsorbable fluid, a mixture of Locke's solution and a vital dye, and then finding the least pressure required to overcome the resistance of the skin to the passage of this fluid through it at the lowest rate measurable with accuracy by the apparatus at hand. In the present paper measurements of this pressure have been termed the interstitial resistance. In normal skin the interstitial pressure, as estimated by measurements of the interstitial resistance, is low, slightly less, on the average, than 1.7 cm. of water in the skin of the mouse, and less than 3.1 cm. of water in human skin. It remains unchanged in states of active hyperemia. In edematous skin the interstitial pressure can be directly measured by determination of the edema fluid pressure. It has been compared with determinations of the interstitial resistance and found to be only 0.5 cm. of water lower in both the mouse and man. Under the conditions of our experiments, in skin rendered slowly edematous by the introduction of irritant chemicals or their topical application, little rise in pressure took place. On the other hand, in rapidly forming edema of the skin the edema fluid pressure and the intradermal interstitial resistance rose and became great enough to hinder materially the further escape of fluid from the blood vessels. The edema fluid pressure rose in proportion to the rapidity with which the edema formed. When a rapidly formed edema subsided, the edema fluid pressure and interstitial resistance fell, but if inflammation and induration followed later, the interstitial resistance became high again. As these conditions subsided the interstitial resistance fell, at times to normal levels, even in the presence of edema. In mouse skin injured by squeezing according to a standard procedure, with result in pronounced edema, the intradermal interstitial resistance rose within a few hours to levels of 10 to 15 cm. of water. In those instances in which the injury progressed to induration, the interstitial resistance rose to such high levels that it seemed impossible that fluid could continue to escape from the capillaries. Such a state of affairs may be of great importance in determining whether necrosis follows trauma.

The changes of intracutaneous pressure in the limbs of mice and human beings have been followed during and after periods of venous obstruction with almost unhindered arterial flow. During the first 30 minutes of obstruction the interstitial pressure in the tense skin of the lower legs of mice, a pressure which is slightly higher than that in the loose skin of the ears, backs, and thighs (21), rose from 2.6 to 4.6 cm. of water to about 32 cm., thereafter remaining constant. It would appear that the escape of fluid from the capillaries is checked at this pressure. In the skin of the arm and leg of man the interstitial pressure rose from 2.5–3.7 cm. of water to 15.0–23.0, within 15 to 27 minutes after venous obstruction had been produced, mounting no higher during the period of observation. When venous obstruction had existed for about 20 minutes or more the subjects sometimes experienced sensations of relief from congestion as if some tissue adjustment or the opening of some venous by-pass in the marrow had occurred, preventing a further rise of pressure. However this may be, the pressures still appeared to be great enough to prevent further escape of fluid from the capillaries, at least for the time being.

Advantage has been taken of the relative transparency of the claw of the mouse to devise a method, here described, to measure the blood pressure in the animal's leg. Direct measurements of the systolic blood pressure from the carotid arteries of anesthetized mice have also been made. Simultaneous blood pressure readings by both these methods applied to the same animal showed close agreement. The systolic pressure ranged from 60 to 126 mm. Hg, according to the conditions.

With the aim of determining the structural conditions which affect fluid movement in the cutaneous connective tissue of mice, various test fluids were brought into contact with it under conditions such that neither blood vessels nor lymphatics were directly entered. Locke's solution, mouse serum, and a mixture of Locke's solution with a dye which causes edema were all employed. At atmospheric pressure, Locke's solution entered the tissues intermittently. When subjected to very low pressures it continued to enter the skin intermittently and at approximately the same rate. At pressures above 4.5 cm. of water, however, the flow became continuous but it did not increase in rate significantly until pressures of about 8.5 cm. were employed. There was no relationship between the rate of flow and the pressure employed. At a pressure of about 8.5 cm. the resistance of the tissues seemed to give way abruptly as if the formed elements had been separated. This has been termed the "breaking point." After it had been reached each further increase of pressure produced a proportionately greater inflow. Under the conditions of our experiments, the dye-Locke"s solution and also the homologous serum failed to enter the tissues at atmospheric pressure. It was necessary to subject these fluids to pressure to force them into the skin at the same rate at which the Locke's solution entered it spontaneously. Under these circumstances the dye-Locke's solution and the serum entered the skin continuously, not intermittently like the plain Locke's solution. As the pressure was gradually raised, no significant increase of flow into the tissues occurred until a point was reached, on the average 8.5 cm. of water, at which fluid suddenly began to enter very rapidly. This point, the "breaking point" already mentioned, was reached at the same pressure irrespective of the character of the fluid employed, showing that the phenomenon was produced by the fluid bulk. Once it had been attained, further increases in pressure caused proportionately greater inflow of fluid. The circulation had nothing to do with the phenomenon, for it occurred in the skin of dead mice." The findings indicate that under normal circumstances the movement of fluid in the interstitial tissue does not take place as though in preexisting channels. The experiments confirm previous observations from this laboratory (13, 14) that in normal skin tissue the state of affairs is such that fluid cannot flow freely. However, when fluid is introduced into the skin under pressure spaces are forcibly opened up. Inflammatory edema in the skin changed the phenomena of fluid entrance into it under pressure. The reason is that there then occurred a separation of the formed elements and the interstitial fluid moved as in preformed channels. Even when very low pressures were employed (3.0 to 7.0 cm. of water), there appeared usually a linear relationship between the pressure and the rate of flow.

Methods have been devised to bring microscopic amounts of fluid into contact with cutaneous connective tissue, under pressure or without pressure, in such a manner that it enters neither blood capillaries nor lymphatics directly. The take-up of fluid brought into contact with the tissues in this way has been measured and its characteristics studied. Elaborate control tests, here described and discussed, have indicated the possible errors in the employment of these methods. Locke's or Tyrode's solutions brought into contact with the cutaneous tissues, by the method described and at atmospheric pressure, pass into the tissue intermittently. Forced into the skin by pressures of 1.0 to 2.0 cm. of water the take-up is still intermittent in character. From this it follows that either the absorption of interstitial fluid from localized regions is periodic or the movements of interstitial fluid are influenced by intermittent physiological changes.

Minute amounts of Locke's or Tyrode's solution have been brought into contact with the interstitial connective tissue of the skin of the living mouse, at atmospheric pressure, in such a manner that the blood or lymphatic vessels are not entered directly. Under such circumstances these absorbable fluids enter the tissue spontaneously. Entrance is strikingly intermittent, not continuous, and so too when very slight pressures are brought to bear on the fluids (1). Hyperemia of the tissues, with accompanying dilatation of the blood vessels, increases the entrance of fluids at atmospheric pressure but it is still intermittent. By contrast, venous obstruction leads to intermittent backflow into the apparatus, but reflex hyperemia, following release of the obstruction, is attended by an increase of flow into the tissues in spite of the great reactive dilatation of vessels. The inflow is also intermittent. If the skin is deprived of circulation, fluid does not enter it at all at atmospheric pressure, though it moves in regularly and continuously if slight pressure is put upon it. Edema-forming fluids, described in the text, also enter in a continuous manner when forced into the skin of either living or dead animals. So too do serum and sperm oil. The findings indicate that the passage of interstitial fluid into the blood vessels may be intermittent under normal circumstances and its escape from them as well. The observed occurrence of intermittent flow in the blood vessels of several tissues (9, 15–25) will go far to account for the intermittent entrance of fluid into the skin.

The escape of a vital dye from the lymphatics of the ears of living mice and its subsequent movement through normal and pathological connective tissue have been observed at high magnification. The dye first appears outside such channels as bristly, wavy lines of color, which can be bent and twisted by pressure with a micro probe and spring back to their original positions when the pressure is removed, as if the dye were fixed upon or between some tissue elements. Our findings indicate that this is the case, that the bristly lines of color are formed by dye moving between connective tissue fibers or along them. With the onset of mild edema, such as the dye induces secondarily, the bristles disappear, the coloration becoming diffuse and freely movable with the micro probe. When edema is induced before dye is introduced into the lymphatics, the character of its escape is wholly different. It first appears as a colored cloud, freely movable in the edema fluid, the manner of its passage into the tissues being completely changed. In the ears of mice partly dehydrated by bleeding, or in those of dead animals, the bristly or wavy lines were more evident than in normal individuals. It was plain that dehydration did not change the mode of transportation of the dye through the tissue but merely emphasized some of the characteristics of its passage. In animals injected intravenously with large amounts of physiological saline, with result in the presence of more tissue fluid than usual, the colored bristles were seldom seen. It is plain that connective tissue fibers serve indirectly as pathways for the transport of substances of large molecule. We have not been able by the dye method to demonstrate the presence of any free fluid in the normal tissues of the mouse ear.

The interstitial movement of several dyes of widely different chemical constitution and diffusibility, in the connective tissues of the mouse ear, has been observed at high magnification. Dye extension seems to be conditioned by the form and structure of the connective tissue fibers. After escaping from the lymphatics of the ears of living mice, each dye appeared first in the tissues as bristly projections of color. These bent and twisted when pressed upon by a micro probe but sprang back into place when the pressure was removed. The present work and the preceding have shown that the lines of color are formed by dye along or between connective tissue fibers. Intermittent external pressure applied to the tissue, squeezes and bends the fibers together and greatly increases the spread of dye along them. The connective tissue fibers assume an important role in the spread of substances through tissues subjected to pressure changes. The experiments have given evidence of the existence of a tissue matrix in the organ but none of the presence of free interstitial fluid. In tissue subjected to irritant stimuli and in frankly edematous tissue, free fluid is readily demonstrated. When it is present the method of extension of dye is completely changed. Dye appears in the tissues as a diffusely colored cloud which can be freely moved by pressure with a micro probe. The bearing of this evidence upon the condition of interstitial fluid and the nature of the interstitial spaces is discussed.

A method has been devised to measure the spread of vital dyes in the skin of mice. Spread is greatly influenced by physiological and pathological changes which add fluid to the tissue or abstract it. Spread is greater in the quiet, living ear than in the ear of an animal just killed. It is equally considerable in the frankly edematous ears of living and dead animals, and not greater in either case than in normal, quiet tissues. During the early stages of edema formation on the other hand, dye spread is notably rapid. It is still greater in the ears of normal animals actively moving about, and is greatest in tissues subjected to very gently intermittent changes in external pressure. The significance of these findings is discussed.

The ears of rabbits were perfused with defibrinated rabbit's blood in such a way that pulsation could be imparted to the perfusate or withheld from it at will. In the absence of pulsation there was almost no lymph flow, whereas when it was present lymph flow was rapid despite the fact that the "systolic" pressure of the perfusate never exceeded the constant pressure in the non-pulsatile instances and the volume flow was far less. Non-pulsatile perfusion led to a slight flow of lymph in ears that were becoming edematous, whereas when it was pulsatile the lymph flow was enormous. The pulse exercises an influence to move fluid into the lymphatics and along them.

The pulsation of blood vessels in the ear of the rabbit greatly increases the rate of the spread of dye introduced into the subcutaneous tissue. The appearance of edema in tissues perfused at a constant pressure leads to very little increase in the rate of dye spread. By contrast, a rapid interstitial spread of dye occurs in tissues becoming edematous while perfused with a pulsatile flow of blood. The significance of these facts is discussed.