The Validity of Blood and Urinary Cytokine Measurements for Detecting the Presence of Inflammation
Lyle L. Moldawer1
The purpose of this review is to assess the utility of urinary and blood cytokine measurements to document the presence of inflammatory stress, particularly as it relates to field studies and to military performance. This is a daunting challenge for two reasons. First, at least 15 different interleukins have been described, and according to Carl Nathan, there are a considerable number of additional humoral and growth factors (Nathan and Sporn, 1991). However, the focus of this discussion will be directed to the detection of three pluripotent cytokines, specifically tumor necrosis factor-a (TNF-a), interleukin-1 (IL-1), and IL-6.
The reasons to restrict this discussion to these three inflammatory mediators are severalfold. Not only have these three cytokines been implicated as the most
proximal mediators in the proinflammatory cascade, but they also induce secondary inflammatory mediators such as IL-8, some of the macrophage inflammatory proteins, and the neutrophil adherence proteins (Fong et al., 1990a). Another reason to focus on TNF-a, IL-1, and IL-6 is that the actions of these three proinflammatory cytokines are both overlapping and distinct. These three cytokines affect nearly every organ system in the host. In fact, the original description of TNF-a, IL-1, and IL-6 function was due in large part to military research conducted in the 1970s by the research group at Fort Detrick, led by William R. Beisel and contributed to by Robert Wannemacher, Jr., and Michael Powanda (reviewed in Beisel, 1975, 1977, 1980). These investigators described and examined the functional group of proteins, designated leukocyte endogenous mediator (LEM), in the plasma of animals and patients after inflammatory challenges (Powanda and Beisel, 1982; Wannemacher et al., 1972).
It is now known that those mediators, originally described as LEM or as other functional monikers, including leukocytic pyrogen or endogenous pyrogen, are for the most part one or more of these three proinflammatory cytokines, TNF-a, IL-1, and IL-6. One could argue whether LEM, as originally described by Beisel's group, is IL-1 or whether it is IL-6 or TNF-a, but suffice it to say that these three cytokines are the primary mediators responsible for the induction of the proinflammatory responses to infection and inflammation. It is now recognized that the majority of innate responses to bacterial and viral pathogens, including fever, anorexia, weight loss, trace mineral redistribution, neutrophilia, and hepatic acute-phase protein synthesis among others are mediated by the release of these three proinflammatory cytokines (Beisel, 1980; Dinarello, 1994).
The second problem with discussing proinflammatory cytokines and the utility of their measurements in regard to inflammation or military performance is the dual nature of these cytokines. As several groups showed in the 1970s, the effects of these cytokines are primarily beneficial to the infected host (Kampschmidt and Pulliam, 1975; Kluger et al., 1975), including the stimulation of fever and nonspecific host immunity, the activation of macrophages, the redistribution of trace minerals, the induction of a hepatic acute-phase response, and their role as comitogens or adjuvants for cell-mediated immunity or immunoglobulin production (Beisel, 1975; Kampschmidt, 1983). Although these cytokines are inherently beneficial, they can and do often have adverse effects primarily associated with either their exaggerated production or their continued production. The former occurs in gram-negative septic shock and the disseminated intravascular coagulopathy that accompanies it (Beutler and Cerami, 1986; Dinarello and Wolff, 1993). The latter is associated with some chronic inflammatory syndromes such as AIDS or some neoplasms, or their inappropriate compartmentalized production as occurs in rheumatoid arthritis, and often can lead to devastating pathological consequences.
The challenge, therefore, is how to meaningfully interpret urinary or blood measurements of these cytokines in order to identify the presence and magni-
tude of an inflammatory response when it is unknown or unclear whether cytokines are mediating a beneficial or an adverse effect. The question at hand is whether blood or urinary cytokine measurements can be used to make decisions regarding therapy or performance in a field setting.
INFLAMMATION, SEPSIS, AND PLASMA CYTOKINES
Following purification and cloning of IL-1, TNF-a, and IL-6 in the early to mid-1980s, immunoassays were developed that readily could detect the presence of these proteins in the blood and tissue fluids of humans. At that time, there was a great deal of enthusiasm for the documentation of TNF-a and IL-1 in the plasma of patients with a variety of inflammatory processes. It initially was believed that these cytokines acted as endocrine factors and that they would be routinely measurable in the circulation of patients with infections or inflammation. However, these conclusions were drawn initially from studies based on patients, rodents, and primates who were administered live bacteria or endotoxin (Fong et al., 1990b; Hesse et al., 1988; Michie et al., 1988). When healthy volunteers or primates were given endotoxin or live E. coli bacteria, respectively, a monophasic rise in plasma TNF-a concentrations was seen, followed in some cases by a monophasic rise in IL-1 and then a later rise in plasma levels of interferon, IL-6, and other cytokines (Fong et al., 1990b; Hesse et al., 1988). The intensity of the cytokine appearance often correlated with the magnitude of the physiologic responses. These data, from patients who received small amounts of endotoxin, were first published by Douglas W. Wilmore's group in the New England Journal of Medicine (Michie et al., 1988). The generalized assumption, thereafter, was that these cytokines circulate frequently in the plasma in an endocrine fashion.
At the same time that these data were being presented, Anders Waage from Norway and other groups were reporting plasma cytokine measurements from septic patients and patients with other inflammatory processes (Waage et al., 1986, 1987, 1989). These investigators for the most part were only able to document an occasional appearance of TNF-a and IL-1 in the circulation of patients with acute inflammation. Unlike the observation that TNF and IL-1 are detected uniformly in animals or patients administered endotoxin or E. coli, these and other authors were able to detect TNF-a and IL-1ß in fewer than 40 percent of patients with sepsis syndromes. This is a very important observation since it suggests that the detection of plasma TNF-a and IL-1 appearance is not essential to the manifestation of an inflammatory response. Previous studies at Cornell University by Michael Marano and Stephen Lowry examined plasma cytokine concentrations in burn patients who sustained greater than 35 percent total-body surface area thermal injury (Marano et al., 1990). Blood was collected every 4 hours from the time of admission for as long as the patients were critically ill, at times extending to periods of up to 60 or 70 days. Retrospective analyses were then performed, examining which patients became septic, which
did not become septic, which survived, and which died, and these were compared with the frequency of being able to detect TNF-a in the circulation.
Statistically, there was a greater frequency of detecting TNF-a in the serum of septic patients compared with nonseptic patients. Patients who eventually died (primarily from overwhelming gram-negative sepsis) had an even higher frequency of detecting TNF-a, but a large number of patients who died from sepsis had no detectable TNF-a. In fact, in only about 40 percent of septic patients was TNF-a detected. IL-1 was detected even less frequently. Only in about 4 to 5 percent of these patients was IL-1ß measurable in the circulation, despite the fact that these patients were critically ill and, in many cases, floridly septic. So, contrary to what the experimental model suggested, it was not a common finding to detect TNF-a or IL-1 in the circulation of critically ill patients, in whom increased production would be expected.
The plasma response of IL-6 to inflammatory stimuli is entirely different, however. In contrast to TNF-a or IL-1, IL-6 levels are frequently elevated in patients with inflammation, and the magnitude of the IL-6 response is often of some prognostic value. The University of Pennsylvania group (Moscovitz et al., 1994) evaluated 100 patients who were presented to the emergency room with the clinical criteria of systemic inflammatory response syndrome (SIRS) and screened them for plasma levels of TNF-a, IL-1, and IL-6. What they found was similar to what this laboratory found in burn patients, in that the frequency of detecting TNF-a and IL-1 in the circulation of patients with SIRS was very low and nonpredictive. However, all these patients had elevated plasma levels of IL-6. In fact, there was a trend toward higher concentrations of IL-6 in patients who eventually died. A retrospective analysis comparing IL-6 concentrations with the probability of the patients being septic showed that if a patient initially presented a plasma IL-6 level of greater than about 1 ng/ml, he or she had a greater than 50 percent probability of becoming septic, or in this case, bacteremic. With an IL-6 level of greater than 10 ng/ml, the likelihood increased to 90 percent. Such findings are consistent with a large body of data by several authors showing that the magnitude of the IL-6 response is often associated with the severity of the physiologic response (Calandra et al., 1991; Damas et al., 1992; Frieling et al., 1995; Rintala et al., 1995; Steinmetz et al., 1995).
The findings to date suggest that, for hospitalized patients with systemic inflammatory response syndromes, plasma TNF-a and IL-1 are not likely to be detected with any regular frequency; therefore, the utility of these plasma measurements is likely to be small. In contrast, the data suggest that IL-6 may be a reliable indicator for the presence of infection or inflammation, and levels may correlate with the severity of injury or outcome.
FACTORS INFLUENCING THE MEASUREMENT OF PLASMA CYTOKINES
A question that arises is why plasma measurements of TNF-a and IL-1 are poor indicators of their production. It must be emphasized that the question is not really whether TNF-a and IL-1 are being produced during sepsis and systemic inflammatory response syndromes. In multiple animal models of infection or inflammation, examining organs or tissues of the reticuloendothelial system reveals evidence of up-regulation of TNF-a and IL-1 gene transcription despite an inability to detect the proteins in the circulation.
The reasons for the failure to detect TNF-a or IL-1 in the circulation are multiple and represent the inherent biology of the proteins. Foremost, the production of TNF-a and IL-1 is episodic. Macrophages exposed to continuous levels of endotoxin, for example, release TNF-a and IL-1 in a single burst over several hours and then become tolerant to repeated or continued exposure (Beutler et al., 1986). This development of tachyphylaxis (the rapid appearance of a progressive decrease in response following repetitive administration of a physiologically active substance) remains a hallmark cytokine response to exposure to bacterial cell products.
Another possible reason for failing to detect TNF-a and IL-1ß is that the plasma half-lives of TNF-a and IL-1ß are generally less than 60 minutes, and random sampling techniques are probably too imprecise to pick up these episodic productions. More importantly, there are inhibitors and antagonists in the blood of septic patients, which often interfere with the immunoassays or bioassays used to quantitate TNF-a and IL-1 (Engelberts et al., 1991). This becomes extremely important because the presence of these inhibitors or antagonists in the plasma may be used to document the presence of these proinflammatory cytokines indirectly (Moldawer, 1994).
Finally, the production of TNF-a and IL-1 is primarily paracrine in nature, so the appearance in the systemic circulation probably reflects those conditions when either the cytokine is being produced directly in the plasma compartment by blood monocytes and neutrophils or is being produced in large quantities in tissue compartments and is only spilling into the blood as an overflow type of phenomenon from the tissues.
The results of a study by Jean Michel Dayer's group in Geneva is extremely illustrative in this regard (Suter et al., 1992). The investigators examined TNF-a appearance in the plasma and bronchoalveolar lavage fluid (BAL) of patients with adult respiratory distress syndrome (ARDS). They evaluated three groups of patients: patients without ARDS, those with early ARDS, and those with late ARDS (24 hours later). While statistically significant plasma concentrations of TNF-a were about 100 pg/ml, BAL levels were significantly higher, on the order of 100-fold higher in early ARDS, providing evidence of compartmentalization of TNF-a in this patient population.
Unpublished data from studies that this laboratory has done with Greg Schultz's group at the University of Florida, looking at the levels of TNF-a in nonhealing wounds (decubitus ulcers or diabetic foot ulcers), are similar (Unpublished data, L. L. Moldawer, University of Florida, Gainesville, Fla., 1996). Very high levels of TNF-a were found compartmentalized in these nonhealing wounds, on the order of 5 ng/ml. At the same time in the peripheral circulation of these patients, there was no detectable TNF-a (< 50 pg/ml). IL-1ß levels also have been examined in these same tissues. A similar phenomenon occurred with no detectable IL-1ß in the circulation of patients with nonhealing wounds, but very high levels of IL-1ß in the wounds themselves.
The second reason identified above for not being able to detect TNF-a or IL-1 in the circulation is the presence of inhibitors that not only bind and inactivate the cytokines but also interfere with their detection. The most common inhibitors of TNF-a are its shed receptors, the p55 receptor and the p75 receptor. During inflammation, these receptors are shed after a protease cleaves their extracellular portions at the level of the cell membrane (Van Zee et al., 1992). The shed receptors retain the ability to bind ligand and therefore compete with remaining cellular TNF receptors for the binding of circulating TNF-a, thus making them natural inhibitors.
Studies performed in this laboratory that were published in 1992 showed that human volunteers given very small amounts of endotoxin, sufficient to produce chills and a mild tachycardia, resulted in a systemic TNF-a response (Van Zee et al., 1992). Also seen subsequent to the appearance of TNF-a (after 2 hours) was an increase in the plasma concentrations of the shed p55 and p75 receptors. Ninety minutes after the administration of endotoxin, TNF-a bioactivity and immunoactivity peaked in the plasma. Thereafter, however, plasma from such patients had net inhibitory activity for TNF-a. If recombinant TNF-a was added to the plasma, it would be bound and inactivated, due in part to the excess quantities of these shed receptors.
Where do these shed TNF receptors come from? The results of a flow cytometry study on monocytes from patients administered endotoxin, performed by Steve Calvano at Cornell University, are illustrative (van der Poll et al., 1995). After endotoxin is administered to healthy volunteers, the number of TNF binding sites on the monocytes (presumably receptors) declines. Accompanying this decline is a corresponding increase in the appearance of soluble p55 and p75 in the circulation.
Following acute inflammation secondary to endotoxin administration, there is a release of shed TNF receptors, which can bind TNF-a and act as antagonists. The same phenomenon occurs for IL-1. One of the IL-1 receptors, p68, is shed and binds to IL-1, thus inactivating it (Colotta et al., 1993). This receptor also is shed during sepsis and in major surgical procedures (Giri et al., 1994; Pruitt et al., 1995). However, the IL-1 system has another inhibitory molecule, IL-1 receptor antagonist (IL-1ra), which is a competitive antagonist for the
functional type I IL-1 receptor and also is released after acute inflammation (Hannum et al., 1990).
In summary, the appearance of the TNF-a inhibitors p55 and p75 can be used as an indirect estimate of the presence and degree of inflammation. The same phenomenon holds true for the IL-1 inhibitors. It has been argued that the plasma concentration of these shed receptors or concentrations of IL-1ra can be used as an indirect estimate of the degree of inflammation present and, probably, of the induction of these cytokines at the paracrine level. Concentrations of p55 and p75 in critically ill patients are routinely elevated (Fischer et al., 1992; Rogy et al., 1994). Concentrations seen in normal volunteers are generally at the lower levels of sensitivity of this assay, while 95 percent of the patients who are critically ill, meeting the criteria for SIRS or sepsis syndrome, have elevated p55 and p75 levels (Rogy et al., 1994).
FIELD MEASUREMENT OF URINARY CYTOKINES
The problem with cytokine measurements in the field setting is that it is often difficult to obtain blood samples. The question of whether cytokine levels in the plasma or in tissue compartments can be documented through noninvasive sampling is important. At least with regard to urine measurements for cytokines, there is now good evidence to suggest that several cytokines that appear in the plasma in the high pg/ml or ng/ml range are excreted immunologically intact in the urine. Importantly, the first descriptions of many of these cytokines, and of most of the cytokine inhibitors, were based on inhibitory activity that was identified in the urine of patients with a variety of inflammatory conditions. For example, p55, p75, and IL-1ra were first purified to homogeneity and sequenced from urine samples of patients with either neoplasms, sepsis, pregnancy, or fevers (Prieur et al. 1987; Seckinger et al. 1987a, b, 1988). So it is known, at least historically, that these cytokine inhibitors can appear in the urine. What is not known is whether proinflammatory cytokines themselves are excreted in the urine and whether their detection can be used to document the presence and degree of inflammation.
Some studies done by Sprenger and colleagues in Germany, in which they looked at blood and urinary cytokine production in conditioned athletes before and after a 20-km run, have a great deal of relevance for the military (Sprenger et al., 1992). The run took about 100 minutes, so the 2-h samples actually reflect blood and urine measurements 20 minutes after the run was completed. IL-6 levels increased in the plasma, consistent with a nonspecific inflammatory response, but significant increases in urinary IL-6 also were documented. In fact, levels peaked very quickly, 2 to 3 hours after the start of the run, and then rapidly returned toward prestress values. Urinary levels were markedly higher than plasma levels, reflecting to some extent the very rapid half-life of the molecule. While conditioned athletes had measurable IL-6 in the urine, normal volunteers and unconditioned athletes did not.
The authors report a similar pattern with TNF-a (Sprenger et al., 1992). This is, in fact, one of the few published studies able to document routinely the presence of TNF-a in the urine. But the same sort of phenomenon is seen as above. TNF-a was detected in the urine of these conditioned athletes before the 20-km run, levels peaked very quickly after this, and then returned to normal. These levels of TNF-a and IL-6 in the urine are well within the range of commercial enzyme-linked immunoabsorption assays (ELISAs), so it is relatively easy to purchase an ELISA kit and measure levels in unconcentrated urine samples.
Another study of military relevance is one by Peter Stein and colleagues at the University of Medicine and Dentistry of New Jersey. These authors looked at why astronauts lose weight during spaceflight and whether it is a manifestation of a nonspecific inflammatory response. Some of these data have been published (Stein and Schluter, 1994), while the remainder are currently in preparation. To determine if urinary cytokine measurements in a setting like this would be helpful to explain some of the weight loss and immunological changes that occur during spaceflight, 24-h urine collections were obtained for 7 days preflight, during the spaceflight, and for 7 days postflight to ascertain whether differences in the urinary excretion of cytokines occurred during spaceflight and whether they would provide some insight into the metabolic changes that were occurring. During spaceflight, there was an acute phase response, indicated by increased fibrinogen synthesis. In addition, there was a remanifestation of this acute-phase response on the first recovery day after return to Earth.
Stein and Schluter (1994) showed that IL-6 excretion increased on the first flight day, and elevations in IL-6 levels corresponded very well to increases in fibrinogen synthesis during that period. This is not unexpected considering that IL-6 is thought to be a primary inducer of fibrinogen synthesis. Thereafter, levels returned to baseline (preflight) values. However, a very interesting phenomenon was seen next. As the astronauts recovered from their spaceflight and the effect of microgravity after reentry and readjusted to normal gravitational loads, there was a progressive increase in the urinary excretion of IL-6. During this recovery period, after days of microgravity (in space), the astronauts often complained of leg pain and weakness in all load-bearing muscles. This secondary increase in cytokine appearance may be due to some direct muscle injury associated with reconditioning to full gravity.
Urinary excretion of IL-1ra, p55, and p75 showed a similar pattern, with the overall profile being similar to that of IL-6. Urinary levels are again elevated on the first flight day and during the postflight recovery period.
Further evidence that urinary cytokine measurements may be used to grade the severity of the inflammatory response can be found in the results of a study that was performed by the author in collaboration with Shogo Yoshido at Kurume University (Unpublished manuscript, Department of Surgery, Kurume, Japan, 1996). Postoperative urinary cytokine levels were measured in patients undergoing either esophagectomy, which is a relatively severe inflammatory
stimulus, or partial gastrectomy, which is a less severe, although still considerable, inflammatory insult. Urinary IL-6, p75, and IL-1ra excretion were significantly higher in patients undergoing esophagectomy, corresponding to the magnitude of surgical injury. The response to gastrectomy was more modest.
In conclusion, plasma TNF-a and IL-1 are not reliable indicators of inflammation. Elevated levels of these cytokines in septic patients, beyond a certain threshold value (about 500 pg/ml), generally are associated with adverse septic events (Waage et al., 1989). However, increased plasma concentrations of the shed TNF receptors, shed IL-1 receptor, or IL-1 receptor antagonist generally are more indicative of a local inflammatory response and can be used, with some reliability, as markers of local TNF-a and IL-1 production.
Plasma IL-6 levels, in contrast, appear to be elevated in a large number of inflammatory processes, and they seem to correlate with physiologic parameters. Urinary TNF-a and IL-1 levels are increased after eccentric exercise but cannot be detected regularly in the circulation or in the urine of a large number of patients. Conversely, urinary IL-6, IL-1ra, p55, and p75 are increased reproducibly under the same conditions.
Finally, plasma and urinary measures of IL-6, IL-1ra, and the shed TNF receptors can be used to detect the presence of metabolic stress. The urine measurements are extremely helpful in avoiding an invasive procedure and can, to some extent, be normalized for random spot collections. For example, a urinary creatinine level may be measured at the same time to achieve some level of normalization for total daily output. In contrast, plasma and urinary measures of TNF-a and IL-1 are less likely to provide meaningful information.
Beisel, W.R. 1975 Metabolic response to infection. Annu. Rev. Med. 26:9–20.
1977 Magnitude of the host nutritional responses to infection. Am. J. Clin. Nutr. 30:1236–1247.
1980 Endogenous pyrogen physiology. Physiologist 23:38–42.
Beutler, B., and A. Cerami 1986 Cachectin and tumour necrosis factor as two sides of the same biological coin. Nature 320:584–588.
Beutler, B., N. Krochin, I.W. Milsark, C. Luedke, and A. Cerami 1986 Control of cachectin (tumor necrosis factor) synthesis: Mechanisms of endotoxin resistance. Science 232:977–980.
Calandra, T., J. Gerain, D. Heumann, J.D. Baumgartner, and M.P. Glauser 1991 High circulating levels of interleukin-6 in patients with septic shock: Evolution during sepsis, prognostic value, and interplay with other cytokines. The Swiss-Dutch J5 Immunoglobulin Study Group. Am. J. Med. 91:23–29.
Colotta, F., F. Re, M. Muzio, R. Bertini, N. Polentarutti, M. Sironi, J.G. Giri, S.K. Dower, J.E. Sims, and A. Mantovani 1993 Interleukin-1 type II receptor: A decoy target for IL-1 that is regulated by IL-4. Science 261:472–475.
Damas, P., D. Ledoux, M. Nys, Y. Vrindts, D. De Groote, P. Franchimont, and M. Lamy 1992 Cytokine serum level during severe sepsis in human IL-6 as a marker of severity. Ann. Surg. 215:356–362.
Dinarello, C.A. 1994 The biological properties of interleukin-1. Eur. Cytokine Netw. 5:517–531.
Dinarello, C.A., and S.M. Wolff 1993 The role of interleukin-1 in disease. N. Engl. J. Med. 326:106–113.
Engelberts, I., S. Stephens, G.J. Francot, C.J. van der Linden, and W.A. Buurman 1991 Evidence for different effects of soluble TNF-receptors on various TNF measurements in human biological fluids [letter; comment]. Lancet 338:515–516.
Fischer, E., K.J. Van Zee, M.A. Marano, C.S. Rock, J.S. Kenney, D.D. Poutsiaka, C.A. Dinarello, S.F. Lowry, and L.L. Moldawer 1992 Interleukin-1 receptor antagonist circulates in experimental inflammation and in human disease. Blood 79:2196–2200.
Fong, Y., L.L. Moldawer, G.T. Shires, and S.F. Lowry 1990a The biologic characteristics of cytokines and their implication in surgical injury. Surg. Gynecol. Obstet. 170:363–378.
Fong, Y.M., M.A. Marano, L.L. Moldawer, H. Wei, S.E. Calvano, J.S. Kenney, A.C. Allison, A. Cerami, G.T. Shires, and S.F. Lowry 1990b The acute splanchnic and peripheral tissue metabolic response to endotoxin in humans. J. Clin. Invest. 85:1896–1904.
Frieling, J.T.M., M. Van Deuren, J. Wijdenes, J.W.M. Van der Meer, C. Clement, C.J. van der Linden, and R.W. Sauerwein 1995 Circulating interleukin-6 receptor in patients with sepsis syndrome. J. Infect. Dis. 171:469–472.
Giri, J.G., J. Wells, S.K. Dower, C.E. McCall, R.N. Guzman, J. Slack, T.A. Bird, K. Shanebeck, K. Grabstein, J.E. Sims, and M.R. Alderson 1994 Elevated levels of shed type II IL-1 receptor in sepsis. Potential role for type II receptor regulation of IL-1 responses. J. Immunol. 153:5802–5809.
Hannum, C.H., C.J. Wilcox, W.P. Arend, F.G. Joslin, D.J. Dripps, P.L. Heimdal, L.G. Armes, A. Sommer, S.P. Eisenberg, and R.C. Thompson 1990 Interleukin-1 receptor antagonist activity of a human interleukin-1 inhibitor. Nature 343:336–340.
Hesse, D.G., K.J. Tracey, Y. Fong, K.R. Manogue, M.A. Palladino, Jr., A. Cerami, G.T. Shires, and S.F. Lowry 1988 Cytokine appearance in human endotoxemia and primate bacteremia. Surg. Gynecol. Obstet. 166:147–153.
Kampschmidt, R.F. 1983 Infection, inflammation, and interleukin-1 (IL-1). Lymphokine Res. 2:97–102.
Kampschmidt, R.F., and L.A. Pulliam 1975 Stimulation of antimicrobial activity in the rat with leukocytic endogenous mediator. J. Reticuloendothel. Soc. 17:162–169.
Kluger, M.J., D.H. Ringler, and M.R. Anver 1975 Fever and survival. Science 188:166–168.
LeMay, L.G., A.J. Vander, and M.J. Kluger 1990 The effects of psychological stress on plasma interleukin-6 activity in rats. Physiol. Behav. 47:957–961.
Marano, M.A., Y. Fong, L.L. Moldawer, H. Wei, S.E. Calvano, K.J. Tracey, P.S. Barie, K. Manogue, A. Cerami, G.T. Shires et al. 1990 Serum cachectin/tumor necrosis factor in critically ill patients with burns correlates with infection and mortality. Surg. Gynecol. Obstet. 170:32–38.
Michie, H.R., K.R. Manogue, D.R. Spriggs, A. Revhaug, S. O'Dwyer, C.A. Dinarello, A. Cerami, S.M. Wolff, and D.W. Wilmore 1988 Detection of circulating tumor necrosis factor after endotoxin administration. N. Engl. J. Med. 318:1481–1486.
Moldawer, L.L. 1994 Biology of proinflammatory cytokines and their antagonists. Crit. Care Med. 22:S3–S7.
Moscovitz, H., F. Shofer, H. Mignott, A. Behrman, and L. Kilpatrick 1994 Plasma cytokine determinations in emergency department patients as a predictor of bacteremia and infectious disease severity. Crit. Care Med. 22:1102–1107.
Nathan, C., and M. Sporn 1991 Cytokines in context. J. Cell Biol. 113:981–986.
Powanda, M.C., and W.R. Beisel 1982 Hypothesis: Leukocyte endogenous mediator/endogenous pyrogen/lymphocyte-activating factor modulates the development of nonspecific and specific immunity and affects nutritional status. Am. J. Clin. Nutr. 35:762–768.
Prieur, A.M., M.T. Kaufmann, C. Griscelli, and J.M. Dayer 1987 Specific interleukin-1 inhibitor in serum and urine of children with systemic juvenile chronic arthritis. Lancet 2(8570):1240–1242.
Pruitt, J.H., E.M. Copeland, and L.L. Moldawer 1995 Interleukin-1, interleukin-1 receptor antagonist and interleukin-1 receptors in sepsis, shock and the systemic inflammatory response syndrome. Shock 3:235–251.
Rintala, E., K. Pulkki, J. Mertsola, T. Nevalainen, and J. Nikoskelainen 1995 Endotoxin, interleukin-6 and phospholipase-A2 as markers of sepsis in patients with hematological malignancies. Scand. J. Infect. Dis. 27:39–43.
Rogy, M.A., S.M. Coyle, H.S. Oldenburg, C.S. Rock, P.S. Barie, K.J. Van Zee, C.G. Smith, L.L. Moldawer, and S.F. Lowry 1994 Persistently elevated soluble tumor necrosis factor receptor and interleukin-1 receptor antagonist levels in critically ill patients. J. Am. Coll. Surg. 178:132–138.
Seckinger, P., J.W. Lowenthal, K. Williamson, J.M. Dayer, and H.R. MacDonald 1987a A urine inhibitor of interleukin-1 activity that blocks ligand binding. J. Immunol. 139:1546–1549.
Seckinger, P., K. Williamson, J.F. Balavoine, B. Mach, G. Mazzei, A. Shaw, and J.M. Dayer 1987b A urine inhibitor of interleukin-1 activity affects both interleukin-1 alpha and -1 beta but not tumor necrosis factor alpha. J. Immunol. 139:1541–1545.
Seckinger, P., S. Isaaz, and J.M. Dayer 1988 A human inhibitor of tumor necrosis factor alpha. J. Exp. Med. 167:1511–1516.
Sprenger, H., C. Jacobs, M. Nain, A.M. Gressner, H. Prinz, W. Wesemann, and D. Gemsa 1992 Enhanced release of cytokines, interleukin-2 receptors, and neopterin after long-distance running. Clin. Immunol. Immunopathol. 63:188–195.
Stein, T.P., and M.D. Schluter 1994 Excretion of IL-6 by astronauts during spaceflight. Am. J. Physiol. 266:E448–E452.
Steinmetz, H.T., A. Herbertz, M. Bertram, and V. Diehl 1995 Increase in interleukin-6 serum level preceding fever in granulocytopenia and correlation with death from sepsis. J. Infect. Dis. 171:225–228.
Suter, P.M., S. Suter, E. Girardin, P. Roux Lombard, G.E. Grau, and J.M. Dayer 1992 High bronchoalveolar levels of tumor necrosis factor and its inhibitors, interleukin-1, interferon, and elastase, in patients with adult respiratory distress syndrome after trauma, shock, or sepsis. Am. Rev. Respir. Dis. 145:1016–1022.
van der Poll, T., S.E. Calvano, A. Kumar, C.C. Braxton, S.M. Coyle, K. Barbosa, L.L. Moldawer, and S.F. Lowry 1995 Endotoxin induces downregulation of tumor necrosis factor receptors on circulating monocytes and granulocytes in humans . Blood 86:2754–2759.
Van Zee, K.J., T. Kohno, E. Fischer, C.S. Rock, L.L. Moldawer, and S.F. Lowry 1992 Tumor necrosis factor soluble receptors circulate during experimental and clinical inflammation and can protect against excessive tumor necrosis factor alpha in vitro and in vivo. Proc. Natl. Acad. Sci. USA 89:4845–4849.
Waage, A., T. Espevik, and J. Lamvik 1986 Detection of tumour necrosis factor-like cytotoxicity in serum from patients with septicaemia but not from untreated cancer patients. Scand. J. Immunol. 24:739–743.
Waage, A., A. Halstensen, and T. Espevik 1987 Association between tumour necrosis factor in serum and fatal outcome in patients with meningococcal disease. Lancet 1(8529):355–357.
Waage, A., P. Brandtzaeg, A. Halstensen, P. Kierulf, and T. Espevik 1989 The complex pattern of cytokines in serum from patients with meningococcal septic shock. Association between interleukin-6, interleukin-1, and fatal outcome. J. Exp. Med. 169:333–338.
Wannemacher, R.W., Jr., H.L. DuPont, R.S. Pekarek, M.C. Powanda, A. Schwartz, R.B. Hornick, and W.R. Beisel 1972 An endogenous mediator of depression of amino acids and trace metals in serum during typhoid fever. J. Infect. Dis. 126:77–86.
DOUGLAS WILMORE: That was a super review. I have a couple of practical questions that relate to application. You alluded to the ability to collect urine, and the issue is raised, in the military context, of people losing sweat and water by other routes, having diarrhea, or maybe not drinking as much water. I was looking at your y axes, body weight, osmolality in urine, and things like that.
Are you sure some of these measurements are not the result of urinary concentration or are not amplified because of urinary concentration? And what are some of the practical guidelines that you can give to people?
LYLE MOLDAWER: You have hit all the technical requisites for the utility of this measurement. I can tell you that the astronauts have been very well described, in that renal function is within 85 percent of normal limits during spaceflight, but that cannot explain the changes that are seen. You are correct in stating that major perturbations in renal function will affect these measurements.
We know, for example, that p55, p75, and IL-1ra are cleared predominantly by the kidney, so if there is significant renal dysfunction it will throw the validity of these measurements into question. Now, as for hydration, per se, in the
Sprenger article (Sprenger et al., 1992), they normalized for changes in the osmolality of the urine. If you look at the caveat that they discuss in the article, it really does not change the direction or the magnitude of the response.
I think if you recognize the fact that these urinary measurements are susceptible to those errors and you can incorporate that into those estimates of renal function and hydration, then they can be of value. Of course, the presence of any renal or bladder infection will invalidate the use of urine measurements for systemic production.
Sweat and saliva are other issues. Unfortunately, my feeling or bias is, and I may be corrected by people in the audience, that these secretions are from individual body compartments and are less likely to represent well-mixed whole-body pools. I think what you find is that, in burn blister fluid, cytokines are produced by keratinocytes in that local microenvironment. Whereas the stable isotope people can use saliva or sweat, you probably cannot use those compartments for cytokine measurement, because they are going to represent the cell population that produces those mediators.
DOUGLAS WILMORE: To continue this discussion, one of the things that might be very helpful, particularly in the field, if we could extrapolate it to the field, would be to measure military personnel and determine whether, in fact, they had some inflammatory component or whether it was sleep deprivation and underfeeding or something like that. Would a urinary analysis allow a sort of off-on determination, a yes-no determination, not necessarily quantitative but just ''it is there or it is not there"?
LYLE MOLDAWER: I think you probably are close. There are two amazing points. One is that you do not need an infection or a tissue injury to induce these cytokines. What the Sprenger article shows and what the NASA data show—and Matt Kluger has shown this before (LeMay et al., 1990)—is that psychological stress will induce some manifestation of a proinflammatory cytokine response.
So yes, and if you read the Sprenger article, what was shown was that if you are a conditioned athlete, urinary cytokines can be measured. Unless we have an infection of some sort or some psychological stress (which is common in our field since much of our time is spent writing grant proposals), you and I do not normally exhibit the presence of urinary cytokines, although we do excrete basal quantities of cytokine inhibitors, such as IL-1ra, p55, and p75.
When I got the letter from Dr. [Bernadette] Marriott, I thought, gee, why are cytokines involved in military performance? Then I started reading these articles, and now I think it is an untouched area. The potential is there, whether the data go with the hypothesis and the theory.
DOUGLAS WILMORE: A final question. C-reactive protein2 pretty much mirrors IL-6 and may be a much easier assay. Does it show up in the urine?
LYLE MOLDAWER: That I do not know, but you have hit on a key issue, which is that these are nonspecific indicators of inflammation. Erythrocyte sedimentation rate may be adequate. Nobody has done the sensitivity-specificity assays that are required to determine the validity of these techniques.
ROBERT NESHEIM: I think what I would like to do is move on because we do have time for discussion after the other two presentations here. It may be that all of this will come together a little bit better, so if we could do that, I would appreciate it.
LYLE MOLDAWER: Just a quick take-home message, which is that urinary cytokine measurements really are an untapped resource. We have focused predominantly on blood and tissue measurements.