A Molecular Basis for Bidirectional :
Communication Between the Immune and Neuroendocrine Systems
J. EDWIN BLALOCK Department of Physiology and Biophysics, University of Alabama at Birmingham, Birmingham, Alabama
HORMONAL FUNCTIONS OF PRODUCTS OF THE IMMUNE SYSTEM
In the preceding section, it was shown that many neuroendocrine pep-tides have diverse immunoregulatory actions. This section deals with the converse, that is, the hormonal functions of peptides and proteins that were originally described as products of the immune system. Immunologic activation of various components of the hypothalamic pituitary adrenal axis is at present the best-studied response, which stemmed in part from the observation of an elevation of circulating glucocorticoid hormone levels that coincided with peak antibody titers during an immune response (14). The elevation of glucocorticoid was subsequently shown to be due, in part, to a f actor(s) of unknown composition that was derived from activated leukocytes (11). It is now known that there are a number of well-defined immunologically derived molecules that can account for the earlier as well as more recent observations (Table 3).
table 3. Neuroendocrine effects of lymphokines and monokines
|IFN-a and/or 0||Adrenal steroidogenesis||17, 117|
|Induction of melanin synthesis||17|
|Enhancement of iodine uptake by||16|
|Excitation of neurons||16|
|Suppression of morphine withdrawal||41|
|Catalepsy and analgesia||20|
|Promotion of slow-wave sleep||82|
|Hypothalamic release of CRF||9,32,121,138,140|
|Pituitary release of ACTH and||8,10, 28, 56, 77,|
|Elevation of glucocorticoid levels||12|
|Thymosin ax||Elevation of ACTH and glucocorticoid levels||62|
|Thymosin (5A||Hypothalamic release of LHRH||114|
|IL2||Pituitary release of ACTH and endorphins||28, 49, 90,132|
|Elevation of glucocorticoid levels||89|
See text for definitions of abbreviations.
Interferons were perhaps the first products of macrophages and lymphocytes to be observed to have a hormonal function. In fact, based on the ubiquitous nature of the IFN receptor and IFN-induced secondary messengers that are shared with hormones, IFN was predicted to have numerous hormonal activities (22). This turned out to be the case, and there are many examples. For instance, IFN was found to induce an in vitro ACTH-like steroidogenic effect on cultured adrenal cells (17). During clinical trials in humans, a corollary to this finding was seen. a-Interferon was observed to cause elevations in circulating cortisol levels (117). If the in vitro findings are considered, the simplest explanation for the in vivo observation would be that IFN-a acted directly on the adrenal glands. Other hormonal activities of IFN include an a-MSH-like induction of melanin synthesis, a TSH-like iodine uptake by thyroid cells, an endorphin-like excitation of neurons, and a glu-cagon-like antagonism of insulin action (for review, see Ref. 16). Although the preceding effects were most likely mediated by an action on the IFN receptor, IFN can also cause hormonal effects through actions on another hormone’s receptor. Human IFN-a bound to mouse brain opiate receptors in vitro and caused an endorphin-like analgesic and catatonic response in vivo (20). Because these effects were blocked by naloxone and because naloxone
did not alter antiviral activity of IFN, this appears to be an example of IFN acting through another hormone’s receptor. A more recent and perhaps related finding is the ability of human IFN-a to block a naloxone-induced withdrawal reaction in morphine-addicted rats (41). Collectively, these observations seem to leave little doubt that this lymphokine/monokine has many hormonal activities that may in part explain its numerous side effects, including neuropsychiatric manifestations, when administered to patients (1).
There is now a growing body of evidence that, in addition to IFN, other monokines and lymphokines may directly or indirectly influence the neuro-endocrine system. One of the early described effects of monocyte-derived IL1 (formerly leukocyte pyrogen or endogenous pyrogen) was the generation of fever (reviewed in Ref. 43). Introduction of IL 1 by intracerebroventricular administration produces a more profound rise in temperature than does peripheral venous injection. The effect of IL 1 on temperature appears to be localized to the anterior hypothalamus. In addition to its fever-inducing characteristics, IL 1 also promotes slow-wave sleep (82).
The administration of recombinant IL 1 to rats also results in a measurable increase in blood ACTH and corticosterone levels in these animals (12). This effect is independent of the fever response and, of course, could result from either an effect of IL 1 on the hypothalamus or on the pituitary gland. In fact, it appears to be due to effects on both, as indicated by the ACTH-re-leasing potential of IL 1 on pituitary cells (8, 10, 28, 56, 77, 139, 144) and corticotropin-releasing factor (CRF)-releasing activity on hypothalami (9, 32,121,138,140). It is interesting to note that those studies that showed an effect on the hypothalamus failed to show IL 1-mediated ACTH release from primary cultures of pituitary cells (9, 32, 121, 140). This is particularly puzzling since the ACTH-releasing activity of IL 1 on AtT-20 cells as well as on primary pituitary cells has been independently observed by a number of laboratories (8,10, 28, 56, 77,139,144). At this point it is difficult to reconcile the differences. However, when it is considered that the effects on both types of target cells have been independently reproduced, it is likely that both observations are correct and the discrepancy is of a technical nature. The possibility that both sites are affected in vivo is also possible, since intravenous administration of IL 1 increased CRF yet antibody to CRF did not completely abrogate the pituitary ACTH response (9,121). Furthermore, the effects of endogenous production as opposed to exogenous administration of IL 1 should be considered. Perhaps local IL 1 production by astrocytes and glial cells (51, 52) would primarily activate hypothalamic CRF release, whereas IL 1-producing peripheral monocytes would preferentially deposit IL 1 in the pituitary gland, since such cells pass through the portal circulation. These would seem to be exciting possibilities for future studies. With regard to the CRF-releasing activity of IL 1, it is interesting to note that another immunologically derived molecule, thymosin /?4, can also act on the hypothalamus. In this case, however, it results in LHRH rather than CRF release (114).
Other lymphokines and monokines also seem to regulate the hypotha-lamic pituitary axis. Thymosin «i, for instance, seems to cause in vivo ACTH release through an action on the central nervous system that in turn stimulates the pituitary gland (62). Interleukin 2 was also recently shown to cause an elevation in circulating ACTH and cortisol levels during clinical testing (89). In this case, the action of IL 2 may well be on the pituitary gland, since this lymphokine can cause POMC production and release from pituitary tumor cells and corticotrophs (28,49,90,132). Furthermore, IL 2 as well as IL 1 function like CRF not only in terms of increased ACTH release but also increased POMC mRNA expression (28, 90). A newly described monokine, hepatocyte-stimulating factor (HSF), has also been shown to have CRF-like activity, except that it is threefold more potent than either IL 1 or CRF-AVP (144). The authors concluded that since both IL 1 and HSF are released during inflammatory processes they may play a very important role in activation of the pituitary adrenal axis. Although these studies are just beginning, they nonetheless seem to clearly show that lymphokines and monokines can modulate virtually all components of the neuroendocrine system. Furthermore, such regulation occurs at different sites, depending on the lymphokine or monokine. Thymosin ax seems to act on the nervous system, IL 1 and thymosin /34 act on the hypothalamus, IL 1 and IL 2 act on the pituitary, and IFN-a acts on the adrenal glands.
PRODUCTION, STRUCTURE, REGULATION, AND PROCESSING OF PEPTIDE HORMONES COMMON TO THE IMMUNE AND NEUROENDOCRINE SYSTEMS
As it turns out, lymphokines and monokines are not the sole mediators of communication from the immune to the neuroendocrine system. Indeed, cells of the immune system are now known to produce peptide hormones that were previously thought to be restricted to the neuroendocrine system (Table 4). It seems that this observation is probably pivotal to a biochemical understanding of how and why there is bidirectional communication between these two systems. Put most simply (discussed in detail in sect, vm), the immune and neuroendocrine systems share a set of hormones and their receptors that are used for inter- and intrasystem communication (21). The following sections describe the evidence for shared peptide hormones.
A. Production and Structure of Leukocyte-Derived Peptide Hormones
1. Adrenocorticotropic hormone, endorphins, and enkephalins
To date, the most thoroughly studied neuroendocrine hormones that are produced by leukocytes are the POMC-derived peptides. Human peripheral blood lymphocytes and mouse spleen cells were initially observed to simultaneously express immunoreactive ACTH and endorphins after virus infection
table 4. Neuroendocrine peptides produced by the immune system
|Peptide||Constitutive||Inducible||Cellular or Tissue Source||References|
|ACTH||+||+||Lymphocytes and macrophages||19, 47, 66, 87, 126, 128|
|GH||+||+||Lymphocytes||71, unpublished results|
leukocytes, mast cells, and PMN leukocytes
leukocytes, mast cells, and PMN leukocytes
See text for definitions of abbreviations.
or interaction with transformed cells or bacterial LPS (19, 66, 126, 128). In contrast to the inducible synthesis of ACTH and endorphins by most lymphocytes, a subpopulation of mouse splenic macrophages, as well as rat lymphocytes in the tunica propria, produced these peptides in a constitutive fashion (47,87). An ACTH-like substance has also been shown to be produced by avian leukocytes (124). Numerous criteria have shown that these leukocyte-derived peptides are identical to pituitary ACTH and endorphins. These include shared antigenicity, as determined with monospecific antibodies against synthetic peptide hormones; identical retention times on reverse-phase high-performance liquid chromatographic (HPLC) columns; identical molecular weights; and shared biological activities (19, 21, 66, 86, 126, 128). More recently, the mRNA for POMC has been identified in lymphocytes as well as macrophages (47, 86, 143; unpublished results), and the amino acid sequences of mouse splenic and pituitary ACTH were found to be identical (E. M. Smith and J. E. Blalock, unpublished results). Taken together, these
results would seem to conclusively demonstrate that the immune system can produce bona fide POMC-derived peptides.
Abundant levels of preproenkephalin mRNA have also been found in mitogen or antigen-activated but not resting Th cells. This mRNA represented from 0.1 to 0.5% of the total mRNA, depending on the particular Th line that was induced. The mRNA was apparently translated, and the product was secreted, since immunoreactive Met-enkephalin was detected in Th cell culture supernates (146).
Thyrotropin was the second de novo synthesized peptide hormone to be found in the immune system. Its induction was initially observed in response to activation of human peripheral blood cells by a particular T-cell mitogen, SEA (131). More recently, TSH has been shown to be constitutively produced by a human T-cell leukemia line, MOLT 4, and such production is increased by SEA (65). In these studies, the lymphocyte-derived TSH was recognized by a monospecific antibody to TSH-/3 and was shown to have the same molecular weight as pituitary TSH. Furthermore, the intact molecule was shown to be composed of two polypeptide chains of the molecular weight of TSH-a and TSH-/3. With regard to identity with the pituitary hormone, TSH-/? mRNA has been observed in human and mouse lymphocytes (65; unpublished observations).
3. Growth hormone/prolactin
Yet another T-cell mitogen, ConA, was recently reported to cause the production of GH and prolactin-related mRNAs (71). Unlike the pituitary, however, the lymphocyte mRNAs were larger than the corresponding precursor and mature RNA species. Apparently, the GH-related mRNA must be translated, since we can detect constitutive immunoreactive (ir)GH production by both T and B lymphocytes. Although a portion of this material has the molecular weight of pituitary GH, it is interesting to note that there are higher-molecular-weight forms. It remains to be determined whether the higher-molecular-weight forms of GH correspond to the aforementioned larger GH-related mRNA. The human lymphocyte-derived irGH is biologically active in the Nb2 lymphoma proliferation assay and competes with 125I-labeled GH for binding to its receptor (D. A. Weigent and J. E. Blalock, unpublished results).
U. Chorionic gonadotropin
Interestingly, mixed lymphocyte reactions (MLR) result in T-cell mito-genesis but unlike SEA do not evoke the production of TSH. Rather, this
allogeneic stimulus results in the production of an immunoreactive chorionic gonadotropin (67). Chorionic gonadotropin production, as monitored by im-munofluorescence with antibody to CG-0, paralleled the blastogenic response of the MLR. Gel filtration of the de novo synthesized lymphocyte-derived CG showed that this material comigrated with the human CG (hCG) standard at a molecular weight of ^58,000. This molecule was apparently glycosylated, since it bound to a ConA affinity column. The lymphocyte-derived CG was also dissociable into two subunits, CG-a and CG-/?, of the molecular weights 32,000 and 18,000, respectively. The material from the MLR was biologically active, since it elicited testosterone production from Leydig cells, and this activity was neutralized by antiserum to CG. The finding that mouse, as well as human, lymphocytes produce CG is important, since controversy lingers as to whether mouse placentas produce the molecule. The results with the mu-rine MLR would seem to support the notion that rodents have the equivalent of hCG.
5. Vasoactive intestinal peptide and somatostatin
Vasoactive intestinal peptide and somatostatin have been detected in platelets, mononuclear leukocytes, mast cells, and PMN leukocytes (40,58,61, 91, 104). These hormones were immunologically detected in the aforementioned cell types and were shown to have the appropriate molecular weight and/or HPLC retention times. It has not, however, been conclusively demonstrated that these peptides were de novo synthesized as opposed to being passively acquired by leukocytes. In the case of PMN-derived somatostatin, the amino acid composition of the peptide is similar but not identical to somatostatin (61).
6. Arginine vasopressin, oxytocin, and neurophysin
Oxytocin, AVP, and neurophysin are immunologically detectable in a lymphoid-associated organ, the thymus, although they are apparently not localized to lymphocytes (57, 93). Thymus-extracted oxytocin and neurophysin eluted in the same positions as reference standards on Sephadex G-75. The authenticity of oxytocin was also confirmed by biological assay and HPLC analysis. Local synthesis was suggested by the finding of a similarity in the molar ratio of oxytocin to neurophysin in thymus and the hypothala-moneurohypophyseal system (57). Immunoreactive AVP coeluted with authentic nonapeptide on reverse-phase HPLC (95).
B. Regulation and Processing of Leukocyte-Derived Peptide Hormones
Although immunostimulants were the first inducers described of leukocyte-derived peptide hormones, we now know that cells of the immune system also respond with fidelity to the classic hypothalamic regulators of pituitary
hormones. For instance, CRF was observed to cause the de novo synthesis and release of leukocyte-derived ACTH and 0-endorphin (130). Although it occurred at —10-fold higher concentrations, AVP alone was also observed to have intrinsic CRF activity on leukocytes. At concentrations that are frequently used on cultured pituitary cells, CRF and AVP together acted in an additive fashion to induce POMC-derived peptides, and such induction was blocked by dexamethasone. Thus leukocytes seem quite similar to cortico-trophs with respect to control of the POMC gene by CRF and AVP and feedback inhibition by a synthetic glucocorticoid hormone. More recently, we have observed an upregulation of T- and B-cell-derived irGH in response to GHRH, and this effect of GHRH was blocked by somatostatin (D. A. Wei-gent, K. L. Bost, W. Wear, and J. E. Blalock, unpublished results). Similarly, TRH has been found to induce the synthesis of TSH in leukocytes, and such induction was blocked by triiodothyronine (T3) and thyroxine (T4) (T. E. Kruger, D. V. Harbour, and J. E. Blalock, unpublished results). The human T-cell leukemia line, MOLT 4, was also observed to have a TSH response to TRH (65). Thus, collectively, cells of the immune system seem quite similar to circulating pituitary cells with respect to their ability to positively respond to hypothalamic releasing factors and to have such responses blunted by the appropriate neuroendocrine factors.
Interestingly, whereas control of the leukocyte POMC gene may be similar to that of anterior pituitary cells, the processing of its products appears somewhat different (Fig. 2). For instance, whereas Newcastle disease virus and CRF cause the production of POMC-derived peptides with the molecular weight of ACTH-(l-39) and /?-endorphin (126,130), bacterial LPS elicits the production of ACTH-(1—24 to 26) and a- and 7-endorphin (66). Although the relative contribution to alternate processing of the stimuli as opposed to the possible different leukocyte types that are responding to the stimuli are
fig. 2. Alternate processing of pro-opiomelanocortin (POMC) by lymphocytes. Corticotro-pin-releasing factor or Newcastle disease virus causes induction and processing of POMC into ACTH-(1—39) and 0-endorphin. Endotoxin causes induction or activation of an acid protease that further cleaves ACTH-(l-39) into ACTH-(l-24 to 26) and /?-endorphin into a- or 7-endorphin.
presently unknown, these findings nonetheless point to alternate proteolytic cleavages of POMC, as have been previously observed in the anterior and intermediate lobe of the pituitary as well as the hypothalamus. Of course, these results also suggest that cells of the immune system differ from virtually all other extrapituitary tissues where the major proteolytic cleavages are similar to those in the intermediate lobe of the pituitary gland (81). For example, although we detect /?-endorphin, we have yet to observe the production of an a-MSH-like peptide. Furthermore, bacterial LPS induction of an immunoreactive ACTH, with a molecular weight of ~2,900, suggests a quite novel processing pathway. In fact, we have recently demonstrated a B-cell-derived enzyme that may be responsible for this alternate processing. This protease is induced by LPS and at pH 5 cleaves ACTH-(1—39) to a form that comigrates with ACTH-(1—24) on polyacrylamide gels (64). Such differential processing points to cells of the immune system having processing pathways that are both unique and in some instances are composites of those seen in the anterior and intermediate lobes of the pituitary gland. A further interesting implication of this work is the suggestion of a possible stimulus-specific B lymphocyte-processing mechanism for POMC. This could be quite different than previously described pathways that are largely determined by the cell type (i.e., intermediate vs. anterior lobe pituitary cells) in which the POMC is processed. Within the immune system, such differential processing could have important immunoregulatory consequences. For instance, a- but not /3-endorphin suppresses in vitro antibody responses, whereas /}- but not a-endorphin enhances T-cell mitogenesis (59, 73). Also, ACTH-(1—39) but not ACTH-(1—24) suppresses in vitro antibody production (73). Thus the type of POMC stimulus may determine the processing pathway and ultimately the specific peptides that result. The specific peptide, in turn, then determines which immunologic cell type and function will be affected. Of course, the end result would be that different stimuli would elicit different responses via the same prohormone.
In addition to having proteolytic processing enzymes, macrophages, at least, must also contain acetylating enzymes. This is based on the finding that although the major endorphin species in mouse spleen macrophages is j8-endorphin [j3-EP-(l—31)], there are smaller amounts of JST-acetylated |3-EP-(1-16) (a-endorphin), /?-EP-(l-17) (7-endorphin), £-EP-(l-27), and /?-EP-(l—31) (86). Although the study of regulatory factors that control leukocyte-derived hormone genes and enzymes that posttranslationally modify their products are just beginning, it is clear that this will be an important and exciting area of future investigation.
children’s ALLERGY CENTER online
JL TAMAN BENDUNGAN ASAHAN 5 JAKARTA PUSAT, JAKARTA INDONESIA 10210
PHONE : (021) 70081995 – 5703646
email : email@example.com\
Information on this web site is provided for informational purposes only and is not a substitute for professional medical advice. You should not use the information on this web site for diagnosing or treating a medical or health condition. You should carefully read all product packaging. If you have or suspect you have a medical problem, promptly contact your professional healthcare provider.
Copyright © 2009, Children Allergy Center Information Education Network. All rights reserved.