Experimental Hematology
Volume 27, Issue 12 , Pages 1715-1723, December 1999

The SOCS-1 story☆

  • Donald Metcalf

      Affiliations

    • The Walter and Eliza Hall Institute of Medical Research, Royal Melbourne Hospital, Victoria, Australia
    • Corresponding Author InformationOffprint requests to: Donald Metcalf, The Walter and Eliza Hall Institute of Medical Research, Post Office, The Royal Melbourne Hospital, Victoria, Australia

Received 1 August 1999; accepted 11 August 1999.

Article Outline

Abstract 

SOCS-1 is an intracellular protein able to block the differentiation of leukemic M1 cells inducible by interferon γ (IFN-γ) or regulators using the gp130 receptor. Its transient production is readily inducible by cytokine stimulation, and SOCS-1 appears to be a negative feedback molecule, modulating or suppressing receptor signaling activated by at least eight cytokines. Mice lacking SOCS-1 develop a lethal neonatal syndrome including liver damage, depletion of T and B lymphocytes, and granulocyte-macrophage infiltration of the liver, lungs, pancreas, heart, and skin. These and the associated hematologic abnormalities in SOCS-1−/− mice can all be mimicked by the neonatal injection of high doses of IFN-γ. The lethal neonatal disease in SOCS-1−/− mice is preventable by injection of antibodies to IFN-γ or by crossing SOCS-1−/− mice with IFN-γ2/− mice, identifying IFN-γ as being essential for the initiation of the neonatal disease and death. IFN-γ appears not to be overproduced in SOCS-1−/− mice, and the lethal disease may arise from hyperresponsiveness of −/− cells to normal levels of IFN-γ. SOCS-1−/− mice allowed to survive the neonatal period by cross-mating with IFN-γ2/− mice may well ultimately develop other disease states, because loss of SOCS-1 potentially renders them hyperresponsive to other cytokine signaling.

Keywords:  SOCS-1, Gene inactivation, Interferon γ, Cytokine signaling, Hematopoietic cells

 

Back to Article Outline

Introduction 

For many years, we have been exploring the action of various hematopoietic regulatory factors on leukemic cell lines, either as proliferative signals, or, more interestingly, for their ability, when using appropriate cell lines, to dramatically suppress self-renewal capacity, with or without maturation of the permanently altered cells. With these cell lines, many naturally expressed or inserted receptors and their corresponding cytokine ligands can induce such suppression, and much has been established regarding the specific cytoplasmic domains of the receptor needed to initiate these responses 1, 2. Intriguing as their responses are, these cell lines do not reflect the reality of primary leukemias in man or mice, where most leukemic populations exhibit growth factor dependency but show little evidence of suppression or maturation induction by such regulators [3].

We therefore have used a genetic modulation approach to attempt to identify cellular mechanisms able to block the types of response exhibited by favorable cell lines such as WEHI-3B D+ or M1 [1]. We concentrated particularly on M1 cells that can be suppressed by interleukin 6 (IL-6), leukemia inhibitory factor (LIF), oncostatin M (OSM), and interferon γ (IFN-γ). Initial work, using mutagens on M1 cells, resulted in the isolation of M1 subclones that were resistant to suppressive cytokine action, but in those characterized to date, the mechanism was somewhat predictable, merely being an acquired failure to express the gp130 receptor chain necessary to mediate the actions of IL-6, LIF, or OSM on M1 cells.

Accordingly, an alternative approach was taken in which M1 cells were transfected with a retroviral library containing cDNAs from the immortalized cell line, FDC-P1. Other than being available for such a study [4], retrospective justification of the use of this particular library could be that immortalized cells may have adapted themselves to overexpress proteins able to block the type of suppressive responses elicited by agents like IL-6 or LIF.

From this study, an M1 clonal subline was obtained that was resistant to suppression by IL-6 or LIF and, on analysis, the inserted cDNA was found to encode a nonsecreted protein of 212 amino acids, now termed by us “suppressor of cytokine signaling-1” (SOCS-1) [5]. When overexpressed in M1 cells, SOCS-1 conferred resistance to the suppressing/maturation action of IL-6, LIF, OSM, and IFN-γ, but the cells remained responsive to differentiation induction by dexamethasone. SOCS-1 was found to be expressed in normal adult mice in the thymus, lungs, and spleen, and its expression was rapidly inducible in M1 cells by IL-6, in the liver by the in vivo injection of IL-6, or in marrow cells by a brief stimulation in vitro by granulocyte-colony-stimulating factor (GM-CSF), IL-3, IL-13, or IFN-γ, but not by macrophage colony-stimulating factor (M-CSF) [5].

Whereas the pattern of constitutive tissue expression was not encouraging for the possibility that SOCS-1 might have a special role in hematopoietic tissues, SOCS-1 was active in leukemic cells, and its expression was readily able to be induced by cytokine action on normal marrow cells. Therefore, SOCS-1 was considered worthy of further study by us as an experimental hematology group.

Back to Article Outline

Structure and biochemical interactions of SOCS-1 

The SOCS-1 protein consists of three domains: a C-terminal SOCS Box, a central SH2 region, and an N-terminal region [5].

SOCS-1 was simultaneously and independently discovered by two other groups whose observations threw considerable light on the possible biochemical interactions entered into by SOCS-1. SOCS-1 also was cloned as JAB, a protein able to bind to and prevent phosphorylation of JAK2 [6], and as SSI-1, a protein inducible by IL-6 via STAT-3 but also inhibiting STAT-3 [7]. Analysis in our laboratory also showed that overexpression of SOCS-1 in M1 cells blocked or reduced IL-6–induced phosphorylation of gp130 and of STAT-3 [5]. The initial picture emerging was that SOCS-1 protein was inducible by IL-6 signaling, but terminated or reduced signaling from the IL-6 receptor, possibly by inhibiting the phosphorylation of JAK kinases and thereby blocking subsequent signaling requiring phosphorylation and activation of STAT-3 (Fig. 1).

  • View full-size image.
  • Figure 1. 

    (A) Production of SOCS-1 is inducible by cytokine signaling, and SOCS-1 may promptly undergo proteasomal degradation by complexing with elongins B and C. (B) SOCS-1 modulates or suppresses cytokine receptor signaling by differing methods according to the receptor involved. For gp130-containing receptors, SOCS-1 blocks phosphorylation of JAK1, gp130, and STAT3. For the IFN-γ receptor, SOCS-1 blocks phosphorylation of JAK1, JAK2, and STAT1

Analysis of SOCS-1 mutants showed that the C-terminal SOCS Box region was not necessary for biologic activity, at least when SOCS-1 was overexpressed. Under these conditions, the SH2 domain was essential, as were the adjacent 20 to 30 amino acids in the N-terminal region 8, 9.

A search of EST databases revealed the recorded existence of 19 other partial cDNAs encoding a broadly similar C-terminal SOCS Box [10]. These C-terminal–related proteins fell into five structural classes. Eight had a generally similar overall structure to SOCS-1 and contained SH2 domains with N-terminal regions of variable lengths. These proteins included the previously described molecule, CIS [11]. Four contained ankyrin repeats, two contained WD-40 repeats, three contained SPRY domains, and there was a class of small GTP-ases that also contained SOCS motifs.

This presented a formidable logistical problem, because it was not feasible for our group to clone full-length cDNAs for all 20 candidate proteins, then establish their actions in vitro and determine the consequences of deletion of the various genes in mice.

The distribution of tissues constitutively expressing these cDNAs might provide some guidance in selecting proteins for priority analysis, particularly if expression was restricted to hematopoietic tissues. Meantime, SOCS-2 to SOCS-6, which are structurally related to SOCS-1, have been cloned and overexpressed by us in M1 cells. SOCS-3 functions in an apparently similar manner to SOCS-1 in this assay system, whereas SOCS-2 and SOCS-5 have a weaker action, and SOCS-6 appears not to have activity [8].

Based on this information, domain swapping was analyzed between SOCS-1, SOCS-2 and SOCS-3, using M1 cells or a luciferase reporter system in 293T fibroblasts [8]. N-terminal and SH2 regions were interchangeable between SOCS-1 and SOCS-3, but not between SOCS-1 and SOCS-2. Although this evidence suggested a close similarity between SOCS-1 and SOCS-3, including an ability to suppress phosphorylation of STAT-3, the precise mechanisms of action must differ because, unlike SOCS-1, SOCS-3 does not inhibit the kinase activity of JAK1 or JAK2 [8].

An intriguing aspect of the lability of SOCS-containing proteins—necessary if they are to function as transient inducible suppressors of signaling—is the ability of the SOCS Box region to bind to elongins B and C 12, 13 and, by this mechanism, to potentially enter into a subsequent interaction with Cullin 2 to target SOCS Box-containing molecules for proteasomal degradation.

Back to Article Outline

Biologic actions of SOCS-1 

The data from SOCS-1 overexpression in M1 cells monitored only actions involving self-renewal and maturation [5]. Such cells generated smaller colonies than parental M1 cells, suggesting that SOCS-1 also might be partially suppressive of cell proliferation. However, M1 cells are autonomous and not dependent on growth stimulation by known cytokines, so no formal testing was possible of interference with cytokine-induced cell proliferation with the M1 cell system. In this context, it has been reported that overexpression of SOCS-1 in factor-dependent mast cell lines reduces proliferative responses to stem cell factor (SCF) [14], indicating that SOCS-1 may be able to suppress or modulate at least certain types of proliferative signaling cascades.

We attempted to characterize the biologic actions of SOCS-1 by generating mice with homozygous deletion of the SOCS-1 gene [15], and another group has described the phenotype of these mice [16].

In our inactivation procedure, the lacZ reporter gene was inserted into the deleted region of the SOCS-1 gene, allowing parallel studies in +/− mice on the cell types transcribing SOCS-1. SOCS-1+/− mice are fertile and phenotypically normal. From cross-mated +/− mice, apparently healthy SOCS-1−/− mice were born in the expected Mendelian ratio. On the day of birth, no histologic abnormalities were apparent. However, SOCS-1−/− mice failed to gain weight at the normal rate, developed a dry, scaly skin; became sick; and died by day 21 to 22 of age. This forced our analytical studies to be performed on less-than-ideal mice of 2 to 4 g body weight during a neonatal period when major changes in hematopoiesis and organ structure occur, but at a somewhat variable rate in individual mice.

Moribund SOCS-1−/− mice are half the weight of normal +/- or +/+ littermates, and there is a severe depletion of T and B lymphocytes, with thymus atrophy, mild granulocytosis, eosinopenia, and a mild degree of thrombocytopenia and anemia. The spleen is of normal weight, but it is markedly depleted of lymphocytes, with no lymphoid follicles, and contains excessive numbers of nucleated erythroid cells. In parallel with the smaller body weight, bone marrow cell numbers are half those in control mice, with, again, as the most obvious anomaly, a major reduction in marrow B lymphocytes15, 16, 17.

The various organ lesions in sick SOCS-1−/− mice are shown summarized in Figure 2. The most striking abnormality is a pale liver that shows extensive areas of fatty degeneration of hepatocytes, often with focal areas of necrosis, and extensive infiltration of the organ by immature macrophages and granulocytes (Fig. 3). Although normal at birth, the thymus becomes markedly atrophic, usually with no demarcated lymphoid cortex. The pancreas exhibits dispersion and atrophy of acinar tissue, with infiltration by macrophages and granulocytes. Similar cell infiltrates are present in the lung, both in the alveolar walls and particularly in perivascular cuffs (Fig. 3). The myocardium commonly shows macrophage and granulocyte infiltration, particularly in atrial regions and sometimes as a pericardial infiltrate. The kidney shows retardation of glomerular and tubular development. The skin is of particular interest, being invariably infiltrated in the dermal regions by macrophages with thickening of the epithelial cell layer and, never present in normal skin, keratinization of the superficial epithelial cells (Fig. 3). Remarkably, the keratinized regions of the skin epithelium commonly contain small focal aggregates of essentially pure populations of mature eosinophils. Tissues less frequently infiltrated by macrophages are the gut and meninges.

  • View full-size image.
  • Figure 2. 

    Mice lacking SOCS-1 die in neonatal life with fatty degeneration and necrosis of the liver, severe T- and B-lymphocyte depletion, and macrophage-granulocyte infiltration of the pancreas, heart, lung, and skin

  • View full-size image.
  • Figure 3. 

    SOCS-1−/− mice show fatty degeneration and necrosis of the liver with infiltrating hematopoietic cells (A), thickening of skin epithelium with keratinization and macrophage infiltration of the dermis (C), granulocyte-macrophage infiltration of the lung (E), and pancreas (G). In contrast, SOCS-1−/− IFNγ−/− mice show no organ pathology in the liver (B), skin (D), lung (F), or pancreas (H)

A reasonable assumption about these mice is that they die primarily from acute liver failure and that the thymus atrophy and possibly the B-lymphocyte depletion (either arrest at the pro–B-lymphocyte stage or selective loss of more mature cells) may be secondary to the severely stressed state of the animals. In this context, SOCS-1−/− lymphocytes have been reported to overexpress Bax and to exhibit accelerated apoptosis [16].

It is intriguing that the organs showing the most obvious abnormalities are those in which SOCS-1 expression is either constitutive (thymus, lung) or readily inducible (liver). This raises the possibility that the various types of tissue pathology may represent the independent dysfunction of those cell types normally expressing SOCS-1 and, therefore, likely to be dependent for normal function on the presence of SOCS-1.

A quite different possibility is that the infiltration of various organs by macrophages, granulocytes, and possibly other hematopoietic cells is not merely a response by normal cells to tissue injury but that organ damage actually is induced by the aberrant behavior of these infiltrating cells.

These alternatives eventually will be able to be explored by transplanting SOCS-1−/− marrow into normal recipients. These experiments are in progress.

Back to Article Outline

Biology of SOCS-1−/− hematopoietic cells 

Use was made of lacZ reporter gene in +/- mice to demonstrate that SOCS-1 is being transcribed in at least one third of progenitor cells, although continuing transcription is not detectable in the maturing progeny [17]. In SOCS-1−/− mice, the situation was quite different, with half of all progenitor cells actively attempting to transcribe SOCS-1, as were 80% of their maturing granulocyte and macrophage progeny. This differing transcription pattern may be due to the absence of the SOCS-1 protein in these cells, or it may be induced indirectly by perturbations in cytokine levels as a result of major tissue damage in other organs.

In cultures of marrow cells from −/− mice stimulated by CSFs or SCF, the most notable feature was the overall normality of the cultures in terms of colony numbers, relative frequencies of various progenitor cell subsets, colony size, and maturation [17]. These data clearly indicate that SOCS-1 is not essential for proliferative and maturation responses to the regulators tested, as monitored in clonal cultures.

Survival of −/− granulocyte-macrophage progenitor cells in cultures initially lacking a growth factor was similar to that of +/+ cells. Cells of both genotypes exhibited a half-life of 15 hours.

Despite this overall pattern of normality and the inability of SOCS-1 cells to exhibit autonomous proliferation, there were consistent abnormalities in SOCS-1 hematopoietic progenitor cells [17]. Macrophage-committed progenitor cells and their cluster-forming immediate progeny were significantly more numerous in −/− mice, providing a potential cellular basis for the greatly increased numbers of macrophages in various organs. Analysis of quantitative responsiveness to regulators documented a consistently increased responsiveness to GM-CSF, as might be expected in cells lacking a negative regulator of signaling from at least some cytokine receptors. Interestingly, however, quantitative responsiveness to IL-3 or M-CSF was normal. Other studies documented reduced responsiveness of mast cell lines to stimulation by SCF when SOCS-1 is overexpressed [14] and we have noted a minor hyperresponsiveness of −/− progenitor cells to stimulation by SCF (D. M., unpublished data, 1999).

In terms of overall progenitor cell numbers, −/− mice contain only half of the total numbers present in +/+ or +/− mice, again paralleling their correspondingly reduced body size; however, aberrantly in −/− mice, one fourth of all such progenitor cells were in the liver, vs 3 to 4% in control mice.

The most striking abnormality of −/− granulocyte-macrophage progenitor cells was their greatly increased susceptibility to inhibition by IFN-γ [17]. IFN-γ is not a proliferative stimulus for granulocyte-macrophage progenitor cells and, when combined with CSFs or SCF, tends to reduce the numbers of granulocytic, granulocyte-macrophage, and macrophage colonies that develop. This inhibition was much more evident with −/− than +/+ cells, implying that SOCS-1 normally significantly reduces signaling from the IFN-γ receptor, which is in agreement with the original experiments in M1 leukemic cells overexpressing SOCS-1. What was of particular interest was that the degree of excess inhibition observed (up to 60-fold) varied widely according to the stimulating factor involved, being least with GM-CSF (to which SOCS-1−/− cells are hyperresponsive) and greatest with M-CSF. These very obvious differences imply strongly that IFN-γ is not exerting its suppressive effects directly, but that signals from the activated IFN-γ receptor interfere to a varying degree with signals from activated growth factor receptors, the latter signals being necessary for survival and proliferation of these cells. Transfer studies using M-CSF–initiated clones showed that the inhibitory action of IFN-γ was a direct effect on the proliferating cells and was more evident with −/− than +/+ cells (Table 1). The aberrant responses of −/− cells to IFN-γ documented in these in vitro studies become relevant in view of parallel in vivo studies on the behavior of SOCS-1−/− mice.

Table 1. Direct inhibition by IFN-γ of M-CSF–stimulated survival and proliferation of developing macrophage colonies
Age of clone at transfer (days) Stimulus in recipient cultureFate of transfered clones
HealthyUnhealthy or dead
GenotpyeM-CSFIFN-γColoniesClustersColoniesClusters
4−/−+−32200
++001431
3−/−+−29600
++01123
+/++−28300
++92526
IFN-γ = interferon γ; M-CSF = macrophage colony-stimulating factor.

Back to Article Outline

IFN-γ is essential for the development of neonatal disease in SOCS-1−/− mice 

During the course of our work with SOCS-1−/− mice, we were intrigued to come across some early studies on IFN-γ that seemed to have dropped out of current reviews on this agent. IFN-γ–injected neonatal mice failed to gain weight, developed fatty degeneration of the liver, and died within 21 days—a disease pattern strikingly resembling to that seen in SOCS-1−/− mice 18, 19. These studies also noted the occurrence of thymus atrophy and subtle structural changes in renal glomeruli.

We repeated these studies using daily injections of 3μg IFN-γ for 2 weeks in neonatal C57BL mice and confirmed these basic observations. Based on our knowledge of other abnormalities in SOCS-1−/− mice, we also established that these IFN-γ–injected mice developed reduced numbers of lymphocytes in the blood, marrow, and spleen, and mild granulocytosis, thrombocytopenia, and anemia. The mice also developed cellular infiltrates in the lungs, pancreas and skin—changes that, apart from the major liver damage, reproduced in milder form all of the abnormalities noted in SOCS-1−/− mice [20].

In parallel, we observed that the daily injection of monoclonal IFN-γ antibodies, but not of IL-6 antibodies or isotype-matched control immunoglobulin, prevented neonatal death in SOCS-1−/− mice and either eliminated or markedly reduced organ histologic abnormalities when such mice were analyzed at 21 days of age [20].

By crossing SOCS-1−/− mice with IFN-γ−/− mice, we found that double knockout mice did not die in the neonatal period but developed into fertile, apparently healthy, adult animals. When examined at 21 days of age, such mice showed no hematologic abnormalities and essentially no pathologic changes other than an unusual tendency to have enlarged medullary regions in the thymus [20].

To date, studies on media conditioned by SOCS-1−/− or +/+ organs have not revealed any consistent evidence that SOCS-1−/− tissues overproduce IFN-γ. Therefore, it may be that SOCS-1−/− tissues are merely hyperresponsive to normal IFN-γ levels. But what triggers such IFN-γ production in neonatal mice, and, specifically, is it a response to inductive signals from microorganisms? To explore this question, we derived germ-free SOCS-1−/− mice and analyzed their fate. Germ-free SOCS-1−/− mice became ill and died at the same age as conventional −/− mice. Histologic and hematologic analysis of sick −/− germ-free mice revealed identical pathology to that seen in conventional −/− mice, including even the eosinophil foci in the keratinized skin of the −/− mice [20]. These observations exclude a necessary involvement of microorganisms from the chain of events leading to fatal IFN-γ–induced organ damage but did not allow the inducing signals for IFN-γ transcription to be established.

This chain of evidence identifies IFN-γ as being essential for the development of neonatal disease and death in SOCS-1−/− mice. The findings do not necessarily document a direct action of IFN-γ on all the various affected cell types. The action of IFN-γ does seem to be direct on hematopoietic progenitor cells but, for other cell types, could well be indirect and involve the induction by IFN-γ of the production of other toxic agents. The findings do not exclude the possibility that other molecules may be necessary for the development of the syndrome. In this context, the administration of TNF to neonatal mice also causes severe liver damage [21]. The findings also do not allow discrimination between the two general possibilities of multiple independent organ failure vs some type of integrated induction of organ failure, possibly involving the action of aberrant hematopoietic cells, including the infiltrating macrophages and granulocytes.

Identification of an essential initiating agent of disease and death in neonatal SOCS-1−/− mice leaves a number of unresolved questions. Why was SOCS-3 not able to compensate for loss of SOCS-1 when its actions on M1 leukemic cells are comparable? If SOCS-1−/− mice can be rescued from neonatal death by IFN-γ antibodies or crossing with IFN-γ2/− mice, will they now remain in good health, despite the continuing lack of SOCS-1? On this latter question, I suspect that events will prove that SOCS-1−/− IFN-γ−/− mice will develop a range of abnormalities. SOCS-1 is a modulator or suppressor of signaling from a variety of receptors already known to include those for IL-6, LIF, OSM, IFN-γ, SCF, IL-2, GM-CSF, and growth hormone 5, 14, 16, 17, 22, 23. Lack of this suppressive action of SOCS-1 potentially makes the mice polytransgenic, being overstimulated by what are, in reality, normal concentrations of these regulators. There is no experience of the potential consequences of a complex multiple transgenic state. Almost any combination of tissue changes could be envisaged as developing with time in these mice, unless the cells can belatedly compensate for this perceived signaling imbalance by using some related suppressor molecule.

It is unlikely that SOCS-1 plays a role that is of major significance only for signaling from the IFN-γ receptor. For example, its actions in suppressing signaling from gp130 are quantitatively more marked in M1 leukemic cells than those in suppressing signaling from the IFN-γ receptor.

The chapter about to commence on the late fate of SOCS-1−/− IFN-γ−/− mice may turn out to be a complex mixture of the consequences of overaction and interaction of a number of possible regulators. The story should continue to be intriguing, the more so if, in fact, nothing unusual happens.

Back to Article Outline

Intracellular biology of SOCS-1 action 

In cells not constitutively transcribing SOCS-1, the temporary transcription of SOCS-1 is inducible by signaling from at least several cytokine receptors, including those for IFN-γ, IL-6, GM-CSF, and IL-3. This induced transcription is of brief duration. Therefore, SOCS-1 can be viewed as a labile inducible molecule with an ability to suppress or modulate activation of certain receptors or their immediate signaling intermediates.

The actual mechanisms involved seem likely to differ somewhat according to the receptor involved. As shown in Figure 1, for gp130-containing receptors, SOCS-1 in sequence blocks phosphorylation of JAK1, gp130, then STAT3. For the IFN-γ receptor, SOCS-1 blocks phosphorylation of JAK1 and JAK2, then STAT1. For the SCF receptor, SOCS-1 does not inhibit the catalytic activity of the kit tyrosine kinase but binds to Grb-2 and vav [14].

A typical cytokine-activated receptor complex is capable of initiating a variety of cellular responses from various regions in its cytoplasmic domains. Present information is incomplete as to whether SOCS-1 can influence all, or only some, of these signaling cascades and whether its action necessarily always is inhibitory in nature. Known consequences of SOCS-1 action are summarized in Figure 4.

  • View full-size image.
  • Figure 4. 

    The action of SOCS-1 is, in general, to modulate or suppress signaling from multiple cytokine receptors. According to the receptor and cell type involved, this can influence cell proliferation, differentiation commitment, survival, or functional activation. Conversely, SOCS-1 may enhance GM-CSF production by lung tissue

The actions of SOCS-1 in M1 cells block cytokine-induced suppression of self-renewal and maturation induction. In mast cell lines, SOCS-1 blocks proliferative stimulation by SCF and, by deduction from progenitor cells lacking SOCS-1, SOCS-1 partially blocks proliferative stimulation by GM-CSF but not by M-CSF or IL-3. The hypersusceptibility of SOCS-1−/− hematopoietic progenitor cells to IFN-γ may be based on an inhibitory action of SOCS-1 on IFN-γ receptor signaling, although the response is likely to be indirect, because it is influenced by the cytokine in use to stimulate cell survival and proliferation. In mature SOCS-1−/− macrophages, loss of the modulating action of SOCS-1 actually enhances the phagocytic activity of mature macrophages in response to IFN-γ [20]. This suggests that the consequences of the IFN-γ–SOCS-1 interaction for cells within a single lineage can depend on the maturation status of the cells.

Conversely, the action of SOCS-1 in some cell types may have positive effects. In current experiments, GM-CSF production by SOCS-1−/− organs, particularly the lung, is markedly subnormal. Although this may be an indirect consequence of organ pathology, it is possible that SOCS-1 might have a positive action in enhancing GM-CSF production by the lung cells involved.

For other cell types, the role being played by SOCS-1 is uncertain. Why does an hepatocyte lacking SOCS-1 develop fatty degeneration in response to IFN-γ? If the action of IFN-γ on hepatocytes is a direct one, which type of signaling from the IFN-γ receptor leads to cellular dysfunction, and what are the mechanisms involved? Although it could be postulated that SOCS-1 inhibits all types of signaling from an activated IFN-γ receptor, this may prove to be a misleading oversimplification. Cellular dysfunction may result from an imbalance of signaling if some types of signaling, but not others, are suppressed.

The action of SOCS-1 in blocking cytokine-induced suppression of self-renewal or maturation induction could encourage the emergence of leukemic cells or, at the very least, of dysplastic cells. No information is yet available on the status of any of the SOCS group of proteins in primary leukemias or myelodysplasias. Any information on this question would be of great interest.

Finally, it cannot be assumed that all members of this family of 20 proteins sharing a SOCS box necessarily have similar functions. Some may have opposite actions or completely unrelated actions. The SOCS box may, in fact, be no more than a targeting sequence ensuring prompt degradation of molecules with unrelated actions.

The first chapter of the SOCS-1 story may have been revealed, but, for the group of proteins sharing a SOCS motif, their stories remain in the future.

Back to Article Outline

Acknowledgements 

The work from the author's laboratory was supported by the Carden Fellowship Fund of the Anti-Cancer Council of Victoria, the National Health and Medical Research Council, Canberra, an Australian Government Cooperative Research Centre Programme Grant, and the Grant CA-22556 from the National Institutes of Health, Bethesda, Maryland.

Back to Article Outline

References 

  1. Smith A, Metcalf D, Nicola NA. Cytoplasmic domains of the common β-chain of the GM-CSF/IL-3/IL-5 receptors that are required for inducing differentiation or clonal suppression in myeloid leukaemic cell lines. EMBO J. 1997;16:451
  2. Dong F, Van Buitenen C, Pouwels K, Hoefsloot LH, Löwenberg B, Touw IP. Distinct cytoplasmic regions of the human granulocyte colony-stimulating factor receptor involved in induction of proliferation and maturation. Mol Cell Biol. 1993;13:7774
  3. Metcalf D (1984) The hemopoietic colony stimulating factors. Elsevier, Amsterdam
  4. Rayner JR, Gonda TJ. A simple and efficient procedure for generating stable expression libraries by cDNA cloning in a retroviral vector. Mol Cell Biol. 1994;14:880
  5. Starr R, Willson TA, Viney EM, Murray LJL, Rayner JR, Jenkins BJ, et al. A family of cytokine-inducible inhibitors of signalling. Nature. 1997;387:917
  6. Endo TA, Masuhara M, Yokouchi M, Suzuki R, Sakamoto H, Mitsui K, et al. A new protein containing an SH2 domain that inhibits JAK kinases. Nature. 1997;387:921
  7. Naka T, Narazaki M, Hirata M, Matsumoto T, Minamoto S, Aono A, et al. Structure and function of a new STAT-induced STAT inhibitor. Nature. 1997;387:924
  8. Nicholson SE, Willson TA, Farley A, Starr R, Zhang J-G, Baca M, et al. Mutational analyses of the SOCS proteins suggest a dual domain requirement but distinct mechanisms for inhibition of LIF and IL-6 signal transduction. EMBO J. 1999;18:375
  9. Narazaki M, Fujimoto M, Matsumoto T, Morita Y, Saito H, Kajita T, et al. Three distinct domains of SSI-1/SOCS-1/JAB protein are required for its suppression of interleukin 6 signaling. Proc Natl Acad Sci U S A. 1998;95:13130
  10. Hilton DJ, Richardson RT, Alexander WS, Viney EM, Willson TA, Spring NS, et al. Twenty proteins containing a C-terminal SOCS box form five structural classes. Proc Natl Acad Sci U S A. 1998;95:114
  11. Yoshimura A, Ohkubo T, Kiguchi T, Jenkins NA, Gilbert DJ, Copeland NG, et al. A novel cytokine-inducible gene CIS encodes an SH2-containing protein that binds to tyrosine-phosphorylated interleukin 3 and erythropoietin receptors. EMBO J. 1995;14:2816
  12. Kamura T, Sato S, Haque D, Liu L, Kaelin WG, Conaway RC, et al. The elongin BC complex interacts with the conserved SOCS-box motif present in members of the SOCS, ras, WD-40 repeat, and ankyrin repeat families. Genes Dev. 1998;12:3872
  13. Zhang J-G, Farley A, Nicholson SE, Willson TA, Zugara LM, Simpson RJ, et al. The conserved SOCS box motif in suppressors of cytokine signaling binds to elongins B and C and may couple bound proteins to proteasomal degradation. Proc Natl Acad Sci U S A. 1999;96:2071
  14. De Sepulveda P, Okkenhaug K, La Rose J, Hawley RG, Dubreuil P, Rottapel R. SOCS1 binds to multiple signalling proteins and suppresses Steel factor-dependent proliferation. EMBO J. 1999;18:904
  15. Starr R, Metcalf D, Elefanty AG, Brysha M, Willson TA, Nicola NA, et al. Liver degeneration and lymphoid deficiencies in mice lacking suppressor of cytokine signaling-1. Proc Natl Acad Sci U S A. 1998;95:14395
  16. Naka T, Matsumoto T, Narazaki M, Fujimoto M, Morita Y, Ohsawa Y, et al. Accelerated apoptosis of lymphocytes by augmented induction of Bax in SSI-1 (STAT-induced STAT inhibitor-1) deficient mice. Proc Natl Acad Sci U S A. 1998;95:15577
  17. Metcalf D, Alexander WS, Elefanty AG, Nicola NA, Hilton DJ, Starr R, et al. Aberrant hematopoiesis in mice with inactivation of the gene encoding SOCS-1. Leukemia. 1999;13:926
  18. Gresser I, Aguet M, Morel-Maroger M, Woodrow D, Puvion-Dutilleul F, Guillon J-C, et al. Electrophoretically pure mouse interferon inhibits growth, induces liver and kidney lesions, and kills suckling mice. Am J Pathol. 1981;102:396
  19. Gresser I. Can interferon induce disease?. Interferon. 1982;4:95
  20. Alexander WS, Starr R, Fenner J, Scott CL, Handman E, Sprigg NS, et al. SOCS-1 is a critical inhibitor of interferon-γ signaling and prevents the potentially fatal neonatal actions of this cytokine. Cell. 1999;98:597
  21. Gresser I, Woodrow D, Moss J, Maury C, Tavernier J, Fiers W. Toxic effects of recombinant tumor necrosis factor in suckling mice. Am J Pathol. 1987;128:13
  22. Adams TE, Hansen JA, Starr R, Nicola NA, Hilton DJ, Billestrup N. Growth hormone preferentially induces the rapid, transient expression of SOCS-3, a novel inhibitor of cytokine receptor signaling. J Biol Chem. 1998;273:1285
  23. Sakamoto H, Yasukawa H, Masuhara M, Tanimura S, Sasaki A, Yuge K, et al. A Janus kinase inhibitor, JAB, is an interferon-γ-inducible gene and confers resistance to interferons. Blood. 1988;92:1668

☆ Presented as the ISEH Donald Metcalf Lecture at the 28th Annual Meeting of the International Society for Experimental Hematology, Monte Carlo, Monaco, July 11, 1999.

PII: S0301-472X(99)00120-4

Experimental Hematology
Volume 27, Issue 12 , Pages 1715-1723, December 1999

Article Tools

* Image Usage & Resolution