| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Articles |
Department of Medicine, Cedars-Sinai Research Institute, University of California School of Medicine, Los Angeles, California 90048
Address all correspondence and requests for reprints to: Dr. Shlomo Melmed, M.D., Division of Endocrinology and Metabolism, Cedars-Sinai Medical Center, 8700 Beverly Boulevard, B-131, Los Angeles, California 90048. E-mail: melmed{at}cshs.org
Abstract
Leukemia inhibitory factor (LIF) exhibits multiple biological activities in various tissues, and we have shown that LIF activates POMC gene transcription in response to immune signals. As higher serum levels of LIF have been reported in septicemia, we measured LIF values in biological fluids by RIA. Immunoreactive LIF was detected in 303 of 428 human serum samples. Circulating LIF detection rates were 69% in acute inflammatory diseases, 83% in chronic inflammatory diseases, 61% in noninflammatory diseases, and 90% in cancer patients. Serum concentrations of human LIF was higher in patients with inflammatory disease than in noninflammatory disease (0.80 ± 0.10 vs. 0.53 ± 0.02 ng/mL; P < 0.05) or in cancer patients (0.44 ± 0.06; P < 0.05). Higher serum human LIF levels were found in septicemia (0.78 ± 0.14 ng/mL), pneumonia (0.80 ± 0.10 ng/mL), acute bronchitis (0.88 ± 0.09 ng/mL), other infections (1.01 ± 0.17 ng/mL), and systemic lupus erythematosus (SLE; 0.79 ± 0.06 ng/mL). In 7 septicemia patients, Gram-negative infection was associated with higher LIF levels (1.06 ± 0.16 ng/mL) than was Gram-positive infection (0.58 ± 0.14 ng/mL). In patients with acute inflammatory disease, serum LIF levels decreased within several days after hospitalization.
To test circulating mouse (m) LIF changes in response to inflammatory stress, lipopolysaccharide (LPS) was injected ip to mice. LPS increased serum mLIF values concordantly with ACTH levels. After ip injection of 80 µg LPS, serum mLIF increased by 144% (P < 0.05), 173% (P < 0.05), and 134% at 30, 90, and 120 min respectively. In vitro, however, LPS did not increase ACTH and mLIF secretion from dispersed mouse primary pituitary cells.
These results suggest that LIF is an important participant in the pathogenesis of the acute inflammatory response. The elevated serum LIF levels observed in inflammation do not appear to originate from the pituitary.
LEUKEMIA inhibitory factor (LIF) is a pleiotropic cytokine that exhibits multiple functions in various tissues and cell types (1). An important function of LIF is to activate POMC gene transcription in response to immune signals (2, 3, 4, 5). Our previous studies have demonstrated that human pituitary (2) as well as mouse hypothalamus and pituitary (3) express both LIF and LIF receptor genes, predominantly in corticotrophs (2). Furthermore, we found that LIF induces POMC transcription, resulting in a significant increase in ACTH secretion from pituitary cells in vitro (2) as well as in vivo (4), and LIF potentially synergizes with both CRH and cAMP in induction of POMC transcription (5). In vivo, we have shown that lipopolysaccharide (LPS) administered to mice induces hypothalamic and pituitary mouse (m) LIF messenger ribonucleic acid (mRNA) and mLIF receptor mRNA, and increases serum ACTH levels (3).
However, the quantitative levels of endogenous LIF under physiological and pathological conditions in vivo are still unclear. Waring et al. found that serum LIF levels were transiently elevated to 2–200 ng/mL in six subjects with Gram-negative septic shock (6). In a similar study, Guillet et al. showed that LIF was detected in 40% of 40 patients suffering from septic shock, with levels varying from 10–1000 pg/mL and with no correlation between serum LIF and oncostatin M (OSM) and interleukin-6 (IL-6) (7). Villers detected LIF in 11 of 34 septic patients (plasma levels of 0.1–34 ng/mL) (8). The central role of LIF in septic shock was suggested by the 50% improved survival in mice after injection of anti-LIF antibody to mice given a lethal dose of LPS (9).
Measurement of human and murine LIF in biological fluids was performed by bioassay, with a detectable level LIF of 2 ng/mL (10). Improved sensitivity was achieved with a RRA, which detected 1 ng/mL human (h) LIF (6). Recently, enzyme-linked immunosorbent assay methods have been developed with detectable hLIF levels varying from 1–150 pg/mL using either monoclonal or polyclonal antibodies (7, 11, 12, 13). To determine the significance of circulating LIF and LIF regulation in physiological and pathological conditions, we established a specific, sensitive, precise, and facile LIF RIA and measured LIF in biological fluids.
Materials and Methods
Radioiodination of LIF
Escherichia coli-derived recombinant hLIF, mLIF, goat polyclonal anti-hLIF antibody, and anti-mLIF antibody were commercially purchased (R&D Systems, Minneapolis, MN). LIF was iodinated using 125iodine (DuPont, Boston, MA) and Iodogen (Pierce, Rockford, IL) (14). Labeled LIF was purified through a P-6 DG desalting column (Bio-Rad Laboratory, Richmond, CA). Incorporation of iodine was determined by trichloroacetic acid precipitation (15), and about 88% of the peak fraction of LIF eluted. The specific radioactivity of labeled LIF measured by displacement analysis (16, 17) was 54–57 µCi/µg.
LIF RIA
RIA was performed by incubation of sample or LIF
standard, antibody, and [125I]LIF in a total volume of
300 µL for 72 h at 4 C. Precipitation of binding pellet was
achieved through centrifugation after the addition of goat IgG (Sigma
Chemcial Co., Louis, MO), rabbit antigoat antiserum [1:6 in
phosphate-buffered saline (PBS); Gemini, Calabasas, CA], and 6%
polyethylene glycol (Sigma) and counting using a 
-counter. The
standard curve ranged from 250–100,000 pg/mL. The anti-hLIF antibody
used in the assay was reported by the manufacturer to exhibit no
cross-reactivity with hIL-1 
, hIL-2, hIL-3, hIL-4, hIL-6, human
tumor necrosis factor (hTNF), human granulocyte colony-stimulating
factor (hG-CSF), human granulocyte/macrophage CSF (hGM-CSF), human
transforming growth factor-ß (hTGF), or fibroblast growth factor. The
anti-mLIF antibody used in the assay was reported to exhibit no
cross-reactivity with TGF, epidermal growth factor, platelet-derived
growth factor, fibroblast growth factor, IL-1, IL-2, IL-3, IL-4, IL-5,
IL-6, IL-7, IL-8, IL-9, IL-10, IL-11, IL-12, CSF, macrophage
inflammatory protein, TNF, OSM, SLPI, interferon (IFN), insulin-like
growth factor I, insulin-like growth factor II, macrophage inflammatory
protein-2, monocyte chemoattractant protein, PTN, vascular endothelial
growth factor, or HGF.
Size-exclusion chromatography
About 500,000–1,000,000 cpm 125I-labeled hLIF (or
mLIF) with or without 40-fold unlabeled hLIF were incubated in 1 mL
human (or murine) serum or 0.01 mol/L PBS overnight at 4 C with
stirring. The mixture was then applied to a 1.5 x 95-cm column of
Ultrogel AcA 22 (IBF-Biotechnis, Villeneuvela Garenne, France), eluted
at 1.5 mL/fraction with 0.01 mol/L PBS, and counted in a
-counter.
Western ligand blot
Five microliters of pregnant mouse (or human) serum were diluted and then applied to 5% SDS-PAGE. Proteins were transferred to nitrocellulose membrane (Amersham, Arlington Heights, IL), incubated with [125I]hLIF, and identified by autoradiography. hLIF is able to bind soluble murine LIF receptor, like LIF-binding protein, with high affinity (18).
Human serum samples
This project was approved by Cedars-Sinai Medical Center Institutional Review Board. Serum samples collected as part of the routine care of patients seen in the Ambulatory Clinic or hospitalized at Cedars-Sinai Medical Center between June 1994 to May 1996 were obtained from the Clinical Pathology Laboratory after the tests ordered by the patients’ physicians had been run. The diagnosis was obtained from patient’s medical records. Collected samples were stored anonymously at -20 C until assayed. Samples in each study were measured in duplicate in single LIF RIA assay.
Murine serum samples
Six-week-old B6D2F1 mice were purchased from Jackson Laboratory (Bar Harbor, ME). Twenty mice were injected ip with 80 µg LPS (Sigma) dissolved in 0.3 mL saline, and three or four mice were killed 30, 90, and 120 min, respectively, after the injection. Blood was collected, and serum was stored at -20 C until measurement of mLIF and ACTH. These experiments were repeated twice. Sera were also obtained from normal and pregnant mice. LIF concentrations were measured in duplicate using a single assay for each experiment.
Pituitary cultures
Murine pituitaries were dissected and washed, followed by digestion with 0.035% collagenase (Sigma) and 0.01% hyaluronidase (Sigma) for 20 min at 37 C, as previously described (2). Cells were preincubated in DMEM with 2.5% FBS for the times indicated, and then exposed to LIF, CRH (American Peptide), LIF plus CRH, or LPS for an additional 24 h. At end of the experiments, conditioned medium was collected, and cells were lysed in 1 mol/L NaOH through freezing and thawing cycles. The collected medium was concentrated 10 times by lyophilization, reconstituted with water, and desalted through a P-6 column (Bio-Rad). Measurement of medium LIF or ACTH was performed in duplicate in a single assay for each experiment. The protein content of the cells was measured by the method of Bradford (19), using BSA as a standard and Bio-Rad protein reagents (Bio-Rad).
ACTH assay
ACTH levels in sera and conditioned medium were measured using a RIA kit (Diagnostic Products Corp., Los Angeles, CA).
Statistics
Data are presented as the mean ± SEM. Student’s t test was used for comparison of two groups.
Results
LIF assays
The LIF RIA had a sensitivity of 50 pg/mL for hLIF and 250 pg/mL
for mLIF, and a midvalue of 3 ng/mL for hLIF and 4 ng/mL for mLIF at a
50% bound/Bo. The slope of the assay was -2.36 for hLIF RIA and
-2.65 for mLIF RIA. No cross-reaction of hLIF was found with a mixture
of IL-1, IL-2, IL-4, IL-6, IL-8, IL-10, IL-12, G-CSF, GM-CSF, IFN, and
TGF at doses ranging from 1.4–131 ng/mL; hOSM at a dose of 1000 ng/mL;
or mLIF at a dose of 100 ng/mL in the hLIF RIA. In the mLIF RIA, no
cross-reaction with mIL-6, hLIF, or hOSM was found (Fig. 1
).
|
, when 1
or 2 ng hLIF were added to 1 mL human serum, 89% or 101% hLIF was
recovered, respectively. Two, 5, or 50 ng mLIF were added to 1 mL
murine serum, and 96–110% of the mLIF was recovered. These results
suggest that no interfering substance in human serum or normal mouse
serum influences LIF measurement in the RIAs. The recovery of mLIF
added to undiluted pregnant mouse serum, however, was higher (135%),
and this elevated recovery was decreased to 105% by dilution of the
pregnant mouse serum with PBS at a ratio of 1:8. After conditioned
medium was concentrated 10-fold by lyophilization, reconstituted with
water, and desalted through a P-6 column, about 80% of the added LIF
was recovered.
|
LIF-binding protein in serum
Iodine-labeled recombinant hLIF exhibited a molecular mass of 20
kDa by size-exclusion chromatography (Fig. 2A). When [125I]hLIF was
incubated with human serum and applied to the column, a labeled complex
with about one sixth of the total applied radioactivity was found in
fractions corresponding to molecular weights greater than 100 kDa. The
quantity of this labeled complex in human pregnant serum was not
greater than that found in nonpregnant serum, as shown in Fig. 2B. A
40-fold excess of cold hLIF could not displace the labeled complex in
human pregnancy serum (Fig. 2B), and Western ligand blot did not show
LIF-binding protein in human pregnancy serum, suggesting that the
binding substance was not specific for hLIF. In contrast, a previously
reported murine LIF-binding protein (
90 kDa) in mouse pregnancy
serum was confirmed by chromatography (Fig. 2C) as well as by Western
ligand blotting.
|
LIF levels were screened in sera obtained from patients with
varying diseases. hLIF was detected in 236 of the 353 samples (67%)
obtained from patients with benign disorders. The LIF detection rates
were not different among patients groups with acute or chronic
inflammatory diseases or noninflammatory diseases (Table 2). The average serum level of LIF in
patients with noninflammatory diseases (0.53 ± 0.02 ng/mL; range,
0.07–1.26 ng/mL) was not different when stratified by system of
involvement: 0.50 ± 0.09 ng/mL for nervous system, 0.62 ±
0.06 ng/mL for cardiovascular system, 0.46 ± 0.03 ng/mL for
pulmonary system, 0.49 ± 0.04 ng/mL for gastrointestinal system,
0.56 ± 0.04 ng/mL for renal system, 0.42 ± 0.01 ng/mL for
reproductive system, and 0.68 ± 0.09 for skin and muscle system.
The average serum LIF level in patients with inflammatory diseases was
0.84 ± 0.05 ng/mL (range, 0.20–4.04), which was higher
(P < 0.001) than that in the noninflammatory disease
group. Each inflammatory subgroup, except rheumatoid arthritis, had
higher (P < 0.05) LIF levels than patients with
noninflammatory diseases (Fig. 3). In
seven patients with septicemia, Gram-negative infection was associated
with higher LIF concentrations (1.06 ± 0.16 ng/mL) than those in
Gram-positive infection (0.57 ± 0.14 ng/mL; P =
0.069). Of interest, two patients with bone fracture had serum LIF
concentrations of 1.36 and 1.51 ng/mL. In acute inflammatory disease
samples, LIF levels obtained on admission were higher (0.97 ±
0.10 ng/mL) than those (0.70 ± 0.05 ng/mL) in the samples
obtained 2–5 days later (P = 0.038), as shown in Fig. 4.
|
|
|
LIF levels in mouse serum
The murine LIF concentration in serum obtained from five pregnant
mice (33.8 ± 3.9 ng/mL) was higher (P < 0.001)
than that in serum derived from six normal mice (5.8 ± 0.3
ng/mL). To examine changes in circulating LIF concentrations in
response to inflammatory stress, mice were injected ip with LPS.
Pretreatment mLIF and ACTH levels in serum were 5.6 ± 0.8 ng/mL
and 22.3 ± 7.0 pg/mL, respectively. LPS increased mLIF
concentrations in mouse serum concordantly with ACTH levels, as shown
in Fig. 5. After injection of LPS, serum
mLIF increased by 144% (P < 0.05), 173%
(P < 0.05), and 134%, ACTH increased by 126%, 218%
(P < 0.05), and 152% at 30, 60, and 90 min,
respectively.
|
To determine a potential source of circulating LIF during
inflammation, we performed several in vitro experiments.
After preincubation for 48 h, primary mouse pituitary cells
treated with LIF showed increased ACTH secretion (Fig. 6). The average yield of LIF from the
cultured cells of each single mouse pituitary was 30 pg during 48
h of culture. In contrast to the in vivo studies, LPS did
not stimulate either mLIF or ACTH secretion from primary pituitary
cells in vitro (Fig. 7). These
results suggest that the increased serum LIF, in response to
stimulation by LPS, was not derived from the pituitary.
|
|
We developed a highly specific, sensitive, and precise RIA for LIF. This assay is easy to perform and economical for screening a large number of samples. However, it should be noted that no LIF standard based upon LIF mass is available. LIF activity has previously been based on LIF bioactivity, which may result in different values when other standards are used. In our study, all samples in each experiment were measured in a single assay. Therefore, LIF levels measured in this RIA should be considered relative immunoreactive values.
We used this assay to measure serum mLIF in pregnant and nonpregnant mice, and the higher mLIF levels noted in pregnancy serum were consistent with a previous report (18). Furthermore, complementary DNA of mLIF-transfected murine pituitary cell line AtT-20 produced large amounts of mLIF compared with wild type of AtT-20 cells (20). These data suggest that the LIF RIA is a useful tool to measure LIF in biological fluids.
The elevated serum recovery of mLIF in pregnant mouse serum (135%)
reflects the high concentration of LIF-binding protein in pregnant
mouse serum that was previously reported to have a circulating
concentration of 32 µg/mL (18). This binding protein competes with
LIF antibody binding to the [125I]LIF, resulting in less
[125I]LIF binding to anti-LIF antibody giving apparent
high serum LIF values. We could not detect a specific LIF-binding
protein in human serum, consistent with a previous report (18).
However, a small amount of nonspecific binding of labeled hLIF in human
serum was detected. This nonspecific binding may result in a slightly
elevated LIF level in human serum. Recovery studies shown in Table 1
,
however, indicated that this material did not significantly influence
LIF values measured by this RIA.
Our results show that higher serum levels of LIF are associated not
only with septicemia, but also with other acute inflammatory diseases,
including pneumonia, acute bronchitis, acute pyelonephritis, pyogenic
arthritis, cellulitis of the foot, and postoperative infection, as well
as bone fracture. In response to infection or injury, a complex series
of reactions (inflammation) is executed by the host. The acute phase
response includes activation of macrophages and other cells, increase
in cytokine release, and stimulation of acute phase plasma protein
production by the liver (21). It has been well established that TNF,
IL-1, and IL-6 are important inflammatory cytokines during the acute
phase response (22, 23). Recently, LIF has been recognized as another
inflammatory cytokine (1, 6, 9). Beside high LIF concentrations in
serum of septic patients (6, 8), LIF levels were also elevated in
circulating fluids of patient with giant cell arteritis (12) and
patients receiving kidney transplants (13), in infectious pleural
effusions (24), and in inflammatory synovial fluid (25). Our data
showing high serum LIF levels in almost all acute inflammatory
diseases strongly suggest that LIF is an important
participator in acute inflammatory responses to immune signals.
This hypothesis was supported by our studies in mice in which LPS
induced both hypothalamic and pituitary LIF mRNA expression (3) and led
to a increase in LIF concentration concordant with that in serum ACTH
levels.
In patients with acute inflammatory diseases, serum LIF levels significantly decreased within several days after hospitalization. Furthermore, mLIF as well as ACTH levels in mouse serum peaked 1 h after LPS injection, then declined to normal 2 h after injection. These results are consistent with the finding that the acute phase response in the human lasts for 24–48 h, followed by a return to normal organ function (21). The limitation of the acute phase response may be due to several factors. First, the stimulator causing cytokine release may be disrupted by treatment or the body defense system. Second, tissue capacity for cytokine production may be limited. Third, the short half-life of many cytokines in the circulation may limit their ability to serve as mediators of the acute phase response. It has been demonstrated that murine LIF injected ip into adult mice has an initial half-life of 6–8 min and a prolonged secondary clearance phase (26). Fourth, cytokines stimulate ACTH production (2, 5), which, in turn, induces cortisol production that induces a negative feedback loop to inhibit cytokine gene expression. In support of this hypothesis, inhibition of LIF expression by glucocorticoids in cultured rat anterior pituitary (27) and a human thyroid carcinoma cell line (28) has been observed. Finally, natural antagonists, such as IL-1 receptor antagonist, soluble TNF receptor, IL-4, and IL-10, markedly interfere with the ongoing cascade of acute inflammation (21).
Cellular sources of circulating LIF in the acute phase response are
unclear. Our previous studies showed that normal human and murine
pituitaries express LIF (2), LPS induces LIF mRNA expression in
pituitary in vivo (3), and anti-hLIF antibody inhibits
endogenous LIF-induced ACTH secretion from AtT-20 cells (2). In the
present study, we observed detectable LIF in the conditioned medium of
cultured pituitary cells, but the yield of about 30 pg LIF from single
mouse pituitary during 48-h culture is unlikely to be a source of the
5–10 ng/mL LIF detected in mouse serum. LPS, at a dose 200 ng/mL,
minimally stimulated LIF mRNA expression in normal and malignant rat
glial cells in vitro (29). However, in our studies, LPS at
wide dose ranges did not stimulate either mLIF or ACTH from cultured
primary mouse pituitary cells or AtT-20 cells (Ren SG and Melmed S;
unpublished data). Although LPS (1 ng/mL) stimulated primary mouse
pituitaries in vitro to produce 300 pg/mL (10 pmol/L) IL-6
(30), IL-6 at this concentration was too low to stimulate LIF
production (2). These data suggest that the elevated serum LIF level
observed in inflammation may not originate from the pituitary. LIF is
produced from multiple cell types, including macrophages, monocytes,
and fibroblasts, and LIF production can be stimulated by TNF, IL-1,
IL-6, and other cytokines and peptides (1, 9, 31). During inflammatory
reactions, TNF- usually appears first in the serum, followed by IL-1
and then IL-6 (23, 32). TNF and IL-1 are stimulators of IL-6.
Therefore, inflammation may initiate a cascade of cytokine production,
which directly or indirectly induces LIF production from multiple
cells, resulting in increased serum LIF levels.
We have not yet correlated LIF levels with other markers of
disease activity, such as fever, acute phase proteins, other cytokine
levels, disease stage, and treatment; therefore, the specific
significance of LIF levels as prognostic indicators for disease
progress remains to be determined. However, a number of studies have
measured serum cytokines in an attempt to provide objective diagnostic
indicators of disease severity and progression. Serum IL-6 has been
correlated with the severity of organ dysfunction (33), and early-onset
neonatal infection (34). A positive correlation between serum LIF
levels and shorter survival was observed in septic patients with shock
(8). The data shown in Fig. 4 suggested that increased LIF levels were
associated with the activity of the disease. Therefore, serum LIF
levels may serve as a potential indicator of the severity and
progression of inflammatory diseases. As LIF exhibits strong synergy
with CRH in stimulating POMC transcription (5, 35), the ACTH response
to inflammation may reflect those induced LIF levels.
Measurements of serum cytokines, especially serum IL-6, in SLE patients have provided ambiguous results (36, 37). We observed elevated serum LIF levels in patients with SLE. Abnormal production of several cytokines in SLE may be due to an intrinsic immune defect. Higher serum levels of cytokines may be associated with the activity of this disease. Patients with rheumatoid arthritis have measurable LIF in synovial fluid (25), but LIF values in their sera are not different from those in patients with noninflammatory diseases. Although a number of carcinoma cell lines produce LIF (1), serum LIF levels in the cancer patients we screened were not higher than those in patients with noninflammatory diseases. Both low production and rapid clearance of LIF during the chronic process in these diseases and in noninflammatory diseases may explain their low serum LIF levels.
In summary, the LIF RIA we have developed is a useful tool to measure LIF in biological fluids; LIF is an important participant in the acute inflammatory process. The elevated serum LIF levels in inflammation do not appear to originate from the pituitary.
Footnotes
1 This work was supported by NIH Grant DK-42792 and the Doris Factor
Molecular Endocrinology Laboratory. 
Received September 10, 1997.
Revised December 10, 1997.
Accepted December 17, 1997.
References
genes in normal an malignant rat
glial cells: regulation by lipopolysaccharide and vitamin D. J
Neurosci Res. 46:360–366.[CrossRef][Medline]
This article has been cited by other articles:
 |
|
 |
|
  L. J. Quinton, M. R. Jones, B. E. Robson, B. T. Simms, J. A. Whitsett, and J. P. Mizgerd Alveolar Epithelial STAT3, IL-6 Family Cytokines, and Host Defense during Escherichia coli Pneumonia Am. J. Respir. Cell Mol. Biol., June 1, 2008; 38(6): 699 - 706. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S.-G. Ren and S. Melmed Pyridoxal Phosphate Inhibits Pituitary Cell Proliferation and Hormone Secretion Endocrinology, August 1, 2006; 147(8): 3936 - 3942. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S.-G. Ren, S. Kim, J. Taylor, J. Dong, J.-P. Moreau, M. D. Culler, and S. Melmed Suppression of Rat and Human Growth Hormone and Prolactin Secretion by a Novel Somatostatin/Dopaminergic Chimeric Ligand J. Clin. Endocrinol. Metab., November 1, 2003; 88(11): 5414 - 5421. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Beishuizen and L. G. Thijs Review: Endotoxin and the hypothalamo-pituitary-adrenal (HPA) axis Innate Immunity, February 1, 2003; 9(1): 3 - 24. [Abstract] [PDF]
|
|
|
|
|
|
A. Ben-Shlomo, I. Miklovsky, S.-G. Ren, W. H. Yong, A. P. Heaney, M. D. Culler, and S. Melmed Leukemia Inhibitory Factor Regulates Prolactin Secretion in Prolactinoma and Lactotroph Cells J. Clin. Endocrinol. Metab., February 1, 2003; 88(2): 858 - 863. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. J. Auernhammer and S. Melmed Leukemia-Inhibitory Factor--Neuroimmune Modulator of Endocrine Function Endocr. Rev., June 1, 2000; 21(3): 313 - 345. [Abstract] [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
