|Home | About | Journals | Submit | Contact Us | Français|
Tristetraprolin (TTP) is an mRNA destabilizing protein that binds to AU-rich elements in labile transcripts, such as the mRNA encoding tumor necrosis factor alpha (TNF), and promotes their deadenylation and degradation. TTP-deficient (KO) mice exhibit an early-onset, severe inflammatory phenotype, with cachexia, erosive arthritis, left-sided cardiac valvulitis, myeloid hyperplasia, and autoimmunity, which can be prevented by injections of anti-TNF antibodies, or interbreeding with TNF receptor-deficient mice. To determine whether the excess TNF that causes the TTP KO phenotype is produced by myeloid cells, we performed myeloid-specific disruption of Zfp36, the gene encoding TTP. We documented the lack of TTP expression in lipopolysaccharide-stimulated bone marrow-derived macrophages from the mice, whereas fibroblasts expressed TTP mRNA and protein normally in response to serum. The mice exhibited a minimal phenotype, characterized by slight slowing of weight gain late in the first year of life, compared to the early onset, severe weight loss and inflammation seen in the TTP KO mice. Instead, the myeloid-specific TTP KO mice were highly and abnormally susceptible to a low dose lipopolysaccharide challenge, with rapid development of typical endotoxemia signs and extensive organ damage, and elevations of serum TNF levels to 110-fold greater than control. We conclude that myeloid-specific TTP deficiency does not phenocopy complete TTP deficiency in C57BL/6 mice under normal laboratory conditions, implying contributions from other cell types to the complete phenotype. However, myeloid cell TTP plays a critical role in protecting mice against LPS-induced septic shock, primarily through its post-transcriptional regulation of TNF mRNA stability.
Tristetraprolin (TTP2) is the prototype member of a small family of RNA binding proteins that are characterized by two nearly identical tandem CCCH zinc finger domains with highly conserved sequences and spacing. Initially discovered as a gene that could be induced rapidly and transiently by the stimulation of fibroblasts with growth factors and mitogens (1–4), it is now well established that TTP is an mRNA destabilizing protein that binds to AU-rich elements (AREs) in its target mRNAs, such as that encoding tumor necrosis factor alpha (TNF) (5, 6). TTP-deficient mice appear normal at birth, but soon develop a complex inflammatory phenotype consisting of cachexia, dermatitis, conjunctivitis, destructive arthritis, myeloid hyperplasia, and autoimmunity (7), which resembles in some respects the inflammatory syndrome seen in TNF overproduction transgenic mouse models reported earlier (8, 9).
As the first described and best studied TTP target transcript, the TNF mRNA serves as a prototype target for post-transcriptional regulation of gene expression by TTP. The TNF mRNA contains several closely spaced and overlapping copies of the nonamer UUAUUUAUU, the optimal TTP binding motif, in its 3’-untranslated region (3’-UTR). After direct binding to these sequence elements in the 3’-UTR of the TNF transcript, TTP then promotes the removal of its poly(A) tail, followed by its accelerated degradation (5, 10). The functional relevance of the TNF ARE in mice was established by Kontoyiannis et al (11), who demonstrated that genetic removal of this element led to stabilization of the TNF transcript and a systemic inflammatory phenotype similar to, but more severe than, that seen in the TTP KO mice. The inflammatory phenotype in TTP-deficient mice could be prevented by repeated injections of anti-TNF antibody, or interbreeding with TNF receptor-deficient mice (7, 12). These findings highlight the fact that TNF itself is by far the most important mediator in the pathogenesis of the mouse TTP-deficiency syndrome, although other TTP target transcripts have been identified subsequently, such as those encoding GM-CSF (13), immediate early response-3 (IER-3) (14), polo-like kinase 3 (PLK3) (15), and others.
TNF can be produced by a wide variety of cell types, of both hematopoietic and non-hematopoietic lineages, including lymphocytes, mast cells and stromal cells (11, 16, 17). Previous studies from our group have identified macrophages as one of the important cellular sources that contribute to the TNF overproduction in TTP KO mice (18). TNF can also promote its own expression; innate immune system stimuli such as LPS can thus stimulate TNF expression both primarily and secondarily through TNF itself (17). In this setting, TTP deficiency, and the resulting stabilization of the TNF mRNA, result in disruption of the normal feedback mediated by TTP, and the state of chronic TNF excess that characterizes the TTP-deficient mice (7, 19). Transplantation of bone marrow from TTP KO mice into RAG-2−/− immunodeficient mice reproduced the complete TTP deficiency inflammatory phenotype after a lag period of several months (18), demonstrating that one or more cell type(s) of hematopoietic origin, probably cells other than lymphocytes, were responsible for triggering the inflammatory responses that eventually lead to the development of the full TTP deficiency syndrome.
In this study, we hypothesized that specific deficiency of TTP in myeloid cells would completely recapitulate the early onset, severe inflammatory phenotype that is characteristic of the TTP KO mice (7). To test this hypothesis, we generated a floxed Zfp36 mouse line, and crossed these mice with mice expressing the Cre recombinase transgene under the control of the Lysozyme M promoter. Surprisingly, mice with myeloid-specific deletion of TTP (M-TTP KO mice) did not recapitulate the early onset, severe inflammatory phenotype of the TTP KO mice. Instead, these mice exhibited increased susceptibility to a low dose LPS challenge, with rapid development of an endotoxemia syndrome with extensive organ damage that was associated with dramatic increases in circulating TNF. Our results demonstrate that myeloid-specific TTP deficiency has much less effect than complete TTP deficiency on C57BL/6 mouse growth and development under normal laboratory conditions. However, myeloid cell TTP appears to be critical for the protection of mice from LPS-induced septic shock, primarily through its ability to regulate TNF expression at post-transcriptional steps.
Heterozygous mice with a conditional floxed Zfp36 allele were generated by gene targeting in embryonic stem (ES) cells by Xenogen Biosciences (Cranbury, NJ). To construct a targeting vector, a 3.6-kb NotI-KpnI fragment and a 4.6-kb MluI-SalI fragment isolated from BAC clone RP23-342K19 were used as 5’ and 3’ homologous regions, respectively (Fig. 1A). A 1.9-kb loxP-flanked MluI-SalI fragment was generated for the conditional KO region, resulting in a vector designed to delete exon 2, including the polyadenylation signal, of Zfp36. A loxP-flanked neor expression cassette was inserted into intron 1 between the 5’-homologous arm and the conditional KO region. The loxP flanked Neo expression cassette, and the diptheria toxin-A gene fragment (DTA) expression cassette in the vector, were used for positive and negative selection in ES cells, respectively. The composition of the final vector was confirmed by restriction digestion and end-sequencing. C57BL/6 ES cells were then electroporated with 30 µg of the SwaI-linearized targeting construct and selected in G418 (200 µg/ml). ES clones with homologous recombination were identified by Southern blotting using a 5’ external probe; three clones were selected for ES clone expansion. Additional Southern confirmation used the Neo probe and 3’ and 5’ external probes, and the successful integration of the third loxP site was identified by PCR screening. Two positive ES clones confirmed for homologous recombination were selected for transient Cre transfection using electroporation for Cre recombinase-mediated excision of the neor expression cassette, and two targeted clones with deletion of the neor expression cassette were identified and confirmed upon expansion by PCR analysis. Blastocyst injections were performed using these two independent targeted ES cell clones, and germline transmission was obtained by further crossing of male chimeras with C57BL/6N Tac wild-type females. The floxed Zfp36 mice were maintained by heterozygous matings. The mice were routinely genotyped by PCR, using the following primer pair (primer 1: 5’-GAA CCC TCT CTC GAT CGG GGA TAC-3’; primer 2: 5’-GGA TGG AGT CCG AGT TTA TGT TCC AA-3’), yielding amplicons of 514 bp for the floxed Zfp36 allele and 327 bp for the wild-type (WT) allele, as distinguished by agarose gel electrophoresis.
M-TTP KO mice were achieved by crossing the loxP-flanked Zfp36 mice (Zfp36flox/flox) with mice expressing Cre recombinase under the control of the murine M lysozyme promoter, which is specific for cells of the myeloid lineage (LysMcre) (20). Homozygous LysMcre mice on a C57BL/6 background were purchased from Jackson Laboratory (Bar Harbor, Maine), and mated with the heterozygous Zfp36flox/+ animals, to generate heterozygous conditional TTP mice with LysMcre (LysMcre/Zfp36flox/+). Heterozygous matings, LysMcre/Zfp36flox/+ mice crossed with Zfp36flox/+ mice, were employed to generate M-TTP KO mice and their littermate control mice. Double floxed mice without LysMcre (+/Zfp36flox/flox), mice with a WT TTP allele but carrying LysMcre (LysMcre/Zfp36+/+), and WT mice were used as “WT” controls. The integration of loxP or the Cre recombinase transgene into the mouse genome did not cause any evident changes in morphology or responses to stimuli, in both cell and intact mouse experiments (data not shown).
For PCR genotyping, genomic DNA from tail clips was extracted as described (21). The LysMcre transgene was detected using three primers: 5’-CCC AGA AAT GCC AGA TTA CG-3’; 5’-CTT GGG CTG CCA GAA TTT CTC-3’; and 5’-TTA CAG TCG GCC AGG CTG AC-3’, as per the Jackson Laboratory’s recommendations. The myeloid-specific deleted Zfp36 allele was examined in multiplex reactions using forward primer 3 (5’-CTG GCT GGA AAT GAG AGA GG- 3’) and reverse primers 2 (as described above) and 4 (5’-CAC CCC TTA CGC CAG AAC TA-3’), which amplified the wild-type (683 bp), floxed (870 bp) or deleted Zfp36 alleles (769 bp, Fig. 2A-B). All of the animal breeding and other procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of National Institute of Environmental Health Sciences.
8–12 week-old male mice were euthanized by CO2 inhalation, and bone marrow cells were isolated from the femurs as described previously (18). After overnight culture in T25 flasks, non-adherent bone marrow cells were collected and cultured in RPMI 1640 medium supplemented with 10% heat-inactivated FBS (Hyclone, Logan, UT), 25 mM HEPES, 2 mM glutamine, 100 U/ml penicillin, 100 µg/ml streptomycin (Gibco, Invitrogen) as well as 30% (v/v) L929 cell conditioned medium. Culture medium was replaced by fresh medium every 3 days. Adherent macrophage monolayers were obtained within 8 to 10 days; these cells were >99% positive for Mac-1+ antibody staining and negative for Gr-1 as determined by flow cytometric analysis, identifying them as mature macrophages. Differentiated bone marrow-derived macrophages (BMDM) were then harvested by gently scraping the cells from the dishes using a rubber “policeman” and seeded onto 100 mm petri dishes for experiments. Cells were subjected to serum starvation with RPMI 1640 medium containing 1% FBS for at least 20 h before stimulation with 1 µg/ml LPS (serotype 055:B5, Sigma Chemical Co., St. Louis, MO) for the times indicated.
LysMcre/Zfp36flox/+ mice were crossed with +/Zfp36flox/flox mice to obtain the myeloid-specific TTP deficient (M-TTP KO) embryos and mice. Littermate Zfp36flox/flox mice without LysMcre were used as controls. Primary cultures of mouse embryonic fibroblasts (MEF) were prepared from embryos at day 15.5 of gestation (E15.5) where E0.5 was the date of detection of the vaginal plug. The genotypes for individual fetuses from each litter were determined by evaluation of tail DNA from each embryo (7). MEF were maintained at 37°C (5% CO2) in DMEM (Invitrogen, Carlsbad, CA) containing 10% FBS (Invitrogen), 100 U/ml penicillin, 100 µg/ml streptomycin, and 2 mM L-glutamine; MEFs at passages two to four were used for the experiments described here. For serum deprivation before stimulation, cells at approximately 70 to 80% confluence were washed once in serum-free DMEM and then incubated in DMEM containing 0.5% (vol/vol) FBS for at least 16 h (14, 15). The cells were then stimulated with 10% (vol/vol) FBS (HyClone, Logan, UT) for the indicated times.
Cytosolic extracts and whole cell lysates were prepared in RIPA buffer (50 mM Tris pH 8.0, 150 mM NaCl, 1% Nonidet P-40, 5 mM EDTA and 5% glycerol) or hypertonic lysis buffer (100 mM Tris-HCl pH 8.0, 0.5 M NaCl, 5 mM EDTA, 1.25% Nonidet P-40 and 5% glycerol) containing protease inhibitors (Complete Mini, Roche, Indianapolis, IN, USA), respectively, cleared by centrifugation, and quantified using the Bradford assay (BioRad, Hercules, CA, USA). Denatured lysates were separated on 10–20% Criterion Tris-HCl precast gels (BioRad) and transferred onto nitrocellulose membranes. Blots were incubated at 4°C overnight with rabbit antiserum raised against a recombinant mouse TTP-maltose binding protein fusion protein (14, 22), followed by incubation with horseradish peroxidase-conjugated goat anti-rabbit (Fab’)2 (Pierce). Immune complexes were detected using SuperSignal West Pico chemiluminescent substrate (Pierce, Rockford, IL, USA).
Adult male mice between 12–15 weeks of age and weighing between 24–31g were used. The mice were housed in a pathogen-free animal facility at 25° C and were illuminated by 12:12-h light–dark cycles. The mice were provided with standard rodent chow and water ad libitum. Experiments were carried out in accordance with NIH guidelines for animal treatment, housing, and euthanasia. Mice were challenged by the intraperitoneal injection of LPS at a low dose of 0.5 mg/kg (O11:B4, Sigma Aldrich) in 200 µl of sterile saline, with total amounts ranging from 12 µg to 15.5 µg of LPS per mouse, and monitored for general condition and survival. Body temperatures were taken with an instant Infrared Ear Thermometer (VET-TEMP Model VT-150, Advanced Monitors Corp., San Diego, CA). Serial blood samples were collected through retro-orbital sinus bleeding at 0, 1.5, 3, and 6 h after LPS injection. At 10 h after the LPS challenge, the mice were euthanized, and serum and organs were collected for chemistry analyses and histological examinations. Serum levels of aspartate aminotransferase, alanine aminotransferase, blood urea nitrogen, and lactate dehydrogenase, were determined to assess the extent of muscle, kidney, and liver damage.
Tissues were fixed in 10% phosphate-buffered formalin and processed using standard procedures. 5 µm paraffin tissue sections were cut and stained with hematoxylin and eosin. Samples of bones and paws were decalcified with RDO solution (Apex Engineering Products, Aurora, IL) for 1 h at room temperature, before being processed and embedded like the others.
For immunostaining, sections were incubated in 3% H2O2 to inactivate endogenous peroxidases, followed by antigen retrieval with heat and pressure in citrate butter (Biocare Medical, CA). Endogenous biotin was blocked with Avidin-Biotin blocking reagents (Vector Laboratories, Burlingame, CA). Sections were then incubated with anti-CD45 antibody (GeneTex, San Antonio, TX) for 60 min at a 1:250 dilution, followed by peroxidase-conjugated streptavidin SS labeling (Biogenex Laboratories, San Ramon, CA) for 30 min. Immunolabeled antigen-antibody complexes were visualized using diaminobenzidine. The sections were lightly counterstained with hematoxylin before analysis.
Spleens and bone marrow cells were removed from freshly euthanized 8–12 week-old female mice. Single-cell suspensions were prepared in ice-cold PBS by passing minced tissue through a 100-µm nylon cell strainer, and then 2 × 106 cells were stained with the respective antibodies for 20 min at 4°C after red blood cell lysis and preincubation with anti-FcγR (2.4G2, BD PharMingen) for 15 min. Antibodies against CD3ε-FITC (145-2C11), CD19-APC-Cy7 (1D3), Gr-1-PE-Cy7 (RB6-8C5), and CD11b-AF 647 (Mac-1) were purchased from BD PharMingen. Anti-CD4-PE (GK1.5) and anti-CD8a-PE-Cy5 (53-6.7) were from eBioscience, and anti-F4/80-R-PE (CI: A3-1) was from Serotec. Stained cells were acquired on a LSR II flow cytometer using FACSDiva (BD Biosciences) software and analyzed with FlowJo software (Tree Star, Ashland, OR). Cell debris and dead cells were excluded from the analysis based on scatter signals and/or propidium iodide fluorescence.
Concentrations of IL-1β and IL-6 in supernatants from primary macrophages were determined by ELISA using the OptEIA mouse kits (BD Biosciences), as per the manufacturer’s instructions. Quantification of TNF levels in serum and culture supernatants was carried out using capture and biotinylated anti-mouse TNF antibodies from BioLegend (San Diego, CA). Recombinant mouse TNF from eBioscience was used as a standard. Peroxidase-conjugated avidin (BioLegend) was applied before the colorimetric development with the 3′, 3′, 5, 5′-tetramethylbenzidine liquid substrate system (Sigma). The reactions were stopped with 5 N hydrogen chloride, and the absorbance at 450 wavelengths (OD 450) was measured with the Tecan Infinite 200 Pro microplate reader with a 570 wavelength (OD 570) set as reference.
Total cellular RNA from differentiated macrophages was isolated using the GE Healthcare illustra RNAspin MiniRNA Isolation Kit according to the manufacturer’s instructions (GE Healthcare). Residual genomic DNA was removed by on-column digestion with RNase-free Dnase I. To determine the decay rates of cytokine transcripts, actinomycin D (5 µg/ml, Sigma) was added to block transcription after BMDM were stimulated with LPS (1 µg/ml) for 1 or 4 h. Total RNA was isolated at the indicated times over a 2 h time course after addition of actinomycin D. TTP and cytokine transcript levels were determined by real-time RT–PCR. First-strand cDNAs were synthesized using oligo(dT)12–18 primers and SuperScript III Reverse Transcriptase (Invitrogen). Real-time PCR was performed using SYBR Green and the ABI Prism 7900 Sequence Detection System (Applied Biosystems). Primers were designed and compared to the current mouse genome reference sequence using BLAST to ensure that no cross-reactivity with other genes would occur. Results were normalized against the β-actin transcript as an internal control, and were then used to calculate expression levels according to the ΔΔCt method (23). All data were expressed in terms of fold change relative to the unstimulated WT sample, which was set as one unless otherwise specified. The primers were validated for their amplification efficiency and specificity prior to being used in the study; sequences are available upon request.
Data are presented as means ± SEM. Statistical differences between wild-type and mutant groups were determined by the two-sided, unpaired Student t test. P < 0.05 was considered to be significant.
To examine the effects of myeloid cell TTP deficiency, we generated a conditional TTP KO mouse line (Fig. 1A) and crossed it with a Cre mouse line that expresses the Cre recombinase transgene driven by the Lysozyme M promoter (LysMcre). Expression of this transgene is limited to myeloid cells including monocytes, mature macrophages and granulocytes, and a small proportion of dendritic cells (20). Cre activity should result in the deletion of exon 2 and the 3’-untranslated region of Zfp36 in the myeloid lineage; exon 2 encodes the tandem zinc finger domain that is responsible for the RNA binding activity of TTP and its family members (Fig. 1A).
The M-TTP KO mice were viable, fertile, and born with the expected Mendelian frequency (data not shown). The growth curves of both male and female M-TTP KO mice under normal laboratory conditions were almost identical to those of their littermate controls until the growth curves began to diverge in females at about 6.5 months of age and in males after about 9 months (Fig. 1B). These growth curves are in marked contrast to those from the total TTP KO mice, which begin to diverge sharply from those of their WT littermates as early as 3–4 weeks of age (7) (Fig. 1B). It should be noted that, although the TTP KO growth data shown in Fig. 1B are from ongoing natural history data from our mouse colony, both the M-TTP KO mice and their littermate controls, and the TTP KO mice shown in Fig. 1B, are on a pure C57BL/6 background, and are therefore directly comparable. Although the growth curves of both the male and female M-TTP KO mice eventually diverged modestly but significantly from those of their WT littermates (Fig. 1B), none of these mice has ever become sick enough to require euthanasia, as is almost universal with the TTP KO mice within the first several months of life (data not shown).
Histological analyses of 8-month old M-TTP KO mice revealed very few of the abnormal findings typical of total TTP KO mice at this age or earlier, including severe polyarticular arthritis, loss of body fat, myeloid hyperplasia, dermatitis, left-sided cardiac valvulitis, and glomerular mesangial thickening (7, 24). An exception to this rule included slightly swollen front paws in some of the mice; one of four mice analyzed pathologically at this age had accompanying mild interphalangeal arthritis and inflammation (Supplemental Fig. S1). This is in contrast to the severe arthritis seen in all TTP KO mice at or before 7 months of age (7). In addition, foci of mild inflammation and fibrosis were observed in the myocardium of 3 of 4 M-TTP KO mice that were examined at 8 months of age; these lesions were absent in control mice (data not shown). Occasional enlargement of lymph nodes and spleens was noted in some cases in the M-TTP KO mice; overall, there were modest increases in the average spleen weights (0.103 ± 0.004 g, n=7) and the ratios of spleen weight/body weight (0.506 ± 0.019 g, n=7) in 3–4 month old female mutant mice when compared with their controls (0.083 ± 0.002 for the spleen weights and 0.407 ± 0.010 for the ratios of spleen weight/body weight in the WT controls, n=8; P=0.0012 and P=0.0003 for WT vs. M-TTP KO mice).
Analysis of peripheral blood from 8-month old M-TTP KO mice revealed normal percentages of blood cells and WBC differential counts (Fig. 3A). FACS analyses of splenocytes and bone marrow cells collected from 3-month old M-TTP KO mice and their corresponding controls exhibited normal percentages of CD3+ T lymphocytes and CD19+ B lymphocytes, and the T cell subsets CD4+ T helper and CD8+ cytotoxic cells in the spleens and bone marrows of M-TTP KO mice were similar to those of WT mice (Fig. 3B). Likewise, the percentages of CD11b+ macrophages, Gr-1+ granulocytes and CD11b+Gr-1+ cells, a subset of immature myeloid cells, were found to be comparable with those of controls (Fig. 3B).
To confirm the TTP deficiency in myeloid cells from the M-TTP KO mice, we cultured bone marrow-derived macrophages (BMDM) from these mice. Successful elimination of the normal Zfp36 alleles in the BMDM from the M-TTP KO mice was confirmed in genomic DNA isolated from these cells (Fig. 2A). Both the floxed and wild-type Zfp36 alleles were barely detectable, and instead, the disrupted fragment resulting from Cre-mediated excision was present (Fig. 2A). We then stimulated these cells with LPS and measured TTP mRNA and protein. The basal TTP transcript levels in the M-TTP KO BMDM were approximately 100-fold lower than those seen in littermate WT control cells (Fig. 2B inset). LPS stimulation of the cells resulted in almost no detectable increase in TTP mRNA levels in the M-TTP KO cells, whereas there was a robust response to LPS in the WT cells (Fig. 2B). Similarly, there was no detectable TTP protein in the M-TTP KO cells after LPS stimulation, despite readily detected LPS-induced immunoreactive TTP in the WT cells (Fig. 2C). These data demonstrate the essentially complete absence of TTP mRNA and protein expression in BMDM derived from the M-TTP KO mice.
The expression of the other two TTP family members, Zfp36l1 and Zfp36l2 mRNAs, was also examined in these LPS-treated BMDM. There was modest induction of both Zfp36l1 and Zfp36l2 transcripts in both WT and M-TTP KO cells within 1 h after LPS challenge; in both cases, these levels then rapidly decreased below basal (Supplemental Fig. S2). For both transcripts, levels in the KO cells were slightly but significantly higher than those found in the WT cells at some time points, but these modest changes were apparently not sufficient to compensate for the TTP deficiency.
We next examined the cell-type specificity of the Cre-mediated TTP disruption in mouse embryonic fibroblasts (MEF), cells known for their highly inducible TTP expression in response to multiple stimuli, including serum (14). MEF were derived from E15.5 M-TTP KO embryos and their littermate controls. In the cells from the M-TTP KO mice, the floxed Zfp36 allele was detectable by PCR of genomic DNA, but the potentially deleted Zfp36 fragment mediated by LysM-Cre excision was not detectable (Fig. 2D). Serum stimulation of these cells after overnight serum deprivation produced virtually identical increases in TTP mRNA levels in the cells from the WT and M-TTP KO mice (Fig. 2E). Similarly, the induction of immunoreactive protein in response to LPS was virtually identical in cells from the WT and M-TTP KO mice (Fig. 2F). These data demonstrate that the myeloid-specific Cre transgene did not excise the floxed allele to any significant extent in MEF, a prototype non-myeloid cell type.
To determine whether macrophages derived from the M-TTP KO mice exhibited similar patterns of TNF transcript stabilization to those seen in the total TTP KO mice, we measured TNF mRNA levels and stability in BMDM after LPS stimulation. In control BMDM, TNF mRNA levels increased rapidly after LPS stimulation, and returned to near basal levels after 24 h (Fig. 4A). Mutant BMDM exhibited increases in TNF mRNA levels that were 2–3 fold higher than control at all times between 1–6 h after LPS stimulation (Fig. 4A). We next examined the stability of the TNF mRNA in control and M-TTP KO macrophages after adding actinomycin D 1 and 4 h after LPS stimulation. As expected, the times to 50% decay of the TNF mRNA were 3-fold greater in the mutant macrophages than in the corresponding control cells when actinomycin D was added after either 1 or 4 h of LPS treatment (Fig. 4B-C). Specifically, the average half-lives of the TNF mRNA were 55 and 67 min in the M-TTP KO macrophages 1 and 4 h after LPS, respectively, compared to 18 and 23 min in WT cells (P=0.03 and 0.001, respectively).
We also measured TNF protein released by BMDM under the same circumstances. In the WT cell cultures, protein levels reached a maximum at about 8 h after LPS stimulation and remained at that level for the duration of the experiment (Fig. 4D, and supplemental Fig. S3). In the M-TTP KO cells, there was an approximately 6-fold increase in the medium TNF after 8 and 24 h LPS stimulation compared to control (Fig. 4D). The differences in protein concentration were statistically significant (P<0.05) at all time points between 2 and 24 h.
IL-6 and IL-1β are two other pro-inflammatory cytokines that are thought to play critical roles in LPS responses, and post-transcriptional regulation has been reported to be one of the important determinants of their expression (25, 26). We therefore examined the expression patterns of IL-6 and IL-1β mRNAs, and their turnover rates, in LPS-stimulated BMDM isolated from the M-TTP KO mice and their littermate controls. After LPS stimulation, there was robust induction of both IL-6 mRNA and protein in BMDM; however, these levels were comparable between the M-TTP KO macrophages and control cells (Fig. 5A). In the case of IL-1β mRNA, there was a slightly higher basal level of the IL-1β transcript in the TTP-deficient BMDM (2.01 ± 0.41 compared to 1.03 ± 0.04 in WT cells; P<0.01), and there was a tendency for higher levels of this transcript between 2 and 6 h after LPS in the mutant cells, which did not reach statistical significance (P=0.20 – 0.81; Fig. 5A). IL-1β protein was not detectable in the culture supernatants from either mutant or WT cells until 24 h after LPS treatment, when a modest, 1.5-fold decrease in IL-1β accumulation was seen in the TTP-deficient macrophages (P<0.05, Fig. 5B).
When the stabilities of the IL-6 and Il-Lβ transcripts were examined in BMDM after adding actinomycin D at 1 and 4 h after LPS stimulation, both transcripts remained stable over the 2 h time course of the experiment in both control and TTP-deficient cells (Fig. 5C-D). Thus, under these experimental conditions, we were unable to detect changes in IL-6 and IL-1β mRNA stability in the TTP-deficient BMDM.
To examine the sensitivity of the M-TTP KO mice to an LPS challenge, we injected the mice with a low dose of LPS (0.5 mg/kg). At this dose in normal mice, LPS stimulates pro-inflammatory cytokine expression, but typically does not provoke significant signs of endotoxemia (27). Unexpectedly, the otherwise normal-appearing M-TTP KO mice at 3 months of age exhibited dramatic and typical signs of endotoxemia, including lethargy, tachypnea, piloerection and marked decreases in body temperature. This became notable as early as 1.5 h following LPS injection; the temperature readings dropped below the lower detection limit after 3 h in all the M-TTP KO mice tested (Fig. 6A). In contrast, neither the WT mice receiving low-dose LPS or vehicle, nor the M-TTP KO mice receiving the vehicle alone, exhibited such endotoxemia signs (Fig. 6A). Per IACUC guidelines, all of the mice were then euthanized at 10 h after LPS administration.
We next examined the serum levels of TNF, as the critical mediator of endotoxin shock, as well as the product of the best-studied TTP target transcript. Elevated basal TNF levels (142 ± 32 pg/ml) were seen in serum samples from the 3-month old M-TTP KO mice before LPS exposure; this was significantly higher than the average value from control mice (8 ± 6 pg/ml; P=0.0087). After the LPS injection, a parallel kinetic pattern of TNF induction was seen in both M-TTP KO and control mice, with peak responses at 1.5 h after LPS followed by a return to baseline after 6 h (average peak values were 67 ± 13 ng/ml in the M-TTP KO mice vs. 612 ± 113 pg/ml in the controls at 1.5 h; P=0.0002; Fig. 6C). The serum levels of TNF in the M-TTP KO mice were remarkably higher than those in the control mice at all time points examined, with the peak response in the mutants approximately 110-fold higher than that seen in the WT mice.
Necropsy of these M-TTP KO mice 10 h after LPS administration revealed enlarged spleens and dilated small intestines, with widespread inflammatory cell infiltrates in liver, kidney and lung (Fig. 7); the infiltrating leukocytes were predominantly localized either inside the lumens of vessels or surrounding the vasculature, suggesting the mobilization and migration of these cells during endotoxemia. Parenchymal infiltration was also remarkable in the mutant mice, reflected by scattered foci of hepatic necrosis, glomerular hypercellularity, and pulmonary alveolitis (Fig. 7 and Supplemental Fig. S4). The M-TTP KO mice treated with saline, and the control mice treated with either LPS or saline, exhibited minimal abnormal pathology (Fig. 7 and Supplemental Fig. S4). Serum samples from the M-TTP KO mice exhibited a 5-fold increase in serum alanine aminotransferase, a 2.2-fold increase in blood urea nitrogen, and a 3-fold increase in lactate dehydrogenase compared to control mice (Fig. 6B), reflecting liver and kidney dysfunction.
The major findings of this study are that (1) myeloid-specific deficiency of TTP results in no overt phenotype in mice during the first several months of life under normal laboratory conditions, in contrast to the early-onset, severe inflammatory syndrome characteristic of total TTP deficiency; and (2) the myeloid-specific TTP (M-TTP) KO mice were nonetheless highly susceptible to activation of the innate immune system by a low dose of LPS that had little or no effect on littermate WT mice. Our overall conclusions are that TTP deficiency in one or more other cell types is likely to be required for the full-blown syndrome to develop in this strain of mice, but that regulated expression of myeloid cell TTP is necessary for the prevention of septic shock in the setting of infection with Gram negative bacteria and perhaps other innate immune system stimuli.
Concerning the first major finding, the principle hypothesis behind this study was that specific deficiency of TTP in myeloid cells would completely recapitulate the early onset, severe inflammatory phenotype that is characteristic of the TTP KO mice (7). This was based on several pieces of supporting evidence, including the facts that macrophages, and to a lesser extent, other myeloid cells, are the major physiological sources of TNF (16, 17); that TTP-deficient macrophages strikingly overproduce TNF after stimulation with LPS (18, 19); and that total bone marrow transplantation can recapitulate the entire, severe TTP deficiency syndrome after a lag period of several months, suggesting that hematopoietic cells other than lymphocytes are the primary drivers of the complete syndrome (18). However, the specific KO of TTP in myeloid cells using a conditional Zfp36 allele and Cre driven by the LysM promoter did not lead to the same phenotype as seen in the total TTO KO mice. Instead, under normal laboratory conditions and on the same genetic background, the M-TTP KO mice were remarkably healthy, with modest weight loss and other signs of TTP deficiency first appearing, in mild form, after six months or more of age.
One potential explanation for these surprising results might be inadequate excision of Zfp36 by this Cre driver in myeloid cells. However, this was tested in primary macrophages derived from these mice at the level of the genome, as well as with basal and LPS-stimulated TTP mRNA and protein production. In all cases, the data confirm the essentially complete excision of this gene in the myeloid cells, as well as the complete loss of gene expression. This was in the setting of totally normal serum-induced TTP expression in fibroblasts, a non-myeloid cell type. Thus, we cannot ascribe the mildness of this syndrome to inadequate Zfp36 deletion in the targeted cells.
Since the entire TTP deficiency syndrome can be transplanted with whole bone marrow (18), these findings suggest that TTP deficiency in one or more other types of hematopoietic cells is also necessary for the full syndrome to evolve in the setting of myeloid TTP deficiency. It may be possible to uncover the roles of other cell types using various types of co-culture experiments, using, for example, WT or TTP KO lymphocytes cultured with the M-TTP KO macrophages under conditions that would allow or prevent direct cell-cell contact. Alternatively, it might be possible to recapitulate the severe TTP deficiency syndrome using conditional TTP deletion in hematopoietic stem cells; this could eliminate non-hematopoietic cell types from playing major primary roles in the pathogenesis of this syndrome.
An informative previous experiment was the use of the TNF “delta ARE” allele, which leads to a greatly stabilized TNF mRNA and a severe inflammatory syndrome that in many respects resembles the syndrome of total TTP deficiency (11). A conditional version of this allele was combined with the myeloid-specific Cre driver used in the present study; in contrast to the severe, early onset inflammatory disorder caused by total deletion of the TNF ARE in both heterozygous and homozygous mice, restricting expression of the delta ARE allele to myeloid cells resulted in mice that were overtly normal for at least the first four months of life (28). After this point, they began losing weight and ultimately developed a milder version of the inflammatory bowel disease that occurs in an earlier onset, more severe form in the heterozygous or homozygous TNF delta ARE mice. Since we believe that most aspects of the TTP deficiency syndrome are caused by excessive production of TNF, this study is another example of how stabilization of the TNF mRNA in myeloid cells alone is apparently not sufficient to lead to illness in the absence of other cell types with the same defective allele.
The second major finding of this study is that normal, inducible TTP expression in myeloid cells is indispensable for protecting mice against an LPS challenge. This conclusion is based on the fact that the M-TTP KO mice exhibited extreme sensitivity to a low dose LPS exposure that led to typical signs and laboratory finding of endotoxemia, under conditions in which control mice were unaffected. The extreme sensitivity of these mice to LPS was presumably caused, at least in part, by the high serum concentrations of TNF that occurred after the LPS injections. These in turn could be attributed to the high levels of TNF secreted by macrophages and other myeloid cells, as demonstrated in culture of primary BMDM; this resulted from increased TNF mRNA levels and stability in the TTP-deficient cells. These results are supported by previous findings that macrophages and neutrophils were the predominant cellular sources of systemic TNF after LPS in mouse models with selective TNF deletion in myeloid cells or lymphocytes (29). Similar observations to ours have been reported in the TNFΔARE mice, in which elevated levels of circulating TNF, and an approximately 50% mortality rate, were seen after a normally sublethal LPS challenge (11).
In contrast with the apparent central role of TTP in regulating the stability of TNF mRNA, and subsequent TNF secretion, in this study, we found little evidence to support a major effect of TTP on the expression of two other critical mediators of endotoxemia, IL-1β and IL-6, at least under these experimental conditions. Transcripts of both cytokines contain AU-rich elements, and a recent study reported that IL-6 and IL-1β transcripts were elevated in the livers of TTP knockout mice (26). Furthermore, the combination of hypoxia and LPS, with subsequent TTP activation, was able to destabilize IL-6 transcripts in Raw 264.7 cells (26, 30). However, in BMDM derived from the M-TTP KO mice, we found little evidence of instability for these transcripts after LPS stimulation and actinomycin D treatment, in contrast to the obvious effects of TTP deficiency on the unstable TNF transcript. There are numerous studies linking other ARE-binding proteins to the stabilities of these transcripts (25, 31), and it may be that they are more important to physiological regulation of IL-1β and IL-6 transcript stability and expression than TTP.
Our studies did not address a parallel role for TTP in the regulation of TNF mRNA translation. Previous studies have suggested that the TNF mRNA ARE can function as a translation repressive element, in addition to its well-known role as a transcript destabilizing element (11). In TTP-deficient macrophages, we have consistently found that TNF protein expression in the TTP-deficient cells in response to LPS is greater than expected from the changes in TNF mRNA stability alone (19). In addition, we have also found that translation of other TTP “target” messages is greater in the setting of TTP deficiency, under conditions in which transcript levels are largely unchanged (W.S. Lai and P.J. Blackshear, unpublished data). This would be in keeping with previous studies with other ARE-binding proteins that are thought to act at least in part by inhibiting TNF mRNA translation through direct ARE binding (32).
Our studies describe the extreme sensitivity of the M-TTP-KO mice to LPS, leading us to conclude that myeloid cell expression of TTP under normal circumstances is a critical aspect of the body’s defense mechanisms when confronted with Gram negative bacterial infections. However, there are many other environmental stimuli to TNF secretion, including both microbial products and non-biological agents, such as UV light and ionizing radiation. It will be interesting to determine whether TTP plays a role in protecting the body against the deleterious effects of TNF excess in response to these other stimuli; the experimental models described here should make it relatively straightforward to test these ideas, both in cells and intact mice.
We thank Drs. Mark Hoenerhoff and Gordon Flake for assistance with analysis of the histological slides, Dee Wenzel for animal husbandry support, Toni Ward for help with sample preparation, and the NIEHS histology facility for tissue processing, immunostaining and photography. We also thank Drs. Donald N. Cook and Michael B. Fessler for helpful comments on the manuscript.
1This work was supported by the Intramural Research Program of the National Institute of Environmental Health Sciences, National Institutes of Health.
2Abbreviations used in the article:
ARE, AU-rich elements
BMDM, bone marrow-derived macrophages
ES cell, embryonic stem cells
LysMcre, lysozyme M promoter driven Cre recombinase
MEF, mouse embryonic fibroblasts
M-TTP KO, myeloid-specific TTP-deficient mice
3’-UTR, 3’-untranslated region
Zfp36, Zinc finger protein 36
The authors have no financial conflicts of interest.