Inhibition of Tumor Necrosis Factor mRNA Translation by a Rationally Designed Immunomodulatory Peptide*

Based on sequences of immunomodulatory peptides derived from the heavy chain of HLA Class I, novel immunomodulatory peptides with increased potency were developed by computer-aided rational design. Allotrap 1258 was characterized in detail and shown to inhibit cell-mediated immune responses in vitro and in vivo . Immunomodulatory activity was associated with the ca-pability of the peptides to modulate heme oxygenase (HO) activity. In this study we analyzed the effect of Allotrap 1258 on cytokine expression. Allotrap 1258 inhibited concanavalin A- and lipopolysaccharide-in-duced human and mouse tumor necrosis factor (TNF) production in vitro and in vivo but had no effect on interleukin (IL)-1, IL-2, IL-4, IL-6, or IL-10 expression. Experiments with HO-1/KO and iNOS/KO mice showed that Allotrap 1258-mediated inhibition of TNF was independent of HO-1 and iNOS. Quantitation of TNF protein expression and mRNA steady state levels demonstrated that Allotrap 1258-mediated inhibition occurred at the translational level. Deletion of the AU-rich element in the 3 * -untranslated region (UTR) of TNF mRNA, a region known to be involved in TNF mRNA translation, had minimal effect on Allotrap 1258-mediated inhibition. trans- ARE targeted animals in a TN-FRI-deficient background were bred and maintained on mixed or genetic backgrounds at the animal facilities the Pasteur under specific pathogen-free conditions. using and to be

Peptides derived from the heavy chain of the HLA Class I molecule have been shown to modulate immune responses in vitro and in vivo (1)(2)(3)(4)(5)(6)(7). In vitro, these peptides inhibited cytotoxicity and differentiation of cytotoxic T cells (1). In vivo, they were shown to prolong allograft survival in rodents (2)(3)(4)(5)(6)(7). Based on these HLA-derived peptide sequences, novel immunomodulatory peptides were developed by computer-aided rational design (8). Initially, the biological activity of a panel of 19 peptides derived from peptide 2702.75-84 (HLA-B2702, amino acids 75-84) as well as other major histocompatibility complex molecules was assessed in a heterotopic mouse allograft model. A rational drug design strategy based on these in vivo data resulted in a series of immunomodulatory peptides. One of these peptides, RDP1258 and its D-isomer, Allotrap 1258, showed enhanced efficacy in comparison with 2702.75-84 both in vivo when tested in rodent allograft models and in vitro in the inhibition of cytotoxicity in human and mouse T lymphocytes.
The mechanism of immunomodulation mediated by these peptides is not fully understood. Observations in various in vivo models indicated that the peptides do not interact directly with T cell receptor or NK cell inhibitory receptors (4,5,9) The observation that both L-and D-isomers of peptide 2702.75-84 prolong graft survival indicates that the immunomodulatory activity of the peptides is not based on presentation by major histocompatibility complex molecules (5). Similarly, the recent hypothesis that these peptides modulate immune responses by binding to Hsc/Hsp70 can be excluded because the peptides D-enantiomers, while being immunomodulatory, do not bind to these proteins (10,11). Affinity chromatography with one of these immunomodulatory peptides resulted in the identification of heme oxygenase-1 (HO-1) 1 as a potential target for these peptides (12). All of the biologically active peptides derived from 2702.75-84 were shown to inhibit HO activity in vitro. The original compounds, however, were 100-fold less potent than the rationally designed Allotrap 1258. Similar to what has been observed with other HO inhibitors, in vivo administration of the peptides to animals resulted in a rapid up-regulation of HO-1 protein expression (12). Subsequent studies with peptides and protoporphyrins specifically affecting HO activity demonstrated that modulation of heme oxygenase activity is associated with the inhibition of several immune effector functions including lymphocyte proliferation and cytokine production (13,14). Currently it is unclear how modulation of HO activity may influence immune effector functions, and experiments with knock-out animals will be necessary to clarify this issue.
In this study, we continued our studies on the immunomodulatory activity of these peptides. Extensive analysis of ConAinduced cytokine profiles in the presence or absence of peptides demonstrated that Allotrap 1258 effectively inhibited the production of TNF. In addition, peptides partially inhibited the production of IFN␥ and IL-12, but not IL-1, IL-2, IL-4, IL-6, or IL-10. Effect of Allotrap 1258 therapy on TNF in vivo was studied in a galactosamine (GalN)/LPS mouse model that has been characterized as a model for systemic inflammatory response syndrome (15,16). A detailed analysis of TNF production in the presence of peptide demonstrated that peptidemediated inhibition of TNF production occurs at the translational level. Experiments with macrophages from transgenic and targeted animals expressing modified TNF genes demonstrated that AU-rich elements (AREs) in the mRNA 3Ј-UTR sequences appear to play a moderate role in mediating the suppressive effects of the peptides, but other as yet unidentified sequences in the 3Ј-UTR are more dominant in mediating the inhibition by the peptide of TNF translation.
TNF exerts a key role in the pathogenesis of many inflammatory diseases such as sepsis, inflammatory bowel disease, rheumatoid arthritis, and Crohn's disease (17)(18)(19)(20). Although originally described as a molecule that could reduce the size of malignant tumors, a massive up-regulation of TNF at a systemic level such as occurs in global infections or severe tissue injury leads to vasodilation, intravascular coagulation, and multiple organ failure (21). A therapeutic benefit of inhibiting TNF in such inflammatory conditions is readily apparent. Clinically, various approaches to reduce TNF levels in patients have been explored, including the use of the phosphodiesterase inhibitor, pentoxifylline, and anti-TNF antibodies. In particular, protein based therapies (i.e. anti-TNF antibody and soluble receptors) have shown considerable clinical success (22). To our knowledge, Allotrap 1258 is the first low molecular weight compound inhibiting TNF production at a translational level.

MATERIALS AND METHODS
Mice-C57BL/6, CBA, and inducible nitric-oxide synthase (iNOS) knock-out mice were purchased from Jackson Laboratory (Bar Harbor, ME). HO-1 knock-out mice were kindly provided by Dr. Poss (23). Mice were housed according to the Animal Welfare Guidelines of the California Department of Health. Tg1278 and Tg197 human TNF transgenic mice (24) and homozygous TNF ⌬ARE targeted animals in a TN-FRI-deficient background (25,26) were bred and maintained on mixed 129SvxC57BL/6 or CBAxC57BL/6 genetic backgrounds at the animal facilities of the Helenic Pasteur Institute under specific pathogen-free conditions.
Effects on ConA-induced Cytokine Levels in Vivo-In initial experiments, the kinetics of cytokine production following intravenous injection of ConA (20 mg/kg) were analyzed. Groups of mice (n ϭ 3/group) were sacrificed 0.5, 1, 1.5, 2, 3, 4, 5, 6, and 7 h post-injection. Plasma was collected and analyzed for IL-1␤, IL-2, IL-4, IL-6, IL-10, IL-12, TNF, and IFN␥ levels. Based on this kinetic analysis, the time point with maximum plasma cytokine concentration was chosen to study the effect of peptide. Mice (n ϭ 3/group) were injected intravenously with 20 mg/kg ConA and 0, 5, 10, or 20 mg/kg peptide in 5% mannitol. Animals were sacrificed at the predetermined time point, and serum was collected. Cytokine levels were determined by sandwich ELISA.
Effects on LPS-induced TNF in Vivo-Male 4 -6-week-old C57BL/6, CBA, iNOS/KO, and HO-1/KO mice were used to study the effects of Allotrap 1258 treatment on GalN/LPS-induced TNF production in vivo. Animals were given a single injection of GalN at 800 mg/kg, followed by LPS at 0.1 mg/kg. Allotrap 1258 or negative control peptide was administered at 10 mg/kg simultaneously with GalN and LPS. All injections were given intraperitoneally. At 60 min, animals were sacrificed, and blood was collected. TNF concentrations in plasma were measured by sandwich ELISA.
Measurement of TNF by ELISA-Concentration of TNF in plasma and cell culture medium was measured by ELISA. Anti-TNF mAb (clone MP6-XT22) was used to capture murine TNF, and anti-TNF mAb (clone mAb1) was used to capture human TNF. Biotinylated polyclonal anti-TNF and biotinylated monoclonal anti-TNF (clone mAb11) were used for detection of captured murine and human TNF, respectively. All antibodies were purchased from Pharmingen (San Diego, CA). Capture antibodies were coated onto 96-well Maxisorb flat bottom plates (Nunc, Naperville, IL) at 2 g/ml in 0.1 M Na 2 HPO 4 , pH 9, overnight. Plates were washed three times with phosphate-buffered saline, 0.05% Tween-20 solution to remove unbound antibodies and remaining sites were blocked with 1% bovine serum albumin/phosphate-buffered saline/ Tween. Following the addition of samples, the plates were incubated for 2 h at room temperature. Captured TNF was detected with biotinylated anti-TNF antibodies and streptavidin-conjugated horseradish peroxidase (Jackson Immunoresearch, West Grove, PA). Bound horseradish peroxidase was detected using o-phenylenediamine dihydrochloride (Sigma) at 1 mg/ml. The colorimetric reaction was stopped by the addition of 1 M HCl solution, and absorbance was measured at 490 nm.
Effects on Survival-Survival studies were performed in the GalN/ LPS mouse model using C57BL/6, CBA, and iNOS/KO mice. Animals were given simultaneous injections of GalN at 800 mg/kg and LPS at 0.1 mg/kg (intraperitoneal). One group each was also treated with Allotrap 1258 or negative control peptide at 10 mg/kg (intraperitoneal) at the time of GalN/LPS injection. Animal survival was monitored twice daily.
Cell Culture-RAW264.7 cells, a mouse macrophage cell line, and THP-1 cells, a human monocyte/macrophage cell line, were obtained from ATCC (Manassas, VA) and were routinely cultured in RPMI 1640 containing 10% fetal bovine serum in 5% CO 2 /95% air at 37°C. RAW264.7 cells were seeded at a density of 3.5 ϫ 10 5 /ml and stimulated for TNF production by the addition of LPS (10 g/ml). THP-1 cells similarly seeded were stimulated in the presence of LPS (5 g/ml) and IFN␥ (100 units/ml). For experiments with transgenic mouse macrophages, total exudate peritoneal macrophages were isolated by peritoneal lavage with phosphate-buffered saline from 10-week-old animals 3 days after a single intraperitoneal injection of 1.0 ml of 4% thioglycolate broth (Difco Laboratories) as described previously (27). Cells were seeded at a density of 5 ϫ 10 5 /ml and stimulated for TNF production by the addition of LPS (1 g/ml) in the presence or absence of Allotrap peptides. Peptides were added at various concentrations ranging from 3 to 100 M. Cells were incubated in 5% CO 2 /95% air at 37°C for 24 h. At the indicated times, culture medium was collected and stored frozen until use. TNF concentrations were determined by sandwich ELISA using anti-mouse or anti-human TNF mAb (see above) Immunoprecipitation-THP-1 cells were seeded at a density of 5 ϫ 106 cells/ml in methionine-deficient RPMI spiked with [ 35 S]Met (3 Ci/well) and were stimulated for TNF production with LPS (5 g/ml) and IFN␥ (100 units/ml). After 24 h, cells were separated from culture medium and were washed in ice-cold phosphate-buffered saline. Cell pellets were resuspended in lysis buffer consisting of 20 mM sodium phosphate, pH 7.5, 500 mM NaCl, 0.1% SDS, 1% Nonidet P-40, 0.5% sodium deoxycholate, and 0.02% sodium azide (immunoprecipitation buffer). Insoluble material was discarded after centrifugation. Monoclonal anti-TNF antibody (clone mAb1) (3 l/ml) was added to culture medium as well as cell extracts. In identically treated control cells, TP-25 (anti-Class I mAb) or mouse anti-human IgG (3 l/ml) were added to culture supernatants and cell extracts. The samples were incubated overnight at 4°C. 10 l of protein G-agarose was added to each sample, and the mixture was incubated for 2 h at room temperature. The agarose beads were washed three times in ice-cold lysis buffer and counted in a scintillation counter (TopCount, Packard Instrument, Meridan, CT).
Northern Blots and RT-PCR-Trizol reagent (Life Technologies, Inc.) was used to extract total RNA from 5 ϫ 10 6 THP1 cells for each condition and time point used in the LPS stimulation assay. The RNA from each sample was adjusted to an equal concentration according to their A 260 values. RNA was reverse transcribed at 42°C for 1 h using oligo(dT) 15 and avian myeloblastosis virus reverse transcriptase (Promega). Quantitative PCR for TNF and GAPDH was performed using a commercially available kit and according to the instructions of the manufacturer (BIOSOURCE). To allow sample comparisons, results were expressed in number of TNF mRNA copies per GAPDH mRNA copy. The TNF mRNA quantitation shown in this study is the result of two independent quantitative PCRs from the same cDNA.
TNF mRNA from transgenic macrophage cultures was extracted using guanidinium isothiocyanate (28) and subsequently treated with RQ1 RNase free DNase (1 unit/g of nucleic acid, Promega). Total cDNA was synthesized from 5 g of cellular RNA at 37°C for 1 h in a final volume of 20 l containing 5 M oligo(dT) primer, 1 mM of each dNTP (New England Biolabs) in the presence of 200 units of Moloney murine leukemia virus reverse transcriptase (Promega). Primer sequences and PCR conditions were identical to those described recently (25). Data were normalized relative to ␤-actin values. For Northern blot of total RNA from targeted mutant macrophages, 15 g of RNA were resolved in 1.2% formaldehyde agarose gel, blotted onto Hybond Nϩ nylon membrane (Amersham Pharmacia Biotech), and hybridized with 32 P-labeled probes as described (25).

Allotrap 1258-mediated Effects on Cytokine Production in
Vivo-Mice challenged with a single intravenous dose of ConA (20 g/kg) showed up-regulation of circulating cytokine levels that peaked at various time points. 5 mg/kg Allotrap 1258 caused a 75% reduction of TNF plasma levels measured 60 min after treatment. Higher doses of Allotrap 1258 had a similar effect. Similarly, administration of 5-20 mg/kg Allotrap 1258 caused a 50% reduction of IFN␥ plasma levels measured 360 min after treatment. In contrast, administration of Allotrap 1258 at 5 mg/kg had little effect on IL-12 plasma levels (30% reduction), whereas 20 mg/kg Allotrap 1258 resulted in a 70% reduction. Other cytokines studied, namely IL-1␤, IL-2, IL-4, IL-6, and IL-10, showed no differences in response to peptide treatment (data not shown). For all cytokines, the effect of Allotrap 1258 was studied at their respective peak time points. Treatment with Allotrap 1258 alone (in the absence of ConA stimulation) did not result in detectable levels of any of the studied cytokines (Fig. 1).
Allotrap 1258-mediated Effects on TNF Production in Vivo-Allotrap 1258 administration effectively inhibited the production of TNF in mice challenged with GalN and LPS (Table I). Treatment with GalN alone at 800 mg/kg had no effect on serum TNF levels in all strains of mice studied. Challenge with a low dose of LPS (0.1 mg/kg) following GalN, however, resulted in significantly higher TNF levels at 60 min. (Table I). Simultaneous administration of Allotrap 1258 prevented the rise in TNF levels in C57BL/6 and CBA mice, which were undetectable. Similar inhibition was observed in iNOS/KO and HO-1/KO mice as well as in the respective wild type controls ( Table I). Inhibition of LPS induced TNF production correlated with increased survival. Mice (C57BL/6 or CBA) challenged with GalN/LPS died because of endotoxic shock 12-24 h postinjection (Table I). Those treated with Allotrap 1258 at 10 mg/kg showed significantly higher survival rates with 80 -100% animals surviving at 24 h. Negative control peptide D2RP had no protective effects (Table I). Allotrap 1258 treatment in iNOS/KO animals similarly challenged with GalN/LPS resulted in protection of these animals from death (70% survival at 24 h). Because of the limited number of HO-1/KO animals available and given their extreme sensitivity to endotoxin in general (23), survival studies were not done in this group.
Peptide-mediated Inhibition of TNF in Vitro-The mechanism of peptide-mediated inhibition of TNF production was investigated in in vitro experiments. First, the toxicity of the peptide was evaluated in 51 Cr release experiments. The human monocyte cell line THP-1 was labeled with 51 Cr and incubated at 37°C in the presence or absence of Allotrap 1258, or vehicle control for 4 and 24 h. 51 Cr released into the medium by labeled cells cultured in the presence of Allotrap 1258 was not different from the spontaneous release of 51 Cr by untreated cells either at 4 or 24 h (data not shown). This demonstrated that the peptides were not cytotoxic and did not induce cell lysis.
Neither THP-1 cells nor the mouse macrophage cell line RAW264.7 produced detectable amounts of TNF in normal tissue culture. In the presence of LPS (10 g/ml), TNF levels in the RAW264.7 cell culture medium increased continuously over 24 h to a level of 2.46 ng/ml ( Fig. 2A). Similarly, THP-1 cells stimulated with LPS (5 g/ml) and IFN␥ (100 units/ml) for 24 h produced 3.04 ng/ml TNF. (Fig. 2B). In the presence of Allotrap 1258 a dose-dependent inhibition of TNF production was seen both in RAW264.7 cells ( Fig. 2A) as well as in THP-1 cells (Fig.  2B). The IC 50 values obtained was 20 M. Negative control peptide had no effect on LPS-mediated TNF levels in either cell line.
To evaluate whether peptide-mediated inhibition of TNF production was on a translational or transcriptional level, we quantitated the amount of TNF protein and mRNA produced in the presence or absence of peptide. Protein production was measured by immunoprecipitation of [ 35 S]Met-labeled TNF. THP-1 cells were cultured in [ 35 S]Met-containing medium and stimulated to produce TNF in the absence or presence of Allotrap 1258 or negative control peptide. Subsequently, TNF and HLA class I was immunoprecipiated from the culture supernatant and the cell lysate separately. A control antibody (goat anti-human) precipitated small amounts of labeled protein (2160 Ϯ 360 cpm) from the culture supernatant (Fig. 3A). (3108 Ϯ 1088 cpm), whereas the control peptide D2RP had no effect (43447 Ϯ 1769 cpm). However, Allotrap 1258 had no effect on the incorporation of [ 35 S]Met into other proteins. In particular, peptide did not change production of HLA class I immunoprecipitated with mAb TP25. Equivalent amounts of class I were immunoprecipitated from stimulated cultures in-dependent of the presence of Allotrap 1258 or D2RP (61808 Ϯ 2093 cpm). Similar results were obtained with cell lysates (Fig.  3B). However, as expected, the amount of TNF in the supernatant was larger than the amount of TNF in the cell lysate, and the opposite was true for HLA class I.
The amount of TNF mRNA was quantitated by RT-PCR. RNA was isolated from THP-1 cells at 2, 4, and 24 h after addition of LPS Ϯ Allotrap 1258. Following reverse transcription, both TNF and GAPDH cDNAs were quantitated by PCR and ELISA (Biosource). Irrespective of the addition of Allotrap 1258 at 100 M, TNF mRNA levels increased in all samples after stimulation (Fig. 4). Relative to GAPDH mRNA copy numbers, the average number of TNF transcripts increased from 0.66 in unstimulated cells to 9.07 and 7.43 in the presence of LPS alone or with Allotrap 1258, respectively. This minor difference was not statistically significant and could not explain the complete inhibition of TNF protein in the presence of Allotrap 1258. Stimulated cells in this set of experiments produced 96.81 pg/ml TNF. Addition of Allotrap 1258 resulted in complete inhibition of TNF production, and TNF was undetectable. In the presence of a control peptide and LPS, cells produced normal amounts of TNF (117 pg/ml) that did not differ from control LPS-stimulated cells. Allotrap 1258-mediated inhibition of TNF translation requires 3Ј-ARE-dependent and ARE-independent mechanisms The post-transcriptional regulation of TNF mRNA has been reported to rely on mechanisms that utilize elements residing on the 3Ј-UTR of this message (25, 29 -31). To determine whether the Allotrap 1258-mediated suppression of TNF translation requires elements in the TNF 3Ј-UTR, the effect of this peptide on TNF biosynthesis by a 3Ј-modified human TNF transgene was analyzed. Thioglycolate-elicited peritoneal macrophages (TEPMs) were isolated from two separate transgenic lines: (a) Tg1278, which contained a wild type human TNF-3Ј-UTR transgene, and (b) Tg197, which contained a 3Јmodified human TNF transgene bearing the 3Ј-UTR of the ␤-globin gene (24). As in the case of macrophage cell lines, addition of Allotrap 1258 to Tg1278 TEPM cultures inhibited LPS induced hTNF (Fig. 5A) and endogenous mTNF (not shown) production in a dose-dependent manner with a halfmaximal inhibition at a concentration of 6.2 g/ml. No change in the levels of the corresponding mRNAs were detected, via semi-quantitative RT-PCR, at 2 h after LPS ϩ Allotrap 1258 stimulation, confirming that the peptide inhibits translation. In sharp contrast, similar treatment of Tg197 TEPM, which affected endogenous mTNF production (not shown), had no effect on either TNF secretion or mRNA accumulation, demon-  2. Inhibition of TNF production by macrophages. Unstimulated RAW264.7 (a) and THP-1 (b) cells produced undetectable amounts of TNF (closed triangle). Upon stimulation with LPS or LPS ϩ IFN␥, in RAW264.7 cells and THP-1 cells, respectively, a continuous increase in TNF concentration was observed. The TNF concentration after 24 h was 2.492 Ϯ 0.305 ng/ml and 5.587 Ϯ 0.04 ng/ml in RAW264.7 and THP-1 cells, respectively (square). Addition of Allotrap 1258 (closed circles) resulted in a dose-dependent inhibition of TNF production. Negative control peptide D2RP (open circles) had no effect. strating a requirement for the 3Ј-UTR in the Allotrap 1258mediated suppression of TNF translation.
To further define whether the minimal ARE that resides in TNF 3Ј-UTR is necessary for the inhibitory effect of Allotrap 1258 on TNF translation, the effect of this peptide on TNF production from mutant mouse macrophages lacking TNF ARE was examined. To eliminate the possibility that the chronic inflammation that occurs in TNF ⌬ARE mice affects the efficacy of the peptide in inhibiting TNF translation, TEPM were isolated from disease free TNF ⌬ARE/⌬ARE ϫ TNFRI Ϫ/Ϫ with TN-FRI Ϫ/Ϫ TEPM as controls (25). Significantly reduced levels of TNF protein were measured from LPS-induced TNF␣ ϩ/ϩ ϫ TNFRI Ϫ/Ϫ TEPM in the presence of Allotrap 1258 reaching a Ն90% inhibition at 25 g/ml (Fig. 6A). As before, no reduction in the levels of steady state TNF mRNA was observed. Interestingly, similar treatment of TNF ⌬ARE/⌬ARE ϫ TNFRI Ϫ/Ϫ TEPM with Allotrap 1258 resulted in a moderate 50 -60% inhibition of TNF protein production, indicating that translational repression could still take place in the absence of ARE elements (Fig. 6, A and B). Taken together, these results point toward a minor, if any, role of the 3Ј-ARE in Allotrap 1258mediated inhibition and indicate that 3Ј sequences other than ARE are required to impose the full inhibitory effect of Allotrap 1258 on TNF translation. DISCUSSION In this study we continued the characterization of rationally designed immunomodulatory peptides. We demonstrated that peptide Allotrap 1258 did not induce cytokine production in mice. Allotrap 1258 inhibited ConA-and LPSinduced production of TNF in vitro and in vivo. In addition, a partial but significant inhibition of IFN␥ and IL-12 synthesis was observed. Allotrap 1258 therapy also resulted in significant and efficient protection from endotoxic shock following LPS challenge. Survival in all groups was closely correlated with circulating TNF levels. Given the rapid kinetics of TNF, complete inhibition of a TNF burst by Allotrap 1258 treatment may have been responsible for higher survival. This also indicates that the inhibitory activity of the peptide may be a result of direct effects and may not involve synthesis of other cellular proteins. However, systemic changes that may determine survival and occur at later time points were not analyzed in our study.
We have previously reported that Allotrap 1258 modulates heme oxygenase activity in vitro and in vivo (8,12). In animals, peptide therapy resulted in up-regulation of heme oxygenase, and we have speculated that the immunomodulatory activity of Allotrap 1258 may be mediated via effects on HO-1. The degradation of heme by hemeoxygenase results in the production of carbon monoxide, which has been shown to directly affect the activity of guanylate cyclase and therefore cGMP production (32)(33)(34). The secondary messenger cGMP has been implicated in cell growth arrest and the production of TNF. Based on these observations one may speculate that the effect of the peptide on TNF production may also involve heme oxygenase. This hypothesis is supported by the extreme sensitivity of HO-1 knockout mice to LPS challenge (23). In addition, up-regulation of HO-1 was shown to inhibit inflammation, whereas inhibition of HO exacerbated it (14). Up-regulation of HO-1 following administration of hemoglobin was also shown to protect rats from LPS-induced death, whereas inhibition of HO-1 reversed the protective effect of hemoglobin administration (38). These results demonstrate that up-regulation of HO-1 provides protection against inflammatory insults including those induced by LPS. Furthermore, CoPP, an inducer of HO-1, was shown to inhibit TNF production by RAW264.7 cells (13), supporting the hypothesis that the protective effect of elevated HO-1 includes inhibition of TNF synthesis. On the other hand, we have seen that the LPS-induced rise in serum TNF levels in mice was rapid (60 min), and it is unclear whether a small up-regulation of HO-1 expression can effectively inhibit TNF production. To test the above hypothesis, further studies in HO-1 knock-out animals were carried out. The inhibition of TNF synthesis in Allotrap 1258-treated HO-1-deficient animals was as effective as in normal mice. This indicates that the protective effects of Allotrap 1258 on TNF levels are not mediated via direct effects on HO-1.
Another closely related molecule, NO, generated by the activity of iNOS and shown to be involved in cGMP-mediated TNF production was studied using iNOS knock-out animals. Similar to the result with HO-1/KO mice, Allotrap 1258 therapy resulted in significantly lower levels of TNF in iNOS/KO mice and protected these animals from death. This further supports our hypothesis that Allotrap 1258 exerts direct effects on TNF production, independent of the presence of HO-1 or iNOS.
The mechanism of inhibition of TNF production was investigated in vitro using macrophage cell lines or those from mice expressing various TNF transgenes. Allotrap 1258 inhibited effectively the production of TNF by mouse and human macrophages. Quantitative RT-PCR analysis and Northern blot analysis of TNF mRNA levels demonstrated that addition of Allotrap 1258 had little or no effect on human and mouse TNF mRNA levels. However, under those conditions the synthesis of TNF was completely suppressed indicating that the peptide inhibited TNF mRNA translation. Addition of peptide resulted in a complete inhibition of TNF synthesis but had little effect on overall protein synthesis as measured by [ 35 S]methionine incorporation. The degree of peptide-mediated inhibition was similar in cell extracts and culture medium, indicating that the peptides did not influence the enzyme-mediated release of TNF into the medium.
It is well established that elements in the TNF mRNA 3Ј-UTR are involved in the regulation of TNF mRNA translation (25,30,31,35). In particular it has been demonstrated that an ARE in the TNF 3Ј-UTR mediates both alleviation and reinforcement of message destabilization and translational silencing. In addition, in the absence of ARE, TNF mRNA is no longer responsive to translational modulation by the stress-activated protein kinases p38/SAPK and JNK/SAPK. Absence of the ARE from TNF mRNA rendered the effect of Allotrap 1258-mediated inhibition of TNF mRNA translation less effective but did not eliminate it. In contrast, replacement of the TNF mRNA 3Ј-UTR completely abolished the inhibitory effect of the peptide. Furthermore, Allotrap 1258 had little effect on the expression of IL-1␤, whose mRNA contains ARE. Similarly, Allotrap 1258mediated inhibition of IFN␥, another mRNA containing ARE, was much less effective than Allotrap 1258-mediated inhibition of TNF expression. Taken together these results clearly demonstrate that the peptide inhibits TNF mRNA translation via mechanisms involving besides the ARE, other as yet unidentified elements in the 3Ј-UTR of TNF mRNA. Such elements have been described previously (30) by analysis of the translation efficiency of a reporter gene linked to various deletion mutants of the TNF 3Ј-UTR. Hel et al. (37,38) further identified a distinct and novel region in the 3Ј-UTR shown to be different from the ARE and characterized cytosolic proteins that formed complexes with the regulatory elements. At this point, we cannot attribute direct binding of Allotrap 1258 to the described regulatory elements in the 3Ј-UTR of TNF mRNA, as the potential mechanism of inhibition. Further studies are necessary to determine whether mechanism of inhibition of translation by Allotrap 1258 involves other cellular protein(s). These observation pose the following possibilities: (a) that mechanisms affected by Allotrap modulation act positively (and in an additive fashion to the ARE functions), to activate TNF translation or (b) Allotrap inhibition acts through a novel negative enforcing mechanism. In either case it seems that the ARE-independent modulation may not involve the utilization of SAPK modules because it has been shown that these modules act through the ARE (39,40).
The ARE-dependent translational modulation of TNF mRNA has been demonstrated to rely on SAPK signals, namely through the p38 and the JNK modules (25,39,40). It can therefore be speculated that the ARE-dependent, Allotrap 1258-mediated, translational inhibition may result from the modulation of these kinases. However, the ARE-independent control may not utilize similar signaling pathways because the ARE dependence for SAPK modulation of the TNF message seems, at least in macrophages, to be absolute. The application of the Allotrap 1258 peptide for the characterization of new 3Ј element(s) as well as the corresponding signaling modules via which the effects of these elements are modulated will provide novel insight into the modes of control of TNF mRNA translation.
TNF is an important cytokine that has been vastly studied for both its inflammatory as well as immunoregulatory properties. TNF is a mediator of inflammation and tissue damage and appears to be involved in the pathology of several diseases including sepsis syndrome (41), rheumatoid arthritis (42,43), inflammatory bowel disease (43,44), and cachexia (45). Recently, anti-TNF monoclonal antibody therapy was shown to be effective in patients suffering from Crohn's disease (a particular form of inflammatory bowel disease) (22). Similar observations have been made in rheumatoid arthritis patients (36). The application of a small non-immunogenic molecule like Allotrap 1258 that effectively lowers TNF serum levels could be of clear clinical benefit in similar situations.
In conclusion our results show that (a) the immunomodulatory peptide Allotrap 1258 displays a strong suppressive capacity on the in vivo and in vitro production of potent pro-inflammatory mediators and predominantly of TNF; (b) peptidemediated inhibition of TNF biosynthesis is independent of the NO and HO pathways and occurs at the level of TNF mRNA translation; and (c) translational inhibition targets cis-elements residing on the 3Ј-UTR of TNF mRNA; more specifically, Allotrap 1258-mediated inhibition of TNF translation requires the function of both ARE-dependent and ARE-independent mechanisms, establishing novel non-ARE elements that modulate translation of TNF mRNA.