The TNF Receptors p55 and p75 Mediate Chemotaxis of PMN Induced by TNFα and a TNFα 36–62 Peptide

The present study was performed to examine whether residues 36–62 of TNFα contain the chemotactic domain of TNFα, and whether the p55 and p75 TNF receptors are involved in TNFα induced chemotaxis. The chemotactic effect of TNFα on PMN was inhibited by the mAbs Hrt-7b and Utr-1, against the p55 and p75 TNF receptors, respectively. Both receptors may therefore be required for mediating the chemotactic effect of TNFcz. The synthetic TNFα 36–62, similar to TNFα, had chemotactic effects on both PMN and monocytes. The chemotactic activity of the TNFα 36–62 peptide on PMN, was inhibited by Htr-7b, Utr-1 and soluble p55 receptor, which shows that the peptide possessed the ability to induce chemotaxis through the TNF receptors. In contrast to TNFα, the peptide did not show a cytotoxic activity against WEHI 164 flbrosarcoma cells. It is suggested that different domains of the TNFα molecule induce distinct biological effects.


Introduction
Circulating polymorphonuclear cells (PMN) and monocytes are activated by chemotactic factors for recruitment to sites of inflammation. The pleiotropic cytokine, tumour necrosis factor-(TNF), has been reported to be chemotactic, as it induces directional locomotion of PMN and monocytes in vitro. '-5 Furthermore, in vivo studies show that TNF0t plays a crucial role in the recruitment of neutrophils at an early stage, and monocytes at a later stage of immune complex-induced inflammatory reactions. 6,7 Many biological effects induced by TNF: -11 have been shown to involve binding to the 55 kDa (p55) and the 75 kDa (p75) TNF receptors, which are expressed on almost all cell types, m However, the involvement of these receptors in the TNFx induced chemotaxis has not been studied.
An interesting approach for the study of distinct TNF{x activities is the use of TNF peptides. Different TNF peptides have recently been reported to induce distinct TNF effects, [2][3][4][5][6][7][8][9][10][11][12][13][14] and inhibit binding of TNF to the TNF receptors. The authors have performed molecular dynamic calculations 15 combined with studies on the three-dimensional structure of TNF0t 16 in order to design TNF0t peptides which could interact with TNF receptors, and induce TNF0t effects. It was found that a peptide including residues 36-62 had conformational properties which could be related to the corresponding parent molecule. This sequence is also one of the most homologous domains between TNF and TNF, which both bind to the TNF receptors. e In the present study, this peptide was investigated for two crucial TNF0t effects, chemotaxis and cytotoxicity. The involvement of p55 an p75 TNF receptors in the chemotactic response of TNF and TNF{x 36-62, was also studied.

Materials and Methods
Molecular modelling: The molecular modelling studies of TNF peptides were performed using the Molecular Simulation Inc. Quanta 3.2/CHARMm 21.2 program package, on a Silicon Graphics Personal IRIS 4D/30 EG (USA). The peptide atom coordinates were obtained from the TNF0t X-ray structure (pdbl tnf), and peptide candidates were minimized by molecular mechanics using 2000 steps of adopted basis set Newton-Raphson minimization, before the minimized structures were subjected to molecular dynamic calculations with a total simulation of 250 ps at 300 K. Synthesis of TNF peptides: TNF: peptides were synthesized as described previously, 7 using Fmoc chemistry on a semi-automatic peptide synthesizer (Milligen, Model 9020). The peptides were purified and analysed using reverse-phase HPLC, and FIB-MS on a VG Tribid MS instrument (VG Analytical, Manchester, UK). Materials: Human recombinant TNF0t (Hr TNF00, with a specific activity of 1.0 x 108 U/mg, was purchased from Boeringer (Mannheim, Germany). The generation of the mAbs Utr-1 an Htr-7b, specific for p75 and p55 respectively, is described elsewhere, TM and soluble p55 was kindly provided by Dr Hansruedi Loetscher, Hoffman-La Roche (Basel, Switzerland). Anti-IL-8 and anti-MCP-1 were purchased from British Biotechnology (UK). Formylmethionyl-leucyl phenylalanine (FMLP) was purchased from Sigma Chemical Co. (St Louis, MO). Endospecy from Seikagaku Co. (Tokyo, Japan) was used to check endotoxin contamination. PMN isolation: Polymorphonuclear cells (PMN) were isolated as follows: 2 ml of freshly drawn heparinized blood (10 U/ml) from healthy adults was applied on top of a bilayer consisting of 3 ml polymorphoprep and 3 ml lymphoprep in polycarbonate tubes (Nycomed Pharma AS, Norway). After centrifugation at 530 x g for 20 min, the PMN band in the polymorphoprep layer was isolated. The cells were washed once with ice cold sterile 0.15 M NaCl, and centrifuged at 185 x g for 10 min. Contaminating erythrocytes in the PMN band were lysed with ice cold 0.2% NaCl for 90 s. The cells were resuspended at 106/ml in ice cold RPMI-1640 and used immediately. The PMN preparation contained at least 95% neutrophils.
Monocyte isolation: The monocyte band was isolated using the method previously described by Byum. 19 In brief, mononuclear cells (PMBC) from healthy adults, from either freshly drawn heparinized blood or buffy coats (10 U/ml), were centrifuged on lymphoprep, isolated and washed with 0.15 M NaCl. The PMBC, resuspended at 106/ml in RPMI-1640 were used directly for chemotaxis studies. Assay for chemotaxis: TNF and TNF peptides were tested for chemotactic activity on PMN and PBMC.
Chemotactic activity was assayed in a 48-well microchemotaxis chamber (Neuro Probe Inc. Cabin John, MD, USA), as described previously, i In brief, the upper wells were filled with 50 l-tl of cells, and 25 l.tl of the compounds tested for chemotaxis were filled in the bottom wells. For checkerboard analysis, the stimulants were also placed in the upper wells. In the pretreatment studies, the anti-TNF receptor antibodies, or other antibodies were added to the cells for 10min at 4C, before they were placed in the upper wells. The soluble p55 was mixed with 0.5 l.tM TNF0t 36-62 in a 1:1 molar ratio at 20C for 10min, before addition to the lower wells. A polycarbonate-polyvinyl pyrrolidone (PVP) filter with 51.tm pore size was used in the PBMC chemotaxis assay, while a PVP-free polycarbonate filter, with the same pore size, was used for the PMN chemotaxis assay. Chemotaxis chamber assemblies could also be recognized in the peptide. The two loops (38-41 and 50-54) located at the base of TNF{x, and important residues surrounding a shallow depression which are suggested to be involved in receptor binding, 6 were all exposed in a similar manner in the peptide as in TNFx. The distance between the carbon in the C-terminal carboxyl and the nitrogen in the N-terminal amino group in TNF0t    (Fig. 3A), when used separately or in combination.
Likewise, the chemotactic effect of TNF0t 362 peptide was also significantly inhibited by Utr-1 or Htr-7b (Fig. 3B) or a combination of both mAb. These mAb alone, or in combination, did not induce chemotaxis on PMN. FMLP induced migration was not inhibited by these mAb alone or in combination. The experiments were repeated four times, and similar results were obtained despite donor variations. Antibodies against IL-8 or MCP-1 did not show any inhibitory effect on either the TNF, TNF  p55 was able to significantly inhibit the chemotactic response of the TNF0t 36-62 peptide on PMN (Fig. 4).
Soluble p55 did not inhibit the chemotactic effect of FMLP (Fig. 4), indicating a specific binding of TNF0t It is shown that the chemotactic activity TNF on PMN involves both p55 and p75 receptors. Even nonredundancy in the function of the two receptors for some effects has been observed, 27 a simultaneous involvement of both p55 and p75 receptors has been reported for a series of TNF0t activities, such as induction of differentiation of ML-1 cells, NF-cB activation, cytotoxicity on U937 cells and IL-6 production by endothelial cells. [27][28][29] Tartaglia et al. 3 have suggested that the high affinity p75 receptor may regulate the rate of TNF0t association with the p55 receptor, by increasing the local concentration of TNF0t through rapid ligand association and dissociation. In contrast to this, Brouckaert et al. 1 have proposed an alternative cooperation between the two receptors where p75 interferes with the p55 signalling pathway. This hypothesis is based on the fact that p75 triggering is not sufficient to initiate the redundant signals and that p75 triggering can diminish p55 mediated c-fos induction. However, it remains to be explored how the two receptors cooperate in TNF0t induced chemotaxis. Similar to TNF0t, the TNF0t 36-62 peptide also induced chemotaxis on PMN and PBMC. The finding that soluble p55 inhibited this effect (Fig. 3), indicates that the peptide possesses the conformation needed for interaction with the receptor. This supports the results of the molecular dynamic calculations (Fig. 1). That TNF{x 36-62 is able to interact with soluble p55, is also in line with the recent work by Ratjen et al. 1  Although these authors did not show that the chemotaxis induced by the peptide was TNF receptor mediated, desensitisation studies suggested involvement of TNF receptors. The present work supports and extends this hypothesis. TNF0 36-62 was not cytotoxic, suggesting the existence of distinct TNFc regions for the cytotoxic and chemotactic effects. It is noteworthy that two other TNF peptides, which overlap with only four residues (54-58), were shown to be cytotoxic, 1 suggesting a critical domain for TNF cytotoxicity. This sequence is included in our TNF 36-62 peptide and the molecular calculations of the peptide suggested that this specific domain encompassing residues 54-58 did not possess the conformation needed for optimal interaction with the TNF receptors. This was due to the attraction between the oppositely charged Cand N-terminals (Fig. 1).
Crystallographic studies on the TNF[3/TNF receptor complex have shown that three TNF]3 monomers bind three TNF receptors in a symmetrical fashion. 2 It has been suggested that a crosslinking of TNF receptor is also necessary for signal transduction leading to TNFo effects, [3][4][5] and is based, on studies with antibodies and TNFot mutants against TNF receptors. [3][4][5][6] Our and other investigations with synthetic TNFot peptide fragments, 1'-14 however, suggest that at least some of the TNFc effects are not dependent on crossreaction of TNF receptors with three TNFot monomers.
In conclusion, different TNFot peptides may induce distinct activities, indicating that TNF possess distinct domains critical for different TNF activities. This property opens the possibility of designing TNFx fragments with specific TNFot effects. We are currently investigating this hypothesis by studying TNFot 36-62 and several other TNF peptides for their bioactivities and specificities to target cells.