Tumor necrosis factor inhibitory protein tip B1 and method of using same

Abstract

A specific, unique, natural biological factor that has the capacity to abate one or more of the toxic effects of TNF. This factor may have therapeutic potential in one or more of the diverse disease states in which TNF is known to have a causative role. The unique, approximately 27 kDa protein is designated TIP-B1. TIP-B1, when purified and added back to TNF sensitive cells, effectively ablates TNF-induced cell lysis and TNF-induced apoptosis; cytolytic effects are central to many TNF-mediated actions. The invention includes a protein designated TIP-B1 purified to homogeneity. The protein has a molecular weight of about 27 kD. The protein is capable of inhibiting the action of TNF upon a cell, when said protein is introduced into extracellular medium surrounding the cell. The protein is free of sequences which interfere with normal cellular TNF binding sites and is free of sequences which directly bind to TNF. The invention further includes a method for using the protein for controlling TNF activity upon a cell.

Description

BACKGROUND OF THE INVENTION

This invention relates to the inhibition of the activity of tumor necrosis factor (TNF). TNF is known to have a causative role in many diseases such as chronic inflammation, certain types of arthritis, late stage wasting disease in cancer (cachexia), auto immune diseases such as lupus, and septic shock.

TNF is a cytokine which possesses multiple activities and affects almost every type of cell in the body. It was initially identified and named because of its ability to cause hemorragic necrosis of certain tumors in mice. It was subsequently recognized as the causative agent in the wasting condition in late stage cancer known as cachexia. A single gene codes for a membrane-bound 26 kD TNF protein; proteolytic cleavage produces circulating 17 kD TNF. Both are believed to be active. TNF is produced by most cells of the immune system, certain other cells and is constituitively produced in the thymus and by many leukemic cell lines.

TNF plays a critical role in the immune system and inflammation which may be ascribed to three different TNF functions. Evidence indicates that TNF is important in 1) facilitating leukocyte adhesion to and migration through endothelial cells by effects on remodeling of the extracellular matrix as well as on regulating cell adhesion molecules and integrins, 2) cell growth and differentiation and 3) programmed cell death (apoptosis) and cytotoxicity. The Immunopharmacology and Therapeutics section of the Department of Experimental Therapeutics (Grace Cancer Drug Center) has made a number of significant contributions toward clarifying the effects of TNF on host defense systems, and is currently using this information to develop effective anticancer therapies.

TNF's multiple activities may be explained in part by the existence of two receptors for TNF (molecular weights of 55 kD, TNFR 1 , and 75 kD, TNFR 2 ). They are members of a TNFR superfamily which share considerable extracellular homology, indicating that they bind to similar ligands, but limited intracellular homology, suggesting that they may initiate different signal transduction/second messenger systems. Consistent with their assignment to this receptor superfamily, the TNFRs: 1) share extracellular homology and bind both TNF and lymphotoxin (TNFβ) with high affinity, and 2) have little intracellular homology, consistent with TNF's effects on a multitude of molecules involved with toxic/detoxification mechanisms and signal transduction. Recent evidence suggests that the cytoplasmic portions of the receptors bind different proteins, and that these proteins may also bind to other TNFR superfamily members.

The relative roles of TNFR 1 and R 2 signaling in TNF effects differ depending on the cells under investigation, and the effect examined. Both receptors are present on most cells, complicating the assignment of a specific effect to a specific TNFR type. Data indicate that TNFR 1 stimulation, including receptor clustering, leads to cell lysis (both necrotic and apoptotic mechanisms have been reported); whether or not TNFR 2 signaling contributes to this lysis is not always addressed. The “death domain”, believed to signal lysis, at least in part through activation of the apoptotic pathway, is present in the cytoplasmic portion of both the TNFR 1 and another TNFR superfamily member, Fas, whose binding also induces apoptosis. It is unclear if TNF and Fas induce apoptosis by the same mechanism. Several studies have linked activation of TNFR 2 , probably including clustering, to stimulation of proliferation, but TNFR 2 's role in TNF mediated lysis is less defined. It does appear that, under certain conditions, TNFR 2 signaling and receptor clustering are responsible for a component of TNF mediated lysis.

A small number of transformed cell lines are sensitive to direct TNF cytotoxicity, but a greater number are TNF resistant. In the large majority of cases, TNF resistance of these cells does not appear to be correlated with a lack of TNF receptors or with defects in the internalization of the TNF-receptor complexes. In fact, many resistant cells are lysed by TNF when a protein synthesis inhibitor is present. This finding suggests that TNF induces/activates (at the mRNA level and/or the protein level) a protein or proteins which protect(s) the cell from TNF mediated lysis. TNF increases mRNAs for molecules related to cell adhesion, contact and substructure (ELAM-1/E-selectin, ICAM-1, VCAM-1, collagen, vimentin, collagenase, and stromelysin), certain cytokines and lymphokines (IL8 and MCAF/MCAP-1/murine JE), and several key transcription factors (p65 subunit of NF-kB, c-myc, c-jun, c-fos, and c-fra). A number of proteins/activities are known to be induced by TNF, although their respective mRNAs are not always found in TNF subtraction libraries. This suggests that TNF may also induce proteins at the translational/posttranslational level. TNF treatment of cells has been shown to result in increased secretion of a number of cytokines and immune factors (including IL6, IL8, IL1, CSF-1, GSCF and GMCSF) and receptors (MHC class I molecules, the p55 IL2 receptor and the EGF receptor).

Several proteins have been suggested, with varying numbers of supporting studies, as proteins which protect cells from TNF mediated lysis. However, their mechanisms of protection and their physiologic significance have yet to be elucidated. These proteins do not appear to protect the cells completely, and the investigators involved are in agreement that more than one protein is probably involved in protection of a given cell. These data are consistent with the hypothesis that multiple mechanisms contribute to TNF initiated lysis, and, therefore, the proteins which are protective to a cell correlate to the specific mechanism(s) by which that cell is lysed by TNF. Proteins which have been linked to protection from TNF induced lysis include: 1) manganese superoxide dismutase (MnSOD), an enzyme which converts toxic superoxide radicals to innocuous products, 2) plasminogen activator inhibitor 2 (PAI 2 ), a cytosolic protein which is a member of the serpin family of protease inhibitors, 3) the ˜80 kD protein encoded by the A20 clone of a TNF subtraction library, 4) heat shock protein (hsp) 70 and hsp 27 , and 5 ) endogenous TNF. Most of the studies implicating these proteins in resistance to TNF cytotoxicity involve transfection and expression of the protein in cells which then show increased resistance, or surveys which show increased levels of these proteins in subclones with increased resistance. However, levels of activity or mRNA of these proteins do not always correlate with the degree of resistance.

U.S. Pat. No. 5,136,021, entitled “TNF-Inhibitory Protein and a Method of Production”, discloses a TNF inhibitory protein named TIP. Subsequent to filing the patent, the amino-terminal amino acid sequence (10 amino acids) of TIP was determined to be V-V-X-A-V-X-L-X-A-H (SEQ. ID. #1). As this protein was in very low abundance and the purification procedure was extremely labor intensive and difficult, it was not possible to obtain sufficient material for internal amino acid sequencing. These sequences were necessary to facilitate the search for the TIP gene. An alternative simpler and more effective method to isolate a “TIP” was therefore needed.

BRIEF DESCRIPTION OF THE INVENTION

In accordance with the invention, a specific, unique, natural biological factor has been discovered that has the capacity to abate one or more of the toxic effects of TNF. This factor, therefore, may have therapeutic potential in one or more of the diverse disease states in which TNF is known to have a causative role. The unique, approximately 27 kDa protein is designated TIP-B 1 . TIP-B 1 , when purified and added back to TNF sensitive cells, effectively abates TNF-induced cell lysis and TNF-induced apoptosis; cytolytic effects are central to many TNF-mediated actions.

More particularly, the invention includes a protein designated TIP-B 1 purified to homogeneity. The protein has a molecular weight of about 27 kD. The protein is capable of inhibiting the action of TNF upon a cell, when said protein is introduced into extracellular medium surrounding the cell. The protein is free of sequences which interfere with normal cellular TNF binding sites and is free of sequences which directly bind to TNF.

The invention further includes a method for using the protein for controlling TNF activity upon a cell.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1A is a graph showing kinetics of induction of protection from TNF+CHX lysis by pre-exposure to low concentrations of TNF.

FIG. 1B is a graph showing kinetics of induction of protection from TNF+CHX lysis by pre-exposure to low concentrations of TNF.

FIG. 2A is a gel showing isolation of three fractions with tip-like activity by isoelectric focusing on a Rotofor® apparatus.

FIG. 2B is a graph showing percent protection of a cell line by exposure to the protein fractions isolated in FIG. 2 A.

FIG. 3A is a graph showing protection by the proteins in Fraction 2+3 further resolved by Rotofor® refocusing—protective activity (protein fractions 9-17) and pH of resolved fractions 8-20.

FIG. 3B is a gel and graph showing the distribution of the resolved 27 kDa and 37 kDa proteins in fractions 9-17.

FIG. 4A shows a two dimensional gel analysis. Protein in pooled fractions 9-11 was resolved by isoelectric focusing (IEF, horizontal direction) followed by PAGE (vertical direction) and then silver stained. Molecular weight standards are visible on the left, the faint smear across the top is a staining artifact. The amino acids obtained from aminoterminal sequencing of pooled fractions 9-11 are shown.

FIG. 4B shows the peptides selected for sequencing following HPLC separation of the products produced by the endoproteolytic cleavage of the ˜27 kDa protein in pooled fractions 9-11 (FIG. 3 ). The amino acid sequences of peptides A, B and C are shown.

FIG. 5 demonstrates that preincubation of U937 cells with TIP-B 1 does not change TNF binding capacity. 5×10 5 U937 cells were incubated without any addition (top panel) or with TIP-B 1 (bottom panel) for 6 hours. The cells were incubated with medium or biotinylated TNF, washed, incubated with Streptavidin-red-670, fixed, and analyzed on a flow cytometer. Unlabeled TNF was capable of competing with biotinylated TNF for binding in a concentration dependent manner (data not shown). U937 51 Cr release assays run in parallel showed TIP-B 1 inhibited TNF mediated lysis in a concentration dependent manner; for the conditions shown, inhibition was 20%.

FIG. 6 is a series of graphs showing that TIP-B 1 protects U937 cells from TNF-induced apoptosis assessed based on percentages of total (10,000) events found in the subdiploid propidium iodide labeled DNA peak.

Panel A: control cells cultured for 24 h in medium, 11.25% subdiploid.

Panel B: cells cultured for 18 h in medium and 6 h in TNF+CHX, 28.78% subdiploid.

Panel C: cells cultured for 18 h with TIP-B 1 and 6 h with TNF+CHX, 13.05% subdiploid.

FIG. 7 is a series of graphs showing that TIP-B 1 protects U937 cells against TNF mediated lysis. Arrows indicate the TdT positive events (also, see Table 3).

FIG. 8A is a map of primers used to generate a TIP-B 1 specific PCR product from a cDNA library from TNF-treated BG-9 cells. The location of degenerate oligonucleotide primers, based on the amino acid sequence of TIP-B 1 internal peptides A and B, that were used to generate a 120 bp PCR product encoding these peptides are shown.

FIG. 8B is a map of primers used to generate a TIP-B 1 specific PCR product from a cDNA library from TNF-treated BG-9 cells. P18 and P27 TIP-B 1 specific oligonucleotide primers were used to generate a 274 bp PCR product from TNF-treated BG-9 cell cDNA.

FIG. 9A is an agarose gel electrophoresis of the 274 bp PCR product derived from various DNA samples: TNF-treated BG9 cDNA library, BG9 second subtraction cDNA library and genomic DNA.

FIG. 9B is an agarose gel electrophoresis of the 274 bp PCR product derived from various DNA samples showing the comparison between the product derived from TNF-treated and control BG9 cells, along with the purified 274 bp fragment, obtained above, used as a reference standard.

FIG. 10 is a Southern blot analysis of the eight putative TIP-B 1 cDNA clones probed with the 274 bp insert encoding the three internal TIP-B 1 peptides.

FIG. 11 is an agarose gel electrophoresis of the restriction enzyme digests of the plasmids derived from bacterial clones used in FIG. 10 revealing the 0.8 kb insert. These six clones have been given the letter designations A, B, C, D, K and S. In addition to the restriction fragments visible on the copy, faint additional bands were observed for C (0.3/0.4 kb) and D (0.9+0.3).

FIG. 12 is a Southern blot analysis of the six putative TIP-B 1 clones (A, B, C, D, K and S). The plasmids were restricted, separated on agarose gels, transferred to a membrane and then exposed to the 32 P-274 bp probe.

FIG. 13 is a nucleotide sequence of Clone A (also called tip-SN) and its encoded amino acid sequence including the sequenced TIP-B 1 peptides A, B, C and consensus sequences for three myristylation sites (M 1 , M 2 and M 3 ), an amidation site, and a casein kinase II site (CKII).

FIG. 14A is a Northern blot analysis of RNA isolated at different times from BG9 cells following treatment with TNF and probed with Clone A insert. RNA samples were resolved by formaldehyde agarose gel electrophoresis, transferred to a membrane which was probed with 32 P-insert from Clone A. Lanes are: 1,2-2 h C; 3,4-2 h TNF; 5,6-6 h C; 7,8-6 h TNF; 9,10-12 h C; 11,12-12 h TNF; 13,14-24 h C; 15,16-24 h TNF. Film exposure time 12 hr.

FIG. 14B is the Northern blot of FIG. 14A with a film exposure time of 36 hrs.

FIG. 15 is an SDS-PAGE, stained with Coomassie blue, of proteins from bacteria expressing the PET vector containing the Clone A insert. Lanes: 1—mol. wt. markers ( 31 , 20 , 17 , 14 , 8 , 6 and 2.5 kDa); 2—10 5 g supernatant from uninduced bacteria; 3—10 5 g supernatant from IPTG induced bacteria; 4—40% NH 3 SO 4 precipitate; 5—centricon (30 kDa) retentate; 6—centricon (10 kDa) retentate; 7—blank; 8—guanidiniumHCl treated; 9—blank; 10—40% NH 3 SO 4 supernatant.

FIG. 16 is a graph showing protection of BG9 cells from TNF+CHX lysis by 10 5 g supernatant from Clone A recombinant bacteria (Abs 570 : controls˜0.600; TNF+CHX˜0.270).

FIG. 17 is a Northern blot hyridization of Clone A with RNA from various cell types.

FIG. 18 shows Western blots of bacterial proteins using unadsorbed and adsorbed antiserum 459. This antiserum was generated against the recombinant partial TIP-B 1 protein (˜14 kDa): the product of Clone A transfected bacteria.

FIG. 19 shows a Western blot analysis of BG9 and HEL Cell 100 kxg supernatants with antiserum 17 . This antiserum was generated against purified TIP-B 1 (fractions 9-11, see FIGS. 3 and 4 ).

FIG. 20 is a graph showing comparison of protective activity of equal amounts of Rotofor® fractions obtained from untreated (&Circlesolid;) and TNF-treated (▪) BG9 cells.

FIG. 21 is a graph showing protection of NHDF cells from TNF+CHX by pretreatment with TNF.

FIG. 22 is a Western blot analysis of NHDF proteins with 18 hr pre-treatment with various concentrations of TNF (U/ml) using antiserum 459 and antiserum 17 .

FIG. 23A is immunocytochemical analysis of paraffin embedded NHDF cells—preimmune untreated.

FIG. 23B is immunocytochemical analysis of paraffin embedded NHDF cells—preimmune—50 U/ml TNF 18 hours.

FIG. 23C is immunocytochemical analysis of paraffin embedded NHDF cells—antiserum 17 untreated.

FIG. 23D is immunocytochemical analysis of paraffin embedded NHDF cells—antiserum 17 —50 U/ml TNF 18 hours.

FIG. 24A is immunocytochemical analysis of NHDF cells grown on coverslips—preimmune—untreated.

FIG. 24B is immunocytochemical analysis of NHDF cells grown on coverslips—antiserum 17 —untreated.

FIG. 24C is immunocytochemical analysis of NHDF cells grown on coverslips—antiserum 17 —50 U/ml TNF 18 hours

FIG. 25A is a graph showing that MCF7 cells are lysed by TNF (72 h, with MTT added for the last 6 h).

FIG. 25B is a graph showing that the MDR subline MCF7/A6 are TNF resistant (72 h, with MTT added for the last 6 h).

FIG. 26 is a gel showing that MCF7/A6 cells (lane 2 ) contain twice as much TIP-B 1 mRNA as MCF7 (lane 1 ) cells.

FIG. 27A is a Western blot analysis of selected cell lines cell proteins with or without 18 hr 10 U/ml TNF pre-treatment—antiserum 459 .

FIG. 27B is a Western blot analysis of selected cell lines cell proteins with or without 18 hr 10 U/ml TNF pre-treatment—antiserum 17 .

FIG. 28A is a Western blot analysis of MCF-7 and MCF-7/A6 untreated and pre-treated with 10 U/ml TNF for 18 hrs—preimmune 17.

FIG. 28B is a Western blot analysis of MCF-7 and MCF-7/A6 untreated and pre-treated with 10 U/ml TNF for 18 hrs—preimmune 459.

FIG. 29 are gels showing anti-TIP-B 1 (#17) and anti-r-TIP-B 1P (#459) antersera have similar reactivities and two different drug resistant cell sublines express an ˜60 kDa protein which reacts with antiserum 17 . Proteins were from MCF7 (lanes 1 , 2 ); MCF7/A6 (lanes 3 , 4 ); REH (lanes 5 , 6 ); REH3.6 (lanes 7 , 8 ). Antisera used were #459 (lanes 1 , 3 , 5 , 7 ) and #17 (lanes 2 , 4 , 6 , 8 ).

DETAILED DESCRIPTION OF THE INVENTION

TIP-B 1 as used herein means a protein having a molecular weight of about 27 kD which acts as an extracellular protein which inhibits the activity of TNF upon a cell, without binding to TNF or interfering with the binding of TNF to the target cell. TIP-B 1 , as discussed in detail below, is obtained from lysed TNF treated cells by isoelectric focusing or an equivalent method.

In a particular example, in accordance with the invention, a simple four step protein purification method yielded a TIP-B 1 protein, having a molecular weight of about 27 kD, which has TNF inhibitory activity and is a unique protein. The amino-terminal amino acid sequence was determined to be A-P-Y-T-V-V-Y-F-P-V-R-G-R-X-A-A-L-R (SEQ. ID. #2) and, therefore, is different from the protein covered by U.S. Pat. No. 5,136,021.

Many cell lines which are lysed by appropriate concentrations of TNF alone or TNF plus macromolecular synthesis inhibitors such as cycloheximide (CHX) or actinomycin D are rendered refractory by pre-exposure to low concentrations of TNF. An example of such an effect is shown on FIGS. 1A and 1B . These data were obtained using the BG9 fibroblastoid cell line. Panel A shows the rapid onset (i.e. within 15 min. of exposure to 2.5 U/mL TNF) of a relatively modest level of protection and its equally rapid decay. Subsequently, a second inhibitory activity develops. It is detectable by 2 h, reaches a maximum by 4 h and is sustained for at least 20 h (FIGS. 1 A and B). A similar TNF concentration dependence was noted for the induction of both the rapid onset and delayed protective activities; namely, the maximally effective concentration was ˜2.5 U/mL and concentrations both lower (e.g. 0.5 U/mL) or higher (e.g. 50 U/mL) were less effective.

Experimental evidence was obtained indicating that: 1) this protective activity did not develop if a protein synthesis inhibitor was present during the period of pre-exposure to low concentrations of TNF, 2) cells pre-treated with the lysate from TNF-treated cells were protected from TNF-induced lysis, 3) the protective activity was present in the 10 5 xg supernatant from sonicated TNF-treated cells.

To obtain a specific TIP-B 1 protein, proteins in the 10 5 xg supernatant from TNF-treated cells were subjected to isoelectric focusing using a BioRad Rotofor® isoelectric focusing apparatus and ampholytes covering the pH range of 3 to 10. The twenty fractions from the Rotofor® were analyzed for pH, protein content and protective activity. There were three peaks of activity. The fractions encompassing each of the three peaks were pooled and analyzed for activity and for molecular size by SDS-PAGE (FIGS. 2 A and 2 B). All three pooled fractions had similar levels of protective activity. The pool of Fractions 2 plus 3 (pI 4 to 4.5) appeared to have only one predominant protein on SDS-PAGE; whereas, the other two pools had a number of prominent proteins. Based on these observations, it was decided to concentrate initially on the protein in the pool containing fractions 2+3.

The protein(s) in the 2+3 fractions combined from 5 separate Rotofor® runs were resolved into 20 fractions by a second Rotofor® separation. The ampholytes in the second separation were only those present in the 2+3 fractions from the initial 5 separations. The pH gradient formed in the second Rotofor® separation, therefore, was very shallow (FIG. 3 A,+−−+). All 20 fractions were analyzed for protein content and only those with detectable protein were analyzed further (fractions 8 and below and 18 and above did not have detectable protein). The protective activity profile for these 9 fractions (i.e. fractions 9-17) is shown in FIG. 3 A and the SDS-PAGE analysis is shown in FIG. 3 B. It can be seen that, when an equal volume of each fraction was tested, all fractions containing detectable protein had activity and the activity peaked in Fraction 13 at 100% protection. There were two proteins visible, a predominent ˜27 kDa one (fractions 9-16) and a 37 kDa one of lower abundance (fractions 13-17). The amounts of the two proteins were estimated by densitometric analysis relative to the known amounts of BSA run on the same gel and are shown by the lines superimposed on the gel (FIG. 3 B). The densitometric analysis indicated that the ˜27 kDa protein peaked in Fraction 13, the same fraction in which the activity peaked (FIG. 3 A). From the information available, it is possible that there is protective activity associated with both proteins; however, based on the coincidence of the peak of activity and the peak of abundance of the ˜27 kDa protein and the lack of detectable 37 kDa protein in fractions 9-12, it was decided to investigate the ˜27 kDa protein first.

Fractions 9-11 were pooled. An aliquot was evaluated by two-dimensional gel electrophoresis and the material migrated as a single spot (FIG. 4 A). A second aliquot was subjected to partial proteolysis with the endoproteinase LysC and the resulting peptides were separated by HPLC. Eight well-resolved peaks were collected. Four were selected for amino acid sequencing and were designated A-D (FIG. 4 B). Sequence information was obtained on the three peptides designated A, B and C; the fourth apparently was amino-terminally blocked. The sequences obtained for the three peptides revealed 13, 27 (one, X, unidentified) and 12 (one, X, unidentified) amino acids, respectively. Namely: S-Q-Q-S-E-V-T-R-I-L-D-G-K (SEQ. ID. #3), X-I-Q-Y-Q-L-V-D-I-S-Q-D-N-A-L-R-D-E-M-R-A-L-A-G-N-P-K (SEQ. ID. #4), X-T-P-P-Q-I-V-N-G-D-Q-Y (SEQ. ID. #5). None of the sequences overlapped and the three internal sequences (a total of 52 amino acids, 2 not identified) had no significant homology with any sequences in either Swiss-Prot or PIR-Protein banks.

A comparison of the mass of the three peptides obtained by mass spectrometry with that derived from the determined amino acid sequence (determined either by amino acid sequencing, peptides A and B, or from the nucleotide sequence of Clone A, peptide C for which only a partial amino acid sequence was obtained) is shown in Table 1. The close agreement between the values determined in two different ways provides evidence that Clone A encodes a portion of TIP-B 1 . Clone A is also designated as tip-SN and the protein product it encodes, i.e. the recombinant partial TIP-B 1 , is referred to as r-TIP-B 1P .

TABLE 1
Comparison of the Mass of HPLC Purified Internal Peptides A, B and C
Obtained by Mass Spectrometry with that Derived Based on Amino Acid
Sequence (Determined or Nucleotide Translated).
				predicted mass
			predicted mass	based on
	# of sequenced	observed	based on amino	translation of
Peptide	amino acids	mass	acid sequence	clone tip-SN
A	13	1461.2 g	1461.54	1461.5
B	26	3115.0 g	3116.5	3116.5
C	11	3761.0 g	1307.5	3788.57

A third aliquot of the pooled fractions (9-11) was analyzed to determine the protein's amino-terminal sequence and an 18 amino acid long sequence was obtained. The amino-terminal sequence, A-P-Y-T-V-V-Y-F-P-V-R-G-R-X-A-A-L-R (SEQ. ID. #2), was confirmed by independently sequencing the material in the pool of fractions 2+3 from a different batch of TNF-induced cells. The amino-terminal sequence corresponds to the first 18 amino-terminal residues of human fatty-acyl-ethyl-ester synthase III (FAEES III), a 26 kDa (pl 4.9) protein which catalyzes the addition of ethanol to fatty acids, and has been postulated to be an important enzyme in the detoxification of xenobiotics. FAEES III shares 98% homology with the placental acidic glutathione transferase (GST). FAEES III can also catalyze glutathione transfer. An aliquot of pooled fractions 9-11 did not have GST activity when assessed using 2,4 dinitrochlorobenzene, a standard substrate. Thus, despite amino-terminal (18 amino acids) homology to two known enzymes, the protein does not have comparable catalytic activity and three non-overlapping, internal peptide sequences (52 amino acids) had no homology with known proteins. It was concluded, therefore, that this was a unique, previously uncharacterized protein and it was named TNF inhibitory protein B 1 (TIP-B 1 ).

As shown in Table 2, cells pretreated in the standard assay with either Fraction 9-11 or Fraction 13 were protected from TNF-induced lysis in a concentration dependent manner (note that, at a 1:8 dilution, Fraction 13 was able to protect 100% of the cells). In contrast, TIP-B 1 added together with TNF did not protect the cells, regardless of whether or not the TIP-B 1 and TNF had been incubated together for 10 h prior to addition to the cells. Thus, TIP-B 1 does not simply neutralize TNF and for cells to be protected from TNF-induced lysis they must be pretreated with TIP-B 1 . TIP-B 1 was demonstrated to be effective in protecting a number of different cell lines, e.g. MLD, U937 and BG9, from TNF-induced lysis.

TABLE 2
The Effect of Preincubation of TIP-B 1 and TNF on TNF plus CHX
Mediated Lysis of MLD Cells in Comparison to TIP-B 1 Induced
Protection in the Standard Assay.
	% Protection a
	Test Agent	STD Assay	TIP-B 1 + TNF (±10 h@37°)
	Frac 9-11 (1:8)	22	0
	Frac 9-11 (1:16)	9	0
	Frac 13 (1:8)	115	5 (NS) b
	Frac 13 (1:16)	60	7 (NS) b
	For the “STD assay”, TIP-B 1 fractions (see FIG. 3 ) or medium were added to MLD cells. 10 hours later medium, TNF, and CHX were added. 18 hours later the cells were stained with crystal violet. Final concentrations were 500 U/ml TNF and 100 μg/ml CHX.
	For the other two assays, TIP-B 1 was either preincubated with TNF for 10 hours at 37° C., and then both were added together with CHX or all three were added together (cells were preincubated with medium alone). Lysis was determined 18 hours later. The results obtained were identical and only one set is shown. All appropriate controls were run in parallel.
	# “N” = 14 for medium alone, 9 for TNF + CHX alone, and 5 for all other groups.
	a % protection was determined using the Absorbance 570 (Abs) readings from a plate reader as follows: 100 [(Abs. TIP-B 1 + TNF + CHX - Abs. TNF + CHX)/(Abs. medium alone - Abs. TNF + CHX)]
	b Based upon the fact that the S.D.s for the mean (n = 5) values for these two observations were quite large, it is considered that they are not different from background.

TIP-B 1 does not alter the binding of biotinylated TNF to the cells (FIG. 5 ). Flow cytometric analysis indicated that the pretreatment of the cells with TIP-B 1 did not alter their ability to bind TNF. Aliquots of the same TIP-B 1 treated cells were protected from lysis and unlabelled TNF was able to compete with biotin-labelled TNF for binding in a concentration dependent manner (data not shown).

TIP-B 1 was effective in protecting against apoptosis. This was shown by assessing apoptosis by the appearance of a subdiploid peak in propidium iodide labeled DNA studies, FIG. 6 (A-C), and in terminal deoxy transferase reactivity studies, FIG. 7 . In the latter study, the effect of TIP-B 1 on apoptosis was compared to its effect on 51 Cr release, a measure of membrane disruption, and protection from both was seen (Table 3).

TABLE 3
TIP-B1 protects U937 cells against TNF mediated lysis (including apoptosis). Lysis
was determined similarly to that described for the “STD assay” in Table 2, except the
U937 cell targets were prelabeled with 51 Cr, and protection was determined by the
radioactivity released into the supernatant. Apoptosis was measured by flow
cytometric analysis of the specific fluorescein label incorporated following fixation in
formaldehyde and incubation with TdT enzyme and fluorescein labeled-dUMP.
	LYSIS	APOPTOSIS
	% Specific 51 Cr Release	Specific TdT + Events
	(% Protection)	(% Protection)
ASSAY	6 hr	10 hr	6 hr
TNF	16	24	1000
TIP-B 1 (1:8) + TNF	ND a	12 (50)	0 (100)
TIP-B 1 (1:16) + TNF	ND	13 (46)	600 (40)
TNF + CHX	40	53	2700
TIP-B 1 (1:8) + TNF + CHX	16 (60)	ND	1300 (52)
TIP-B 1 (1:16) + TNF + CHX	18 (55)	ND	1800 (33)
a ND not determined

Double stranded phagemid (ZAP®) cDNA libraries were constructed using mRNA from TNF-treated BG9 cells and untreated BG9 (control) cells following an assymetric PCR-based subtraction protocol. Briefly, single strand phagemid recombinants were produced from the two libraries. The control library was amplified (by PCR) in the presence of biotinylated UTP, and the biotinylated PCR-amplified single stranded DNA inserts were hybridized to the single strand phagemid recombinants from TNF-treated cells. The recombinants that did not hybridize were removed and split into two aliquots; one aliquot was used to transform an appropriate strain of E.coli to produce the first cDNA subtraction library, the other aliquot was rehybridized to another batch of biotinylated PCR-amplified product derived from the control library and the recombinants that did not hybridize were separated and used to transform E.coli to produce the second cDNA subtraction library.

Using degenerate primers based on the amino acid sequence of peptides A and B, PCR analysis of the cDNA library from TNF-treated cells led to the generation of a 120 bp fragment (FIG. 8 A). The formation of this fragment suggested that peptides A and B were oriented in tandem (head to tail) as shown in FIG. 8 A. Subsequently, PCR amplification primers (Primers P 18 and P 27 , FIG. 8B ) were selected based on the partial sequence provided by an outside commercial laboratory. PCRs were conducted on the first subtraction library (containing 1000 cDNA clones), the second subtraction library (containing 30 cDNA clones), a cDNA library from TNF-treated cells (containing approx. 10 6 independent clones) and BG9 genomic DNA. A 274 bp fragment was generated from the first subtraction library (a faint band, not shown), the cDNA library from TNF-treated cells (a strong band) and BG9 genomic DNA, but not from the second subtraction library (FIG. 9 A). This 274 bp fragment was subcloned and then sequenced. The sequence revealed that it contained the coding region for the 52 amino acids of the three TIP-B 1 internal peptides (A, B and C) in tandem (FIG. 8 B). The two amino acids not identified by amino acid analysis (page 14) were now deduced based on the nucleotide sequence to be arginine and alanine.

To evaluate the potential for TNF induction of the corresponding message, and to estimate the abundance of this message in untreated cells and TNF-treated cells, the same primers as above (P 18 and P 27 ) were used, and PCR analyses of cDNA libraries from BG9 cells that were either not treated or treated with TNF for 16-18 hrs were carried out. The library from the TNF-treated BG9 cells, as expected, produced the 274 bp band, while the control library yielded a very faint band (FIG. 9 B), suggesting that TNF induces TIP-B 1 message. To confirm this, a northern blot prepared with very limited amounts of mRNAs from TNF-treated and untreated cells was probed with the labeled 274 bp fragment. The results, normalized to G3PDH message, revealed ˜1.6-fold induction. After having established, as detected by the 274 bp probe, the presence of more TIP-B 1 message in TNF-treated cells than in untreated cells, the cDNA library from the TNF-treated cells was screened in an effort to isolate the TIP-B 1 cDNA clone(s).

The 274 bp insert was used to screen 150,000 plasmid-containing bacterial colonies derived from unidirectional phage recombinants that contained cDNA inserts corresponding to mRNAs from TNF-treated BG9 cells. On the third rescreening with increasing stringency, 8 colonies that hybridized with the 274 bp fragment were picked for stock cultures. Six colonies released an ˜0.8 kb fragment upon digestion with EcoR1 (E) and Xho (X) (or E alone in some cases). These 0.8 kb fragments hybridized to the 274 bp fragment on Southern analysis (FIG. 10 ). These 6 colonies were designated as A, B, C, D, K and S and, on the basis of preliminary restriction enzyme analysis carried out on mini-plasmid preparations, were characterized as follows: Clone A released a ˜0.8 kb fragment; Clones B and S each released two fragments of ˜0.8 kb and 0.9 kb; Clone C released a 0.8 kb fragment and a faint 0.3/0.4 kb fragment; Clones D and K released two fragments (0.8 kb and 0.9 kb) and a small (0.3/0.4 kb) faint fragment (FIG. 11 ). These results suggested that the cDNA inserts of these clones, most likely, were of the following sizes: A, ˜0.8 kb; B and S, ˜1.7 kb; D and K, ˜2.0 kb; and C, ˜1.1 kb. Southern blots of these inserts probed with the 274 bp fragment are shown in FIG. 12 . As can be seen, the ˜0.8 kb fragments from all these clones hybridized to the 274 bp fragment. The ˜0.8 kb Clone A insert was sequenced (780 bp), and was found to contain the 274 bp sequence that encodes the 52 amino acids of TIP-B 1 internal peptides (FIG. 13 ). It was concluded, therefore, that Clone A is a partial message for TIP-B 1 . Subsequent evaluation of Clones B, S, C, and D indicated that they were not TIP-B 1 message. FIG. 13 also shows the consensus sequences for three myristylation sites (M 1 , M 2 , and M 3 ), an amidation site, and a casein kinase II site (CKII).

To determine the size of the complete TIP-B 1 message, the insert from Clone A (0.8 kb) was used to probe northern blots of RNAs from control and TNF-treated BG9 cells harvested at various times (up to 24 hrs) following treatment with TNF. It was found that the 0.8 kb insert from Clone A hybridized with three messages, 1.1, 1.7 and 3.0 kb (FIGS. 14 A and 14 B). It strongly hybridized with 1.1 kb (FIG. 14 A), and weakly hybridized to 1.7 and 3.0 kb, as indicated by the longer exposure time required for their visualization (FIG. 14 B). These results suggest that the Clone A insert recognizes three messages, and that possibly the 1.1 message is much more abundant than the 1.7 and 3.0 kb messages or that the insert shares more homology with the 1.1 message than with the other two. As had been suggested earlier with the 274 bp as probe, careful analysis, relative to G3PDH message, indicated that the 1.1 kb message was induced 1.7 fold by TNF after 20-24 h treatment of BG9 cells (Table 4).

TABLE 4
Induction of TIP-B 1 -Specific Message Following Treatment of BG9 Cells with TNF a
	BG9	TNF-Treated BG9	TNF
	(Arbitary Units b)	(Arbitary Units b)	Induction
TIME	A Insert	G3PDH	Ratio (X)	A Insert	G3PDH	Ratio (Y)	Y/X
12 hr.	1807	4636	0.389	2216	4970	0.446	1.15
24 hr.	1941	6078	0.319	3252	5873	0.554	1.73
a BG9 cells were treated with TNF (1000 U/ml medium) for the indicated times. Total RNA was isolated from the cells and used in northern blot analysis. Briefly, the RNA was resolved by formaldehyde gel electrophoresis, transferred to Genescreen plus membranes and probed with 32 P-labeled Clone A insert. The membranes were stripped and reprobed with 32 P-labeled G3PDH probe and the autoradiograms were scanned using a laser densitometric scanner.
b Integrated volume points obtained from laser densitometric scan.

Using the Clone A insert as a probe, northern blots of mRNA from a number of cell types including those shown in FIG. 17 below, were examined. Such examined cell lines include: human embryonic lung fibroblasts HEL, mammary adenocarcinoma MCF7, histocytic lymphoma U937, promyelocytic leukemia HL-60, acute lymphoblastic leukemia REH, lymphoblastic T cell leukemia Molt 4, normal human diploid fibroblasts NHDF, and malignant melanoma A375. As was shown in FIG. 14 , three size messages were seen. Twenty cell types in total have been examined to date, including 2 from mice, 2 from rats and 16 from humans; in every case, a 1.1±0.2 kb mRNA band reacted with the Clone A probe.

The Clone A insert has been placed, in frame, in a pET® vector and expressed in E. coli.. The bacteria have been induced with IPTG, lysed and 10 5 xg supernatant produced. An abundant protein of the expected size for Clone A (˜14 kDa) is seen on SDS-PAGE ( FIG. 15 ) and the 10 5 xg supernatant protects against TNF-induced lysis (FIG. 16 ).

Using purified recombinant partial TIP-B 1 protein, r-TIP-B1 p , and purified TIP-B1, polyclonal antibodies were generated in NZW rabbits. The two antisera generated, one to TIP-B 1 and the other to r-TIP-B 1p were designated #17 and #459, respectively. In testing for the specificity of the antisera, antiserum 459 was found to react with some proteins present in the 10 5 xg supernatant from bacteria containing vector alone (first lane, left panel, FIG. 18 ) but also with a 14 kDa protein found only in the 10 5 xg supernatant of bacteria containing the vector with Clone A (tip-SN) insert (the 2nd lane, left panel) or in the pool of purified r-TIP-B 1p used to immunize the rabbit (3rd lane, left panel). Antiserum 459 absorbed with bacteria containing vector alone no longer reacted with proteins in the 10 5 xg supernatant from bacteria containing vector alone but still reacted with proteins in two other samples (lanes 1 , 2 and 3 , center panel, respectively). Adsorption of the antiserum with both bacteria containing vector alone and purified r-TIP-B 1p (pool 1 ) resulted in the removal of all reactivity, other than that assumed to be non-specific immunoglobulin binding to very low molecular weight material (right panel, FIG. 18 ). The western blots shown in FIG. 19 , developed with pre-immune and immune serum 17 , show that the immune serum reacts specifically with proteins of approximately 18, 32 and 34 kDa in 10 5 xg supernatants from both BG9 and HEL cells. The 32 kDa protein appears to be several fold higher in the supernatant from BG9 cells treated with TNF, suggesting that it may be induced. The activity data, obtained with the Rotofor® fractions from cells with and without TNF treatment, are also consistent with the possibility of induction of the TIP-B 1 protein (FIG. 20 ). Thus, proteins from TNF treated cells found in the Rotofor® fractions in the pH range of 4.5 to 5.0 apparently had considerably more protective activity than the parallel fractions from untreated cells.

NHDF cells had been found to have considerable message for TIP-B 1 ( FIG. 17 ) and are fibroblasts as were the MLD and BG9 cells originally used as targets. Therefore, NHDF were examined for TNF plus cycloheximide sensitivity and TNF-induced protection (FIG. 21 ). Without pretreatment with TNF, NHDF are essentially ablated by an 18 h exposure to TNF plus CHX at the indicated concentrations. Following a 6 h pretreatment with as little as 1 U/ml TNF, however, the majority of the cells are protected. Maximal protection was achieved by pretreatment with 50 U/ml TNF. These findings suggested that NHDF might be both a suitable target cell for bioassay of TIP-B 1 and also a good source of TIP-B 1 . In further investigations, NHDF cells, cultured in the presence of 0 to 10,000 U/ml TNF, were lysed and fractionated by centrifugation. The large cellular particulate fraction was removed following 12×10 3 xg centrifugation. The remaining material was separated at 10 5 xg into supernatant and pellet fractions which were analyzed on parallel western blots utilizing the two antisera (FIG. 22 ). The patterns of reactivity seen with the two antisera are somewhat different. The 459 antiserum detected a 27 kDa protein while the 17 antiserum failed to do so. Conditions have just recently been optimized which allow for routine detection of a 27 kDa protein with the 17 antiserum. Previously, i.e. for all blots shown, the conditions used were not optimal. Antiserum 17 reacted with two proteins, a 52 and an 18 kDa, that 459 did not detect (the preimmune 17 reacts with a 46 kDa protein, see FIG. 28A , below, and therefore it is likely that this is nonspecific). Interestingly, the 18 kDa band was barely detectable in the sample not pretreated with TNF and appeared to be strongly induced. In contrast, there wasn't evidence of induction of the 52 kDa protein, but there was evidence of possible TNF-induced translocation of the protein from the cytosolic to the particulate fraction.

The possibility of translocation of antibody reactive proteins was investigated further in immunocytochemical studies ( FIGS. 23A-D and 24 A-C). For the analysis shown in FIGS. 23A-D , NHDF cells, with and without TNF treatment, were harvested, pelleted, formalin fixed, embedded in paraffin blocks, sectioned, counter stained with hematoxylin (blue nuclear stain), and reacted with either preimmune or antiserum 17 followed with biotinylated goat anti-rabbit antiserum and then strepavidin labeled horseradish peroxidase and substrate. For the immunocytochemical analysis shown in FIGS. 24A-C , NHDF cells grown on coverslips were either treated with TNF or not treated, air dried, frozen (−70° C.) overnight, acetone fixed, and developed as given for FIGS. 23A-D . Little or no positive reactivity was seen with the preimmune, however, considerable reactivity was seen with antiserum 17 in both experiments. In the cells which had not been treated with TNF the antiserum reactive proteins appeared to be diffusely dispersed in the cytoplasm. In contrast, in cells treated with TNF, the antiserum reactive proteins were found to be localized to distinct areas, particularly near the plasma membrane as well as perinuclear.

These findings support a possible translocation step for TIP-B 1 in the induction of protection seen following exposure to low concentrations of TNF.

The MCF7 cell line, human breast adenocarcinoma, was found to be sensitive to TNF but its MDR variant, MCF7/A6, was not (FIGS. 25 A and 25 B). MCF7/A6 cells contain twice as much TIP-B 1 mRNA as MCF7 cells (FIG. 26 ). Based upon these results, western blots assessing 10 5 xg supernatants from nontreated and TNF treated cells were carried out using NHDF as a positive control ( FIGS. 27A and 27B , see FIGS. 28A and 28B for results with preimmune sera). A 27 kDa protein was detected in the MCF7 but not the MCF7/A6 supernatants. A 34 kDa protein was detected in both, but its TNF-inducibility appeared to be different between the two cell lines. Finally, using the acute lymphoblastic leukemia REH and an ˜8-fold resistant variant, REH3.6, developed in our laboratory, it was found that the two antisera detected similar proteins in the two parent/MDR pairs (FIG. 29 ). Antiserum 17 also detected a higher molecular weight protein which was seen in the MDR variants but not in the parental wild type.

7 1 10 PRT HOMO SAPIENS UNSURE 3,6,8 Xaa at position 3 is either E or L. Xaa at position 6 is either P or D. Xaa at position 8 can be any amino acid. Xaa residues at positions 3,6, and 8, in conjunction with the remainder of the sequence is the amino terminus of a protein which inhibits TNF-induced cytotoxicity 1Val Val Xaa Ala Val Xaa Leu Xaa Ala His 1 5 10 2 18 PRT HOMO SAPIENS UNSURE 14 Xaa residues at position 14 can be any amino acid and in conjunction with the remainder of the sequence is the amino terminus of a protein which inhibits TNF-induced cytotoxicity. This sequence is homologous to the amino terminal sequence of two other known proteins GST(pi)and FAEESIII 2Ala Pro Tyr Thr Val Val Tyr Phe Pro Val Arg Gly Arg Xaa Ala Ala 1 5 10 15Leu Arg 3 13 PRT HOMO SAPIENS 3Ser Gln Gln Ser Glu Val Thr Arg Ile Leu Asp Gly Lys 1 5 10 4 27 PRT HOMO SAPIENS UNSURE 1 The Xaa residue at position 1 can be any amino acid and in conjunction with the remainder of the sequence in the peptide is a sequence derived from internal peptides of TIP-B1, a protein which inhibits TNF-induced cyto-toxicity and apoptosis 4Xaa Ile Gln Tyr Gln Leu Val Asp Ile Ser Gln Asp Asn Ala Leu Arg 1 5 10 15Asp Glu Met Arg Ala Leu Ala Gly Asn Pro Lys 20 25 5 12 PRT HOMO SAPIENS UNSURE 1 The Xaa residue at position 1 can be any amino acid and in conjunction with the remainder of the sequence in the peptide is a sequence derived from internal peptides of TIP-B1, a protein which inhibits TNF-induced cyto-toxicity and apoptosis 5Xaa Thr Pro Pro Gln Ile Val Asn Gly Asp Gln Tyr 1 5 10 6 783 DNA HOMO SAPIENS This sequence encodes the TIP-B1 peptides of SEQ. ID. #3, SEQ. ID. #4 and SEQ. ID. #5 6ggcacgagca cggcggcggc gtcgtctccc ggcagtgcag ctgccgctac cgccgccctc 60tgcccgccgg cccgtctgtc tacccccagc atgagcggcc tgcgcgtcta cagcacgtcg 120gtcaccggct cccgcgaaat caagtcccag cagagcgagg tgacccgaat cctggatggg 180aagcgcatcc aataccagct agtggacatc tcccaggaca acgccctgag ggatgagatg 240cgagccttgg caggcaaccc caaggccacc ccaccccaga ttgtcaacgg ggaccagtac 300tgtggggact atgagctctt cgtggaggct gtggaacaaa acacgctgca ggagttcctg 360aagctggctt gagtcaagcc tgtccagagt tcccctgctg gactccatca ccacactccc 420cccagccttc acctggccat gaaggacctt ttgaccaact ccctgtcatt cctaacctaa 480ccttagagtc cctcccccaa tgcaggccac ttctcctccc tcctctctaa atgtagtccc 540ctctcctcca tctaaaggca acattcctta cccattagtc tcagaaattg tcttaagcaa 600cagccccaaa tgctggctgc ccccagccaa gcattggggc cgccatcctg cctggcactg 660gctgatgggc acctctgttg gttccatcag ccagagctcg ccaaaggccc cgcagtccct 720ctcccaggag gaccctagag gcaattaaat gatgtcctgt tccattgaaa aaaaaaaaaa 780aaa 783 7 123 PRT HOMO SAPIENS This is the amino acid sequence predicted from translation of the DNA sequence 4 (=clone A=rtip-b1p cDNA) and which contains the TIP-B1 peptides of SEQ. ID. #3, 4, AND 5 7Gly Thr Ser Thr Ala Ala Ala Ser Ser Pro Gly Ser Ala Ala Ala Ala 1 5 10 15Thr Ala Ala Leu Cys Pro Pro Ala Arg Leu Ser Thr Pro Ser Met Ser 20 25 30Gly Leu Arg Val Tyr Ser Thr Ser Val Thr Gly Ser Arg Glu Ile Lys 35 40 45Ser Gln Gln Ser Glu Val Thr Arg Ile Leu Asp Gly Lys Arg Ile Gln 50 55 60Tyr Gln Leu Val Asp Ile Ser Gln Asp Asn Ala Leu Arg Asp Glu Met 65 70 75 80Arg Ala Leu Ala Gly Asn Pro Lys Ala Thr Pro Pro Gln Ile Val Asn 85 90 95Gly Asp Gln Tyr Cys Gly Asp Tyr Glu Leu Phe Val Glu Ala Val Glu 100 105 110Gln Asn Thr Leu Gln Glu Phe Leu Lys Leu Ala 115 120

Claims

1. A protein designated Tumor Necrosis Factor Inhibitory Protein-B1 TIP-B1) purified to homogeneity, said protein having a molecular weight of about 27 kD and being capable of inhibiting the action of TNF upon a cell when said protein is introduced into extracellular medium surrounding the cell, and said protein being free of sequences which interfere with normal cellular TNF binding sites and being free of sequences which directly bind to TNF, said protein have a terminal amino acid sequence of A-P-Y-T-V-V-Y-F-P-V-R-G-R-X-A-A-L-R.

2. The protein of claim 1 wherein said protein has the 52 amino acid internal peptide sequence SQQSEVTRILDGK (SEQ. ID. #3), XIQYQLVDISQ DNALRDEMRALAGNPK (SEQ. ID. #4), XTPPQIVNGDQY (SEQ. ID. #5).

3. A method for inhibiting the action of TNF upon a cell which comprises exposing the cell to an effective amount of the TIP-B1 protein of claim 2 to inhibit the action of TNF upon a cell.

4. A method for inhibiting the action of TNF upon a cell which comprises exposing the cell to an effective amount of the TIP-B1 protein of claim 1 to inhibit the action of TNF upon a cell.