Protein Engineering, Vol. 15, No. 5, 413-418,
May 2002
© 2002 Oxford University Press
Enhanced proliferative effects of a baculovirus-produced fusion protein of insulin-like growth factor and 
1-proteinase inhibitor and improved anti-elastase activity of the inhibitor with glutamate at position 351
Endocrine Laboratory, McGill University Health Centre, 687 avenue des pins, ouest, Montreal, Canada H3A 1A1
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionReferences |
|---|
1-Proteinase inhibitor (API) was coupled at the C-terminus of a human insulin-like growth factor (IGF) analog to facilitate its production in insect cells. This fusion protein significantly increased thymidine incorporation into HL-60 cells as compared with the incorporation observed with an equivalent molar mixture of the IGF analog and API. The M351E variant of API has been previously shown to reduce aggregate formation in prokaryotic expression systems. When the oxidation-sensitive methionine 351 of the inhibitor was changed to glutamate, the M351E variant was secreted in larger amounts from insect cells than the corresponding fusion protein with wild-type API. The M351E fusion protein and the corresponding chimera containing the wild-type API were tested for their capacity to inhibit human neutrophil elastase. The M351E variant was a more potent elastase inhibitor than the fusion protein containing the wild-type analog, whereas the proliferative activity of both chimeras was identical. The described mitogenic effect of the chimera and the improved anti-elastase activity of the M351E variant are two ideal properties for therapeutic agents acting in pathological situations where cell proliferation and inhibition of neutrophil elastase have to take place simultaneously, such as during wound healing.
Keywords: 1-antitrypsin/infection/neutrophil elastase/serpin polymerization
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
1-Proteinase inhibitor (API), also known as 1-antitrypsin, is an important member of the serpin family of proteinase inhibitors and its deficiency leads to various diseases, such as emphysema, that are associated with high levels of neutrophil elastase (MacDonald et al., 1995; Heiden et al., 1996
; Perlmutter, 1998). Many patients with emphysema are currently treated with a purified API concentrate of human plasma, which will hopefully be replaced with recombinant APIs produced in transgenic animals (Colman, 1999). Another less known property of API or API-like molecules is their cell proliferating activity in tumor-infiltrating T-lymphocytes (Packard et al., 1995a), or in lung fibroblasts and epithelial cell lines, when administered in conjunction with insulin (She et al., 2000). In T-lymphocytes, the mitogenic activity seems to be mediated by a cell surface expressed, elastase-like protein (Packard et al., 1995b). We have previously linked human API at the C-terminus of BOMIGF, a chimera of the insect insulin-like growth factor (IGF) bombyxin and human IGF II. This chimera had both thymidine incorporation stimulating activity into fetal erythroid cells and anti-elastase activity (Curtis et al., 2002). The amounts of recombinant protein produced was low, a problem previously found in other expression systems for API. The first and most common complication for all expression systems is the tendency for serpins to form non-secreted, intracellular polymers, in the endoplasmic reticulum (Mikus and Ny, 1996). Overexpression increases the intracellular concentration and facilitates the formation of inactive aggregates. Schulze et al. (Schulze et al., 1994) tested single amino acid substitutions within the active site loop and found that the M351E variant resulted in a significant aggregate reduction of API synthesized as a cytoplasmic protein in Escherichia coli. To the best of our knowledge this variant has not been tested in eukaryotic expression systems. The second difficulty of recombinant production is the need of proper glycosylation. As reviewed by Luisetti and Travis (Luisetti and Travis, 1996), non-glycosylated recombinant API has a plasma half-life in hours compared to days for the native protein. Furthermore, glycosylation seems to improve the inhibitory action towards several proteases. Therefore, the use of expression systems of higher eukaryotes, such as insect or mammalian cells, represent a significant advantage over prokaryotic cells systems. The third problem of natural and recombinant APIs is the major decrease of the inhibitory activity due to oxidation of methionine 358 at the active site. Therefore, all studies on improving the activity of API have been focused on mutations at position 358 (Luisetti and Travis, 1996). In this report we show that the substitution of methionine 351 to glutamic acid in the fusion protein of IGF and API, had as a consequence, an increased production in insect cells and in a more effective anti-elastase activity as compared to the IGF chimera containing the wild-type protein, without changing the proliferative effects, as measured by thymidine incorporation into HL-60 cells. Regardless of the method of expression, the M351E mutation represents an attractive alternative to natural API for therapeutic use. In particular, the BOMIGF–API M351E mutant described here, with its dual function as elastase inhibitor and growth factor, could be an attractive candidate for local applications in wound healing or tissue regeneration under conditions of infection. |
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
The baculovirus transfer vector 265 (DiFalco and Congote, 1997
) has a multiple cloning site attached with a flexible linker to the C-terminal end of human IGF II. The N-terminal section of this growth factor has been modified by inserting the signal peptide and the first nine amino acids of bombyxin (an insect insulin-like peptide) to facilitate the proper folding and secretion of the recombinant protein in insect cells (Congote and Li, 1994). API was coupled at the C-terminus of a human IGF analog to facilitate its production in insect cells (Curtis et al., 2002). The vector 866 has the following structure, starting from the 5' end: promoter of the baculoviral protein polyhedrin; bombyxin-human IGF II chimera (BOMIGF); flexible polyglycine linker; enterokinase cleavage site; API; factor Xa cleavage site; hexahistidine. This vector was used as the template to obtain the M351E mutant using the QuikChangeTM site-directed mutagenesis kit by Stratagene (La Jolla, CA). Methionine 351 (ATG) was changed to glutamic acid (GAG) using a forward and a reverse primer of 39 nucleotides coding for the API region between amino acids 345 and 358 with the codon GAG in the middle. Briefly, these two primers were used to amplify the DNA template 866 with Pfu DNA polymerase. The original (methylated) 866 template was digested with the endonuclease DpnI. Cells transformed with the amplified DNA were selected for the presence of the API variant and the mutation M351E was confirmed by DNA sequencing (Sheldon Biotechnology Centre, McGill University, Montreal, Canada). The transfer vector was then transfected together with linearized Baculovirus DNA (Baculogold; Pharmingen, Mississauga, Ontario, Canada) and a large preparation of recombinant baculovirus was prepared as previously described. The fusion protein BOMIGF–API and the corresponding M351E variant were isolated from supernatants of Trichoplusia ni infected cells 2 days post-infection using Ni-chelating Sepharose (Amersham-Pharmacia, Baie d'Urfe, Quebec, Canada) and ultrafiltration in Centricon-20 tubes (Millipore, 50 kDa molecular weight limit). Purity was assessed by SDS–PAGE (Phast System; Amersham-Pharmacia) and concentrations were measured by amino acid analysis or by enzyme-linked immunosorbent assay (ELISA), using preparations calibrated by amino acid analysis.
Identification of cell lines with surface elastase was done by a method adapted from established determinations of surface markers for flow cytometry or immuno-cytochemistry (Pollice et al., 1992; Gutensohn et al., 1996; Vrecl et al., 1998). A total of 3x106 cells of the hematopoietic cell lines TF-1, Sup-T1, U-937, TUR and HL-60 (ATCC, Manasas, VI) were washed with PBS and fixed with ice-cold 2% (w/v) paraformaldehyde in PBS for 20 min. Cells were washed with PBS containing 1% (w/v) bovine serum albumin used as a blocking agent. After blocking, the cells were incubated for 1.5 h with a 1:1000 dilution of anti-elastase antibody (ICN, Costa Mesa, CA) in 1% albumin in PBS. Cells were washed three times and incubated for 30 min with donkey anti-sheep IgG conjugated with horseradish peroxidase. After three additional washings with PBS, 1:1 serial dilutions of the cells were prepared and incubated with the peroxidase substrate o-phenylenediamine hydrochloride (Sigma, Mississauga, ON, Canada). The absorbance at 490 nm was measured with a microplate reader.
Thymidine incorporation into HL-60 and TF1 cells was done as previously indicated (Congote et al., 1989; Congote and Li, 1994). Briefly, 10 000 cells were starved overnight in a serum-free medium [RPMI 1640, supplemented with 300 µg/ml bovine serum albumin (fatty acid-free, tissue culture grade; Sigma) and 30 µl/ml bovine transferrin (ICN)] and subsequently incubated with the recombinant proteins for 17 h followed by a 1 h incubation with [3H]thymidine (ICN). Radioactivity incorporated into trichloroacetic acid-insoluble materials was measured as described (Congote et al., 1989). Cell growth was measured by taking advantage of the capacity of viable cells to reduce the soluble, non-toxic, tetrazolium-like indicator Alamar blue, as previously indicated (DiFalco and Congote, 1997). Briefly, the cells were incubated with a 10% (v/v) Alamar blue solution (Biosource, Medicorp, Montreal, Canada) in serum-free medium and the increased absorbance at 570 nm was measured with a microplate reader.
To compare the efficiency of secretion of the wild-type and M351E variant, monolayers of 6x105 T.ni cells (Davis and Wood, 1995) in 12-well plates (Costar) were infected with both recombinant baculoviruses at a multiplicity of infection of 5 and the amount of recombinant API in the supernatants was measured with an ELISA using rabbit antitrypsin antibodies (Behring Diagnostics, Kanata, Ontario, Canada) and horseradish peroxidase-linked anti-rabbit IgG. In some experiments the amount of non-secreted API was evaluated in cell lysates as previously described (Curtis et al., 2002).
The capacity of the recombinant proteins to inhibit human neutrophil elastase (EC 3.4.21.37; Calbiochem, San Diego, CA, USA) was measured by incubating 0.6–10 µg/ml of the inhibitors (9–140 pmol/ml) in 100 µl of 0.1 M Tris–HCl buffer (pH 8.0) with elastase (30–60 pmol/ml) for 10 min at 37°C. Fifty micrograms of elastase substrate N-methoxysuccinyl-Ala-Ala-Pro-Val-p-nitroanilide was added and the increase in absorbance at 405 nm was followed using a microplate reader. In some experiments the volumes were reduced by half, using A/2–96 microplates (Costar; Fisher Scientific, Nepean, ON, Canada). In some of the experiments described here, the recombinant BOMIGF–API inhibitors were previously kept at room temperature for 1 day in a solution of 50% (v/v) Tris buffer (indicated above) and 50% phosphate-buffered saline containing 0.01% (w/v) sodium azide in siliconized polypropylene tubes.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
The recombinant proteins produced in cell cultures of T.ni cells were purified by affinity chromatography and ultrafiltration and analyzed by SDS–PAGE. Figure 1
shows the electrophoretic profile of the proteins used in the present study. The commercial preparation of API was remarkably pure. The fusion proteins of BOMIGF with the wild-type API or with the M351E mutant were also pure, devoid of degradation products. A total of three different preparations with identical profiles of both recombinants were used throughout this investigation.
|
We have previously shown that the chimera of BOMIGF with wild-type API stimulated thymidine incorporation into fetal bovine erythroid cells, a typical target cell for IGFs. This effect was similar to that of BOMIGF alone. API did not have any effect on thymidine incorporation in these cells (Curtis et al., 2002
). The first indication that API alone may have mitogenic activities came from studies of Packard et al. (Packard et al., 1995a,b). They showed that the mitogenic action on tumor-infiltrating lymphocytes was mediated by a cell surface elastase-like protein. These results raised the possibility that BOMIGF–API chimeras may have a synergistic effect on cells containing both IGF receptors and immunoreactive surface elastase. Therefore, we analyzed several hematopoietic cell lines of myeloid and lymphoid origin for the possible presence of immunoreactive elastase. The cells were fixed first to avoid interactions of the anti-elastase antibodies with elastase-like molecules inside of the cell and the relative amounts of elastase present were assessed using horseradish peroxidase-conjugated antibodies (Figure 2). It is evident that among the different cell lines tested, HL-60 was by far the cell line with the highest levels of immunoreactive elastase, which could eventually mediate the mitogenic action of API. We were unable to detect elastase enzymatic activity in these cells (data not shown), suggesting that the enzyme is present in very low amounts, or is in an inactive form. However, API seems to have mitogenic activity in HL-60 cells, as measured with the Alamar blue technique (Figure 3A). The effect observed at the highest concentration tested (2.7 nM) was very modest (9.5% increase over control cell cultures), but was highly significant (p < 0.001). API did not increase Alamar blue reduction in TF-1 cells, which are used as an example of cell lines with very low immunoreactive elastase. The possible synergistic action of BOMIGF–API chimeras was tested using the thymidine incorporation technique (Figure 3B). For these experiments, the action of the chimera was compared with that of an equivalent mixture containing exactly the same molar concentrations of BOMIGF and API. In the case of HL-60 cells, the chimera (closed squares) significantly increase thymidine incorporation as compared with the equimolar mixture of API and BOMIGF (open squares). In the control cell line (TF-1), the equimolar mixture (open circles) seems to increase thymidine incorporation more efficiently than the chimera (closed circles). However, these changes were not statistically significant. As far as thymidine incorporation-stimulating activity is concerned, both the wild-type chimera (closed squares) and the M351E mutant (open diamonds) are equipotent (Figure 3C).
|
|
Figure 4A
shows the amounts of immunoreactive API present in the supernatants of T.ni cell cultures after infection with recombinant baculoviruses for BOMIGF–API (circles) and the M351E variant (squares). To eliminate conditions which may modify the results, such as temperature and number of subcultures, the incubations were done simultaneously for both proteins. We had previously found that the maximal production of the wild-type BOMIGF–API takes place 3 days post-infection at 28°C and a cell density of 2x106 cells. In the experiments of Figure 4A (22–25°C, 6x105 cells) the maximum production took place at day 4. At 2 days post-infection, the amounts of the secreted recombinant M351E were already larger than those in supernatants infected with the baculovirus coding for the wild-type protein, and this difference was significant after 5 days of incubation. Although the amounts of BOMIGF–API reached a maximum at 4 days and started to decline at day 5, there was a continuous production of the M351E mutated fusion protein. The amounts of non-secreted API recombinants was evaluated at days 1, 3 and 5 post-infection (Figure 4B). There was a modest, but significantly higher accumulation of non-secreted API in cells infected with the wild-type baculovirus than in cells infected with the M351E variant 3 days after infection (p < 0.01). At day 5, both the secreted (p < 0.05) and intracellular API (p < 0.01) continued to increase in cells producing the M351E variant, whereas there was a significant decline in the intracellular wild-type chimera (p < 0.05).
|
It is worth noting that the best amino acid substitution found to prevent aggregation by Schulze et al. (Schulze et al., 1994
) involved methionine, because this amino acid is susceptible to oxidation to its sulfoxide derivative and its is well known that oxidation of M358 results in decreased inhibitory activity. Therefore, we compared the inhibitory activity of API using a commercial preparation of neutrophil elastase. Figure 5 shows that the variant was more active than the wild-type recombinant. After 30 min at 70°C, a temperature above the transition state of API (Dong et al., 2000), there was a substantial elimination of the anti-elastase activity in both recombinants. This condition is not likely to occur in vivo, but short-term exposure at temperatures between 20 and 37°C could take place in potential topical applications. Therefore, the two recombinants were left at room temperature for 1 day and then their capacity to inhibit elastase was measured. Figure 6A shows an experiment in which the recombinant M351E (squares) was able to completely inhibit elastase and Figure 6B shows the mean inhibition of four different experiments, each run in duplicate. BOMIGF–API was the best inhibitor of elastase and this inhibition was significantly different from that obtained with the wild-type protein at concentrations of 1.25–10 µg/ml. The regression coefficients of the inhibition caused by both recombinants were also statistically significant.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
The utilization of the signal peptide and the N-terminal sequence of the insect insulin-like peptide bombyxin as the first segment of a chimera with human IGF II resulted in the production of a secreted, properly folded and biologically active IGF in insect cells (Congote and Li, 1994
). If the glycosylated cytokine interleukin-3 is attached to the C-terminal end of the IGF analog, the production of the recombinant chimera was 10–100-fold higher than that previously observed with the IGF analog alone (DiFalco et al., 1997). This result suggested that the same technique could be applied for the synthesis of other therapeutically useful glycoproteins such as API, which are particularly difficult to produce in large amounts with prokaryotic or eukaryotic expression systems. The technique is not only attractive from the point of view of high yields of production, but has the additional advantage of delivering a bifunctional protein having at the same time growth-promoting and anti-elastase activities, two properties which could be very useful at sites of infection and would healing (Galiano et al., 1996). Recently, investigators trying to identify the mitogenically active component of a preparation of placental alkaline phosphatase (She et al., 2000) or secreted from an epithelial carcinoma cell line (Packard et al., 1995a) found that the partial amino acid sequence of the growth-promoting activity corresponded to API, a common component of human serum. These results are interesting, because they show that under certain conditions, API alone or together with growth factors can fulfil at the same time the dual function of growth-promoting and anti-elastase activity. We have previously found that in target cells for both IGFs and interleukin-3, the chimera of IGF–interleukin 3 was far more potent than the single components added together in stimulating cell proliferation (DiFalco and Congote, 1997). Therefore, it was reasonable to assume that in cells where both IGFs and API are mitogenic, the linkage of IGF and API could result in a similar synergistic effect. We tried to identify some potential target cells for API, taking advantage of the observation made by Packard et al. (Packard et al., 1995b), who demonstrated the involvement of elastase-like molecules in the mitogenic response of the serpin in tumor-infiltrating lymphocytes. Among different hematopoietic cell lines tested for the presence of elastase, only the cell line HL-60 had a convincing elastase-like immunoreactivity (Figure 2
). Therefore, we tested the possible mitogenic activity of API in these cells, using the Alamar blue technique (Figure 3A). API is not a very potent mitogen in these cells, but the small increase in reduced Alamar blue was highly significant. Under the same conditions, no effect was observed in TF-1 cells. Figure 3B shows that, in HL-60 cells, the BOMIGF–API chimera significantly increased thymidine incorporation as compared with the equimolar mixture of BOMIGF and API. This result is the second example of synergism caused by the linkage of BOMIGF with a mitogenic compound. This effect was nor observed in TF-1 cells. In these cells, the mixture of BOMIGF and API seems to be more effective than the BOMIGF–API chimera. It is possible that the API moiety of the chimera may have a steric hindrance for the interaction with IGF receptors in cells devoid of surface elastase. We have also observed the same phenomenon in fetal bovine erythroid cells with BOMIGF–API (Curtis et al., 2002). Both the BOMIGF–API and the chimera with the M351E mutant had identical thymidine incorporation-stimulating activities (Figure 3C), suggesting that the capacity to inhibit elastase does not play a role for the mitogenic response. The synergistic effect of the chimera on thymidine incorporation could be an advantage in cells which are targets for both API and IGFs. However, very little is known about the nature of these cells and their importance in pathological conditions such as inflammation and wound healing. From the results of the present study and the mentioned observations made by other investigators on the mitogenic activity of API, it can be concluded that the action of elastase-like molecules on matrix degradation and tissue remodeling and its relationship with API is far more complex than a simple homeostatic effect due to enzyme inactivation. API could be involved in tissue proliferation and indirectly in cell migration, by preventing degradation of matrix proteins such as fibronectin, an important substrate for elastase (Grinell and Zhu, 1994; Curtis et al., 2002). All these cellular processes are particularly important in wound healing.
The successful utilization of recombinant BOMIGF–API chimeras depends on a convenient method of production and in an increase stability of the labile API moiety of the chimera. Although there was a considerable production of secreted BOMIGF–API fusion protein (Figure 4, circles), the synthesis stops abruptly after 3–4 days, depending on the culture conditions. The experiments in Figure 4 were done at 22–25°C. In previous experiments at the optimal temperature of 28°C, the maximal production was reached after 3 days of infection. This is due to the accumulation of an immunologically detectable, but biologically inactive form of the inhibitor inside the cell (Curtis et al., 2002). Experiments on the synthesis of API as a cytoplasmic protein in E.coli by Schulze et al. (Schulze et al., 1994) also showed that the protein accumulated in inclusion bodies is in an inactive form. These authors found that aggregate formation was reduced with a M351E mutant. This seems to be valid as well for the baculovirus expression system 3 days after infection (Figure 4B). Surprisingly, the amounts of the M351E mutant produced inside or outside the cell continue to increase 5 days post-infection, whereas the production of the chimera with wild-type API declines. It is possible that the accumulation of the polymerized wild-type API in the ER is detrimental for further secretion of the recombinant protein. The shape of the production curve of M351E is very similar to that of the IGF–interleukin-3 chimera using T.ni cells (DiFalco et al., 1997). Although it would seem appropriate to purify the protein from cell culture supernatants 5 days post-infection, we preferred to do the BOMIGF–API preparations 2 days post-infection. At this time the production is not optimal, but the amount of contaminants and degradation products is very low, facilitating tremendously the preparation of purified recombinant proteins, as shown in Figure 1. Already at these early post-infection stages, it is possible to obtain approximately twice as much M351E recombinant protein than its wild-type counterpart. Furthermore, the M351E recmobinant was more effective in inhibiting elastase than the chimera with wild-type API (Figure 5). Figure 6 shows that the M351E variant, previously left at room temperature for 1 day, was able to inhibit elastase more efficiently than the wild-type recombinant protein. This property is likely to play a major difference for many potential therapeutic applications. The M351E substitution could be an alternative solution to the widely studied mutation of the active site methionine 358. Luisetti and Travis (Luisetti and Travis, 1996) reviewed several 358 mutants which were stable and could inhibit elastase very effectively. Nevertheless, the authors pointed out that the methionine 358 substitution could be a disadvantage as well, because oxidation in vivo may be one of the natural mechanisms of API inactivation. Therefore, by substituting methionine 351 alone, the natural mechanisms of inactivation may still take place, which could be very important in cases of long-term administration.
In summary, the BOMIGF–API variant M351E described here could eventually find an application for local, short-term use, whereas the introduction of the same mutation of the recombinant protein in other expression systems designed for long-term administration should be explored as a useful alternative to the better known mutations of the active site methionine 358.
|
Notes |
|---|
1 To whom correspondence should be addressed. E-mail: luis.f.congote{at}muhc.mcgill.ca
|
Acknowledgments |
|---|
The authors would like to thank Dr Roger Palfree for his helpful advice with immunological methods. This work was supported by the Canadian Institutes of Health Research and the Bayer Blood Partnership Fund.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Colman,A. (1999) Genet. Anal., 15, 167–173.[Medline]
Congote,L.F. and Li,Q. (1994) Biochem. J., 299, 101–107.
Congote,L.F., Theberge,J., Mazza,L. and Li,Q.A. (1989) J. Tissue Culture Methods, 12, 4752–4758.
Curtis,H., Sandoval,C., Oblin,C., DiFalco,M.R. and Congote,L.F. (2002) J. Biotechnol., 93, 35–44.[CrossRef][Web of Science][Medline]
Davis,T.R. and Wood,H.A. (1995). In Vitro Cell. Dev. Biol., 31, 659–663.
DiFalco,M.R. and Congote,L.F. (1997) Biochem. J., 326, 407–413.
DiFalco,M.R., Bakopanos,E., Patricelli,M., Chan,G. and Congote,L.F. (1997) J. Biotechnol., 56, 49–56.[CrossRef][Web of Science][Medline]
Dong,A., Meyer,J.D., Brown,J.L., Manning,M.C. and Carpenter,J.F. (2000) Arch. Biochem. Biophys., 383, 148–155.[CrossRef][Web of Science][Medline]
Galiano,R.D., Zhao,L.L., Clemmons,D.R, Roth,S.I., Lin,X. and Mustoe,T.A. (1996) J. Clin. Invest., 98, 2462–2468.[Web of Science][Medline]
Grinnell,F. and Zhu,M. (1994) J. Invest. Dermatol., 103, 155–161.[CrossRef][Web of Science][Medline]
Gutensohn,K., Serke,S. and Kretschmer,V. (1996) Infusionsther. Transfusionsmed., 23, 1–23.
Heiden,M., Seitz,R. and Egbring,R. (1996) Semin. Thromb. Hemostasis, 22, 497–501.[Medline]
Luisetti,M. and Travis,J. (1996) Chest, 110, 278S–283S.
MacDonald,J.L. and Johnson,C.E. (1995) Am. J. Health-Syst. Pharm., 52, 481–489.
Mikus,P. and Ny,T. (1996) J. Biol. Chem., 271, 10048–10053.
Packard,B.Z., Lee,S.S., Remold-O'Donnell,E. and Komoriya,A. (1995a) Biochim. Biophys. Acta, 1269, 41–50.[Medline]
Packard,B.Z., Mostowski,H.S. and Komoriya,A. (1995b) Biochim. Biophys. Acta, 1269, 51–56.[Medline]
Perlmutter,D.H. (1998) Semin. Liver Dis., 18, 217–225.[Web of Science][Medline]
Pollice,A.A., McCoy,J.P.,Jr, Shackney,S.E., Smith,C.A., Agarwal,J., Burholt,D.R., Janocko,L.E., Hornicek,F.J., Singh,S.G. and Hartsock,R.J. (1992) Cytometry, 13, 432–444.[CrossRef][Web of Science][Medline]
Schulze,A.J., Degryse,E., Speck,D., Huber,R. and Bischoff,R. (1994) J. Biotechnol., 32, 231–238.[CrossRef][Web of Science][Medline]
She,Q.B., Mukherjee,J.J., Crilly,K.S. and Kiss,Z. (2000) FEBS Lett., 473, 33–36.[CrossRef][Web of Science][Medline]
Vrecl,M., Anderson,L., Hanyaloglu,A., McGregor,A.M., Groarke,A.D., Milligan,G., Taylor,P.L. and Eidne,K.A. (1998) Mol. Endocrinol., 12, 1818–1829.
Weber,E. (1967) Grundriss der biologischen Statistik. VEB Gustav Fischer Verlag, Stuttgart.
Received February 22, 2001; revised February 8, 2002; accepted February 8, 2002.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. R. DiFalco, S. Ali, and L. F. Congote The improved survival of hematopoietic cells cultured with a fusion protein of insulin-like growth factor II (IGF-II) and interleukin 3 (IL-3) is associated with increases in Bcl-xL and phosphatidylinositol-3 kinase activity J. Leukoc. Biol., February 1, 2003; 73(2): 297 - 305. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||