WO1989010403A1

WO1989010403A1 - Human gm-csf variants

Info

Publication number: WO1989010403A1
Application number: PCT/AU1989/000177
Authority: WO
Inventors: Mathew Alexander Vadas; Angel Francisco Lopez; Mary Frances Shannon
Original assignee: Medvet Science Pty. Ltd.
Priority date: 1988-04-21
Filing date: 1989-04-21
Publication date: 1989-11-02
Also published as: AU615550B2; JPH03503897A; EP0413721A1; EP0413721A4; AU3549489A

Abstract

A human granulocyte-macrophage colony-stimulating factor (GM-CSF) variant or mutant, characterised in that amino acid 20 (gln) and/or amino acid 21 (glu) of human GM-CSF is/are replaced by another amino acid, and alternatively or additionally, a substantial number of the amino acids of human GM-CSF are deleted.

Description

HUMAN GM-CSF VARIANTS

This invention relates to the production of variants or mutants of human granulocyte-macrophage colony-stimulating factor which have new and useful properties. Human granulocyte-macrophage colony-stimulating factor (h GM-CSF) is a glycoprotein of 19,O00MW produced by a number of normal and transformed cells. The gene coding for this molecule has been cloned and the cDNA used to transfect eukaryotic and prokaryotic cells to produce recombinant h GM-CSF (rh GM-CSF) (1).

The range of actions of GM-CSF extends over the whole lineage of neutrophils, eosinophils and monocytes. Specifically GM-CSF stimulates the progenitors of these cells to proliferate and differentiate to become mature cells (2). In addition it stimulates mature cells to greater function. The stimulation of mature cells results in greater capacity to phagocytose and kill micro-organisms, kill antibody-coated tumor cells and generate superoxide anions (0₂ ^~) in response to various stimuli (3). The purpose of this activation is presumed to enable the mature cells become better effector cells in inflammatory reactions. Therapeutically, the main indications for GM-CSF are for its effects on progenitor cells or mature cells. Using its effects on progenitor cells, GM-CSF is used in the treatment of bone marrow failure as seen in aplastic anaemia on chemotherapy or AIDS-induced marrow suppression. Although found to be an excellent therapeutic agent, some toxicity is associated and is mainly due to stimulation of mature cells causing blood vessel damage or thrombosis. The eosinophilia caused by GM-CSF appears especially damaging in this regard.

In the treatment of infections, the capacity to stimulate mature cells is especially relevant, and local as well as systemic application of GM-CSF can be envisioned. The capacity of GM-CSF-activated neutrophils and eosinophils -to kill tumor cells that have bound antibody is especially remarkable and could be used in tumor therapy.

It should be noted that certain tumors have receptors for GM-CSF and GM-CSF stimulates their growth (4,5). The administration of GM-CSF in these cases might stimulate tumor growth. Thus, mutant molecules with selectivity for either mature cells or progenitor cells would be useful and have their specific clinical indications. In addition, some mutants may be able to stimulate differentiation without growth, and may also be therapeutically useful in malignancies.

Based on these considerations, site-directed mutagenesis reactions have been performed on the GM-CSF molecule to generate variants or mutants with desired therapeutic properties. This work took into account previous observations of differences between mature and immature myeloid cells in their interaction with h GM-CSF, namely differences in dose-response, signalling pathways and binding characteristics (6) . With that approach, lack of amino acids 1-24 resulted in the complete loss of GM-CSF activities and absence of amino acids 1-18 greatly decreased the effect of GM-CSF with no dissociation of activities. The site-directed mutagenesis approach allowed the manufacture of homogenous GM-CSF analogs and a more precise study of the 14-24 region. This allowed establishment of amino acids 20 and 21 as critical for GM-CSF function. Accordingly, in a first aspect of the present invention, there is provided a human GM-CSF variant or mutant, characterised in that amino acid 20 (gin) and/or amino acid 21 (glu) of human GM-CSF is/are replaced by another amino acid. Preferably, the substitute amino acid for one or both of amino acids 20 and 21 is alanine (ala). Accordingly, the present invention includes the following preferred mutant hGM-CSFs: h GM-CSF-ala²⁰ h GM-CSF-ala²¹ h GM-CSF-ala²⁰'²¹ In work leading to the present invention it has been established that:

1. amino acids at position 20 and 21 are essential for h GM-CSF activity (their deletion results in complete loss of biological activity); and

2. substitutions of amino acids at positions 20 and 21 lead to alterations in the potency of the h GM-CSF molecule and to dissociation between proliferation/differentiation of progenitor cells and functional activation of mature cells. In particular, proli eration and differentiation of myeloid cells is greatly reduced while maximal stimulation of mature cells is still achieved. Accordingly, in a further aspect of the present invention, there is provided a pharmaceutical composition comprising a human GM-CSF variant or mutant as described above. The use of such a composition may avoid some of the side effects of GM-CSF or provide new properties including the ability to block leukemic growth.

As demonstrated with GM-CSF-ala²-, substitution of amino acid gin at this position may replace GM-CSF as a more potent therapeutic molecule. Clearly higher doses of any therapeutic agent are not desirable especially when the subcutaneous route appears to be the preferred way of administering this agent.

As demonstrated with GM-CSF-ala²¹ and GM-CSF-ala²⁰'²¹, substitutions of glu at position 21 and both gin at 20 and glu at 21 will selectively result in stimulation of mature cells. This may be partly due to the different affinity of GM-CSF receptors on mature cells and is reflected in the intermediate capacity of these mutants to inhibit binding of rh GM-CSF. This type of molecule will have special use in all cases when patients bear tumors whose stimulation should be avoided, and also in cases where control of the infectious process is of primary importance. Further substitution of amino acids 20 or 21 may alter the molecule to block function or to change the spectrum of action of the molecule. Blocking mutants will have use in malignancies that depend on GM-CSF for growth, and may be excellent agents to be used as adjuncts in chemotherapy.

In addition to the above "substitution" variants or mutants, it has also been established that a mutant h GM-CSF with deletion of certain amino acids stimulated submaximal number of colonies after 14 days when tested for stimulation of human bone marrow cells in a proliferation/differentiation assay but that maximal stimulation of granulocyte function could be obtained with this deletion mutant as measured in ADCC and superoxide anion release assays.

Accordingly, in accordance with another aspect of the present invention there is provided a human GM-CSF variant or mutant, characterised in that a substantial number of the amino acids of human GM-CSF have been deleted with retention of maximal stimulation of mature cell function. Since amino acids 20 and 21, or replacement mutants thereof, have been-established as critical for GM-CSF function, it is preferred that the deletions in accordance with this invention do not include amino acids 20 or 21. In one embodiment of the invention described in greater detail below, the deletion of amino acids 14-18 of h GM-CSF has been found to be possible with retention of maximal stimulation of mature myeloid function.

Accordingly, in a further aspect of the present invention, there is also provided a pharmaceutical composition comprising a human GM-CSF "deletion" variant or mutant as described above.

Further details of the present invention, including the production of the human GM-CSF variants or mutants by site-directed mutagenesis and expression in COS cells or Escherichia coli. are illustrated by way of example in the following detailed description. In addition, the following description shows that a chemically synthesized GM-CSF peptide containing ala²- has similar properties to the mutant containing this substitution made by site- directed mutagenesis.

EXAMPLE 1

MATERIALS AND METHODS Oliσonucleotide directed mutagenesis Synthetic mutagenic primers were synthesized on an Applied Biosystems 381A DNA synthesizer, acrylamide gel-purified and 5' phosphorylated using T4 polynucleotide kinase. The DNA template for mutagenesis was mpl9GM, a bacteriophage carrying the GM-CSF cDNA. Oligonucleotide- primed site-directed mutagenesis was performed as described (7) . Briefly, the kinased mutagenic primer and M13 universal sequencing primer were hybridized to the single-stranded DNA template. The product was then extended and ligated, using the Klenow fragment of DNA polymerase and T4 ligase, for 16-20hrs at room temperature. Part of the resulting mixture was used to transform competent JM101 cells. The mutagenic primer was then 5* ³²P-labelled and used as a probe to identify plaques with the desired mutation. Positive clones were analysed by DNA sequencing using the chain termination method (8) to confirm the presence of the mutation.

The mutated GM-CSF cDNA sequences were cloned into a ρJL4 expression vector and used for transfection of COS cells by electroporation. After transient expression, the COS cell supernatants containing mutant GM-CSF protein were analysed for GM-CSF activity.

Certain of the mutated GM-CSF sequences were cloned into a pAL181 vector and transformed into E.coli GI586. Mutant GM-CSF was expressed in these clones following temperature induction and released from the cells by disruption. Crude cell lysates were analysed for GM-CSF activity.

A GM-CSF peptide containing ala²¹ was chemically synthesized by Dr. Ian Clark-Lewis (Biomedical Research Centre, University of British Columbia, Vancouver, Canada) on behalf of the inventors, using automated stepwise solid-phase methods as described (9). Clonal Proliferation assays:

Bone marrow cultures.

Myeloid colony-forming unit assay.

Light density non-adherent bone marrow cells were obtained by separation on a Ficoll-Paque (Pharmacia,

Sweden) density gradient followed by monocyte depletion using carbonyl iron. Cells were cultured in 0.3% agar (Difco, U.S.A.) with Iscove's Modified Dulbecco's Medium (Commonwealth Serum Laboratories, Australia), 30% fetal calf serum (FCS) (Gibco, U.S.A.), 0.66% bovine serum albumin (Fraction V, Sigma, U.S.A.), and mercaptoethanol 2 x 10^~~M at a concentration of 5 x 10⁴ cells/1 ml culture and rh GM-CSF (lOOng/ml) or appropriate mutant GM-CSF proteins. Aggregates of more than 40 cells were scored as colonies after 14 days incubation. The agar discs were then fixed with 3% glutaraldehyde and transferred onto individual 5 x 8cm glass slides. The discs were dried at room temperature and stained with luxol fast blue and a combined specific and non-specific esterase stain.

Mature cell function assays:

Purification of human neutrophils. eosinophils and monocytes. Peripheral blood of healthy volunteers was centrifuged on a hypertonic gradient of Metrizamide (Nyegaard, A/C, Oslo) as previously described (10) after dextran sedimentation. The purity was greater than 95% for neutrophils and greater than 92% for eosinophils. The cells were resuspended in Eagle's Minimal Essential Medium supplemented with 10% FCS, 20mM HEPES buffer, and antibiotics.

Monocytes were obtained from the peripheral blood of normal donors after density gradient centrifugation on lymphoprep (Nyegaard Oslo, Norway). The monocytes were further purified by countercurrent elutriation in a Beckman JE-6B elutriator using a Sanderson chamber as described (11). Cytocentrifuge preparations were stained with Giemsa and only preparations judged to be 90% monocytes were used. Antibodv-deoendent cell-mediated cvtotoxicity assay (ADCC) .

40μl of ⁵¹Cr-labelled, trinitrophenyl (TNP)-coupled P815 cells (4 x 10³) were incubated with 24μl of rabbit IgG anti-TNP (Miles-Yeda, Rehovot, Israel), 80μl of purified human neutrophils or eosinophils (1.3 x 10⁵) as effector cells, and 16μl of rh GM-CSF for 2.5h at 37°C in V-bottom microtitre plates. Percent cytotoxicity was calculated as described previously (12). Superoxide production.

Purified neutrophils or eosinophils were incubated with medium or with gIL-3, rh GM-CSF or medium for various times at 37°C. 150μl of cells were then added to a mixture of lOOμl freshly prepared cytochrome C (Sigma, type VI, 12.4 mg/ml), lOOμl of FMLP (Sigma) and made up to lml with medium. The mixtures were incubated at 37°C for 5 min, after which the cells were rapidly cooled, pelleted at 4°C and the supernatants transferred to plastic disposable cuvettes. Superoxide production was measured in duplicate by the reduction of cytochrome C as described (3). In each experiment superoxide dismutase (Sigma, St.Louis, MO) inhibited all 0₂~ generation. Measurement of monocvte adherence.

Monocyte adherence to plastic was measured by an isotopic method. After labelling with ⁵¹Cr, 3 x 10⁵ cells were plated into microtitre wells either with or without the addition of stimulators in a total volume of lOOμl. The cells were incubated at 37°C 5% C0₂ for 16h. Following incubation in the microtitre wells, aliquots of supernatants were counted to provide a measure of spontaneous ~~Cr release. Remaining supernatants were then removed by aspiration and the cells washed three times to remove non-attached monocytes. The cells were lysed by the addition of lysis buffer containing lOmM tris-HCl, 0.15M NaCl and 1% NP40. The contents of each well was then counted in a Packard gamma counter. The percent adherence was determined: % Adherence - ~~Cr cpm in lysate/total cell associated --Or cpm added x 100 where total cell associated ⁵¹Cr cpm is calculated as the total -~Cr cpm added minus the -~Cr cpm spontaneously released.

Binding of -^-^I-rh GM-CSF to neutrophils. rh GM-CSF was radioiodinated by the two-phase method with no detectable change in biologic activity. ~~~I-rh GM-CSF specifically bound to human neutrophils with a single affinity of sites with apparent KD 7+1 x 10^~~~M and 140+50 molecules bound per cell (mean ± SEM, m = 6) . For assessment of binding by mutant GM-CSF proteins, competition experiments were performed with lOOpg -²-I-rh GM-CSF and 4 x 10- neutrophils in 0.15ml incubated at 10°C for 3hr. Cell bound radiolabel was determined by centrifuging through fetal calf serum.

RESULTS

Clonal proliferation and differentiation of myeloid cells.

The stimulation of human bone marrow cells with rh GM-CSF resulted in the formation of colonies after 14 days of culture. The full length GM-CSF (1-127) was as active as the recombinant molecule, and the mutants with deletions of amino acids 1-24, 7-24, 14-24 and 20-21 were inactive (Table I). The mutant gin 20 ala stimulated maximal number of colonies similar to the full length GM-CSF. From dose response curves this mutant appeared to be more powerful than the parent molecule, but at this stage activity per mg of protein is not known. The other mutants stimulated submaximal number of colonies in the order: deletion 14-18; substitution glu 21 to ala; substitution gin 20 and glu 21 to ala, ala. Stimulation of human granulocyte function.

Granulocyte function was stimulated by the recombinant h GM-CSF as measured in the ADCC and superoxide anion release (Table I) assays. Similarly to the proliferation experiments, the full length (1-127) GM-CSF was active while the mutants with deletion of amino acids 1-24, 7-24, 14-24 and 20-21 were inactive. In startling contrast, however, maximal stimulation could be obtained with the three mutants, deletion 14-18, substitution glu 21 to ala, and substitution gin 20, glu 21 to ala, ala, all of which stimulated submaximal number of colonies in the proliferation/differentiation assay. The mutant substitution gin 20 to ala stimulated maximal activity, like the full length 1-127 GM-CSF.

TABLE I - Properties of GM-CSF mutants

GM-CSF mediated function (% maximum-) Stimulus Day 14 colony ADCC^b FMLP stimulated Displace- formation 0₂ ^~ production⁰ ment of rh GM-CSF binding"

Nil- 0 O O O rh GM-CSF 100 (100)9 100 100

1-127 99±6^e 100 70±45^h 61±3

Δ 1-24 6±6 9±9^h 0 16±2 Δ 7-24 0 0 23±23 14±1

Δ 14-18 32±13 80±9 88±2 27±7

Δ 14-24 6±1 0 14±14 ND gin²⁰ ala 101±32 114*1 80±10 62±3 glu²¹ ala 15±8 86±18 67±12 29±9 gln²⁰glu²¹ ala, ala 4±2 80±1 63±19 15±3

Δ 20,21 O±O 0 7±7 13±1

COS cell supernatant- 2±2 0 0

Note: Δ indicates deletion.

(a) In order to compare experiments, results are normalised to plateau level of rh GM-CSF.

(b) Antibody-dependent cell-mediated cytotoxicity.

(c) 30 min preincubation at 37°C before stimulation with 10^"7M f-met-leu-phe (FMLP) .

(d) Competition for binding to neutrophils between a 1/2 dilution of supernatant and lOOpg ¹²-I-rh GM-CSF.

(e) Arithmetic mean ± SEM of three experiments performed on different bone marrow donors.

(f) Supernatant from untransfected COS cells.

(g) Untransfected COS cells contained some capacity to activate in this assay. Comparisons are therefore between 1-127 (maximum) and COS cell supernatant (minimum) . (h) Arithmetic mean of two experiments each performed in triplicate.

EXAMPLE 2

MATERIALS AND METHODS

As in Example 1 above, RESULTS

Table II shows the results of tests performed with GM-CSF (1-127) and the mutant with substitution of glu²¹ for ala produced by different methods. It will be seen that the mutant ala²¹ produced by either of the three methods described here does not maximally stimulate day 14 colony formation although it can maximally stimulate neutrophil and monocyte function.

TABLE II - Properties of GM-CSF ala^-l produced by different methods

GM-CSF mediated function (% maximum-B->

FMLP- Displace- Monocyte

Stimulus Day 14 ADCC° stimulated ment of rh adherenc

0₂~ pro¬ GM-CSF duction⁰ binding¹-

Nil 0 0 0 O 0 rh GM-CSF 100 (100)9 100 100 100

Site-directed mutagenesis and expression in:

COS cells 1-127 99:t6^f 100 70±45^h 61±3 100 ala 21

EtCOli 1-127 138+40 ND 76±12 ND ND ala²- 23±7 ND 71+10 ND ND

Chemical 1-127 163±48 94±5 ND 100 100 synthesis ala²¹ 28±16 91±4 ND 74 80±9 (a) In order to compare experiments, results are normalised to plateau level of rh GM-CSF.

(b) Antibody-dependent cell-mediated cytotoxicity.

(c) 30 min preincubation at 37°C before stimulation with 10^_7M f-met-leu-phe (FMLP) .

(d) Competition for binding lOOpg ¹²⁵I-rh GM-CSF to neutrophils.

(e) Purified human monocytes were incubated for 16h at 37°C in microtitre plates and the percent adherent monocytes calculated after washing three times.

(f) Arithmetic mean SEM of three experiments performed on different bone marrow donors.

(g) Untransfected COS cells contained some capacity to activate in this assay. Comparisons are therefore between 1-127 (maximum) and COS cell supernatant (minimum) .

In addition a substitution of glu²- for arg increased monocyte adherence by 42.4% indicating that the replacement of glu at 21 not only by ala but also by another residue alters the activity of the GM-CSF molecule.

REFERENCES:

1. Wong, G.G., J.S.Witek, P.A.Temple, K.M.Wilkens, A.C.Leary, D.P.Luxenberg, S.S.Jones, E.L.Brown, R.M.Kay, E.C.Orr, C.Shoemaker, D.W.Golde, R.J.Kaufman, R.M.Hewick, E.A.Wang, and S.C.Clark. (1985) Science (Wash.DC) . 223.:810-815.

2. Metcalf, D. , C.G.Begley, G.R.Johnson, N.A.Nicola, M.A.Vadas, A.F.Lopez, D.J.Williamson, G.G.Wong, S.C.Clark, and E.A.Wang. (1986). Blood. £2:37-45.

3. Lopez, A.F., D.J.Williamson, J.R.Gamble, C.G.Begley, J.M.Harlan, S.J.Klebanoff, A.Waltersdorph, G.Wong, S.C.Clark, and M.A.Vadas. (1986). J-.C1in.Invest. 1__:1220-1228.

4. Dedhard, S., Gaboury, L., Galloway, P. and Eaves, C. Proc.Na l.Acad.Sci.USA 15.:9253-9257, (1988).

5. Berdel, W.E., Danhauser-Riedl, S., Steinhauser, G. and Winton, E.F. Blood 73:80-83. (1989).

6. Lopez, A.F., S.J.Hardy, J.Eglinton, J.Gamble, L.B.To, P.Dyson, G.Wong, S.Clark, A.W.Murray, and M.A.Vadas. (1988). In Progress in Leukocyte Biology (in press) .

7. Zoller, M.J. and M.Smith. (1983). Methods Enzvmol. • Ifiϋ:468-500.

8. Sanger, F., S.Nicklen and A.R.Coulson. (1977). Proc.Natl.Acad.Sci.USA.74:5463-5467. 9. Clark-Lewis, I., Lopez, A.F., To, L.B., Vadas, M.A., Schrader, J.W., Hood, L.E. and Kent, S.B.H.

J.Immunol. 111:881-889, (1988).

10. Vadas, M.A., J.R.David, A.E.Butterworth, N.T. Pisani, and T.A.Siongok. (1979). J.Immunol. 122: 1228-1236.

11. Sanderson, R.J., Shepperdson, F.T., Vatter, A.E. and Talmage, D.W. J.Immunol. 111:1409-1414, (1977).

12. Vadas, M.A. , N.A.Nicola, and D.Metcalf. (1983). J.Immunol. 11^:795-799.

Claims

QL__l___:

1. A human granulocyte-macrophage colony- stimulating factor (GM-CSF) variant or mutant, characterised in that amino acid 20 (gin) and/or amino acid 21 (glu) of human GM-CSF is/are replaced by another amino acid, and alternatively or additionally, a substantial number of the amino acids of human GM-CSF are deleted.

2. A human GM-CSF variant or mutant according to claim 1, wherein the replacement amino acid for one or both of amino acids 20 and 21 of human GM-CSF is alanine (ala.

3. A human GM-CSF variant or mutant according to claim 1 or claim 2, wherein amino acids other than amino acid 20 (gin) and amino acid 21 (glu), or replacements thereof, are deleted.

4. Human GM-CSF-ala²⁰.

5. Human GM-CSF-ala²¹.

6. Human GM-CSF-ala²⁰'²¹.

7. A human GM-CSF variant or mutant according to claim 3, wherein amino acids 14-18 are deleted.

8. - A pharmaceutical composition comprising a human GM-CSF variant or mutant according to any of claims 1 to 7, in association with one or more pharmaceutically acceptable carriers or diluents.

9. Use of a human GM-CSF variant or mutant according to any of claims 1 to 7, in therapy.

10. Use of a human GM-CSF variant or mutant according to any of claims 1 to 7, in preparation of a pharmaceutical composition for use in therapy.