EP1012321A2 - Viral infection of cells using viral vectors - Google Patents

Abstract

The invention relates to an in vitro method for infecting one or more cells with a viral vector capable of transporting exogenous or recombinant nucleic acid into the cells to be infected, comprising the steps of: a) collecting the cells to be infected; b) treating the collected cells with at least one cytokine; and c) infecting the collected cells with a viral vector such as adenovirus. A method of infecting rheumatoid synovial cells and other cells comprising high levels of cell surface integrins is also disclosed.

Description

VTRAL INFECTION OF CFT.T.S

This current invention relates to an improved in vitro method for infecting cells with a viral

vector capable of transporting recombinant nucleic acid to those cells. The invention

produces an improved level of viral infectivity. The invention .also relates to cells infected

by such a method.

The use of viruses such as human adenovirus or retroviruses such as Human Immunodeficiency Virus (HIV) derivatives as vectors to deliver exogenous or recombinant nucleic acid, such as foreign genes, into cells is well known. Replication-defective forms of human adenovirus have been used to deliver foreign genes into diverse cells types including liver, muscle, nerve .and airway epithelial cells . The success of adenoviruses as a gene delivery vector is based in part on their highly efficient mode of cell entry and their lack of requirement for host cell replication. A major difficulty in using viral vectors to transport recombinant nucleic acid into cells has been the difficulty in achieving high levels of infectivity in some cells. Where the level of infectivity and hence gene transfer is low, any effect of the transferred gene on the cells may be difficult to observe as a result of the background of non-infected cells. A high level of gene transfer would allow the study of, for example, intracellular mechanisms , without any requirement to remove a non-infected population of cells. This would

allow the regulation of, for example, endogenous chromosomal genes to be studied and precludes the need to use ectopically expressed reporter gene constructs which may be regulated differently. Alternatively, high levels of infection would obviate the need for cloning of infected cells.

When adenovirus is used to infect monocytes, for example, the maximum level of infection achieved has been ap_Oroximately 50% which required a multiplicity of infection (M.O.I. ), as plaque forming units per cell, of up to 1000 ₍Huang, 1995 J. Virol. 69: 2257-63. Haddada^, 1993 Bioche . Biophys . Res. Commun. 195:1174-83). Typically monocytes are prepared by isolating mononuclear cells from healthy adult donors using density gradien^ts ^, washing the mononuclear cell fraction by low-speed centrifugation and resuspending the mononuclear ce_ls obtained in growth medium. The ceils are then allowed to

adhere to plastic tissue-culture flasks. The adherent cell population comprises typically 90% monocytes. The

monocytes obtained may then be treated with a cyto ine, such as a granuiocyte-macrophage colony stimulating factor (GM-CΞF) or acrophaσε colony stimulating factor ⁽M-CSF) , prior to infection with the viral vector. These

cytokines increase the expression of integrins (άVβ3 and άVβ5) on the cell surface and are required for viral entry into cells.

The inventors have identified an improved method of infecting ceils with viral vectors which produces infectivity races of approximately 100% with considerably lower concentrations of viral vector. This removes the necessity to remove non-infected ceils and means that less recombinant viral-vector need be used. Consequently the regulation of cellular pathways can be studied.

The inventors realised that inflammatory diseases such as rheumatoid arthritis,

respiratory diseases and inflammatory diseases of the bowel are manifest as a result of the activity of activated monocytes which are likely to express high levels of integrins,

άVβ3 and aVβ5. Such cells have not previously been infected with viral vectors. The

inventors have demonstrated that such cells may be successfully infected with adenovirus. This allows biochemical mechanisms within cells to be selectively studied and new targets for therapeutic intervention to treat the disease to be identified. This also means that such diseases may be treated by gene therapy. Adenovirus is

known to be used in gene therapy but this is the first time that the use of adenovirus to

trea_t inflammatory disease has been identified.

Accordingly, the invention provides an in vitro method

for infecting one or more ceils with viral vector caτ.able of transporting exogenous or recombinant nucleic acid into the ceils to be infected, comprising the steps

of:

a) collecting the cells to be infected; b ₎ treat.ing the collected ceils with at least one

■cytokine: .and c) infecting the col lected cel l s with a viral vector.

Preferably the cells are subjected to elutriation prior to collection.

Preferably centrifugal elutriation is used. This combines centrifuging a sample to sediment out particles with elutriation, the process of separation by washing. Typically a suspension of cells is pumped into a funnel shaped chamber at a preset flow rate. As fluid travels through the chamber its velocity decreases as the chamber gets wider thus creating a velocity gradient from the narrow end of the chamber to its widest part. Cells migrate to positions in the velocity gradient where the effects of both the centrifugal force field and the fluid velocity are balanced. Smaller cells are at equilibrium at the elutriation boundary where the centrifugal force field and the velocity are low. Larger cells will remain near the inlet to the chamber where the centrifugal force field and the velocity are high. By gradually increasing the flow rate, cells can be washed out according co size.

Cytokines have been found to increase the concentration of integrins, such as άVβ3 or άVβ5,

on the cell surface. The inventors realised that rheumatoid synovial cells have such integrins

in high levels on their cell surface. They have therefore found that viral vectors may,

unexpectedly, be infected into rheumatoid synovial cells in high concentrations. This allows

the mode of action of pathways within such cells to be modified by the insertion of

exogenous nucleic acid into the cells.

Accordingly, a further aspect of the invention provides an in vitro method for infecting one

or more rheumatoid synovial cells with a viral vector capable of transporting exogenous or

recombinant nucleic acid into the cells to be infected comprising the steps of:

a) collecting the rheumatoid synovial cells; and

b) infecting the collected cells with a viral vector.

Other cells obtainable from, for example, brocheoalveolar lavage from inflammatory

diseases, such as asthma and chronic obstructive pulmonary disease (C.O.P.D.), and gut biopsies of inflammatory bowel diseases such as Crohn's disease and ulcerative colitis, may

also be infected by the method of the invention.

Preferably the viral vector is an adenovirus, especially a replication-deficient adenovirus.

The M.O.I._/ expressed as plaque forming units per cell, may be 10-1000, preferably 35-100, especially 50-100. Preferably the cells to be infected are monocytes. Such cells may be isolated from, for example, single donor plateletphoresis blood residues . The cells may be partially purified by density centrifugation to provide the suspension of cells prior to elutriation.

The isolated cells may be cultured in a suitable growth medium, but are preferably treated with cytokine substantially immediately after collection from the elutriator with cytokine.

Cytokine is preferably added to the medium in which the cells axe cultured and may be incubated with the cells for between 24 and 96 hours, typically 65-75 hours, preferably about 72 hours. The cytokine may be M-CSF. The concentration of cytokine used is preferably 100ng-lμg/ml-

Preferably the exogenous or recomhiπant nucleic acid within the viral vector is DNA. The recombinant nucleic acid preferably comprises a sequence encoding a gene product of interest operably linked to suitable regulatory sequences to enable the gene to be expressed within the infected cell.

Preferably the foreign gene encodes a modulator for one or more biochemical pathways within the cell to allow the cell pathways to be studied. -A example of a modulator is the I_/cB protein, an endogenous regulator, especially IκBα, an inhibitor of NF^KB. Alternatively a mutated enzyme, such as one in which catalytic activity has been destroyed,

(a dominant negative construct) may be used.

Alternatively, a single chain antibody or antibody fragment, or antisense DNA may be

encoded by the foreign gene.

The invention also relates to cells infected by virus by means of the methods of the invention.

The invention will now be described by way of example only with reference to the following figures:

Figure 1 - Differential induction of integrins αV£5 on

monocytes by GM-CSF and M-CSF. Monocytes were untreated or treated with GM-CSF (500U/ml) , or M-CSF (lμg/ml) or both cytokines for 48 hours. The cells were then

analysedby FACS forthe expression of a.Vβ2> (broken line) or aVβ5 (dotted

line) , staining with control Ig (solid line) is also shown.

Figure 2 - The effect of cytokine treatment on adenoviral infection of primary human monocytes . Monocytes were treated with GM-CSF (500U/ml) or M-CSF ( lμg/ml⁾ for 72 hours followed by infection with 0-galactosidase (shown by dotted line) containing adenovirus or control virus (solid line) at 100 m.o.i. (a) for various m.o.i. (b) . After 72 hours cells were analysed by FACS for expression of β-galactosidase.

Figure 3- M-CSF treatment is not required for infection of R_AW 264.7 cells. RAW 264.7 cells were incubated with or without M-CSF (lug/ml) for 48 hours followed by infection with control adenovirus or a vector containing β-galactosidase. Following culture for 4 days the cells were assayed for β-galactosidase expression by FACS. Key: control virus, no M-CSF (thick line) , control virus + M-CSF (thin line), β-galactosidase virus, no M-CSF

(dashed line) , β-galactosidase virus + M-CSF (broken line) .

Figure 4 - Expression of IkBα and IkBα adenovirus

infected monocytes and RAW 264.7 cells. RAW 264.7 cells (a) or M-CSF treated monocytes (b) were untreated (un) or infected with IkBα adenoviral vector at the given m.o.i.

After 72 hours cell lysates and post-nuclear supernatants were analysed for IkBα expression by immunoblotcing.

Equivalent amounts of protein were loaded on each track for the given cell type. (c) Nuclear lysates were prepared from M-CSF treated human monocytes that had been infected with IkBα adenovirus or control virus at the

given m.o.i. for 72 hours. IkBα expression was determined as above: equivalent amounts of protein were loaded on each track.

Figure 5 - Expression of NFkB and IkB proteins in virus infected monocytes . Monocytes were treated with M-CSF for 48 hours. The cells were then infected with a control virus or an IkBα containing adenoviral vector at

a m.o.i. of 50. A third group were uncreated. Following a further 3 days in M-CSF the ceils were washed and half of "each group was treated with lOng/ml LPS for 30 minutes . The cells were then lysed and cytosoi and nuclear extracts were prepared which were then analysed for IkBα/β expression (cytosoi) or p65/p50 expression

(nuclear) by immunowestern blotting. 150μg protein and 50μg protein was analysed for each cytosoi or nuclear sample respectively.

Figure 6 - Inhibition of NFkB function in IkBα virus

inf cted monocytes . Monocytes were uninfected or infected with IkBα or control adenovirus followed by LPS

(lOng/ml) stimulation. Nuclear extracts were prepared and 20μg protein analysed for NFkB (upper panel) or AP-12

(lower panel) activity by EMΞA. Figure 7 - IkBα virus -infection inhibits LPS-induced TNF

mRNA and protein. M-CSF treated cells were untreated or

infected with IkBα virus or control virus at various

m. o . i . After three days cells were harvested and stimulated with LPS (10 ng/ l) . (a) After 2 hours cells were harvested and mRNA extracted and analysed for mRNA by Northern blotting with the indicated probes or (b) after 24 hours the culture supernatants were removed and

assayed for TNFα by ELISA, the cells were examined for

any toxic effect by MTT assay. Dat a is given as percentage TNF production of non- infected control cells . The results represent the combined data of cells from five different donors (=S .D . ) . No significant toxicity was observed as judged by the MTT assays (not shown) .

(c) Cells from rheumatoid synovium were cultured in 48 well plates at 1 x 10 cells and infected with either AdvIrcBα or AdvO at a m.o.i., of 40 to 1. The supernatants were taken

for ELISA assay at 48 hours; pooled data from the 5 patients is shown (+ s.d.). The null hypothesis that there was no change in TNF production whether cells were infected with AdvIκB or AdvO could be rejected (p <0.05) using a Wilcoxon's signed rank test on paired

differences compared with untreated cells on each patient. Figure 8: Effect of AdvIκBα infection on LPS induced IL-1, IL-6 and IL-8

expression by human macrophages. Effect of infection of human monocyte-derived

macrophages with various titers of AdvIκBα or AdvO on the LPS-inducεd production of

IL-lβ (A), IL-8 (B) and IL-6 (Q, expressed as a percentage of production of the cytokine

in question induced by LPS (10 ng/^'ml) in uninfected ceils. Error bars indicate standard

error of the mean (n = 7-10).

Figure 9: Effect of exogenous TNF on IL-1/? and IL-6 production. Pretreatment with

20 ng/ml or 100 ng/ml TNFα abrogates the inhibitory effect of AdvIκBα infection on LPS-induced IL-1/? production (A). TNFα-induced (20 ng/ml) IL-6 is potently inhibited by

AdvIκ:Bα infection (B), but TNFα-induced IL-1 is only moderately affected (C). Error bars indicate standard error of the mean (n = 6-8).

Figure 10 -.Inhibition of NF-κB has only minor effects on IL-10 production. Effect of

infection of human monocyte-derived macrophages with various titers of AdvI Bα or

AdvO, on IL-10 production induced by LPS (10 ng ml), expressed as a percen^tage of IL- 10 production induced by the same stimulus in uninfected cells (A). In (B), pre^treatment

wi_th 20 ng ml TNFα abrogates the inhibitory effect of AdvIκBα infection on LPS-

induced IL-10 production (pg ml). Error bars indicate standard error of the mean (n = 7-

10).

Fig isure 11 : Effect of IκBα infection on IL-lra production. Effect of infection of human

monocyte-derived macrophages with various titers of AdvIκBα or AdvO on the LPS-

induced production of IL-lra, expressed as a percentage of the production of the same

_* cy_tokine in uninfected cells challenged with the s∑ime stimulus (A). Pretreatment with IL-

1, or TNFα (20 ng/ml) alone does not affect, but treatment with both cytokines

moderately reverses the inhibitory effect of AdvIκBα infection on the LPS-induced

production of IL-lra (B). Error bars indicate standard error of the mean, n = 7-8.

Figure 12: AdvIκBα inection inhibits LPS-mediated soluble TNF receptors. Effect of

infection of human monocyte-derived macrophages with various titers of AdvIκBc or

AdvO on the LPS-induced production of the p55 (A) and p75 (B) soluble TNF receptors, expressed as a percentage of the production of the same cytokine in uninfected cells challenged with the same stimulus. Error bars indicate standard error of the mean, n = 7-

8. Figure 13: Inhibition of NF-κB selectively inhibits TNFα mRNA- Infection with 40:1

of AdvIκBα, but not 40:1 of AdvO, inhibits TNFα mRNA expression (upper psmeis)

induced by PMA (A) or UV light (B) as .assessed by Northern analysis. Zymosan-induced

TNFα mRNA expression (C) is unaffected by infection with 80:1 of AdvI Bα. Lower

paneis contain GADPH expression. This is a representative of three complete experiments.

Figure 14: PSl inhibits LPS-mediated TNFα where not affecting zymosan-induced

TNFα. Effect of the prote^ome inhibitor PSl on LPS-induced (■) and zymosan-induced

(•) TNFα production. Error bars indicate standard error of the mean, n = 6-7.

Figure 15: In excess of 90% of rheumatoid synovial cells can be infected with -adenovirus. β-gaiactosidase .activity in total rheumatoid synovial ceil cocultures (A), or with T lymphocyte (B), macrophage ( or synoviocyte (D) sub- populations detected through double gating for size/gπmuimty and cell surface markers, in cultures infected with 40: 1 of either AdvO (f 111 ed line) or Advβgal (solid linei A representative of five experiments. In (E), the rne-ans of β-gal.actosidase-positive ceils are plotted +/- stan ard error of the mean; n=5.

Figure !6:AdvIκBα infection inhibits the spontaneous production of proinfiammatory cytokines from rheumatoid synovial cells. Effect of infection of rheumatoid .synovial ceils with 40: 1 of either AdvO or AdvδcBα on the sponraneous production of TL-1 (A), IL-8 (B) and IL-6 (Q. Error bars indicate standard error of the mean; n=6.

Figure 17:I Bα overexpression permanently inhibits TNFα .and IL-6 production. Effect of infection of rheumatoid synovial ceils with 40: 1 of either AdvO or AdvI Bα on die spontaneous production of TNFα (A) or IL-6 (B) over time. Symbols denote uninfected Q, AdvO-infected () or AdvIκBα-infected () ceils. A representative of three independent experiments.

Figure 18: Inhibition of NF-κB through AdvI Bα infection has only marginal effects on the spontaneous production of anti-infl∑uπmatory medi.ators. Effect of infection of rheumatoid synovial cells with 40: 1 of either AdvO or AdvIκBα on the spontaneous production of IL-10 (A), the IL-1 receptor antagonist (B) and the p75 soluble TNF receptor (C). Error b∑irs indicate staηriarri error of the mean; n=ό. Figure 19: IκBα overexpression potently inhibits MMP-1 .and MMP-3 production, but slightly potentiates the production of TIMP-1. Effect of infection of rheumatoid synovial cells with 40: 1 of either AdvO or AdvIκBα on the spont.aneous production of MMP-1 (A), MMP-3 (B) and TIMP-1 (C). Error bars indicate standard error of the mean; n=6.

Fisure 20: Effect of AdvIκB infection on the balance between MMP production versus the production of TEMP-l over time. Effect of infection of rheumatoid synovia! cells with 40: 1 of either AdvO or AdvI Bα on the spontaneous production MMP-1 (A), MMP-3 (B) and TTMP-l (C) over time. Symbols denote uninfected—, AdvO-infected φ or AdvIκBα-infected ■ cells. A representative of three independent experiments.

1. INFECT ION OF HUMAN MONOCYTE-DERIVED MACROPHAGES

MATERIALS AND METHODS

Isolation of Peripheral Blood Monocytes . Single donor plateletphoresis blood residues were purchased from North London Blood Transfusion Service (Colindale UK) . Mononuclear cells were isolated by Ficoll-Hypoque centrifugation ( specific density) , 1 . 077 g/ml) preceding monocyte separation in a Bec man JEL elutriator .

Cells were usually collected from the Ficoll gradient resuspended in Hanks and centrifuged at ©2000 rpm in a centrifuge for 10 minutes. The pellets were resuspended to 50ml in a Falcon tube and respun ©lOOOrpm for 15 minutes .

The resulting pellet was resuspended in RPMI with 10% FCS up to 50ml before injecting into the elutriator. A maximum of 1000xl0⁶cells were loaded into the chamber. Typically the cells were elutriated according to the manufacturers instructions .

The elutriator head was assembled and 200ml 1% Etoxaclean was flushed through. This was followed by 1.5L of sterile distilled water. 300ml of elutriation buffer (RPMI with P/S and 1% low endotoxin FCS) was then pumped through. The sample containing cells to be purified was then aspirated through the elutriator at 2000rpm and 10' C. Fractions containing the cells of interest were collected in Falcon tubes .

Monocyte purity was assessed by flow cytometry using directly conjugated anti-CD45 and anti-CD14 antibodies (Leucocyte, Becton Dickinson, UK) and was routinely greater than >90%. All media used in separation and culture of monocytes was tested for endotoxin using the Limulus amoebocyte lysate test (Bio Whittaker Inc. Bethesda MD) and were rejected if endotoxin contamination exceeded 0.1 unit/ml.

Recombinant Adenovirus Vectors . The recombinant replication-deficient adenovirus vectors encoding E.coli β-galactosidase or having no insert (rAdø) was provided by Dr. A. Byrnes (Oxford UK) . A second virus encoding E.coli β-galactosidase gene (Ad/β-gal) [Watanabe, 1996, Blood 87:5032-9] was generously provided by Canji, Inc. (San Diego CA) and the vector encoding IkBα (rAD

IkB) [Wrighton J. Exp. Med. 183:1013-22] by Dr. de Martin (Sandoz, Vienna Austria) . The viruses were prepared, purified, titered as previously described [Watanabe, supra] . Virus was produced in the 293 human embryonic kidney cell line and purified by ultracentrifugation through two caesium chloride gradients. The titers of viral stocks were determined by a limiting dilution plaque assay on 293 cells, and were 2.9 to 5.8 x 10 ¹⁰ infectious units/ml as measured before dilution for use. Viruses were suspended in a buffer solution of 2% sucrose and 2mmol/L MgCl ₂ and stored at -70OC.

Cvtokine treatment and adenoviral infection of monocytes . Human monocytes or murine RAW 264.7 were cultured at 5 x 10 ⁵ /ml in RPMI 1640 (Bio Whittaker inc.) supplemented with 5% (v/v) heat inactivated foetal calf serum containing 10 units/ml penicillin/streptomycin and 2mM glutamine at 37θC. Monocytes were treated with GM-CSF (5U/ml) or M-CSF (lμg/ml) for 72 hours. Expression of integrins was assayed by indirect immunofluorescence and

FACS analysis using monoclonal antibodies to αvβ3 (100

μg/ml LM609, provided by IXSYS Corp. (San Diego CA USA),

or αvβ5 (undiluted active supernatant kindly provided by

J. Gamble Smith) . The commercially available monoclonal antibody 0X14 (lOOμg/ml) was used as a negative control.

Following cytokine treatment monocytes were washed in RPM 1640 and resuspended at 5 x 10 ⁵ cells/ml. The cells were infected with virus at the indicated plaque forming units per cell (multiplicity of infection - m.o.i.) . β-galactosidase expression in individual cells was measured by FACS .

Preparation of cvtosolic proteins: 4 x 10⁶ cells were plated on petri dishes (Nunc) , and cultured overnight. They were stimulated, washed in ice cold PBS, removed from dishes by scraping, and lysed (20mM Tris pH 8.0, 137mM NaCl, ImM MgCl ₂ 0.1% NP-40 10 minutes at 40C) . Lysates were spun (120Ox g 10 minutes 4θC) and the supernatant was retained and assayed for protein concentration by the Commercially available BCA method.

Preparation of nuclear proteins. Nuclear extracts were prepared as previously described [Whiteside, 1992, Nuc . Acids Res. 20:1531-8 ]. Briefly cells were washed with ice cold PBS then spun (1200x g 30 second 4θC) and the aspirated pellet was resuspended in low salt lysis buffer (lOmM Hepes, 1.5mM MgCl₂ lOmM KCl 0.5 mM DTT ImM PMSF lOμg/ml aprotinin 30 μg/ml leupeptin) at 4θC for 5 minutes, NP-40 was added to a final concentration of 0.125% and cells were vortexed immediately. The samples were spun (120Ox g 10 minutes 4θc) , the cytosolic fraction was removed and the nuclear pellet was resuspended in nuclear extraction buffer (5mM Hepes 25% w/v glycerol, 0.4M NaCl 1.5ml MgCl₂ 0.2mM EDTA ImM DTT ImM PMSF lOμg/ml aprotitin, lOμg/ml pepstatin 30μg/ml leupeptin) and left at 4θC for 1 hour. The samples were spun (120Ox g 10 minutes 4θc)and the supernatant was

retained and stores at -70oc.

_Western blotting of proteins . Samples were mixed with an equal volume of 2x gel sample buffer (62.5mM Tris pH 6.8, 2% SDS, 5% β-mercaptoethanol, 10% glycerol, 0.002% bromoohenol blue) and boiled for 3 minutes. Equal amounts of protein were loaded in each lane, separated by SDS-PAGE on a 10% w/v polyacrylamide gel and transferred to nitrocellulose (Boehringer Mannheim) . Membranes were blocked with a 5% milk powder solution (20mM Tris 150mM NaCl pH 8.0.0.1% Tween 20) and then proteins were detected with rabbit polyclonal antibodies and horseradish peroxidase conjugated donkey anti-rabbit F(ab)₂ fragments (Amersham UK), both diluted in milk powder, and visualised by enhanced che iluminescence (Amersham UK) .

Electrophoretic mobility shift assays. EMSA's were performed as previously described (Clarke, 1995, Eur. J. Immunol. 25: 2961-6). Data was analysed using a Ξiorad GS-670 densito eter.

_Measurement of TNFα production. RAW 264.7 cells were

resuspended at 2x10 ⁵/ml and plated on 24 well plates at 500ml per well. They were incubated for 16 hours in the presence or absences LPS (lOng/ml) and with various concentrations of cytokines . The supernatants were assayed for TNF levels by WEHI 164 (Clone 13) bioassay [Espevik, 1986, J. Immunol. Methods 95: 99-105].

Preparation of cytosolic mRNA. 15x10 ⁶ cells were plated overnight, stimulated for 2 hours with LPS and cytokine, washed in ice cold PBS and resuspended in 400μl lysis buffer (lOmM tris pH 7.9 150mM NaCl 1.5mM MgCl₂ 0.65% NP-40 10 minutes 4θc. Lysates were spun (1200Ox g, 5 minutes, room temp.). EDTA (lμl 0.25M) and SDS (20μl 10% in H₂0) were added to the required supernatant and the tube was gently mixed. mRNA was purified by extraction, with 1:1 phenol chloroform, and precipitation (-20OC, 30 minutes) with sodium acetate (pH 5.2) to a final concentration of 0.3M and 2 columns of ethanol. The precipitate was spun down (120Ox g, 15 minutes, 40C) , washed with freezing 70% ethanol, then re-dissolved in water.

Northern blottinσ. mRNA (lOμl, Iμg/μl) was mixed with 2μl H0, 5μl lOx northern buffer (200mM MOPS, 50mM sodium acetate), lOmM EDTA pH 7.0), and 25μl deionised formamide and heated to 60OC for 5 minutes. lOμl 6x loading buffer (0.25% bromophenol blue, 0.25% xylenol cyanol 25% glycerol) was added and the samples were run on an agarose gel (1% agarose, 20mM MOPS, 5mM sodium acetate, ImM EDTA, 6.5% formaldehyde pH 7.0, 100mA 4 hours). mRNA was transferred to a Hybond-N membrane (Amersham UK) by capillary action and fixed by baking (80oc, 2 hours) . The membrane was pre-incubated in hybridisation solution (50% formamide, 5x SSC, 0.05M sodium phosphate pH 6.6 0.5x Denhardts solution O.lmg/ml salmon sperm DNA) for 4 hours at 42θC and hybridised (12 hours, 420C) with l-3x 10 ⁶ cpm/ml of ³²P labelled cDNA probe dissolved in hybridisation solution. The membrane was washed in 0.2xSSC 0.1% SDS at 42oc and exposed to hyperfilm MP at -70OC for 1-3 days.

RESULTS

Efficient gene transfer into primary macroohage following treatment with M-CSF. Using an adenoviral vector encoding the E.coli β-galactosidase gene the effectiveness of this approach to transfer genes into primary human monocytes obtained from peripheral blood was investigated. As previously shown by others (Haddad, supra, Huang, Supra) using freshly prepared monocytes we were unable to detect expression of β-galactosidase in these cells following exposure to the virus even at m.o.i. > 1000 (results not shown) . It has been

previously reported that OV integrins are required as

cofactors for virus infection [Huang, 1995 supra] .

Expression of αvβ3 and OVβ5 was found to be low on

resting monocytes (Figure 1) , however treatment with GM-CSF or M-CSF for 48 hours unregulated the expression

of _OVβ3 or αVβ5 respectively (Figure 1) . Treatment with

a combination of both cytokines resulted in integrin expression that resembled GM-CSF alone (Figure 1) . With reference to these observations, the infectivity of monocytes with adenovirus was re-examined. Prior treatment with GM-CSF or M-CSF for 72 hours, which was found to be the optimal treatment, respectively resulted in 52% or 90% of the cells expressing the ^β-galactosidase gene following treatment with virus at 100 m.o.i. (Figure 2) . As with integrin expression, treatment with the combination of cytokines gave a result (52% of cells positive) similar to GM-CSF alone (Figure 2) . Further studies were therefore pursued using only M-CSF. Reducing the m.o.i. to 50 from 100 had only a small effect on efficiency of infection. 92% compared to 98% (Figure 3); however at 25 m.o.i. infectivity was greatly reduced (20%) . Infection of monocytic/macrophage cell line was also investigated. The human cell lines U937 and THP-1 were totally refractory to infection by adenovirus even following treatment with GM-CSF or M-CSF,

neither cytokine was able to induce OVβ3 or αVβ5

expression in these cell lines (results not shown) . In contrast, the murine macrophage cell line RAW 264.7 was infected by adenovirus β-galactosidase to a degree similar to human monocytes at 100 m.o.i. (Figure 4), and still showed >60% infection at m.o.i. of 10 (results not shown) . These cells did not require prior exposure to M-CSF (Figure 4) .

Adenoviral transfer of IkBα into primary monocytes. The

success of the gene transfer experiments on human monocytes led the inventors to investigate the use of this system for the introduction of genes that would modify important intracellular signalling pathways. For this purpose the inventors used an adenoviral vector

expressing IkBα [Baeuerle, 1996, Cell 97:13-20] under

the control of CMV promoter, rAD.IkB [Wrighton, Supra] .

IkB is known to inactivate NFkB a transcription factor within cells. Inhibition of NFkB allows the suppression of proteins whose expression is dependent on the factor to be studied. Infection of both RAW 264.7 cells of M-CSF treated human primary macrophages result in the prominent over-expression of IkBα (Figure 5) . No changes in IkBα expression were evident following

infection with control virus. The anti-IkBα antibody

used did not recognise murine IkBα, hence the failure to

detect endogenous IkBα in the RAW 264.7 cells. As a

control, Western blots were reported for p42MAPK that showed no changes to expression following adenovirus infection (Figure 5) . As the IkBα construct also

contained a nuclear localisation sequence, the presence

of nuclear IkBα could also be detected (Figure 5c) .

Infection with rAd.IkB inhibits NFkB activity in human macrophages . Infection with the virus resulted in the expression of high levels of IkBα even after LPS

treatment, unlike uninfected cells or cells infected with control virus (Figure 5) , where IkBα was degraded.

Studies on nuclear extracts from the same cells showed that nuclear translocation of NFkB p65 subunit, and to a lesser extent the p50 subunit, was inhibited in virus infected cells. There was also some inhibitory effect on p50 nuclear translocation in cells infected with control

virus; the reasons for this are unclear. Unlike IkBα,

the levels of IkBβ expression were not effected by viral infection. No LPS-induced degradation of IkBβ (Figure 5b) was observed, a finding the inventors have also consequently observed in human monocytes (results not shown) .

The effect of IkBα virus infection on nuclear NFkB DNA

bind activity was also investigated by EMSA (Figure 6) . Infection of monocytes with the virus resulted in total ablation of nuclear NFkB DNA binding activity. No such effect was observed in uninfected or control virus treated cells. The activity of AP-1 was also investigated. Unlike NFkB, AP-1 appeared to be constitutively active in the M-CSF treated cells and no LPS induced activity was detectable. However it was noted that there was a some low level reduction in nuclear AP-1 binding activity in IkBα virus infected

cells compared with the control cell populations .

Infection of monocytes with IkBα adenovirus inhibits LPS

induced TNF expression.

Activation of NFkB is thought to be essential for the

expression of TNFα although this hypothesis has never

been tested in primary human cells . Infection of

monocytes with IkBα virus resulted in inhibition of LPS induced TNFα mRNA expression (Figure 7a) . No effect was

observed on the levels of GADPHα mRNA. A similar effect

was also observed on the expression of TNFα protein

where a clear close dependency relating to the m.o.i. was

observed (Figure 7b) . The IkBα virus was also used on

synoviocytes obtained from rheumatoid joints. These cells constitutively express TNF and other cytokines . Infection with virus resulted in the inhibition of TNF production by these rheumatoid synoviral cells (Figure 7C).

Lack of Aooptosis using IkB

As NFkB has been reported to be a key anti-apoptotic mechanism and that inhibition of NFkB induced by TNF resulted in cell death, it was possible that cell depletion may account for the inhibitory effects of the IkB virus . However the inventors failed to observe a toxic effect of the virus at the m.o.i. used, and

furthermore treatment of IkBα adenovirus infected

macrophages did not undergo apoptosis following treatment with exogenous TNF (results not shown). DISCUSSION

The results show that the method of the invention produces a high efficiency of infection in cells, such as human primary macrophages by adenovirus .

The inventors have demonstrated approximately 100% infection with m.o.i. of 50-100. Previous studies [Haddada 1993 supra and Huang, 1995 supra] only obtained approximately 50% infection which required a m.o.i. of 1000. The inventor's observation that high levels of gene transfer can be achieved with the conditions of the invention has important consequences for the study of intracellular mechanisms of macrophages as it precludes any requirement to remove a non-infected population and obviates the need for cloning . This allows the regulation of endogenous chromosomal genes to be studied and precludes the need to use ectopically expressed reporter constructs which may be regulated differently.

High levels of infectivity of adenovirus allowed the

insertion of genes expressing IkBα intracellularly to study its effect on the NFkB to TNFα synthesis to be

successfully studied.

2. DEMONSTRATION OF SELECTIVE REGULATIONS OF CYTOKINE INDUCTION

MATERIALS AND METHODS

Isolation of peripheral blood monocytes. Single donor plateletphoresis residues were purchased from North London Blood Transfusion Service (Colindale UK). Mononuclear cells were isolated by Ficoll-Hypaque centrifugation preceding monocyte separation in a Beckman JEL elutriator.. Monocyte purity was assessed by flow cytometry and was routinely greater than 90%.

Adenoviral Vectors. Recombinant, replication-deficient adenoviral vectors encoding E.

coli β-galactosidase or having no insert (AdvO) were provided by Drs A. Byrnes and M.

Wood (Oxford, UK). An adenovirus encoding porcine IκBα with a CMV promoter and a

nuciear localization sequence (AdvIκBα) was provided by Dr R. de Martin (Vienna.

Austria). Viruses were propagated in the 293 human embryonic kidney cell line and purified by ultracentrifugation through two cesium chloride gradients. The titers of viral stocks were determined through a plaque assay on 293 cells, as described (24).

Infection Techniques. The elutriated human monocytes were incubated at approximately 2 x lO'/ml in RPMI 1640 with 25 mM HEPES and 2 mM L-glutamine, supplemented with 5% (v/v) heat inactivated fetal calf serum and 10 units/ml penicillin/streptomycin. To optimize infection purified human monocytes were pretreated with M-CSF (100 ng ml; obtained from the Genetics Institute. Boston MA) for 48 h to

allow upregulation of integrin αVβ5, which has previously been shown to be essential for adenovirus infection of monocytes (25). The cells were then repiated on 100 mm Petri

dishes and infected for 2h with a m.o.i. of beuveen 10: 1 and 120:1 (in most experiments,

20: 1. 40: 1 or 80: 1 was used) of either AdvIκBα or AdvO, in serum-free RPMI 1640.

Cells were then incubated in RPMI 1640 supplemented as above for 48 h to allow for

significant over-expression of IκBα, as assessed . During the changes of medium

involved, non-adherent cells were discarded, resulting in a further purification of

monocyte-derived macrophages.

Cytokine Analysis. For Northern analysis experiments, cells were repiated at 5-10 x 10⁶

cells per 100 mm Petri dish and stimulated with LPS (10 ng/ml), PMA (10 nM), zymosan

(30 μg. ml), ionomycin (,1 μM), or by UV irradiation (2000 J). After 4 h. cells were

harvested. mRNA extracted and subjected to Northern analysis as in Buchan et al_. ( Cl i n. Exp. Immunol . 73 : 449 ( 1988) .

In the assays for cytokine production, cells were repiated at 5 x 10' cells per well on a 96-

well dish, and stimulated as above for 4 or 16 h. Supernatants were analysed for TNFα

(27), IL-lβ. IL-6 and IL-8 (28), IL-10 (29, 30), IL-lra and the p55 and p75 soluble TNF

receptors (31) by ELISA. The proteosome inhibitor Cbz-Ile-Glu(O-t-Bu)-Ala-leucinal

(PSl) was obtained from Calbiochem (Nottingham, England).

Statistical methods. All statistical testing was performed using a paired comparison,

one-sided Student's t test. RESULTS

Cytokine production in response to various stimuli.

LPS is capable of inducing all the cytokines (TNFα, IL-lβ, IL-6, IL-8) and inhibitors (IL-

10, EL- Ira and the soluble TNF receptors) assayed in this study (Table I; the results are

means of 4-11 experiments). The induction of TNFα mRNA by zymos.an has previously

been reported to be far weaker than LPS-induced TNFα mRNA in mouse peritoneal

macrophages (32), but the TNFα response to zymosan in M-CSF-treated human monocytes was equal to that of LPS (Table I).

LPS was a much stronger inducer of IL-lβ than any other stimulus used. In

agreement with observations in mouse macrophages Bondeson J. and Sundler R.

(1995, Biochem Pharmacol 50 : 1753). IL-6 was induced equally well by LPS and zymosan. IL-8 could be induced by the entire array of stimuli. Zymosan and. LPS were die only stimuli to induce IL-10 and the p55 soluble TNF receptor (Table I) whereas

PMA could induce IL-lra and p75 soluble TNF receptor.

Does IκBα over-expression inhibit LPS-induced pro-inflammatory cytokines?

We previously observed that infection at a multiplicity of 20-80:1 with the IκBα

adenovirus produced high levels of IκBα expression. This resulted in a potent inhibition

of LPS-driven TNFα induction in human macrophages by blocking NFKB function (7). This was not due to loss of cells through apoptosis or other causes of cell death (7). The induction of IL-lβ and IL-8 by LPS also appears to be strongly NF-t B dependent (Figure 8A-B), and there is potent inhibition of both these cytokines already at 20:1 of AdvIκBα. TABLE I

CYTOKINE PRODUCTION INDUCED BY VARIOUS STIMULI

CO c m co

H

H

C m co x m m

H c r- m ro

Results using supernatants harvested after 4 h of incubation were similar to those using

16 h of incubation (not shown). In contrast to the other cytokines studied here, there was

a potentiation of IL-6 production of about 20-30 % in AdvO-infected cells, with LPS as

well as with other stimuli (Figure 8C), although adenovirus infection alone has no effect.

Nevertheless, IL-6 expression was' strongly inhibited by infection with AdvIκBα in

human macrophages (Figure 8C).

Since TNFα can induce the synthesis of other cytokines, e.g. IL-lβ, IL-6 and IL-8, the

potent inhibition by AdvIκBα might be secondary to inhibition of TNFα. Culturing with

exogenous TNFα partly abrogated (just 20% inhibition with 100 ng/ml TNFα, compared

with 62% inhibition with no cytokine) the effect of IκBα over-expression on IL-lβ

production (Figure 9A), indicating at least a partial dependence on TNFα. However,

similar experiments showed no major role for TNFα in LPS induced expression of IL-6

and IL-8 (results not shown).

To further investigate the signaling mechanisms involved, we sought to characterize

whether TNFα-driven IL-lβ might be less NF- B dependent than the IL-lβ response

induced by LPS. One difficulty in doing this was the fact that TNFα is a weaker inducεr

of IL-lβ in our system than LPS (Table I). Yet the IκBα-induced inhibition of TNFα-

induced IL-lβ was relatively less potent than that seen with TNFα-induced IL-6 (Figure

9B-C), as expected from the restoration of IL-lβ synthesis by TNFα in cells infected with

AdvIκBα. Does IκBα over-expression affect LPS-induced anti-inflammatory cytokines?

The major anti-inflammatory cytokine produced by macrophages is IL-10 . In

LPS-stimuiated cells, there was gradual inhibition with increasing virus titers, but

.statistically significant inhibition (p<0.05) was only noted at 60:1 or 80:1 of AdvI Bα

(Figure 10A⁾. This inhibition was still quite modest (30% at most), and since the human

EL- 10 promoter lacks KB sites, we investigated whether the inhibitory effects of

AdvI Bα infection was indirect, via its effects on pro-infhmmatory cytokines tint are

known to influence IL-10 expression . A combiimtion of TNFα and LPS,

considerably potentiated the IL-10 response and partially abrogated the inhibition by

AdvI Bα (Figure 10B) . In contra^ IL-1 failed to have any significant effect (results not shown).

The LPS-induced production of IL-lra was slightly inhibited (-30%) by I Bα over-

expression (Figure 11 A) .This response was only modestly affected by adding back IL-1

and TNFα (Figure 1 IB⁾ . In human macrophages LPS induces die production of the p75

soluble TNF receptor and to a lesser extent p55 soluble TNF receptor (Table I). Infection

with AdvIκBα w.as observed to inhibit the LPS-induced production of both these soluble

receptors (Figure 12 );this response w.as unaffected by adding back IL-1 or TNFα (results not shown). Does IκBα over-expression inhibit pro-inflammatory cytokines induced by PMA or

UV light?

The induction of TNFα by PMA is very potently (~ 90%) inhibited by IκBα over-

expression, even more strongly than in LPS- (~ 80%) or even UV light-stimulated

ceils (~ 80%) (Table H). Again, ^'results using supernatants harvested after 4 h of

incubation were similar to those using 16 h of incubation (data not shown). A series of

Northern analysis experiments demonstrated mat, as for LPS-induced TNFα , the

TNFα mRNA expression in response to PMA and UV light was ablated by IκBα over-

expression (Figure 13A-B) .

As shown for TNFα, the induction of IL-lβ and IL-6 by PMA or UV light was NF- B

dependent (Table II). Once again, the inhibition of IL-lβ and IL-6 production by the

AdvIκBα infection appears to be more potent in PMA-treated cells than in cells treated

with LPS or UV light.

There was also inhibition (50-60 %) of the IL-8 response when cells infected with

AdvIκBα were stimulated with PMA or UV light (Table II). Ionomycin, a stimulus that

did not induce significant amounts of the other cytokines of interest, induced a

discemable IL-8 response, which was also inhibited by IκBα over-expression (data not

shown).

PMA did not induce detectable amounts of IL-10, but induced a detectable p75 soluble

TNF receptor response, which was inhibited (75% at 40:1 of AdvIκBα) by IκBα over- co c

CD CO

H

H

O

o

expression, as was the PMA-induced production of IL-lra (40% inhibition at 40:1 of

AdvIκBα).

Does IκBα over-expression affect zymosan-induced cytokines?

In contrast to the prior stimuli when the macrophages were activated with zymos.an,

infection with AdvIκBα had no effect whatsoever on TNFα protein (Table H) or mRNA

expression (Figure 13C) ,even at m.o.i of 80: 1. The induction of IL-lβ and IL-8 by

zymosan was also unaffected by the Ii Bα over-expression (Table II) but was a small (10-

15% compared with uninfected cells) inhibition of IL-6. Zymosan-induced IL-10, IL-lra (Table H) and the soluble TNF receptors (not shown) was also refractory to inhibition by

IκBα over-expression (Table I).

The independence from NFKB of zymosan induced TNF was further emphasised by

studies with the proteosome inhibitor PSL Inhibition of proteosome function inhibits

IκBα degradation thus preventing NFKB nuclear translocation. PSl was very

effective in blocking LPS induced TNF production but did not affect the response to

zymosan (Figure 14) .

DISCUSSION

The inventors have shown that adenoviral gene transfer into macrophages provides a reliable, reproducible and convenient method of studying intracellular signalling

pathways. The natural inhibitor of NF-κB, IκBα was used to establish the principle.

Transfer of the IκBα effectively inhibits NF-κB activity in human macrophages, mainly through the over-expression of IκBα inhibiting nuclear translocation of the p65/p50 subunits of NF-κB. This blocked LPS-induced TNFα both at the mRNA and the protein level. Even more interesting, it was found that the endogenous production of TNFα from rheumatoid synovial mononuclear cell cultures was also inhibited.

The fact that LPS-induced TNFα. IL-lβ and IL-6 were all NF-κB dependent cytokines

could be expec_ted from the majority of data from murine cells .and monocyte/macrophage cell lines. The results on LPS-induced cytokines also agree well with the results of Makarov et al (Gene Therapy (1997) 4 : 46) on LPS-induced IL-1/?, IL-6 and IL-8 in monocytic THP-1 cells

stably modified through retroviral gene transfer of IKB. Among the LPS-induced, pro- inflammatory cytokines studied here, IL-6 was most potently inhibited (> 85%) by over-

expression of IκBα. whereas there was always some residual production of TNFα or IL-

lβ even in LPS-stimuiated cells infected with high titers (120:1) of AdvIκBα (not

shown). This may reflect a certain amount of preformed cytokine mRNA. but this could not be demonstrated in unstimulated cells (data not shown), and furthermore LPS , PMA or UV induced (Figure 13)TNFα mRNA was profoundly downregulated by IκBα. However it could not be excluded that this residual cytokine production emanated from the few uninfected cells still present, and work is in progress to elucidate this question usins intracy_toplas ic staining for cytokines. Another not exclusive hypothesis (discussed in more detail below), is that TNFα and other pro-inflammatory cytokines can

be induced through both NF-κB dependent and NF-κB independent pathways.

Our finding that UV light induces a whole spectrum of pro-inflammatory cytokines in a NF-κB-

dependent manner is novel. It is in contrast to the earlier report that UV-induced TNFα in RAW 264.7 cells does not involve NF-κB (Bazzoni, J. Clin. Invest. (1994), 93 : 56). Similarly, the finding that the PMA-induced induction of TNFα and other pro-inflammatory cytokines is profoundly downregulated by IκBα over-expression, disagrees with several earlier studies in stably transformed

human cell lines (Makarov, Supra). In our hands, this stimulus was actually the one most strongly dependent on NF-κB, as judged by the percentage of inhibition, reproduced in seven separate experiments. These discrepancies between results obtained with human primary cells and those from various transformed cell lines indicate that, at least in some instances, the latter are questionable models for studying cytokine cell signallⁱng occurring in primary cells, as is the case in vivo. In a way, this is not surprising, since mere are interactions between the enzymes and transcription factors of the cell cycle machinery and the resulation of cytokine genes, eg Rb regulates as, which is involved in cytokine activation (Bassi* A. and Leiden O .M. , Adv. Immunol . (1997) 64 : 65) .

With regard to macrophage signal tr∑insduction, one of the most remarkable findings in the present study was that zymosan, although a very powerful macrophage activator, does ot appear to require NF-KB for the induction of either pro- or anti-inflammatory

cy_tokines. These findings would imply that there are, in human macrophages, both NF^¬

_KB dependent and F- LB independent pathways of cytokine induction involved in the

induc_tion of TNFα and other pro-inflammatory cytokines. The modest (15%) of

questionable significance inhibition of zymosan-induced IL-6 observed in cells over-

expressing IκBα may reflect the observation that this cytokine was the most potently

affected bv IκBα over-expression, irrespective of stimulus (see Table II).

Another finding of importance is that in human macrophages. IL-10 is under complex control, and in LPS-stimuiated cells, it appears to be at least partially driven via LPS-

induced TNFα and IL-1. It is interesting to note that, even at 40:1 of AdvIκBα. when

LPS-driven TNFα is abrogated by more than 60%, IL-10 is still not significantly

inhibi_ted (Figure 10). At higher virus titers. resulting in even stronger inhibition of TNFα,

therε is some effect also on IL-10, but never, even with 80: 1 of the virus, exceeding 30%.

This is completely reversible by adding back TNFα, implicating that LPS-induced IL-10

is partly driven secondarily by TNFα. This finding agrees well with previous reports,

and indicates that autocrine interactions can take place, even in short term (16

hour) cultures such as these.

Another intriguing finding was that the IκBα-inducεd inhibition of the LPS-induced

produc_tion of IL-l β (Figure 9A), but not the production of IL-6 or IL-8, or ^;hε soluble

TNF recε ors (data not shown), was also somewhat abrogated when TNFα was restored indicating that TNF induced IL-lβ is mainly independent of NF- B. This finding also

may suggest that IL-lβ is also, although to lesser extent than IL-10, driven partly by LPS-

induced TNFα. This result echoes, although to a much more moderate extent, the

previous work in RA joint cell co-cultures, where TNFα blockade was found to inhibit

the production of IL-1, and subsequently of TL-6, IL-8, IL-10 .and GM-CSF

which has led to die concept of a TNFα-dependent 'cytokine cascade' in infl.aιnιnatory

sites such as the rheumatoid synovium.

The dissection out of signaling pathways in normal primary cells is necessary as there .are differences from cell lines (see .above), and this is now possible within htimsn macrophages, from either normal or pathological specimens, using this adenoviral

technique. The model of human macrophages infected with Adv cBα has, from a

cytokine point of view, provided results similar to those from infecting human synovia! cocultures with the same virus ( see be! o ) . The data

.suggest that NF-κB is an important therapeutic target in chronic infI.ammatory diseases,

.allowing profound dowπregulatiαn of macrophage-produced pro-mfi∑imπiatory cytokines, while not directly affecting the most important .anti-inflammatory cytokines, IL-10 .and IL-lra. This would redress the disturbed equilibrium between these medi.ators.

³- DEMONSTRATION OF ADENOVIRUS INFECTION OF RHEUMATOID SYN0VIA₁ P_{F1 1 S .}

Rheumatoid synovial cells produce high levels of cytokines and express high levels of integrins on their surface (Feldman M. et al Ann. Rev. Immunol (1996) vol. 14). They have been used to demonstrate mat in vivo activated cells, with spontaneous high levels of integrins and cytokines allow improved viral infection rates to be .achieved. A use of such a system is also demonstrated. METHODS

Cells

Synovium from patients with rheumatoid arthritis undergoing joint surgery was dissociated by cutting into small pieces, and digested with collagenase and DNAse(). The total cell mixture was cultured at 37°C in RPMI 1640 with 25 mM HEPES and 2 M L-glutamine. supplemented with 5% heat inactivated fetal bovine serum, on 24-well or 48-well plates.

Adenoviral Vectors

Recombinant, replication-deficient adenoviral vectors encoding E. coli ?-galactosidase (Adv gal) or having no insert (AdvO). An adenovirus encoding porcine -Bα with a CMV promoter and a nuclear localization sequence (AdvI/cBα) was provided as above. Viruses were propagated in the 293 human embryonic kidney cell line and purified by ultracentrifugation through two cesium chloride gradients. The titers of viral stocks were determined through a plaque assay on 293 cells. All viruses used were plaque purified from a master stock, in order to prevent contamination with wild form adenovirus.

Infection Techniques

For experiments concerning cytokine or matrix metalloproteinase synthesis, freshly prepared rheumatoid synovial cells were resuspended in 1 ml serum-free RPMI 1640 on a 12- well plate at 2.5-4 million cells per well. After incubation for lh. they were either left uninfected, or infected with Adv Bα or AdvO at a multiplicity of infection of 40: 1. After 2h, the supernatants were removed and replaced with 0.5 ml RPMI 1640 supplemented with 5% heat inactivated fetal bovine serum. The nonadherent cells were carefully spun down and reintroduced to the rheumatoid cell coculture in 0.5 ml RPMI 1640 supplemented with 5% heat inactivated fetal bovine serum (total volume thus 1 ml).

In some experiments, rheumatoid synovial cells were infected as above, but at a density of 0.4 million cells per well in a final volume of 0.2 ml, on a 96-well plate. Four wells of cells from each cathegory - uninfected, AdvO-infected and AdvIκBα-infected - were plated. After incubation with adenovirus for 2h, one well from each cathegory was untreated, one treated with 20μg/ml of human recombinant IL-1 receptor antagonist (Cambridge Bioscience), one treated with 20μg/ml of the A2 anti-TNFα antibody (Centocor, Malvem, US), and one treated with both these cytokine inhibitors.

For infectibility experiments, cells were resuspended in 0.6 ml serum-free RPMI 1640 on a 24-well plate at 1 million cells per well. After incubation for lh, they were either left uninfected, or infected with Adv gal or AdvO at a multiplicity of infection of 40: 1.

After 2h, me supernatants were removed and replaced with 0.4 ml RPMI 1640 supplemented with 5% heat inactivated fetal bovine serum. The nonadherent cells were carefully spun down and reintroduced to me rheumatoid cell coculture in 0.2 ml RPMI 1640 supplemented with 5% heat inactivated fetal bovine serum (tot∑d volume 0.6 ml).

Analysis of Infectibility

Cells were scraped off the plates in the culture medium 48 h after infection, spun down and washed in FACS staining solution . Each batch of uninfected,

AdvO-infected or Adv gal-infected cells were then resuspended in 25 μl staining solution and incubated with 125 ng of anti-CD3 PerCP and 500 ng of anti-CD 14 PE (both from Becton Dickinson, San Jose CA), in a total volume of 45 μl for 45 min at 4°C. They were then incubated at 37°C for 10 min, before 45 μl of a 2 mM solution of Fluorescein di-(/?-D-galactopyranoside (Sigma) was added for 1 min. Addition of

excess (lOx) ice-cold staining solution was used to stop the reaction. Cell fluorescence

was analyzed by FACS.

Analysis of Cvtokines and Metalloproteases

In the assays for cytokine production, supernatants were taken off and nonadherent cells removed 48 h after infection. They were analysed for TNFα , IL-1 ?. IL-6 and IL-8 , IL-10 (), IL-lra and the p55 and p75 soluble TNF receptors by ELISA. The production of MMP-1 (Collagenase), MMP-3 (Stromelysin), MMP-13 and the tissue inhibitor of metalloproteases (TIMP-1) was analysed by ELISA kits purchased from Amersham (Little Chalfont, Bucks., UK).

In some experiments, aliquots of 50 μl were taken off after 12h. 24h, 48h. 72h and 1 lOh. After the nonadherent cells had been removed, and reintroduced to the cells in the same amount of medium, these aliquots were analysed for cytokines and metalloproteases as described above.

RESULTS

Infectibility of Rheumatoid Svnovial Cells with Adenovirus Infectibility of rheumatoid synovial cells was investigated using an adenovirus transferring the /?-galactosidase gene (Adv/JGal). Since 40:1 of Adv Gal infected 95- 100% of hum∑m macrophages and 30: 1 of Adv/JGal infected ne∑trly 100% of hum.an fibroblasts, we decided to use 10: 1 and 40: 1 of Adv/?Gal in initial experiments. It turned out that 40: 1 of Adv/?Gal infected nearly 100% of the entire rheumatoid synovial cell population (Figure 15A), but that 10: 1 infected only 60-70% (not shown), as evidenced by FACS analysis. All subsequent experiments thus use 40: 1 of Adv Gal and other adenoviruses.

It was of interest to discriminate between the different cell types in the rheumatoid synovial cell cocultures, and to separately study the infectibility of the main constituents - macrophages, T lymphocytes and synoviocytes. This was performed through preincubating the cells with fluorescent antibody markers to CD3 and CD 14, which allowed identification of CD3+/CD14- (T lymphocytes), CD14+/CD3- (macrophages) and CD3-/CD14- (synovial fibroblast) populations. By using human peripheral blood T cells and monocytes, and human skin fibroblasts, as controls, forward scatter/side scatter characteristics of these cell populations could be identified. Double gating was then used to ensure that the rheumatoid synovial cell populations chosen to represent T lymphocytes, macrophages and synoviocytes had the correct size/granularity and surface marker characteristics. As could be expected from earlier results , both macrophages (Figure 15B) and fibroblasts (Figure 15C) were easy to infect: more surprisingly in excess of 90% of the T lymphocytes were also infected by 40: 1 of Adv Gal (Figure 15D).

I/cBα Overexpression Inhibits the Production of Proinflamrnatorv Cytokines We have previously demonstrated that the spontaneous production of TNFα, as assessed two days after infection, was inhibited by 70% in cells infected with AdvIκBα, as compared with mock-infected cells or cells infected with AdvO . The spontaneous production of both IL-1 ? (Figure 16A) and IL-8 (Figure 16B) was also inhibited by I_/cBα

overexpression, but to a more modest degree (40%, as compared with uninfected or

AdvO-infected cells). In contrast, IL-6 was very potently inhibited (85%; Figure 16C).

These results agree with previous findings pointing out IL-6 as the most strongly NF-

/ B-dependent proinflammatory cytokine.

In a separate series of experiments, aliquots of medium were removed after 12h. 24h,

48h, 72h and 1 lOh, and cytokine analysis performed. It was seen (Figure 17) that

although the production of TNFα and IL-6 from uninfected or AdvO-infected cells

gradually increased between day 1 and day 5, the cytokine production from

AdvI/cBα-infected cells remained almost static, indicating that once sufficient

overexpression of cBα had been achieved to inhibit NF-κB activation, only very little

TNFα and IL-6 is produced.

I/ Bα Overexpression does not affect the Production of the Major Anti-inflammatorv Cvtokines

Previous results from human macrophages - major producers of IL-10 in the rheumatoid

joint - have indicated that both IL- 10 and the IL-1 receptor antagonist are

NF-/ B- independent in these cells, although the profound downregulation of TNFu and

IL-1/? induced by NF-/ B downregulation causes some secondary diminuation. In

rheumatoid synovial cell cocultures, the spontaneous production of both IL- 10 and the

IL-1 receptor antagonist are unaffected by overexpression of Bα (Figure 18A-B ). The

production of the p75 soluble TNF receptor is moderately (45%) inhibited, however

(Figure 18C). I/cBα Overexnression inhibits the Production of Metalloproteases

It is seen (Figure 19A-B) that IκBα overexpression potently inhibits the spontaneous production of both MMP-1 and MMP-3 in rheumatoid synovial cells. In contrast, the spontaneous production of TIMP-1 is slightly potentiated after 2 days (Figure 19C).

Furthermore, while studying a timecourse of MMP-production (Figure 20A-B), it appears that although uninfected and AdvO-infected rheumatoid synovial cell cocultures continue producing these enzymes throughout the incubation period, d e

AdvIκBα-infected cells produce very little of either MMP-1 or MMP-3 after significant cBα overexpression has been achieved. Again, there is a remarkable contrast to the relatively strong potentiation of TTMP-l production after 3 or 5 days (Figure 20C).

These results indicate a direct dependence of NF-κB for the production of both MMP-1 and MMP-3, but not for TTMP-l . Both IL-1 and TNFα are known inducers of MMPs, however, and to rule out any influence of the downregulation of these two cytokines on MMP production, a series of experiments was performed with addition of IL-lra and/or a TNF neutralizing antibody just after infection of cells. It was seen (Figure 21) that the relative inhibition of MMPs by Bα overexpression was largely unaffected by neutralization of eitiier IL-1 or TNFα. This speaks in favour of the regulation of MMPs being directly NF-κB regulated; the inhibition of TNFα production in our cultures adds to it, since the MMPs are partially TNFα-driven, but the major part of the reduction of MMPs can be explained by a direct effect on their induction.

Discussion

These results strongly indicate the NF-κB is a key therapeutic target in rheumatoid arthritis, and probably other inflammatory diseases. The unique spectrum of effects achieved by inhibiting this transcription factor includes a potent inhibition of

proinflammatory cytokines, particularly TNFα and IL-6, but no effect on the key

anti-inflammatory mediators EL- 10 and the IL-1 receptor antagonist. This would serve

to readjust the disturbed balance between pro- and anti-inflammatory mediators in the

diseased joint tissue. In addition, there is potent inhibition of the major destructive

enzymes MMP-1 and MMP-3, but instead a potentiating effect on their major inhibitor,

TIMP-1, indicating that NF-κB inhibition serves also to redress the disturbed balance

between MMPs and their inhibitors.

Claims

1. An in vitro method for infecting one or more cells having an elevated integrin level with a viral vector capable of transporting exogenous or

recombinant nucleic acid into me cells to be infected comprising the steps of: a) collecting the cells to be infected; and

b) infecting the collected cells with a viral vector.

2. A method according to claim 1, wherein the cells are treated with at least one cytokine prior to infection with a viral vector.

3. A method according to claim 2 wherein the exogenous cytokine is one or both

of GM-CSF and/or M-CSF.

4. A method according to any preceding claim wherein the cells are suspended

and subjected to elutriation prior to collection.

5. A method according to claims 1 or 4, wherein the source of cells having an elevated cytokine level is selected from rheumatoid synovial cells, broncheoalveolar

lavage from inflammatory respiratory diseases such as asthma and chronic obstructive respiratory disease, and gut biopsies of inflammatory bowel diseases such

as Crohn's Disease and ulcerative colitis.

6. A method according to any preceding claim, wherein the cells are selected

from one or more of macrophages, T-lymphocytes, synoviocytes, dendritic cells, chondrocytes, fibroblasts, epithelial cells and monocytes.

7. A method according to any preceding claim, wherein the viral vector used is

an adenovirus.

8. A method according to claim 7, wherein the viral vector used is a replication-deficient adenovirus.

9. A method according to any previous claim wherein the m.o.i. is 10-1000.

10. A method according to any previous claim wherein the exogenous or recombinant nucleic acid is DNA.

11. A method according to any previous claim wherein the exogenous or

recombinant nucleic acid comprises a nucleic acid sequence encoding a gene product

of interest, operably linked to one or more regulatory sequences to enable the gene to be expressed witiiin the infected cell.

12. A cell infected with a viral vector capable of transporting exogenous or

recombinant nucleic acid into the cells infected obtainable by a method according to any preceding claim.

13. Use of an in. vitro method or a cell according to any previous claim to study the effect of one or more exogenous compounds on the virally infected cell.

14. Use of a method according to any preceding claim for gene merapy.

15. A method of gene therapy comprising infecting one or more cells having an elevated

integrin level with a viral vector containing exogenous or recombinant nucleic acid operably linked to a promotor into the cell to be treated.

16. Use of a viral vector containing exogenous or recombinant nucleic acid in the manufacture of a medicament to treat rheumatoid arthritis, inflammatory respiratory diseases

such as asthma and chronic obstructive pulmonary disease, and inflammatory bowel diseases such as Crohn's disease and ulcerative colitis.