JOURNAL OF INTERFERON RESEARCH 14:81-«5 (1994)Mary Ann Liebert, Inc., Publishers

Induction of Interferon and Other Cytokines by Imiquimodand Its Hydroxylated Metabolite R-842 in Human

Blood Cells In Vitro

CHARLES E. WEEKS and SHEILA J. GIBSON

ABSTRACT

The low-molecular-weight ¡mmunomodulator drug candidate, imiquimod (R-837), and its hydroxylated me-tabolite R-842, induce interferon-a (IFN-a) in human blood cells in vitro when tested in concentrations of 0.5|xg/ml or more. The amounts of IFN-a found increased with time from 2-6 h of incubation up to 24-48 h, andwere dependent on cell number and drug concentration. These two chemicals yielded more IFN-a in humanperipheral blood mononuclear cell (PBMC) cultures than other known inducers tested in parallel. They alsoinduced detectable amounts of interleukin (IL)-l, IL-6, IL-8, and tumor necrosis factor-a in human PBMCcultures in vitro.

INTRODUCTION

IMIQUIMOD (U.S. adopted name) or R-837 is a low-molecu-lar-weight imidazoquinolinamine that induces interferon

(IFN) formation in laboratory animals and in man when givenorally.(1~5) Imiquimod also demonstrates potent antiviral andantitumor activity in many animal models.(6~' " We have stud-ied the effects of this and of its principal active hydroxylatedmetabolite R-842 in human blood cells in vitro. Initial resultsshowed that these compounds induce the formation of IFN andother cytokines in cells.(212) In this manuscript, we confirmand expand these preliminary results, showing that these imid-azoquinolinamines are exceptional inducers of IFN-a and othercytokines in human peripheral blood mononuclear cells(PBMC) in vitro.

MATERIALS AND METHODS

Chemicals: R-837 imiquimod, (l-(2-methylpropyl)-lH-im-idazo[4,5-c]quinolin-4-amine) and R-842, (4-amino-a,a-di-methyl lH-imidazo[4,5-c]quinolin-l-ethanol) are proprietarychemicals of 3M Pharmaceuticals (St. Paul, Minnesota); tilo-rone (2,7-bis[2-(diethylamine) ethoxy]fluoren-9-one) was ob-tained from Richardson Merrell (Cincinnati, OH); 10-car-boxymethyl-9-acridone from Hoffman-LaRoche (Nutley, NJ);

phytohemaglutinin (PHA), concanavalin A (Con A), phorbolmyristate acetate (PMA), and lipopolysaccharide (LPS) fromSigma (St. Louis, MO); polyinosinic-polycytidylic acid (polyrLpoly rC) from Pharmacia (Piscataway, NJ). 2-Amino-5-bromo-6-phenyl-4(3H)pyrimidinone (ABPP) was synthesizedby Dr. John F. Gerster (3M Pharmaceuticals).

Other Reagents: The reference human recombinant leuko-cyte IFN-a (1,000 IU), A549 cells, and encephalomyocarditis(EMC) virus used for the IFN bioassay have been described.'13)Human lymphoblastoid IFN-a and human fibroblast IFN-ßwere from Lee BioMolecular Research, Inc. and usedfor immunochemical characterization assays.n3) EDTA (0.10ml or 15% EDTA(K3) solution per 10-ml tube) and heparin(143 USP units of sodium heparin per 10-ml tube) vacutainerswere obtained from Becton Dickinson (Rutherford, NJ). Peni-cillin-streptomycin solution containing 10,000 units/ml penicil-lin and 10 mg streptomycin/ml in 0.9% NaCl was obtained fromSigma (St. Louis, MO). RPMI-1640 medium containing 25mM HEPES (/V-2-hydroxyethylpiperazine-/v"-2-ethanesulfonicacid), 2 mM L-glutamine, and 2% penicillin (100 units/ml)-streptomycin (0.1 mg/ml) came from GIBCO (Grand Island,NY).

Cell Culture: Human whole blood was collected by veni-puncture from healthy volunteers into EDTA or heparin vacu-tainer tubes. Whole blood was diluted 1:10 with RPMI-1640.

3M Pharmaceuticals, Department of Pharmacology, 3M Center, 270-2S-06, St. Paul, MN 55144.

81

82 WEEKS AND GIBSON

PBMC were separated from whole blood with LeucoPREP cellseparation tubes (Becton Dickinson Labware, Lincoln Park,New Jersey) or Ficoll-Paque solution (Pharmacia). The PBMCswere suspended at 1 x 106/ml in RPMI complete mediumcontaining 10% heat-inactivated (56°C for 30 min) autologousserum.

IFN Induction Assay: Two hundred microliter portions ofthe PBMC suspension or whole blood were added to 96-wellflat-bottomed plates (Becton Dickinson Labware). The IFNinducers were solubilized in ethanol or dimethyl sulfoxide atfinal concentrations of <1%, or in tissue culture-grade water(used for the imidazoquinolamines), and then diluted in water,or 0.01 N sodium hydroxide or hydrochloric acid. Test com-pound was added in a volume of « 50 u.L to the wells contain-ing 200 pL of diluted whole blood or of PBMC in medium.Solvent and/or medium were added to control wells (i.e., withno test compound) to give a final volume of 250 u.1. The plateswere covered with plastic lids, vortexed gently, and incubated,usually for 48 h at 37°C in a 5% C02 atmosphere. The plateswere then covered with parafilm and centrifuged at 1,000 rpmfor 10-15 min at 4°C. Medium (approximately 200 u.1) was

removed from four to eight replicate wells, pooled, and frozenat —70°C until analyzed for cytokine levels.

IFNAssay: Amounts of IFN were determined at Lee BioMo-lecular by bioassay as described previously.<13> Briefly, IFNdilutions and A549 human lung carcinoma cells were incubatedat 37°C for 12-24 h, and then challenged with EMC virus. Afterfurther incubation, the viral cytopathic effect was quantified bystaining, followed by spectrophotometric absorbance measure-

ments. Results are expressed as IFN-a IU/ml in terms of theNIH human IFN leukocyte standard.

Type of IFN Induced: The induced IFN was identified as

essentially all IFN-a by testing in checkerboard neutralizationassays with rabbit anti-human IFN-ß and goat anti-humanIFN-a and A549 cell monolayers challenged with EMC virus.

Assays for Other Cytokines: Interleukin-la (IL-la), IL-2,IL-6, and IL-8 levels were determined by RIA (Advanced Mag-netics, Inc., Cambridge, MA); IL-lß levels by RIA (AmgenBiologicals, Thousand Oaks, California); tumor necrosis factor(TNF)-a by RIA [Centocor, Malvern, PA; in this test, there isno detectable cross-reactivity with lymphotoxin (TNF-ß)].Granulocyte-macrophage colony-stimulating factor (GM-CSF)was measured in an ELISA (Genzyme Corporation, Cam-bridge, MA). Levels of IL-la, IL-lß, IL-6, IL-8, and TNF-awere also verified by ELISA (R and D Systems, Minneapolis,MN).

RESULTS

Induction of IFN in whole blood cultures

Initially we modified the methods of Pauly et a/.(14) andKirchner et al.(,5) to establish a system to detect IFN inductionby R-837 and R-842 in vitro. IFN inductions in a whole blood

culture system incubated with imiquimod, R-842, and a knownIFN inducer poly rLpoly rC are shown in Table 1. R-842yielded 6-96 IFN-a IU/ml when added to these cultures at 2-10u.g/ml and the response was similar to that induced by imiqui-mod at 1-5 u.g/ml. Poly rbpoly rC yielded less IFN even whenadded at concentrations of 50-250 p-g/ml. Other chemicalsknown to induce IFN in certain animals, such as Tilorone,ABPP, acridone acetic acid, PHA, Con A, and PMA, did notinduce detectable IFN this test system.

IFN yields from individual donors

R-842, imiquimod, or poly rLpoly rC were tested in wholeblood cell cultures from volunteer donors (3 males and 2 fe-males). The results are shown in Table 2. R-842 at 3.0 and 6.0u.g/ml yielded from 86 to 330 IFN-a IU/ml and imiquimod at1.5 and 3.0 u-g/ml gave similar results. Poly rLpoly rC at 150p,g/ml induced 28 -84 IFN-a IU/ml. There was some variabil-ity between donors with regard to the amounts of IFN induced.

Induction of IFN in PBMC culturesTable 3 shows results of the IFN production using human

PBMC cultures stimulated with R-842 and imiquimod. R-842induced 660-3,400 IFN-a IU/ml with drug concentrations of1.0-5.0 p-g/ml being active. Imiquimod induced 6.2-1,500IFN-a IU/ml with drug concentrations of 0.5-5.0 u.g/ml. Ingeneral, PBMC cultures (containing 1 x 106/ml mononuclearcells) stimulated with R-842 and imiquimod provided higherIFN values at lower drug concentrations than the whole bloodcultures (containing approximately 2.5 x 105/ml mononuclearcells). PBMC cultures with cell concentrations less than 1 x

106/ml yielded smaller amounts of IFN (data not shown).When media obtained from human whole blood cultures and

PBMC cultures treated with R-842 or imiquimod were analyzed

Table 1. IFN Levels After 24 H in Medium from aHuman Whole Blood Culture

Compound (\xglml) IFN-a (IU/ml)ControlPoly rl: poly rC 50a

100150200250

Imiquimod

R-842

1234523456810

<2.718182442426.2

425555556.2

324255739673

a Previous studies demonstrated no IFN induction with <50.0 u.g/mlPoly rl: poly rC (unpublished results).

CYTOKINE INDUCTION IN VITRO BY IMIQUIMOD 83

Table 2. IFN Levels after 24 H in Medium from Human Whole Blood Cultures

IFN-a levels (IU/ml)'

Compound (\Lglml)A

maleB

maleC

maleD

femaleE

femaleImiquimodImiquimodR-842R-842Poly rLpoly rCControl

1.53.03.06.0

150

63648711028<1.8

61618611025<l.i

11011011011061<l.i

11014097

29061<1.8

1101659733084<1.8

"Mean of duplicate analyses on each of two replicate samples.

by antibody neutralization methods, it was shown that 95% or

more of the IFN formed was IFN-a.

Kinetics of IFN induction

In PBMC's incubated with imiquimod, IFN could be de-tected in the culture media between 2 and 6 h of culture, peakedbetween 24 and 48 h (cf. Table 3) and declined by 48-72 h. Insome experiments, there was no apparent linear dose response.

PBMCs were incubated with R-842 and imiquimod at con-centrations of 2.5 and 5.0 p-g/ml for periods of 0.5, 2, or 6 h,washed twice and then incubated for an additional 48 h. TheIFN yields are shown in Table 4. Two hours of exposure toR-842 was sufficient for cells to produce significant amounts ofIFN, and in this experiment, cells produced some IFN after just0.5 h of exposure to imiquimod.

Induction of additional cytokines by R-837 and R-842Table 5 shows the amount of IFN-a, IL-la and -ß, IL-6,

IL-8, and TNF-a induced in PBMC cultures by imiquimod,"R-842, and LPS, each tested at 0.5, 2.5, and 5 p-g/ml. Bothimiquimod and R-842 induced IL-la, IL-lß, IL-6, and IL-8 ina dose-dependent manner. Induction of TNF-a by both drugswas inconsistent, and in particular, imiquimod on occasionprovided detectable levels only at 3=5.0 u-g/ml. Except for IFNinduction, LPS was a more powerful cytokine inducer thaneither imiquimod or R-842.In our PBMC model, neither IL-2 nor GM-CSF were de-

tected in response to imiquimod or R-842 added at 0.5, 2.5, and5.0 p-g/ml. PMA/PHA yielded 10-15 units/ml in the IL-2 assayafter 24 h and 0.26-0.58 ng/ml GM-CSF after 48 h (data notshown).

Table 3. IFN Levels in Medium from HumanPBMC Cultures

Compound (\xglml)R-842

Imiquimod

ControlR-842

Imiquimod

Control

S0.51.03.05.0S0.30.51.03.05.0

S0.51.03.05.0S0.30.51.03.05.0

Culture time(h)

242424242424242424244X484848484848484848

IFN-a(IU/ml)

<2.06603400870<2.06.2

<2.015001100<2.0<2.06602600660<2.018<2.0660870<2.0

Table 4. IFN Levels at 48 H in Medium fromMononuclear Cell Cultures after Drug Exposure for

0.5, 2, and 6 H

Compound (\xglml)R-842 2.5

R-842 5.0

Imiquimod 2.5

Imiquimod 5.0

Control 0

Hours ofexposure

a

0.5260.5260.5260.5260.526

IFN-a(lU/mL)<98437=£94304306419014014057048<9<9<9

aCells were incubated with the drug solution for the time indicated,then washed twice and the medium sampled later at 48 h for IFN assay.

84 WEEKS AND GIBSON

Table 5. IFNa, IL-la, IL-lß, IL-6, IL-8, and TNFa Levels in Human PBC CuLTURESab

Sample/compound IFN-a IL-la IL-lß IL-6 IL-8 TNF-aconcentration (\kglml) (IU/ml) (pglml) (pglml) (pglml) (pglml) (pglml)ControlImiquimod

LPS

R-842

0.52.55.00.52.55.01.02.55.0

<12250330430

-28«28750750

97144146247338344422205214269

22041042571498611901290544663918

214132361802

l,855d

1,950N.D.C395820

3362,834e7,83312,53324,489

10,3853,2642,22811,680

N.D.CN.D.C

147477

2,7452,4902,817

88f321

1,555a Samples were collected after 24 h of incubation with drug in culture.bIL-la and -ß, IL-6, and TNF-a assays performed by RIA; IFN by bioassay; IL-8 by ELISA.CN.D. indicates value was below limits of detection from standard curve.d Value obtained using 1.0 (Jig/ml LPS.e Value obtained using 1.0 u.g/ml imiquimod.'Value was at the border line of detection.

DISCUSSION

Much research effort has focused on finding defined chemi-cal inducers of IFN, and several have been tested in man.

These include polyribonucleotides, tilorone, pyrimidinones,acridines, and flavone acetic acids (FAA). Despite being activein rodents, few have proved to be potent IFN inducers inman."6-18' Similarly, in vitro screening models for IFN or

cytokine induction have been used in attempts to define struc-

ture-activity relationships, mechanism(s) of action, and po-tency in the hope of obtaining data in a model that may bephysiologically relevant for drug development and/or clinicalutility."4'15-'9'Imiquimod, R-842, and other analogs were developed on the

basis of guinea pig and mouse models as well as in vitro humanPBMC assays. Here, we have shown that imiquimod and R-842are potent in vitro cytokine inducers in a model that may bephysiologically relevant. In initial clinical trials, levels of imi-quimod or R-842 in the serum exceeding 0.3-0.5 u,g/ml in-duced circulating IFN.(4) In contrast, poly rbpoly rC inducesIFN and other cytokines in human donor monocytes,<20) but isan ineffective or a weak inducer in man in vivo.{,6) In our

system, poly rLpoly rC is a very weak IFN inducer comparedwith imiquimod/R-842. Similarly, FAA and derivatives haveantitumor activity in animal models but have not shown a clini-cal effect in man and do not induce IFN in human PBMCs.(21)

We found that imiquimod did not induce IFN formation incells of the HL-60, U-937, THP-1, Vero, MRC-5, or Namwalalineage. Other cell lines, explored by Rosztoczy et al. ,(22) were 1

also nonresponsive. Thus far, imiquimod and R-842 seem to bethe most potent inducers of IFN-a in human blood cell cultures.The profile of cytokines induced by imiquimod and R-842 is

not unlike that observed for LPS stimulation of monocytes/macrophages, which, however, then produce very little IFN-a.It is therefore plausible to speculate that imiquimod and R-842 3are mainly stimulating monocytes. Monocyte/macrophage-likecell lines such as THP-1 produce IL-1, TNF-a, and IL-6 in

response to imiquimod (unpublished observation) but no detect-able IFN-a. Therefore, cells other than monocytes and mac-

rophages may also be contributing to the production of IFN-a.Indeed, Rosztoczy et al. have implicated the B lymphocyte as

partially responsible for IFN production.'22' Further studies are

in progress to determine which cells give rise to IFN-a whentreated with imiquimod, and also which give rise to the othercytokines formed.

In summary, we have shown that this in vitro PBMC modelcan be used to screen drugs for activity, and that imiquimod andR-842 are potent cytokine inducers in this model.

ACKNOWLEDGMENTS

The authors thank Dr. Lee Kronenberg of Lee BioMolecularfor his analytical work and helpful discussions. We also thankDr. John F. Gerster for synthesis of the inducers, Dr. RoyMcQuinn for metabolite profile assistance, and Drs. Mark To-mai and Richard Miller for help with this manuscript; also a

special thanks to Ms. Mary Nilan for her secretarial work.

REFERENCES

GIBSON, S.J., ELROD, S.V., MILLER, R.L., and WEEKS,C.E. (1990). Oral R-837 induces alpha interferon in cynomolgusmonkeys. J. Interferon Res. 10, S124.WEEKS, C.E., GIBSON, S.J., IMBERTSON, L.M., REITER,M.J., MILLER, R.L., and GERSTER, J.F. (1990). PreclinicalPharmacology and efficacy of immunomodulator R-837. J. Inter-feron Res. 10, S89.MILLER, R.L., IMBERTSON, L.M., REITER, M.J.,SCHWARTZMILLER, D.H., PECORE, S.E., and GERSTER,J.F. (1985V Inhibition of herpes simplex virus infection in a guinea

CYTOKINE INDUCTION IN VITRO BY IMIQUIMOD 85

pig model by S-26308. ASM, Interscience Conference on Antimi-crobial Agents and Chemotherapy, Minneapolis, MN, p. 235 (ab-stract).

4. WICK, K.A., KVAM, DC, WEEKS, CE., DIXON, RM.,HAFSTAD, K., and WITT, P.L. (1991). Oral R-837 inducesinterferon in healthy volunteers. Proc. Amer. Assn. Cancer Res.32, 257.

5. BORDEN, E., WITT, P., KVAM, D., WICK, K., RITCH, P.,ANDERSON, T., OWENS, M., SIDKY, Y., and GROSSBERG,S. (1991). An effective oral inducer of interferon in humans. J.Interferon Res. 11, S92.

6. HARRISON, C.J., JENSKI, L., VOYCHEHOVSKI, T., andBERNSTEIN, D.I. (1988). Modification of immunological re-

sponses and clinical disease during topical R-837 treatment ofgenital HSV-2 infection. Antiviral Res. 10, 209-224.

7. BERNSTEIN, D.I., and HARRISON, C.J. (1989). Effects of theimmunomodulating agent R-837 on acute and latent herpes sim-plex virus Type 2 infections. Antimicrobial Agents and Chemo-therapy 33, 1511-1515.

8. HARRISON, C.J., STANBERRY, L.R., and BERNSTEIN, D.I.(1991). Effects of cytokines and R-837, a cytokine inducer, on

UV-irradiation augmented recurrent genital herpes in guinea pigs.Antiviral Res. 15,315-322.

9. SIDKY, Y.A., BRYAN, G.T., WEEKS, CE., HATCHER,J.M., and BORDEN, E.C (1990). Effects of treatment with an

oral interferon inducer, imidazoquinolinamine (R-837), on thegrowth of mouse bladder carcinoma FCB. J. Interferon Res. 10,S123.

10. SIDKY, Y.A., BORDEN, EC, and WEEKS, CE. (1992). Inhi-bition of tumor-induced angiogenesis by the interferon inducerimiquimod. Proc. Amer. Assn. Cancer Res. 33, 458.

11. BORDEN, E.C, SIDKEY, Y.A., and WEEKS, CE. (1991).Mechanisms of anti-tumor action of the interferon inducer R-837.Proc. Amer. Assn. Cancer Res. 32, 258.

12. WEEKS, CE., and GIBSON, S.J. (1989). Alpha interferon in-duction in human blood cell culture by immunomodulator candi-date R-837. J. Interferon Res. 9, S215.

13. BRENNAN, G.L., and KRONENBERG, L.H. (1983). Auto-mated bioassay of interferons in micro-test plates. BioTechniques1, 78-82.

14. PAULY, J.L., SOKAL, JE., and HAN, T. (1973). Whole-bloodculture technique for functional studies of lymphocyte reactivity to

mitogens, antigens, and homologous lymphocytes. J. Lab. Clin.Med. 82, S500.

15. KIRCHNER, H., CLEINICKE, C, and DIGEL, W. (1982). Awhole-blood technique for testing production of human interferonby leukocytes. J. Immunol. Meth. 48, 213-219.

16. HAWKINS, M.J., LEVIN, M., and BORDEN, E.C. (1985). Aneastern cooperative oncology group phase I-II pilot study of poly-riboinosinic-polyribocytidylic acid poly-L-lysine complex in pa-tients with metastatic malignant melanoma. J. Biol. ResponseModif. 4, 664-668.

17. LITTON, G.J., HONG, R., GROSSBERG, SE., VECHLEKAR,D., GOODAVISH, C.N., and BORDEN, E.C. (1990). Biologicaland clinical effects of the oral immunomodulator 3,6-bis(2-piperi-dinoethoxy)acridine trihydrochloride in patients with malignancy.J. Biol. Response Modif. 9, 61-70.

18. RIOS, A., STRINGFELLOW, DA., FITZPATRICK, FA.,REELE, S.B., GUTKNECHT, G.D., and HERSH, E.M. (1986).Phase I study of 2-amino-5-bromo-6-phenyI-4(3H)-pyrimidinone(ABPP), and oral interferon inducer, in cancer patients. J. Biol.ResponseModif. S, 330-338.

19. DESCH, CE., KOVACH, N.L., PRESENT, W., BROYLES,C, and HARLAN, J.M. (1989). Production of human tumor ne-crosis factor from whole blood ex vivo. Lymphokine Res. 8, 141-146.

20. HAYES, M.P., ENTERLINE, I.C., GERRARD, T.L., andZOON, K.C (1991). Regulation of interferon production by hu-man monocytes: Requirements for priming for lipopolysaccharide-induced production. J. Leuk. Biol. 50, 176-181.

21. FUTAMI, H., EADER, LA., KOMSCHLIES, K.L., BULL, R.,GRUYS, M.E., ORTALDO, JR., YOUNG, H.A., and WIL-TROUT, R.H. (1991). Flavone acetic acid directly induces expres-sion of cytokine genes in mouse splenic leukocytes but not inhuman peripheral blood leukocytes. Cancer Res. 51, 6596-6602.

22. ROSZTOCZKY, I., RAJ, N.B.K., and PITHA, P.M. (1992).Studies on the in-vitro interferon-inducing activity of the smallmolecular weight compound R-837. J. Interferon Res. 12, SI41.

Address reprint requests to:Dr. Charles E. Weeks

Sandoz Nutrition1541 Park Place Blvd.

Minneapolis, MN 55416

Received 15 June 1993/Accepted 9 February 1994