Compounds Chloroquine, and Structurally Related Cpg

Home , Amodiaquine, Chloroquine

Antagonism of Immunostimulatory CpG-Oligodeoxynucleotides by Quinacrine, Chloroquine, and Structurally Related Compounds This information is current as of October 2, 2021. Donald E. Macfarlane and Lori Manzel J Immunol 1998; 160:1122-1131; ; http://www.jimmunol.org/content/160/3/1122 Downloaded from References This article cites 42 articles, 11 of which you can access for free at: http://www.jimmunol.org/content/160/3/1122.full#ref-list-1

Why The JI? Submit online. http://www.jimmunol.org/

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on October 2, 2021 Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 1998 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Antagonism of Immunostimulatory CpG-Oligodeoxynucleotides by Quinacrine, Chloroquine, and Structurally Related Compounds1

Donald E. Macfarlane2 and Lori Manzel

Phosphorothioate oligodeoxynucleotides containing CpG (CpG-ODN) activate immune responses. We report that quinacrine, chloroquine, and structurally related compounds completely inhibit the antiapoptotic effect of CpG-ODN on WEHI 231 murine B lymphoma cells and inhibit CpG-ODN-induced secretion of IL-6 by WEHI 231. They also inhibit IL-6 synthesis and thymidine uptake by human unfractionated PBMC induced by CpG-ODN. The compounds did not inhibit LPS-induced responses. Half- maximal inhibition required 10 nM quinacrine or 100 nM chloroquine. Inhibition was noncompetitive with respect to CpG-ODN. Quinine, quinidine, and primaquine were much less powerful. Quinacrine was effective even when added after the CpG-ODN. Downloaded from Near-toxic concentrations of ammonia plus bafilomycin A1 (used to inhibit vesicular acidification) did not reduce the efficacy of the quinacrine, but the effects of both quinacrine and chloroquine were enhanced by inhibition of the multidrug resistance efflux pump by verapamil. Agents that bind to DNA, including propidium iodide, Hoechst dye 33258, and coralyne chloride did not inhibit CpG-ODN effect, nor did 4-bromophenacyl bromide, an inhibitor of phospholipase A2. Examination of the structure- activity relationship of seventy 4-aminoquinoline and 9-aminoacridine analogues reveals that increased activity was conferred by bulky hydrophobic substituents on positions 2 and 6 of the quinoline nucleus. No correlation was found between published http://www.jimmunol.org/ antimalarial activity and ability to block CpG-ODN-induced effects. These results are discussed in the light of the ability of quinacrine and chloroquine to induce remission of rheumatoid arthritis and lupus erythematosus. The Journal of Immunology, 1998, 160: 1122–1131.

acterial DNA is increasingly being recognized as a pow- an unknown mechanism (12, 13). These drugs bind to DNA by erful modulator of immunity (1, 2), stimulating the poly- intercalation (14), and they are weak bases and partition into acidic B clonal proliferation of B cells and the production of cy- vesicles. At high concentration, chloroquine can collapse the pH tokines by monocytes and other cells (3–5). This activity is gradient and disrupt the action of endosomal hydrolytic enzymes attributed to unmethylated CpG sequences in bacterial DNA, and the trafficking of receptors (15). by guest on October 2, 2021 which are uncommon in vertebrate DNA (6). Selected single- Here we report that quinacrine, chloroquine, and similar 9-ami- stranded oligodeoxynucleotides (especially those synthesized with noacridine and 4-aminoquinoline compounds completely block the a nuclease-resistant phosphorothioate backbone) that include the immunostimulatory action of CpG-ODN at concentrations much dinucleotide CpG (CpG-ODN)3 mimic many of the actions of bac- below those needed for other reported anti-inflammatory effects of terial DNA (7) and powerfully inhibit the induction of apoptosis in these compounds in vitro. We discuss whether inhibition of im- mouse WEHI 231 B cells by Ab to cell surface IgM (5, 8). This mune responses triggered by CpG-ODN could account for the abil- system is a convenient and reproducible model to study responses ity of these agents to induce remissions of rheumatoid arthritis and to CpG-ODN. systemic lupus erythematosus, and whether further drug develop- Bacterial DNA immobilized on beads does not stimulate im- ment optimizing this activity might lead to more useful remittive mune responses, suggesting that internalization of the DNA is re- and anti-inflammatory agents. quired for activity (6). DNA and oligodeoxynucleotides are endo- cytosed into acidic vesicles and are then transported to the cytoplasm and nuclei of cells (9–11). Materials and Methods Chloroquine, hydroxychloroquine, and quinacrine induce remis- Methods sions of systemic lupus erythematosus and rheumatoid arthritis by Cell culture. WEHI 231 cells (ATCC CRL 1702; American Type Culture Collection, Rockville, MD) were grown in log phase in medium consisting Department of Medicine, Veterans Affairs Medical Center and University of Iowa, of RPMI 1640 medium supplemented with 10% FBS (heat inactivated at Iowa City, IA 52242 65°C for 1 h), 0.3 g/L L-glutamine, 50 mg/L gentamicin, and 0.05 mM 2-ME in a 37°C humid atmosphere with 5% CO . Cells were diluted with Received for publication June 23, 1997. Accepted for publication October 16, 1997. 2 0.4% trypan blue in PBS, and live cells were counted in a hemacytometer The costs of publication of this article were defrayed in part by the payment of page chamber. charges. This article must therefore be hereby marked advertisement in accordance Human unfractionated PBMC were prepared from blood anticoagulated with 18 U.S.C. Section 1734 solely to indicate this fact. with heparin. The mononuclear cells were isolated using Histopaque 1077 1 This work was supported by a Merit Review Grant to D.E.M. from the Department (Sigma Chemical Co., St. Louis, MO) and resuspended at 2 ϫ 106/ml in of Veterans Affairs. RPMI 1640 medium supplemented with 10% autologous serum (heat in- 2 Address correspondence and reprint requests to Dr. Donald Macfarlane, Internal activated 56°C, 1 h). Medicine, University of Iowa, Iowa City, IA 52242. E-mail address: donald- Growth curves. Cells were resuspended in medium at 2 ϫ 105/ml in 24- [email protected] well tissue culture plates (Costar, Cambridge, MA). Appropriate additions 3 Abbreviations used in this paper: CpG-ODN, oligodeoxynucleotides containing were made, and the cells were incubated and counted daily. Cells were CpG motif; ; ␣-sIgM, goat anti-mouse surface IgM. diluted with fresh medium if the concentration exceeded 7 ϫ 105/ml.

[3H]Thymidine uptake. Cells (200 ␮l in duplicate or triplicate) were in- (substituted 4-methanolquinolines) and primaquine (a substituted cubated in a 96-well plate (Costar). Additions were made as indicated. 8-aminoquinoline) had little effect on CpG-ODN action at concen- 3 After incubation, [methyl- H]thymidine (2 Ci/mmol, DuPont/NEN, Bos- trations less than 10 ␮M (Fig. 1). Quinacrine and chloroquine were ton, MA) was added to each well, and the cells were incubated for 4 h (WEHI 231 cells) or 18 h (human PBMC). Cellular DNA was then cap- selected for further study. tured onto glass fiber filters (no. 934AH, Whatman Ltd., Maidstone, U.K.) using a cell harvester (Brandel, Gaithersburg, MD). The filters were dried, Effect on growth and apoptosis placed into plastic scintillation vials with 10 ml scintillation mixture (EconoSafe, Research Products International, Mount Prospect, IL), and Quinacrine alone influenced neither cell growth nor cell death in- counted in a liquid scintillation counter (no. LS-3133 T, Beckman Instru- duced by ␣-sIgM. CpG-ODN protects cells against ␣-sIgM-in- ments, Fullerton, CA). The 3H recovered on the filter was expressed as a duced apoptosis. Quinacrine reversed this protection by ODN 3 fraction of the H added. 1760, revealed by measurement of both cell growth (Fig. 2A) and DNA fragmentation assay. A DNA fragmentation assay was performed as previously described (8). In brief, WEHI 231 cells were incubated with DNA fragmentation (Fig. 2B). Similar results were obtained with [3H]thymidine and resuspended in fresh medium. After incubation for chloroquine (data not shown). These data show that the effect of 16–18 h, with additions as indicated, the cells were centrifuged, washed, antimalarials on thymidine uptake is a true reflection of their effect and resuspended in 20 ␮l PBS. This suspension was drawn into a pipette on cell growth and protection from apoptosis. tip containing 25 ␮l 2% SDS, and the mixture was immediately ejected into a well of a 0.8% agarose TBE gel containing 0.5 ␮g/ml ethidium bromide. The gel was electrophoresed1hat65V,andthelanes of the gel were cut Noncompetitive kinetics into three approximately equal portions. The first portion included the gel In the experiment shown in Figure 3, we measured the inhibition well and the first 2 mm of the running lane. The second and third pieces 3 by ␣-sIgM of thymidine incorporation in the presence of a range of were the remainder of the lane. The amount of fragmented DNA ( Hinthe Downloaded from second and third gel pieces) was expressed as a percentage of the total in concentrations of ODN 1760 and quinacrine. The data show that the lane. increasing the concentration of ODN 1760 does not overcome the Assay of IL-6. WEHI 231 cells (2 ϫ 105/ml) were incubated with or inhibitory effect of quinacrine, suggesting that the antimalarials do ␮ without ODN 1760 (6 g/ml) and with or without chloroquine or quina- not compete with CpG-ODN for a common site, such as a receptor, crine. After 24 h, the IL-6 content of the supernatant was assayed by ELISA using two mAbs in a sandwich technique as described by the man- an enzyme active site or a transport protein. ufacturer, using the standards supplied (PharMingen, San Diego, CA). IL-6 production by PBMC was assayed similarly, using Abs supplied by R&D Subsequent addition http://www.jimmunol.org/ Systems (Minneapolis, MN). We suspected that antimalarials might interfere with the transport Materials of ODNs to an intracellular site of action, in which case antima- CpG-ODN 1760 has the sequence 5Ј-ATAATCGACGTTCAAGCAAG-3Ј larials would prevent, but not to reverse, the effect of (nuclease- synthesized with a phosphorothioate or (when indicated) with a phosphodi- resistant) CpG-ODNs. This prediction does not appear to be cor- ester backbone. The fluorescent ODN was ODN 1760 with fluorescein rect. Figure 4 shows that quinacrine blocks the action of the linked to the 5Ј terminus. The oligonucleotides were purchased from phosphorothioate CpG-ODN even when added 8 h after the CpG- Genosys (The Woodlands, TX). ODN, suggesting that quinacrine interferes with the mechanism The following were all purchased from Sigma Chemical Co.: goat anti- mouse surface IgM (␣-sIgM, cat. M8644), LPS (Escherichia coli serotype the cell uses to recognize CpG-ODN (or with the subsequent sig- by guest on October 2, 2021 0127:B8 cat. L3129), amodiaquine, chloroquine, quinacrine, primaquine, nal-response coupling) rather than transport. quinidine, quinine, coralyne chloride, and thapsigargin. Hydroxychloro- quine was purchased from Copley Pharmaceutical, Inc. (Canton, MA). Ve- Action on phosphodiester ODN rapamil was from American Reagents Laboratory (Shirley, NY). C2-ceramide was purchased from Matreya, Inc. (Pleasant Gap, PA). Acridine The CpG-ODN used in typical experiments was synthesized on a homodimer (bis-(6-chloro-2-methoxy-9-acridinyl)spermine) and 9-amino- nuclease-resistant phosphorothioate backbone. The phosphodiester 6-chloro-2-methoxyacridine were purchased from Molecular Probes (Eu- ODN with the same base sequence as ODN 1760 is considerably less gene, OR). The other antimalarials listed in Tables I and II were generous gifts from Dr. Jill Johnson (National Cancer Institute). potent than ODN 1760. Figure 4 shows that quinacrine is effective in blocking the action of phosphodiester ODN, although the inhibitory Results effect is less complete than with phosphorothioate ODNs. Suppression of apoptosis by CpG-ODN Effect on CpG-ODN uptake WEHI 231 murine B cells are killed by engagement of ␣-sIgM. This phenomenon is believed to simulate the deletion of self-re- We evaluated the effect of chloroquine on the cellular incorpora- active clones during B cell ontogeny. We and others have shown tion of fluorescein-labeled ODN 1760 by fluorescence microscopy that single-stranded phosphorothioate CpG-ODNs reverse the in- of WEHI-231 cells after they were fixed with paraformaldehyde. hibition by ␣-sIgM of thymidine incorporation by WEHI 231 Preliminary experiments showed that this fixation had little effect cells (5, 8). on the distribution or the intensity of the fluorescence compared In Figure 1, we show that several antimalarial drugs added si- with live cells. Fixation was used to eliminate the potential of multaneously with the phosphorothioate CpG-ODN block the ef- chloroquine to alter fluorescence intensity by altering vesicular pH. fect of the CpG-ODN. WEHI 231 cells were incubated with The fluorescein-labeled CpG-ODN is taken up into numerous ␣-sIgM, ODN 1760, and the indicated concentration of the anti- small perinuclear structures, and this uptake is not inhibited by malarial. After 16 h, [3H]thymidine was added. In the absence of chloroquine; indeed, uptake appeared to be enhanced (Fig. 5). other additions, ␣-sIgM powerfully inhibits [3H]thymidine incorporation. This inhibition is almost completely relieved by ODN Inhibition of acidification 1760. Figure 1 also shows that quinacrine, hydroxychloroquine, 4-Aminoquinoline antimalarials partition into acidic vesicles (15). chloroquine, and amodiaquine (substituted 4-aminochloroquino- We wondered whether preventing this concentrating effect might lines) at nanomolar concentrations completely inhibited the CpG- reduce the potency of the antimalarials. We used a combination of ODN effect. Antimalarials added by themselves did not inhibit ammonium acetate (a permeable cation that collapses proton gra- 3 [ H]thymidine incorporation unless added at concentrations in dients, at concentrations ranging up to 10 mM) and bafilomycin A1 well in excess of 30 ␮M (data not shown). Quinidine and quinine (an inhibitor of the vacuolar type of proton pumping ATPase (16, 1124 ANTIMALARIALS AND CpG-OLIGONUCLEOTIDES

FIGURE 1. Effect of antimalarials on ODN protection against ␣-sIgM- induced inhibition of thymidine uptake. The compounds were added at FIGURE 3. Inhibition of ODN effect by quinacrine is noncompetitive. ␮ ␮ indicated concentrations together with 6 g/ml ODN 1760 and 10 g/ml ␣-sIgM (10 ␮g/ml) and the indicated concentrations of ODN 1760 and ␣ ϫ 5 -sIgM to WEHI 231 cells at 2 10 /ml. The cells were incubated 16 h quinacrine were added simultaneously to WEHI 231 cells. Thymidine in- ␮ 3 Downloaded from at 37°C before the addition of 0.5 Ci [ H]thymidine. After an additional corporation was estimated at 16 h as described for Figure 1. 4 h incubation, the cellular DNA was harvested onto glass fiber filters. The dry filters were counted in a liquid scintillation counter. In the absence of additions, cells incorporated about 20% of the added radioactivity. This was reduced to about 1% by the addition of ␣-sIgM. In the absence of DNA binding agents antimalarials, ODN 1760 restored the uptake to about 15% of added radioactivity. The vertical axis shows the ratio of the uptake of cells incu- Quinacrine and chloroquine bind to dsDNA by intercalating be- bated with CpG-ODN and ␣-sIgM divided by uptake of cells incubated tween adjacent bases. We wondered whether other agents known http://www.jimmunol.org/ with CpG-ODN, each being measured with the indicated concentration of to bind to dsDNA could influence the action of CpG-ODN. We test compound. Each experimental point is the mean of at least duplicate found that neither the intercalating agent propidium iodide nor the determinations. minor groove binding Hoechst dye 33258 (both added at 5 ␮g/ml) influenced the action of CpG-ODN on WEHI 231 cells. Quinacrine also stabilizes triple-helix DNA, but coralyne chloride at 1 ␮M 17), up to 5 nM) to try to block vesicular acidification. This com- (which stabilizes triple-helix as effectively as quinacrine) (18) did bination of agents at the higher concentrations is toxic to the cells, not inhibit CpG-ODN responses (data not shown). reducing thymidine uptake. It marginally reduced the efficacy of by guest on October 2, 2021 the CpG-ODN to protect from apoptosis (without affecting the No effect of quinacrine on LPS response ability of ␣-sIgM to induce growth arrest), but the combination did The B cell repertoire is regulated by a number of factors that mod- not block the action of quinacrine (data not shown). ulate clonal deletion, including LPSs (endotoxin, LPS, derived

FIGURE 2. Quinacrine reverses blockade by ODN of ␣-sIgM-induced growth-arrest and DNA fragmentation. A, WEHI 231 cells (2 ϫ 105/ml) were incubated with or without 6 ␮g/ml ODN 1760, 10 ␮g/ml ␣-sIgM, or 0.1 ␮M quinacrine as indicated. Live cells were counted daily. The cells were diluted in medium if their concentration exceeded 7 ϫ 105/ml. Note that the ODN substantially protects cells from killing by ␣-sIgM and that quinacrine reverses this protection. B, 3H-Labeled WEHI 231 cells were incubated with no addition, or with simultaneous additions as indicated (10 ␮g/ml ␣-sIgM, 6 ␮g/ml ODN, 0.1 ␮M quinacrine). After 16 h, cells were loaded into wells of 0.8% agarose gel by lysing with 1% SDS, and DNA was electrophoresed1hat65 V. The percentage fragmentation was estimated as described in Materials and Methods. In this and other ﬁgures, error bars indicate 1 SD. The Journal of Immunology 1125

FIGURE 6. Lack of effect of quinacrine on protection by LPS. LPS was Downloaded from added at the indicated concentration with or without 10 ␮g/ml ␣-sIgM and 0.1 ␮M quinacrine. Note that the small degree of protection against ␣-sIgM-induced inhibition of thymidine uptake afforded by LPS is not FIGURE 4. Quinacrine time course and effect on phosphodiester ODN. inﬂuenced by quinacrine. WEHI 231 cells were incubated with ␣-sIgM and ODN 1760. At times thereafter, 0.1 ␮M quinacrine was added (left). Cells were incubated with 10 ␮g/ml ␣-sIgM and ODN 1760 (6 ␮g/ml), phosphorothioate, or a phos- http://www.jimmunol.org/ phodiester ODN with the same sequence as 1760 (PD ODN) (30 ␮Mat of LPS. Thus (at the concentration we used), quinacrine does not time0hand8h)(right). Thymidine incorporation was estimated at 16 h. block the detection of LPS and subsequent signal-response coupling in the same way as it does with CpG-ODN. from Gram-negative bacteria), which promote B cell growth and CpG-ODN blockade of apoptosis induced by other agents inhibit B cell apoptosis. Chloroquine and other antimalarials have We have previously shown that CpG-ODNs block apoptosis in been reported to block LPS-induced responses in mononuclear WEHI 231 cells induced by stimuli other than ␣-sIgM (8). To cells (19, 20). In the experiment shown in Figure 6, we examined determine whether the action of antimalarials is restricted to the inﬂuence of LPS on ␣-sIgM-induced suppression of thymidine ␣-sIgM-induced apoptosis, we examined the effect of quinacrine by guest on October 2, 2021 incorporation. Even at high concentrations, LPS has much less on the protection by ODN 1760 from cell death induced by other ␣ ability than ODN 1760 to reverse -sIgM-induced growth inhibi- agents. C2-ceramide mediates receptor-induced cell death (21) and tion in WEHI 231 cells. Quinacrine at a concentration that com- inhibits thymidine incorporation. Figure 7A shows that this effect pletely reverses the effect of ODN 1760 had no effect on the action of C2-ceramide is reduced by ODN 1760, and that quinacrine

FIGURE 5. Cellular uptake of fluorescein-labeled ODN 1760. WEHI 231 cells were incubated for 2 h with fluorescein-labeled ODN 1760 (6 ␮g/ml) with (right) or without (left) chloroquine (1 ␮M). The cells were washed twice with PBS and cytospun onto microscope slides (500 rpm, 5 min). The adherent cells were fixed by immersion of the slide in 2% paraformaldehyde for 1 h, dried, and photographed in mounting medium with a Bio-Rad (Richmond, CA) MRC-1024 confocal laser scanning imaging system using identical settings. Note that chloroquine increases the fluorescence. The photographs shown are typical of many fields, and essentially identical results were obtained on three separate occasions. 1126 ANTIMALARIALS AND CpG-OLIGONUCLEOTIDES Downloaded from http://www.jimmunol.org/

FIGURE 7. Quinacrine reverses protection by ODN of killing induced FIGURE 8. Inhibition of multidrug-resistance efﬂux pump. Cells (2 ϫ 5 by thapsigargin and C2-ceramide. A, WEHI 231 cells were incubated with 10 /ml) were incubated with verapamil, ODN 1760, and chloroquine (top) by guest on October 2, 2021 ␮ ␮ C2-ceramide as indicated with or without 6 g/ml ODN and 0.1 M quin- or quinacrine (bottom) at the concentrations shown for 20 h before the acrine, and thymidine incorporation was estimated. B, WEHI 231 cells measurement of thymidine incorporation. Note that verapamil increased were prelabeled with 0.5 ␮Ci/ml [3H]thymidine for 3 h and resuspended in the potency of the chloroquine and quinacrine. Each point is the mean of fresh medium containing 10 nM thapsigargin, 10 ␮g/ml ␣-sIgM, 6 ␮g/ml triplicates. ODN, and/or 0.1 ␮M quinacrine as indicated. Cells were lysed 24 h later, and the DNA was electrophoresed as described in Materials and Methods to estimate the percentage fragmentation of DNA. Note that quinacrine Structure-activity relationship reverses the protection by ODN against killing by thapsigargin or A large number of analogues of antimalarial compounds have been C2-ceramide. synthesized, many with structures similar to chloroquine or quinacrine (23, 24). We determined the activity of several of these compounds using the thymidine incorporation technique. Each blocks this protective effect. Thapsigargin induces an increase in compound yielded dose-response curves similar to those in Figure intracellular calcium, resulting in massive DNA fragmentation. 1. From these curves, we estimated the concentration of the test Figure 7B shows that quinacrine also blocks the protective effect of compound required for half maximal blockade of the CpG-ODN ODN 1760 against thapsigargin-induced DNA fragmentation. effect. The results are shown in Tables I and II, revealing that Thus, quinacrine seems to block the protective effect of phospho- bulky substituents on positions 2 and 6 of the quinoline ring tend rothioate CpG-ODN generally, not just the protection against at to enhance activity. ␣-sIgM-induced apoptosis. Figure 9 illustrates how some of these substituents on the quinoline nucleus impact on this activity. Inhibition of multidrug-resistance transport Several attempts have been made to determine the structural Multidrug resistance is caused by the active transport of drugs out requirements for antimalarial action (25–27). We tested seven of the cell. This transport mechanism is inhibitable by verapamil. compounds in common with these studies. We found no correla- Chloroquine-resistant variants of malaria accelerate efﬂux of the tion between the reported antimalarial activity of these seven com- drug, an effect reversed by verapamil (22). In Figure 8 we dem- pounds and their ability to inhibit CpG-ODN-driven responses. onstrate that verapamil increases the potency of chloroquine (Fig. 8, top) or quinacrine (Fig. 8, bottom). This results suggests that Effect on IL-6 synthesis these compounds act intracellularly, in a compartment of the cell Our strain of WEHI 231 cells does not produce substantial from which they are pumped by the multidrug-resistance mecha- amounts of IL-6 in response to LPS. We were able to select a clone nism. Verapamil has little effect on the action of CpG-ODNs at of WEHI 231 (designated 231M) by limiting dilution that was 10 ␮M. more responsive to LPS. Figure 10 shows that IL-6 production by The Journal of Immunology 1127

Table I. Activities of quinacrine analogs a

C1 C2 C3 C5 C6 C7 C8 9-Sidechain EC50(nM) Quinacrine H OMe H H Cl H H NHCHMe(CH2)3NEt2 14.1 26 -OMe---- - 10.2

25------NHCH(C6H5)(CH2)3NEt2 19.1 24------NHCH(3-pyridyl)(CH2)3NEt2 30.2 23------NHCH(C6H4F)(CH2)3NEt2 42.7 30------NHCH(C6H4OMe)(CH2)3NEt2 52.4

29------NHCH(1-naphth)(CH2)3NEt2 57.5 27------NHCH(2-thienyl)(CH2)3NEt2 60.3 28------NHCH(C6H4Cl)(CH2)3NEt2 60.3 59------NH2 110 58 Acridine homodimer 417 Downloaded from

a The structures are listed as changes from quinacrine or chloroquine. The symbol - indicates no change. Me and Et indicate -CH3 and -CH2CH3, respectively. Aromatic http://www.jimmunol.org/ substituents are in the para position except as indicated. The EC50 is the concentration (nM) required for half-maximal inhibition of CpG-ODN effect on thymidine uptake in the presence of ␣-sIgM (see Fig. 1). The left most column indicates our reference number. this clone induced by ODN 1760 is inhibited by chloroquine (Fig. when exposed to ODN 1760. Both effects are inhibited by quin- 10, top) and quinacrine (Fig. 10, bottom) at a much lower con- acrine and chloroquine. centration than IL-6 production induced by LPS. These data show that the effect of the antimalarials is speciﬁc for CpG-ODN. Discussion Effect on human PBMC Our results show that the antimalarials, quinacrine, chloroquine, by guest on October 2, 2021 The effects of the test compounds was not restricted to immortal and many structurally related 4-aminoquinolines and 9-aminoacri- mouse cells. Unfractionated human peripheral blood cells in- dines are extremely potent inhibitors of the action of CpG-ODNs. corporate thymidine (Fig. 11A) and produce IL-6 (Fig. 11B) Quinacrine, added in low nanomolar concentration, completely

FIGURE 9. Structure-activity relationship. The data are from Tables I and II. Each box contains our identiﬁcation number, the structural change, and ␣ the activity of the compound (EC50) The EC50 of each compound is the concentration (nM) at which the ODN effect on -sIgM in WEHI 231 cells is inhibited 50%. Substitutions in position 7 replace chlorine. 1128 ANTIMALARIALS AND CpG-OLIGONUCLEOTIDES

Table II. Activities of chloroquine analogs

C2 C3 C5 C6 C7 C8 Sidechain EC50 (nM) H H H H Cl H NHCHMe(CH2)3NEt2 110 Chloroquine

34 ------NHCH2CH(C6H4OMe)(CH2)2NEt2 21.9 43 - - - SCH2C6H5 -- - 24.0 44 C6H5 - - - OMe - - 30.2 33 ------NHCH2CH(C6H4Cl)(CH2)2NEt2 37.2 ¢ 69 C(Me) C - - NMe2 H- - 38.0 (Me)C6H4Cl 32 - Me - Me - - - 42.7

48 C6H5 ------51.3 51 - - - C6H5 H- - 67.6 60 - Me - - - - NH(CH2)4N(C4H9)2 70.8 42 - - - OC6H5 H- - 74.1

56 ------NHCH(CH2Cl)(CH2)3NEt2 75.1 7 ------NHCH2CH(C6H5)(CH2)2NEt2 79.4 18 ------NHC6H3POoNEt2 100 Amodiaquine 15 ------110 Chloroquine

36 - Me - - - - NHCH2CH(C6H5)(CH2)2NEt2 115 Downloaded from

35 - Me - - - - NH(CH2)4NEt2 135 ¢ 53 ------NHCHMeCH CHCH2NEt2 135 68 ------NH(CH2)3C(Me)2NMe2 135 61 - Me - - - - NHCH2CH(C6H4OMe)(CH2)2NEt2 138 9 - - Cl - H - - 145

50 - - - - OC6H5 - - 148 4 - - - - Br - - 151 Bromoquine http://www.jimmunol.org/

10 ------NH(CH2)4NMe2 151 12-C6H5 ---- - 158 64 - - - - OC6H5 - - 178

65 - Me - - - - NHCHCH(C6H3Cl2)(CH2)2NEt2 191 1 - Me - - - - - 200 Sontoquine

67 ------NH(CH2)2CH(OH)CH2NEt2 209 41 - Me - - OMe - - 229 11 - - Cl - - - - 275

62 - Me Br - H - - 309 by guest on October 2, 2021

16 ------NHCH(CH3)(CH2)3NEtEtOH 407 Hydroxy-chloroquine 13 - - - - F - - 417 40 - Me Me - H - - 479 46 - Me - Me H - - 525 14 - - - - OMe - - 631 3 - - - - Me - - 692 63 - Me - - H OMe - 1120 71 - Me - OEt H - - 1260

45 ------NCH2CH(C6H4OH)(CH2)2NEt2 1740 39 Me Me - - H - - 2630 37 - - - Me Me - - 3160 47 ------Ncyclopentane2-OH 5620 ϩ 8 ------NCHMe(CH2)3N CH3Et2 6010 2------CH2NHCHMe(CH2)3NEt2 10000

See Table I legend for details.

blocks the ability of CpG-ODN to protect WEHI 231 cells from other pro-apoptotic agents. The antimalarials inhibit CpG-ODN- growth arrest, DNA fragmentation, and inhibition of thymidine induced IL-6 production by WEHI 213 cells and by unfractionated incorporation induced by the engagement of its surface IgM and by human PBMCs, and they inhibit the mitogenic effect of CpG-ODN The Journal of Immunology 1129 Downloaded from http://www.jimmunol.org/

FIGURE 10. Inhibition of IL-6 production. A clone of WEHI 231 cells FIGURE 11. ODN-induced effects on human mononuclear cells. A, Hu- 6 (231 M) that is LPS sensitive was incubated with or without ODN 1760 man mononuclear cells (2 ϫ 10 /ml) from three different donors were by guest on October 2, 2021 and the additions indicated for 24 h. The IL-6 content of the supernatant incubated with or without ODN 1760 and the indicated concentration of was measured. quinacrine or chloroquine for 4 days. Thymidine incorporation (18-h pulse) was then measured. The figure indicates 3H uptake as a percentage of ODN treated cells. The 3H uptake per donor for control samples is as follows: donor 1, 4.31%; donor 2, 2.76%; and donor 3, 6.39%. Thymidine incorporation for untreated samples was less than 0.05%. Samples from three on human PBMC. This inhibitory effect is noncompetitive with donors were averaged, and error bars indicate 1 SD. B, Human PBMC from respect to CpG-ODN, and it occurs even when quinacrine is added three donors were incubated with or without ODN 1760 and the indicated after the CpG-ODN. The potency of chloroquine and quinacrine is concentration of test compounds. After 24 h, the concentration of IL-6 was magnified by inhibition of the multidrug-resistance pump, suggest- determined in the supernatant. The figure indicates percentage of IL-6 in ing that the site of their action is intracellular. The specificity of the reference to ODN treated cells. The IL-6 amounts for control sample per effect toward CpG-ODN is illustrated by the fact that quinacrine donor are as follows: donor 1, 203 pg.ml; donor 2, 71.5 pg.ml; and donor does not influence the ability of LPS to inhibit apoptosis in, or 3, 465 pg/ml. IL-6 measured from untreated cells was 0–21 pg/ml. Results induce IL-6 production by, WEHI 231 cells. These results can be from the three donors were averaged, and error bars indicate 1 SD. reviewed in the light of the known biologic effects in vitro of quinacrine and chloroquine.

Inhibition of lysosomal function Uptake of ODNs Being a diprotic weak base, chloroquine (and probably many of the DNA and oligonucleotides bind to cells and are internalized via other compounds we use) partitions into the acidified vesicles such acidified vesicles (9–11), and this internalization is thought to be as lysosomes (28). At high concentration, chloroquine can collapse required for their effect (6). Two lines of evidence make it unlikely the pH gradient of lysosomes (29, 30) and induce their swelling that the antimalarials acts by inhibiting this transport. Phosphoro- (30). Chloroquine delays the recycling of proteins to the cell sur- thioate ODNs persist in cells in undegraded form for many hours face from lysosomes, resulting in altered trafficking of lysosomal (11), but we find that quinacrine blocks the action of CpG-ODN enzymes and receptors (15). These effects on lysosomal function after internalization, suggesting that quinacrine blocks a later step. are seen only when cells are exposed to higher concentrations of More directly, chloroquine did not decrease (rather, it appeared to chloroquine than we use here, so it is improbable that the suppres- increase) the cellular uptake of fluorescein-labeled ODN. sion by the antimalarials of CpG-ODN responses can be attributed solely to these relatively gross actions, even though our experi- DNA binding ments with ammonia plus bafilomycin A1 do suggest that the CpG- Chloroquine and its congeners bind to dsDNA. Modeling suggests ODN response requires vesicular acidification. that the planar nucleus of the drugs intercalate between adjacent 1130 ANTIMALARIALS AND CpG-OLIGONUCLEOTIDES base residues, and the side chain amines interact with the phos- drugs inhibit the release of cytokines, they could be useful in treat- phate residues of opposing strands (14). The binding affinity of ing septic shock, inflammatory bowel disease, respiratory and chloroquine for DNA is not high (of the order of 104 MϪ1), and it other infections, and graft-vs-host disease amplified by virus or has little affinity for ssDNA. We found that the more powerful other infection. DNA intercalator, propidium iodide, is not an effective anti-CpG- ODN agent, and that acridine homodimer, which has extraordinar- Acknowledgments ily high affinity for DNA especially at acid pH (31), is less effec- We thank Dr. Arthur Krieg for his encouragement and many valuable tive than chloroquine. Quinacrine stabilizes triplex DNA, but discussions. We thank KWM for donating blood. We also thank Dr. Jill another triplex stabilizer, coralyne chloride (18, 32, 33) does not Johnson (Drug Synthesis and Chemistry Branch, National Cancer Institute) inhibit CpG-ODN effects. These experiments do not lend support for providing many of the antimalarial compounds used in this study and to the supposition that the antimalarials operate by binding to Ms. Jennifer Shultz for her technical assistance. CpG-ODN, although an action to block an intracellular CpG-ODN recognition site (at which duplex DNA could be formed) is not Note added in proof: Chloroquine and other inhibitors of vesic- improbable. ular acidification are recently shown to inhibit the generation of reactive oxygen species and NF␬B activation in leukocytes treated Antimalarial activity with CpG-ODN. A-K. Yi, R. Tuetken, T. Redford, M. Wald- The mechanism of the antimalarial action chloroquine and other schmidt, J. Kirsch, and A. M. Krieg. CpG motifs in bacterial DNA 4-aminoquinolines and 9-aminoacridines has been intensively activate leukocytes through the pH-dependent generation of reactive oxygen species. Journal of Immunology, in press. studied (34, 35). These agents are effective in the erythrocytic Downloaded from stage of malaria. Chloroquine binds with high affinity to hematin References (ferriprotoporphorin IX, a toxic metabolite generated during the digestion of hemoglobin by the parasite) and interferes with its 1. Krieg, A. M. 1996. Lymphocyte activation by CpG dinucleotide motifs in pro- karyotic DNA. Trends Microbiol. 4:73. polymerization and detoxification (35). The structure-activity re- 2. Pisetsky, D. S. 1996. Immune activation by bacterial DNA: a new genetic code. lationship we report does not correlate with published antimalarial Immunity 5:303. 3. Ballas, Z. K., W. L. Rasmussen, and A. M. Krieg. 1996. Induction of NK activity

activity (25, 26), suggesting that the inhibition of CpG-ODN- http://www.jimmunol.org/ in murine and human cells by CpG motifs in oligodeoxynucleotides and bacterial driven responses involves a different mechanism than inhibition of DNA. J. Immunol. 157:1840. the malarial parasite. 4. Klinman, D. M., A.-K. Yi, S. L. Beaucage, J. Conover, and A. M. Krieg. 1996. CpG motifs expressed by bacterial DNA rapidly induce lymphocytes to secrete Remittive action IL-6, IL-12, and INF-␥. Proc. Natl. Acad. Sci. USA 93:2879. 5. Yi, A.-K., P. Hornbeck, D. Lafrenz, and A. M. Krieg. 1996. CpG DNA rescue of The antimalarial drugs chloroquine, hydroxychloroquine, and murine B lymphoma cells from anti-IgM induced growth arrest and programmed cell death is associated with expression of c-myc, c-myb, myn, and bcl-2. J. Im- quinacrine induce remissions of rheumatoid arthritis and systemic munol. 157:4918. lupus erythematosus (12, 13). The beneficial effect of these anti- 6. Krieg, A. M., A.-K. Yi, S. Matson, T. J. Waldschmidt, G. A. Bishop, R. Teasdale, malarials becomes apparent after several weeks of treatment. To G. A. Koretzky, and D. M. Klinman. 1995. CpG motifs in bacterial DNA trigger direct B-cell activation. Nature 374:546. account for this therapeutic action, a number of investigators have 7. Krieg, A. M., S. Matson, and E. Fisher. 1996. Oligodeoxynucleotide modifica- by guest on October 2, 2021 explored the effects of antimalarials on immune and inflammatory tions determine the magnitude of B cell stimulation by CpG motifs. Antisense responses in vitro, revealing a wide range of inhibitory actions. Nucleic Acid Drug Dev. 6:133. 8. Macfarlane, D. E., L. Manzel, and A. M. Krieg. 1997. Unmethylated CpG-con- Quinacrine inhibits phospholipase A2 (36–38), blocks ion chan- taining oligodeoxynucleotides inhibit apoptosis in WEHI 231 B-lymphocytes in- nels (39, 40), binds to receptors (41, 42), and inhibits FMLP-induced by several agents: evidence for blockade at a distal signaling step. Immu- nology 91:586. duced superoxide production and enzyme release by granulocytes 9. Zhao, Q., S. Matson, C. J. Herre, E. Fisher, H. Yu, and A. M. Krieg. 1993. (43). Chloroquine and quinacrine have little influence on other Comparison of cellular binding and uptake of antisense phosphodiester, phos- functional assays of neutrophils (44). Chloroquine inhibits cyto- phorothioate, and mixed phosphorothioate and methylphosphonate oligonucleotides. Antisense Res. Dev. 3:53. kine release from mononuclear cells induced by endo- and exo- 10. Bennet, R. M., G. T. Gabor, and M. M. Merritt. 1985. DNA binding to human toxins (19, 20, 45) and inhibits binding of inositol 1,4,5-trisphos- leukocytes: evidence for a receptor-mediated association, internalization, and phate (IP ) to its intracellular receptor (46). Each of these degradation of DNA. J. Clin. Invest. 76:2182. 3 11. Tonkinson, J. L., and C. A. Stein. 1994. Patterns of intracellular compartmen- described effects requires concentrations of chloroquine or quina- talization, trafficking, and acidification of 5Ј-fluorescein labeled phosphodiester crine markedly higher than the concentration we use to inhibit and phosphorothioate oligodeoxynucleotides in HL60 cells. Nucleic Acids Res. 22:4268. CpG-ODN driven responses in WEHI 231 cells and none of them 12. Fox, R. I. 1993. Mechanism of action of hydroxychloroquine as an antirheumatic satisfactorily explains the remittive effect of these drugs on auto- drug. Semin. Arthritis Rheum. 23:82. immune disorders. 13. Wallace, D. J. 1994. Antimalarial agents and lupus. Rheum. Dis. Clin. North Am. 20:243. CpG-ODNs are powerful activators of immunity in vitro. The 14. Hahn, F. E. 1974. Chloroquine (Resochin). In Antibiotics III: Mechanism of Ac- role of DNA from pathogens in the regulation of immunity is un- tion of Antimicroal and Antitumor Agents. J. Corcoran, and F. Hahn, eds. defined (2, 47), but it has been speculated that autoimmunity may Springer, Berlin, p. 58. 15. Chou, S.-C., K. A. Conklin, M. R. Levy, and D. C. Warhurst. 1974. Surrogate be influenced by circulating pathogen DNA (48). Indeed, bacterial models for antimalarials. In Antibiotics III: Mechanism of Action of Antimicroal DNA provokes the synthesis of anti-mammalian dsDNA Abs re- and Antitumor Agents. J. Corcoran and F. Hahn, eds. Springer, Berlin, p. 281. 16. Bowman, E. J. A., A. Siebers, and K. Altendorf. 1988. Bafilomycins: a class of sembling spontaneous autoantibodies (49). The exquisite sensitiv- inhibitors of membrane ATPases from micoorganims, animal cells and plant ity of CpG-ODN-driven responses to quinacrine, chloroquine, and cells. Proc. Natl. Acad. Sci. USA 85:7972. their analogues prompts us to speculate that the remittive effects of 17. Tapper, H., and R. Sundler. 1995. Bafilomycin A1 inhibits lysosomal, phagoso- mal, and plasma membrane Hϩ-ATPase and induces lysosomal enzyme secretion this class of drugs may be due to the blockade of the effect of in macrophages. J. Cell. Physiol. 163:137. pathogen DNA. If this is correct, our in vitro assays provide the 18. Wilson, W. D., S. Mizan, F. A. Tanious, S. Yao, and G. Zon. 1994. The inter- basis for a search for more useful drugs for treating rheumatoid action of intercalators and groove-binding agents with DNA triple-helical structures: the influence of ligand structure, DNA backbone modifications and se- arthritis and systemic lupus erythematosus. In addition, drugs that quence. J. Mol. Recognit. 7:89. inhibit B cell production (and hence Ab production) could be use- 19. Mattson, E., H. Van Dijk, K. Van Kessel, J. Verhoef, A. Fleer, and J. Rollof. 1996. Intracellular pathways involved in tumor necrosis factor-alpha release by ful in conditions caused by humoral immunity including autoim- human monocytes on stimulation with lipopolysaccharide or staphylococcal pep- mune hemolytic anemia and acquired hemophilia. If the resulting tidoglycan are partly similar. J. Infect. Dis. 173:212. The Journal of Immunology 1131

20. Jacobsen, P. H., T. Staales, K. Bendtzen, and D. Sturcher. 1995. The antimalarial 36. Lapetina, E. G., M. M. Billah, and P. Cuatrecasas. 1981. The phosphoinositol drug, Ro 42–1611 (arteﬂene), does not affect cytoadherence and cytokine-induc- cycle and the regulation of arachidonic acid production. Nature 292:367. ing properties of Plasmodium falciparum malaria parasites. Trop. Med. Parasitol. 37. Roberts, P. J., S. L. Williams, and D. C. Linch. 1996. The regulation of neutrophil 46:88. phospholipase A2 by granulocyte-macrophage colony-stimulating factor and its 21. Hannun, Y. A. 1996. Functions of ceramide in coordinating cellular responses to role in priming superoxide production. Br. J. Haematol. 92:804. stress. Science 274:1855. 38. Tollet, P., M. Hamberg, J. A. Gustafsson, and A. Mode. 1995. Growth hormone 22. Pussard, E., and F. Verdier. 1994. Antimalarial 4-aminoquinolines: mode of ac- signaling leading to CYP2C12 gene expression in rat hepatocytes involves phos- tion and pharmacokinetics. Fundam. Clin. Pharmacol. 8:1. pholipase A2. J. Biol. Chem. 270:12569. 23. Wiselogle, F. Y. 1946. A Survey of Antimalarial Drugs, 1941–1945. 39. Mironov, S. L., and H. D. Lux. 1992. The selective action of quinacrine on J. W. Edwards, Ann Arbor. high-threshold calcium channels in rat hippocampal cells. Br. J. Pharmacol. 106: 24. Peters, W., and W. H. G. Richards. 1984. Antimalarial drugs. Handbook of Ex- 751. perimental Pharmacology. Springer-Verlag, Berlin, 68/1, 68/2. 40. Sakuta, H., and I. Yoneda. 1994. Inhibition by SKF 525A and quinacrine of 25. Bass, G. E., D. R. Hudson, J. E. Parker, and W. P. Purcell. 1971. Interpretation endogenous glibenclamide-sensitive Kϩ channels in follicle-enclosed Xenopus of antimalarial activity in terms of regression analyses, molecular orbital calcu- oocytes. Eur. J. Pharmacol. 252:117. lations, and theory of DNA-drug binding. Prog. Mol. Subcell. Biol 2:126. 41. Tamamizu, S., A. P. Todd, and M. G. McNamee. 1995. Mutations in the M1 26. Geary, T. G., A. A. Divo, and J. B. Jensen. 1987. Activity of quinoline-containing region of the nicotinic acetylcholine receptor alter the sensitivity to inhibition by antimalarials against chloroquine-sensitive and -resistant strains of Plasmodium quinacrine. Cell. Mol. Neurobiol. 15:427. falciparum in vitro. Trans. R. Soc. Trop. Med. Hyg. 81:499. 42. Stokke, M., E. M. Hagelin, C. Poulsson, R. Patel, Y. Haile, and O. Brors. 1992. 27. Manallack, D. T., and D. J. Livingstone. 1994. Structure-property correlations of Inhibition by amiloride and quinacrine of speciﬁc [3H]nitrendipine binding to rat a series of antimalarial chloroquine derivatives. Monatsh. Chem. 125:1259. cardiac membranes. J. Pharmacol. Exp. Ther. 260:1366. 28. Ohkuma, S., and B. Poole. 1981. Cytoplasmic vacuolation of mouse peritoneal 43. Foldes-Filep, E., and J. G. Filep. 1992. Mepacrine inhibits fMLP-induced acti- macrophages and the uptake into lysosomes of weakly basic substances. vation of human neutrophil granulocytes, leukotriene B4 formation, and fMLP J. Cell Biol. 90:656. binding. J. Leukocyte Biol. 52:545. 29. Poole, B., and S. Ohkuma. 1981. Effect of weak bases on the intralysosomal pH 44. Ferrante, A., R. Rowan-Kelly, W. K. Seow, and Y. H. Thong. 1986. Depression in mouse peritoneal macrophages. J. Cell Biol. 90:665. of human polymorphonuclear leukocyte function by antimalarial drugs. Immu-

30. Myers, B., P. S. Tietz, J. E. Takara, and N. F. LaRusso. 1995. Dynamic mea- nology 58:125. Downloaded from surements of the acute and chronic effects of lysosomotropic agents on hepato- 45. Landewe, R. B., A. M. Miltenburg, F. C. Breedveld, M. R. Daha, and cyte lysosomal pH using ﬂow cytometry. Hepatology 22:1519. B. A. Dijkmans. 1992. Cyclosporine and chloroquine synergistically inhibit the 31. Capelle, N., J. Barbet, P. Dessen, S. Blanquet, B. Roques, and J-B. Le Pecq. 1979. interferon-gamma production by CD4 positive and CD8 positive synovial T cell Deoxyribonucleic acid bifunctional intercalators: kinetic investigation of the clones derived from a patient with rheumatoid arthritis. J. Rheumatol. 19:1353. binding of several acridine dimers to deoxyribonucleic acid. Biochemistry 18: 46. Misra, U. K.,G. Gawdi, and S. V. Pizzo. 1997. Chloroquine, quinine, and quin- 3354. idine inhibit calcium release from macrophage intracellular stores by blocking 32. Lee, J. S, L. J. P. Latimer, and K. J. Hampel. 1993. Coralyne binds tightly to both inositol 1,4,5-triphosphate binding to its receptor. J. Cell. Biochem. 64:225. T-A-TϪ and C-G-Cϩ containing DNA triplexes. Biochemistry 32:5591. 47. Krieg, A. M. 1996. An innate immune defense mechanism based on the recog- http://www.jimmunol.org/ 33. Latimer, L. J. P., N. Payton, G. Forsyth, and J. S. Lee. 1995. The binding of nition of CpG motifs in microbial DNA. J. Lab. Clin. Med. 128:128. analogues of coralyne and related heterocyclics to DNA triplexes. Biochem. Cell 48. Krieg, A. M. 1995. CpG DNA: a pathogenic factor in systemic lupus erythem- Biol. 73:11. atosus? J. Clin. Immunol. 15:284. 34. Wellems, T. E. 1993. How chloroquine works. Nature 355:108. 49. Wloch, M. K., A. L. Alexander, A. M. Pippen, D. S. Pisetsky, and G. S. Gilkeson. 35. Sullivan, D. J., I. Y. Glunzman, and D. E. Goldberg. 1996. Plasmodium hemo- 1997. Molecular properties of anti-DNA induced preautoimmune NZB/W mice zoin formation mediated by histidine-rich proteins. Science 271:219. by immunization with bacterial DNA. J. Immunol. 158:4500. by guest on October 2, 2021