Punica granatum (Pomegranate) Juice Provides an HIV‐1 Entry Inhibitor and Candidate Topical Microbicide

For ∼24 years the AIDS pandemic has claimed ∼30 million lives, causing ∼14,000 new HIV‐1 infections daily worldwide in 2003. About 80% of infections occur by heterosexual transmission. In the absence of vaccines, topical microbicides, expected to block virus transmission, offer hope for controlling the pandemic. Antiretroviral chemotherapeutics have decreased AIDS mortality in industrialized countries, but only minimally in developing countries. To prevent an analogous dichotomy, microbicides should be acceptable, accessible, affordable, and accelerative in transition from development to marketing. Already marketed pharmaceutical excipients (inactive materials of drug dosage forms) or foods, with established safety records and adequate anti‐HIV‐1 activity, may provide this option. Therefore, fruit juices were screened for inhibitory activity against HIV‐1 IIIB using CD4 and CXCR4 as cell receptors. The best juice was tested for inhibition of: (1) infection by HIV‐1 BaL, utilizing CCR5 as the cellular coreceptor, and (2) binding of gp120 IIIB and gp120 BaL, respectively, to CXCR4 and CCR5. To remove most colored juice components, the adsorption of the effective ingredient(s) to dispersible excipients and other foods was investigated. A selected complex was assayed for inhibition of infection by primary HIV‐1 isolates. The results indicate that HIV‐1 entry inhibitors from pomegranate juice adsorb onto corn starch. The resulting complex blocks virus binding to CD4 and CXCR4/CCR5 and inhibits infection by primary virus clades A to G and group O. Therefore, these results suggest the possibility of producing an anti‐HIV‐1 microbicide from inexpensive, widely available sources, whose safety has been established throughout centuries, provided that its quality is adequately standardized and monitored.

2004, there were 4.9 million new HIV infections and 3.1 million AIDS deaths. 1 To date, the number of individuals living with human immunodeficiency virus type 1 (HIV-1) infection/AIDS has reached 39.4 million, 1 and ∼28 million people have already died from AIDS since the beginning of the pandemic. 1,2 Most new infections have been acquired by the mucosal route, heterosexual transmission playing the major (∼80%) role. Although the incidence of transmission per unprotected coital act is estimated to be low (0.0001-0.004), but strikingly increased when acutely infected individuals are involved, 3,4 the cumulative effect is overwhelming.
Anti-HIV-1 vaccines applicable to global immunization programs are not expected to become available for many years. Therefore, other prevention strategies are urgently needed. This includes educational efforts and the application of mechanical and/or chemical barrier methods. The latter correspond to microbicides, that is, topical formulations designed to block HIV-1 infection (and possibly transmission of other sexually transmitted diseases) when applied vaginally (and possibly rectally) before intercourse. 3,[5][6][7] Conceptually, it is preferred that the active ingredient(s) of microbicide formulations (1) block virus entry into susceptible cells by preventing HIV-1 binding to the cellular receptor CD4, the coreceptors CXCR4/CCR5, and to receptors on dendritic/migratory cells (capturing and transmitting virus to cells that are directly involved in virus replication), respectively, 3,[8][9][10][11] and/or (2) are virucidal. The formulations must not adversely affect the target tissues and should not cause them to become more susceptible to infection after microbicide removal. 12,13 Treatment with antiretroviral drugs has decreased mortality from AIDS in industrialized countries but so far has had a minimal effect in developing countries. 14 To avoid a similar dichotomy with respect to microbicides, they should be designed and selected to become affordable and widely accessible, while shortening the time between research and development and their marketing and distribution as much as possible. This would be facilitated if mass manufactured products with established safety records were found to have anti-HIV-1 activity. Qualifying candidates to be considered for microbicide development may possibly be discovered by screening pharmaceutical excipients ("inactive" ingredients of pharmaceutical dosage forms) and foods, respectively, for antiviral properties.
While exploring the possibility that chemical modification of food proteins may lead to the generation of compounds with anti-HIV-1 activity, we discovered in 1994 that bovine β-lactoglobulin (the major protein of whey) modified by 3-hydroxyphthalic anhydride (3HP-β-LG) blocked infection by HIV-1 and herpesviruses, both in vitro and in animal model systems. [15][16][17][18][19][20][21][22][23][24] Considering its antiviral potency, ease of preparation, and practically unlimited and inexpensive source (the worldwide production of whey is approximately 86 billion kg annually), 3HP-β-LG appeared to represent an excellent candidate microbicide for prevention of the sexual transmission of HIV-1. By coincidence, an epidemic of bovine spongiform encephalopathy (BSE) was ongoing at the same time in the United Kingdom and considered to cause a new variant of Creutzfeldt-Jakob disease (vCJD) in humans. This raised questions related to the safety of bovine milk. However, scientific research results indicate that BSE cannot be transmitted by cow's milk even if the milk comes from a cow with BSE, because no detectable infectivity in milk from BSE-infected animals could be demonstrated. Evidence from other animal and human transmissible spongiform encephalopathy (TSE) studies suggests that milk does not transmit these diseases.
Milk and milk products, even from countries with a high incidence of BSE, are therefore considered safe. 25 Consequently, the United States Food and Drug Administration has exempted milk-derived products from restrictions applied to their use as pharmaceutical ingredients. 26 In accordance with this, some bovine milk-derived products are being generally recognized as safe (GRAS). 27 In addition, the current risk of acquiring vCJD from eating beef and beef products is approximately 1 case per 10 billion servings in the United Kingdom and likely to be smaller in other countries. 28 Notwithstanding these unequivocal conclusions, the World Health Organization recommends that the pharmaceutical industry should avoid the use of materials from animal species in which TSEs naturally occur. 29 Based on these negative recommendations from the WHO, further development of 3HP-β-LG as a topical microbicide had to be abandoned.

Anti-HIV-1 Activity of Pomegranate Juice
Serial twofold dilutions of juices (apple, black cherry, blueberry, coconut milk, cranberry, elderberry, grape [red], grapefruit, honey, lemon, lime, pineapple, pomegranate, and red beet [10% reconstituted dry powder]) were assayed for inhibition of infection by HIV-1 IIIB of cells expressing the CD4 and CXCR4 receptors and coreceptors. Most juices (diluted fourfold) had no inhibitory activity, except blueberry, cranberry, grape, and lime juice, respectively (endpoints for 50% inhibition of infection [ED 50 ] between 1/16 and 1/64). Consistently, pomegranate juice (PJ) from distinct geographic areas had the highest inhibitory activity (FIG. 1; vertically shaded area). Since HIV-1 viruses utilizing CCR5 as coreceptor (= R5 viruses) are predominantly transmitted sexually, 3,55 it was important to test whether PJ can inhibit not only infection by HIV-1 IIIB, a virus utilizing CXCR4 as coreceptor (= X4 virus), but also infection by an R5 virus, HIV-1 BaL. Results in FIGURE 1 (horizontally shaded area) show that infection by the latter virus is also inhibited, albeit less effectively, than that by HIV-1 IIIB.
Blocking virus entry is a primary target for microbicide development. 3,8-11 Therefore, it was of interest to determine whether PJ inhibited the binding of the HIV-1 envelope glycoprotein gp120 to CD4, the common receptor for both X4 and R5 viruses. Pretreatment of both gp120 IIIB and BaL by PJ inhibited subsequent binding of soluble labeled CD4 (FIG. 2). This suggested that one or more PJ ingredients bound strongly or irreversibly to the CD4 binding site on gp120. These results, obtained in an enzyme-linked immunosorbent assay (ELISA) using gp120 immobilized on polystyrene plates, were confirmed in another assay in which both  Inhibition of CD4 binding to recombinant gp120 IIIB and BaL, respectively, by pomegranate juice (PJ). Recombinant gp120 coated wells were incubated with dilutions of the PJ for 1 h at 37°C. After removal of the juice and washing the wells, biotinyl-CD4 was added, and its binding to the wells was measured by ELISA. gp120 and CD4 were in soluble form (data not shown). In reverse experiments, pretreatment of CD4 with PJ failed to block subsequent gp120 binding. Other juices having anti-HIV-1 activity (blueberry, cranberry, grape, and lime) failed to block gp120−CD4 binding.

Separation of Anti-HIV-1 Inhibitor(S) from Pomegranate Juice
Pomegranate juice is intensely colored; therefore, it cannot be directly formulated into a microbicide because it would stain clothing, which is unacceptable. Attempts were made to separate or isolate the active ingredient(s) from PJ. After striving . Inhibition by pomegranate juice (PJ) of binding to gp120 of antibodies to synthetic peptides from the gp120 sequence. Wells of polystyrene plates coated with gp120 IIIB were incubated with fourfold diluted PJ for 1 h at 37°C. After removal of PJ, the wells were washed, and 50-fold diluted anti-peptide antisera 106 were added. Bound IgG was quantitated by ELISA. PJ was not added to control wells. Decreases of absorbance, as compared to the respective control wells, are plotted.
intermittently for over 4 years to accomplish this, it was discovered that the inhibitor(s) of gp120-CD4 binding can be adsorbed effectively (≥99%) onto a selected brand of corn starch, PURITY® 21 corn starch NF grade (National Starch and Chemical Company, Bridgewater, NJ; S21) (FIG. 5), resulting in a nearly colorless product, designated as PJ-S21. PJ-S21 suspended in water or unbuffered 0.14 M NaCl had a pH of 3.2 (due to adsorbed PJ ingredients because starch provides a neutral pH) compatible with the acidic vaginal environment in which it would remain stabile after application (see below). Inhibitors of gp120−CD4 binding could be eluted from PJ-S21 by extraction with ethanol/acetone 6:4. Drying of the extract followed by gravimetry indicated that the extract contained 3.17 mg solids per gram of PJ-S21.
PJ-S21, to the same extent as the original PJ, inhibited the binding of gp120 IIIB-CD4 complexes to cells expressing CXCR4, as determined by flow cytometry (FIG. 6). Similarly, binding of a gp120 BaL-CD4 fusion protein to cells expressing CCR5 was blocked by PJ and PJ-S21, as detemined by a cell-based ELISA 67 (FIG. 7). Therefore, PJ-S21 is an inhibitor of both X4 and R5 virus binding to the cellular receptor CD4 and coreceptors CXCR4/CCR5. PJ-S21 also inhibited gp120 binding to peripheral blood mononuclear cells as determined by flow cytometry (FIG. 8). To confirm that FIGURE 4. Location on the gp120 structure of segments corresponding to antipeptide antibodies whose attachment to gp120 is inhibited by ≥50 % in the presence of pomegranate juice (gray) and of amino acid residues involved in CD4 and CXCR4/CCR5 coreceptor binding, respectively. Black portions of the structure correspond to anti-peptide antibodies whose attachment to gp120 is not significantly inhibited by PJ. The CD4 domains and the antigen-binding fragment of a neutralizing antibody were excised from the structure of the gp120-CD4-antibody complex 58 (1gc1 retrieved from the Protein Data Bank (pdb) [http:// www.rcsb.org/pdb/). The V3 loop, generated by homology modeling, was added to the gp120 structure as described 31 The figure was generated by Molscript 107 and Raster3D. 108,109 The locations of gp120 variable loops (V1 -V5) and of the N-and C-termini of the sequence are indicated.

FIGURE 5.
Adsorption onto corn starch of gp120-CD4 binding inhibitor(s) from pomegranate juice (PJ). Corn starch (PURITY® 21, NF grade; 200 mg/ml) was added to PJ prefiltered to remove particulates. After mixing for 1 h at ∼20°C, the starch was allowed to settle and the supernatant fluid was removed by aspiration. The pellets, resuspended (200 mg/ml) in phosphate-buffered saline, and the supernatant fluids were tested at serial dilutions for inhibition of CD4 binding to gp120 IIIB as described in the legend for FIGURE 2. The inhibitory activity of the resuspended pellet against gp120 BaL-CD4 binding was then confirmed. Control starch did not inhibit gp120-CD4 binding. . Inhibition by pomegranate juice (PJ) and PJ-S21, respectively, of gp120 IIIB-CD4 complex binding to cells expressing CXCR4 coreceptors. HIV-1 IIIB gp120 (5 µg) and biotinyl-CD4 (2.5 µg) were added to 100 µl phosphate-buffered saline (PBS) containing 100 µg bovine serum albumin (BSA) (PBS-BSA) and PJ (final threefold dilution) or PJ-S21 (67 mg corresponding to 212 µg solids from PJ adsorbed onto starch). After 1 h at 20°C, the respective mixtures were added to 10 6 MT-2 cells. After 30 min, the cells were washed 3 times with PBS-BSA and PE-streptavidin (a fluorescent label specific for biotin; 0.1 µg) was added. After 20 min, the cells were washed and fixed by 1% formaldehyde in PBS. Flow cytometry analysis was performed in a FACSCalibur flow cytometer (Becton Dickinson Immunocytometric Systems, San Jose, CA). The median relative fluorescence values for cells exposed to gp120-CD4; gp120-CD4 + PJ; gp120-CD4 + PJ-S21; and control cells were: 13.7, 4.0, 4.3, and 2.1, respectively. Inhibition by pomegranate juice (PJ) and PJ-S21, respectively, of FLSC binding to CCR5 expressing Cf2Th/synCCR5 cells. FLSC is a full-length single chain protein consisting of BaL gp120 linked with the D1D2 domains of CD4 by a 20 amino acid linker. The inhibitory effect was quantitated using a cell-based ELISA. 48 The starting concentration of PJ-S21 was 200 mg/ml, corresponding to 634 µg/ml solids adsorbed onto starch from PJ.  FIGURE 9 demonstrate that PJ-S21 interferes with early steps of the virus replicative cycle.

PJ-S21 functions as a virus entry inhibitor, the complex was added to cells at time intervals before and after infection of cells by HIV-1 IIIB and BaL, respectively. Results shown in
To be considered as a topical microbicide, PJ-S21 must be formulated to withstand storage in a tropical environment. Accelerated thermal stability studies revealed that a water suspension of PJ-S21 maintained only 4, 11, and 33%, respectively, of its original activity (measured by inhibition of gp120-CD4 binding) when stored for 30 minutes at 60°C and 1 week at 50°C or 40°C. However, a dried PJ-S21 powder remained fully active after storage at 50°C for 12 weeks (the longest time used in the evaluation). Consequently, anhydrous formulations should be preferred for further development.
Three such formulations were prepared: two kinds of suppositories, melting at 37°C, and a tablet. The inhibitory activity of PJ-S21 was fully preserved after 12 weeks of storage at 50°C within tablets and at 30°C within the suppositories (the highest temperature considered to prevent melting). Data showing the inhibition of infection by HIV-1 IIIB and BaL, respectively, by PJ-S21 and its formulations (except the tablets that also contain anti-HIV-1 inhibitors other than PJ-S21, that is, Carbopol 974P 33 ) are summarized in FIGURE 10. Their inhibitory activities against HIV-1 IIIB and BaL were similar, unlike the inhibitory activities of the original PJs (FIG. 1). These formulations were also virucidal, albeit at concentrations higher than those sufficient for inhibition of infection. These experiments also revealed that PJ-S21 was not cytotoxic under the experimental conditions used. The inhibitory/ virucidal activities were maintained in the presence of seminal fluid at a 1:1 (w/w) ratio of seminal fluid to PJ-S21 (data not shown).
A microbicide can be considered potentially successful only if it displays antiviral activity against primary virus isolates belonging to distinct virus clades and phenotypes. PJ-S21 meets this requirement, because it inhibited infection by primary HIV-1 strains of all clades tested having R5 and X4R5 (dual-tropic) phenotypes (TABLE 1).

DISCUSSION
Pomegranates have been venerated for millennia for their medicinal properties [68][69][70][71][72][73] and considered sacred by many of the world's major religions. In deference to pomegranates, the British Medical Association and several British Royal Colleges feature the pomegranate in their coat of arms. The Royal College of Physicians of London had adopted the pomegranate in their coat of arms by the middle of the sixteenth century. 68 The best known literary reference to the contraceptive power of pomegranate seeds is classical Greek mythology. Ironically, this report shows that FIGURE 10. HIV-1 inhibitory and virucidal activity of PJ-S21 and its formulations. Inhibition of infection by HIV-1 IIIB and BaL, respectively, was determined as described in the legend for FIGURE 1. To measure virucidal activity, the respective viruses were mixed with graded quantities of PJ-S21 for 5 min at 37°C. After low speed centrifugation, the viruses were separated by precipitation with polyethylene glycol (PEG 8000) and centrifugation. The resuspended pellets and control untreated viruses were serially diluted, and the dilutions assayed for infectivity. The concentration range given on the abscissa corresponds to 0.31-1,268 µg solids adsorbed from PJ to starch. pomegranate juice contains HIV-1 entry inhibitors targeted to the virus envelope corresponding to a class of antiretroviral drugs still scarce in development. 74 Pomegranate juice contains several ingredients 75,76 that, isolated from natural products other than PJ, were reported to have anti-HIV activity, such as caffeic acid, 77 ursolic acid, 78 catechin, and quercetin 79,80 and also anti-herpes simplex virus (HSV) activity. 81,82 However, these compounds, in purified form, obtained commercially, did not block (at 200 µg/ml) gp120-CD4 binding as measured by the ELISA as just described and did not adsorb to corn starch, unlike the entry inhibitor(s) from PJ. In fact, the supernatant after treatment of PJ with starch and removal of the entry inhibitors retained anti-HIV-1 activity and also inhibited HSV-1, whereas the HIV-1 entry inhibitors that adsorbed onto starch did not inhibit HSV. Thus, the antiviral activities in the supernatant appeared to be nonspecific and probably similar to those of extracts from pomegranate rind 83,84 and were not characterized further. Additional information [85][86][87][88] has revealed that the findings apply to crude extracts from pomegranate rind prepared at elevated temperatures under conditions that destroy the HIV-1 entry inhibitor described here.
The inhibitor(s) interfering with gp120 binding to CD4 (FIGS. 2 and 5) blocked additional sites on gp120 (FIG. 3) involved in interaction with the CXCR4/CCR5 coreceptors (FIGS. 4, 6, and 7). This was not completely expected and can be explained either by the presence of multiple inhibitors with distinct or overlapping specificities in PJ-S21 or by induction of gp120 conformational changes 89 resulting in blockade of both CD4 and CXCR4/CCR5 binding sites on gp120. Similar effects have been noted for other small molecule inhibitors. 90 Simultaneous blocking of more than a single site on HIV-1 involved in virus entry is expected to increase the effectiveness of candidate microbicides. 11 The target sites for the inhibitor(s) are likely to be located within the protein moiety of gp120, because binding of labeled Galanthus nivalis lectin (specific for terminal mannose residues 91 ) and other lectins to gp120 oligosaccharides was not diminished in the presence of PJ or PJ-S21 (data not shown). Blocking of CD4 binding sites on HIV-1 gp120 by monoclonal antibodies or a CD4-IgG2 recombinant protein has been shown to be sufficient to inhibit HIV-1 infection of human cervical tissue ex vivo 11 and in preventing virus transmission to macaque monkeys when applied vaginally. 92 Therefore, it seems likely that PJ-S21 will be similarly effective, an expectation that remains to be confirmed in an in vivo macaque model system and in human clinical trials, as a candidate topical microbicide. This anticipation would be strengthened if drinking of PJ decreases the HIV-1 viral load in already infected individuals, an issue to be explored.
The application of PJ-S21 as a topical anti-HIV-1 microbicide requires reasonable uniformity among batches produced at distinct times and locations. Similarities in gp120-CD4 binding inhibitory activity among distinct freshly prepared and commercial juices stored for unknown periods (FIG. 2) suggest that this should be feasible. Pasteurization of juice for 30 seconds at 85°C resulted in complete loss of inhibitory activity. A commercial PJ concentrate exposed to 61°C and two other concentrates, presumably prepared by evaporation at elevated temperatures, had no or drastically diminished activity. The gp120-CD4 inhibitory activity from PJ3 (juice with fructose and citric acid added) failed to bind to starch. Separate experiments revealed that these compounds interfere with inhibitor binding to corn starch. Therefore, PJs intended for production of the PJ-S21 complex must be sterilized by filtration and be free of additives.
Particular attention must be devoted to the selection of starch, a pharmaceutical excipient generally used in vaginal formulations, 93 for effective binding of the virus entry inhibitors from PJ. Among a dozen starches tested, the best results were obtained with S21. With other brands, the adsorption of the inhibitors was either incomplete or their binding did not result in a complex having activity in the ELISA measuring gp120-CD4 binding inhibition (ARGO® corn starch), presumably, because of irreversible binding of the PJ inhibitors. Interestingly, only a few references are available regarding the use of starch as an adsorbent for different compounds: flavors, 94,95 dyes, 96-98 low-molecular mass saccharides, 99 lipids, 100,101 proteins, 102 and iodine. 103 The intended dose of PJ-S21 for vaginal application is 1.0 to 1.5 g (= 3.17 -4.76 mg solids from PJ adsorbed onto starch), that is, ≥100-fold higher than the dose needed for blocking HIV-1 infection in vitro (FIG. 10, TABLE 1) and therefore expected to meet requirements for likely in vivo protection against vaginal challenge. 104 This quantity of PJ-S21 is produced from 5 to 7.5 ml of PJ, that is, ≤ 5% of a single (150 ml) serving of juice, attesting to the safety, feasibility, and economy of this proposed candidate topical microbicide.
In an alternative approach to formulation development, PJ-S21 can be incorporated into a water-dispersible film (similar to the widely available "breath control" strips) or into water-dispersible sponges, 105 which are converted into a gel following topical application. 34 Each of the aforementioned formulations would meet the following requirements: (1) minimization of waste disposal problems associated with the use of applicators needed for delivery of microbicidal gels/creams; (2) simplicity; (3) small packaging and discretion related to purchase, portability, and storage; (4) low production costs; (5) amenability to industrial mass production at multiple sites globally; and (6) potential application as rectal microbicides. Furthermore, it would remain possible to produce for local use PJ-S21-based gel formulations with a limited shelf life, avoiding the costs of producing dry PJ-S21 powders via appropriate low temperature drying processes. Whichever of these formulations is selected, adequate quality control will be needed to assure uniform anti-HIV-1 activity of the final product(s) as well as to establish reproducible conditions for manufacture.

CONCLUSIONS
PJ-S21 can be classified as an AAAA candidate microbicide: acceptable; accessible; affordable; and accelerative in transition from development to marketing. Therefore, PJ-S21 would be expected to circumvent some problems associated with antiretroviral drugs and possibly some of the other candidate microbicides, that is, uncertainty related to potential side effects, investment and time needed to establish inexpensive large scale production, and monopoly of supply.