The Root Extract of the Medicinal Plant Pelargonium sidoides Is a Potent HIV-1 Attachment Inhibitor

Global HIV-1 treatment would benefit greatly from safe herbal medicines with scientifically validated novel anti-HIV-1 activities. The root extract from the medicinal plant Pelargonium sidoides (PS) is licensed in Germany as the herbal medicine EPs®7630, with numerous clinical trials supporting its safety in humans. Here we provide evidence from multiple cell culture experiments that PS extract displays potent anti-HIV-1 activity. We show that PS extract protects peripheral blood mononuclear cells and macrophages from infection with various X4 and R5 tropic HIV-1 strains, including clinical isolates. Functional studies revealed that the extract from PS has a novel mode-of-action. It interferes directly with viral infectivity and blocks the attachment of HIV-1 particles to target cells, protecting them from virus entry. Analysis of the chemical footprint of anti-HIV activity indicates that HIV-1 inhibition is mediated by multiple polyphenolic compounds with low cytotoxicity and can be separated from other extract components with higher cytotoxicity. Based on our data and its excellent safety profile, we propose that PS extract represents a lead candidate for the development of a scientifically validated herbal medicine for anti-HIV-1 therapy with a mode-of-action different from and complementary to current single-molecule drugs.


Introduction
As of August 2012, 23 single-molecule drugs were approved for anti-HIV-1 therapy in the USA by the FDA [1]. The continuous need for the development of new therapeutic anti-HIV-1 agents arises from the rapid emergence of viruses resistant to these drugs (reviewed in [2], [3]), the necessity for continuous life-long treatment [4], the challenges of providing antiretroviral treatment in resource-limited settings [5] and the need for novel drugs with fewer adverse effects [6].
Natural products and herbal medicines are a promising source of new therapeutic agents and for the development of complementary and alternative medicines to conventional drug regimens [7]. While medicinal plants have been reported to display anti-HIV-1 activity (reviewed in [8,9]), herbal preparations are currently not part of conventional therapeutic regimens. Antiviral potencies and modes-of-actions of medicinal plants are poorly understood. Therefore they have been considered mainly as sources for the isolation of single anti-HIV-1 hit molecules by conventional drug-discovery approaches [9,10]. However, herbal preparations may also contain unique mixtures of molecules that act in concert to display novel bioactivities [11]. Herbal preparations have many potential benefits for anti-HIV therapy, including the complementation of existing drug therapies, improvement of anti-HIV treatment in resource-limited settings and reduction of the risk of emergence of viral resistance. Furthermore, they may display novel modes-of-action, which are different from current single-molecule drugs. Thus it is worthwhile to perform detailed and rigorous experimental investigations to evaluate anti-HIV-1 activities of herbal extracts.
Pelargonium sidoides (PS) is an indigenous medicinal plant of South Africa which has been used as a traditional medicine for the treatment of various ailments for over a century [12]. A proprietary extract from PS roots known as EPsH7630 or UmckaloaboH has been evaluated in numerous clinical trials for safety and alleviation of symptoms associated with acute bronchitis and is licensed in Germany as herbal medicine for the treatment of upper respiratory tract infections. PS extract contains numerous different metabolites [13] and has been shown to inhibit viruses associated with respiratory diseases like influenza viruses [14,15] and herpes virus [16].
The proven safety profile, richness in metabolites and demonstrated activities against various viruses led us to evaluate PS extract for anti-HIV-1 activity. We demonstrate that PS extract potently inhibits infection by HIV-1 strains with different tropisms. Anti-HIV-1 activity of PS extract is based on a new mode-ofaction that diminishes infectivity of virus particles and prevents their attachment to host cells. Chemical analysis indicated that anti-HIV activity is mediated by multiple polyphenolic compounds. These results support PS extract as a lead candidate for the development into an herbal medicine with a novel mode of anti-HIV-1 activity.

Virus production
Virus stocks were produced by HEK293T cells transfected with proviral plasmids and analyzed for infectivity, signal induction, absence of cell toxicity and p24 quantity as described in [17].

Generation of crude PS extracts and isolation of polyphenols
Crude PS extract was generated from dried plant roots obtained from Amarelo (www.amarelo24.de). Roots were ground with a ball mill (Minimill Pulverisette 23; Fritsch, Idar-Oberstein). 10 g powdered roots were stirred in 50 ml water (ddH 2 0) for ,2h at room temperature (RT). Boiling, while not necessary, did not affect anti-HIV activity. The extract was cleared by centrifugation (4000 g). Aliquots of the extract were evaporated in an Eppendorf Vacuum Concentrator, the dry mass weighed and ddH 2 0 added to generate extract stock solutions with 10 mg dry mass per ml. All PS extract stock solutions were sterilized by filtration (0.45 mM syringe-filter; Merck Millipore Ltd) and stored at -20uC until use.
Enrichment of PS-derived polyphenols (PSPP). 2.5 g of polyvinylpyrrolidone (PVPP; Sigma-Aldrich) was added to 50 ml of PS extract stock solution (10 mg/ml), vortexed, and polyphenols allowed to adsorb to PVPP at RT for 15 min before centrifugation at 4000 g. The supernatant represented the polyphenol-depleted fraction and the pellet the polyphenol-enriched fraction. The pellet was washed 3 times with 25 ml ddH 2 0 and polyphenols eluted by washing 3 times with 25 ml 0.5 N NaOH. The pooled eluates were adjusted to pH 7 with 2 N HCl. The polyphenol-depleted and polyphenol-enriched fractions were purified by solid phase extraction (SPE) (Agilent Bond Elut C18). SPE-purified eluates were evaporated and dissolved in ddH 2 0 to generate stock solutions with 10 mg per ml.

Assays
Assays for inhibition of HIV-1 infection of LC5-RIC cells and primary human cells are described in Materials and Methods S1.
HIV-1-cell attachment assays were performed with the HIV-1 NL4-3 Gag-iGFP reporter virus and LC5-RIC-R5 cells in the presence of the HIV-1 fusion inhibitor T-20 (50 nM). LC5-RIC-R5 cells (2610 5 ) were seeded onto 24624 mm cover slips placed in 6-well plates. 24 hours later, the culture medium was replaced by fresh medium containing inhibitor compounds (100 mg/ml PSextract or polyphenol enriched PS-extract, or 5 nM Griffithsin reference compound) and virus inoculum (8.5 pg p24/cell). After 4 h incubation at 37uC, the cells were washed once with PBS and cell membranes stained with DiD (Invitrogen) according to the manufacturer's protocol. Cells were then fixed in 2% paraformaldehyde (PFA) in PBS at room-temperature for 10 minutes and treated with DAPI-staining solution for 5 minutes at room temperature for staining of cellular DNA. The cells were washed, coverslips fixed onto glass-slides with Moviol and dried in the dark overnight. Samples were analyzed by fluorescence microscopy (Nikon TiE equipped with Perkin Elmer UltraView Vox System). Exposure-times: GFP: 100 ms, DAPI: various, DiD: various, Brightfield: various. Counting of virus particles (GFP spots defined as all pixels within a radius of 1 mm around the brightest spot, intensity threshold 755) was performed using Volocity 6.2.1software (Perkin Elmer). DiD/DAPI-positive cells were counted manually.

Quantification of HIV-1 DNA and RNA levels
HIV-1 DNA copy numbers and viral RNA were determined by quantitative PCR as described in Materials and Methods S1.
Chromatographic Separation of PS Extract was performed by UPLC (Ultra Performance Liquid Chromatography) of PS extract purified by solid phase extraction (for details see Materials and Methods S1).
Ultrahigh Resolution Mass Spectrometry was performed by ICR/FTMS (Ion Cyclotron Resonance/Fourier Transformation Mass Spectrometry). Details, including description of methods for evaluation of complex mass data, are provided in Materials and Methods S1.

Statistical Analysis
EC 50 /CC 50 values were calculated with GraphPad Prism v5, using the equation for sigmoidal dose-response with variable slope. Statistical significances were determined by one-way Anova with Bonferroni post-hoc test. **** p,0.0001. Initial testing of PS-root extract for anti-HIV-1 activity with EASY-HIT technology [17] revealed dose-dependent inhibition of HIV-1 infection for all extract preparations examined, including the commercial herbal medicine EPsH7630 ( Figure S1). The EASY-HIT assay gauges antiviral effects by measuring infection parameters associated with two temporally distinct steps of the HIV-1 replication cycle (i.e. expression of early viral proteins Tat and Rev = step 1 and release of infectious virions = step 2, resp.) with an HIV-1 reporter cell line [17]. Multiple tests showed that PS extract was similarly effective at reducing infection parameters at both steps, with half-maximal inhibition (EC 50 ) at 8.13 mg/ml (64.82; n = 19) for step 1 and 8.00 mg/ml (62.26; n = 2) for step 2 ( Figure 1). This indicates that PS extract interferes with HIV-1 replication at an early step of the HIV-1 replication cycle [17].
To determine anti-HIV activity of PS extract with primary HIV-1 target cells, we used ex vivo cultures of human peripheral blood mononuclear cells (PBMC) and monocyte-derived macrophages (MDM). Both primary HIV-1 target cell types yielded EC 50 values comparable to those measured for the reporter cells in the EASY-HIT system ( Figure 1). Inhibition of infection of PBMCs by PS extract was determined by measuring infectious virus titers of supernatants of cultures exposed to HIV-1 LAI and different concentrations of PS extract (EC 50 = 5.7062.7 mg/ml; n = 4). PS-mediated inhibition of infection of MDMs was analyzed with an R5 reporter virus that coexpresses Nef and EGFP from a bicistronic mRNA (R5 HIV-1NL4-3 IRES-eGFP [20]) and analyzing proportions of EGFP-positive cells by flow cytometry (EC 50 = 8.2762.75 mg/ml; n = 3).
We also investigated whether PS extracts can inhibit infection by clinical HIV-1 isolates. Clinical isolates were derived from serum of a chronically infected individual (P-891; [17]) and from blood of acutely (CH077) or chronically (STCOr1) infected individuals [22]. As shown in Figure 1E, PS extract was active against all clinical isolates investigated, yielding EC 50 values below 6 mg/ml.
In all assays evaluating PS-mediated inhibition of HIV-1 infection, viability of cells exceeded 80%, confirming that inhibition of infection was not caused by PS-induced cytotoxicity.
Together, these results demonstrate that PS extract inhibits HIV-1 infection of various HIV-1 target cell types, including primary cells, and acts against various HIV-1 isolates, including clinical HIV-1 isolates.

PS extract interferes with HIV-1 entry into the host cell.
To narrow down the stage of the HIV-1 replication cycle at which PS extract inhibits HIV-1 replication, we first performed time-of-addition (TOA) assays with PS extract [17,28]. The HIV-1 entry inhibitors Griffithsin [27] and T-20 [29] and the reverse transcription inhibitor AZT [30] were used as reference compounds. The TOA profiles indicate that PS extract acts at a very early stage of HIV-1 replication which precedes reverse transcription ( Figure 2A). For further confirmation, we investigated whether treatment with PS extract reduced levels of input viral RNA. Experiments were performed in the presence of reverse transcriptase inhibitor (Efavirenz) to maximize accumulation of input viral RNA. RNA was isolated from cells exposed to the virus +/-PS extract for 4 hours and levels of input RNA quantified by RT-PCR using HIV-1 specific primers. As shown in Figure 2B, PS treatment led to a strong reduction of input viral RNA levels in virus-exposed cells, compared to levels of HIV-1 RNA in cells exposed to the virus without PS treatment. As expected, PS treatment also strongly diminished levels of HIV-1 DNA in cells exposed to the virus without Efavirenz ( Figure 2C). Together, the facts that PS extract behaves like validated entry inhibitors in TOA experiments and the dramatic reduction of levels of input HIV-1 RNA and HIV-1 DNA in virus-exposed cells caused by treatment with PS extract indicate that PS blocks HIV-1 entry into host cells.
Entry is mediated by the HIV-1 envelope protein. Therefore we investigated the influence of viral envelope proteins on the antiviral activity of PS by comparing inhibitory effects of PS extract on infection by HIV-1 particles pseudotyped with different envelope proteins. Virions carrying the vesicular stomatitis virus G-envelope protein on their surface were on average 10-fold less sensitive to inhibition by PS extract than HIV-1 with native envelope proteins ( Figure 2D). In contrast, PS extract was similarly efficient at inhibiting infection of virus particles with different HIV-1 envelope proteins (i.e. X4 (LAI) and R5 (AD8 and JRFL)). These results demonstrate that PS extract inhibits HIV-1 entry by interfering with the function of the envelope proteins independent of their coreceptor tropism.
Since PS extract perturbs functionality of the HIV-1 envelope, we investigated whether PS extract interferes with infectivity of HIV-1 particles. Virus preparations were pre-incubated with PS extract without target cells prior to assaying their infectivity. Preincubation assays were also performed with the reference compounds Griffithsin, the CXCR4 antagonist AMD3100 [31], the fusion inhibitor T-20 and the reverse transcriptase inhibitor Efavirenz [32]. As shown in Figure 3, virus pre-incubation with PS extract resulted in stronger inhibition of infection (PS panel, green curve, EC 50 3.960.3 mg/ml) than simultaneous addition of virus and PS extract to LC5-RIC cells (PS panel, blue curve, EC 50 14.361.1 mg/ml). Virus pre-incubation also enhanced the inhibitory activity of Griffithsin, but not of the other reference compounds (other panels, green and blue curves).
Next we tested whether anti-HIV-1 activity of PS extract involves association of extract ingredients with target cells LC5-RIC target cells were pre-incubated with PS extract or the reference inhibitors for 4 hours, and subsequently exposed to HIV-1 LAI virus in the absence of inhibitor. While PS extract displayed no antiviral activity under these conditions ( Figure 3, PS panel, black curve), all reference inhibitors showed measurable anti-HIV-1 activities (other panels, black curves), although the compounds were removed and cells were washed after the 4-hour preincubation period.
Together these results show that PS extract is an HIV-1 entry inhibitor that directly diminishes the infectivity of viral particles.

PS extract inhibits attachment of HIV-1 to host cells.
PS extract might target a pre-or post binding step to diminish HIV-1 infectivity and inhibit entry. We thus investigated whether PS extract affects the attachment of HIV-1 to target cells. Assays were performed with GFP-labeled infectious R5 virus particles (R5 HIV-1 NL4-3 Gag-iGFP; [20]) and LC5-RIC-R5 cells. Spinning disc confocal fluorescence microscopy was used to visualize and quantify GFP-labeled HIV-1 particles associated with target cells.  . Treatment with Griffithsin resulted in a small although significant increase in the number of cell-associated virus particles (3664.2 GFP-spots per cell), which is in line with previous evidence for the enhancement of virus binding to the CD4 receptor by Griffithsin [33]. Thus, PS extract prevents HIV-1 attachment to target cells, which is a novel mode of entry inhibition.  Anti-HIV-1 activity of PS extract is mediated by polyphenol compounds.
The high complexity and diversity of the metabolic pool of the PS plant [13] raises the possibility that anti-HIV-1 activity of PS extract may be exerted by multiple compounds. For chemical characterization of anti-HIV-1 activity, PS extract was separated by UPLC-chromatography and anti-HIV-1 activity of fractions determined with the EASY-HIT system. Chemical structures within the fractions were identified by ICR/FTMS (Ion Cyclotron Resonance/Fourier Transformation Mass Spectrometry). This technology allows to distinguish between several thousands of ions with extremely high mass accuracy [34] and enables prediction of the molecular formula of detected compounds and their putative annotation [35].
Several fractions dispersed throughout the chromatogram displayed anti-HIV-1 activity ( Figure 5A). This indicates that multiple chemical compounds in the PS extract have anti-HIV-1 activity. ICR/FTMS analysis combined with structural datamining, statistical analysis and mass network analysis revealed that fractions with anti-HIV-1 activity were selectively enriched for various aromatic and oxygen rich compounds ( Figure 5B). These were annotated as polyphenolic compounds, mainly belonging to the classes of flavonoids and leuco-anthocyanidins.
To investigate involvement of polyphenols in anti-HIV activity of PS extract, we used adsorption of polyphenols to polyvinylpolypyrrolidone [36] to produce a PS-derived polyphenol fraction (PSPP). Specific enrichment of polyphenols in this fraction was confirmed by ICR/FTMS ( Figure S2).
In sum, the antiviral activity of PS extract can be attributed to polyphenolic secondary metabolites belonging to the classes of flavonoids and leuco-anthocyanidins. The potent anti-HIV-1 activity and reduced cytotoxicity of the PSPP fraction indicate that enrichment of polyphenols from PS root extracts eliminates cytotoxic molecules, while retaining HIV-1 inhibitors.

Discussion
This study demonstrates that PS root extract, including the commercial herbal medicine EPsH7630, is a highly reliable source of robust anti-HIV-1 activity. PS extract prevents attachment of virus particles to host cells and hence HIV-1 entry. Our results suggest that HIV-1 inhibitory molecules in PS extracts target HIV-1 envelope proteins, since virus particles bearing the heterologous VSV-G protein instead of HIV-1 proteins in their envelopes are much less sensitive to inhibition by PS extract. The mode-of-action of PS extract exhibits a combination of features that set it apart from approved anti-HIV-1 drugs and from investigational entry inhibitors including lectins [37,38], polyanionic compounds [39,40] and synthetic anti-lipopolysaccharide peptides (SALPs) [41], of which several are under development as microbicides and drug candidates. For one, PS extract prevents attachment of virus particles to host cells. In contrast, the lectin Griffithsin, which was assayed as reference compound, did not inhibit HIV-1 attachment and therefore blocks entry at a later stage. Other plant derived rich lipid compounds (encircled in grey). The graph at the right shows mass networks in highly active fractions (i.e. fractions 5-10, 12, 13, 15, 16; red) versus poorly active fractions (1, 2, 4, 11, 14, 17, 19-21, yellow). The networks include a cluster of masses predominantly associated with high anti-HIV activity (framed in violet) that map to the group of polyphenolic compounds in the van Krevelen diagram (violet arrow). doi:10.1371/journal.pone.0087487.g005 HIV-1 entry inhibitors epigallocatechin gallate (EGCg) and theaflavin also act after attachment [42,43,44] and in this respect are similar to Griffithsin but different from PS extract. Second, PS exerts potent anti-HIV-1 activity independent of the viral coreceptor tropism. This is in contrast to polyanions that mainly target CXCR4-using HIV-1 strains [45], and to compounds that act by blocking single HIV-1 cell surface co-receptors (e.g. AMD3100 [31]) or CCR5 (e.g. Maraviroc; [46]). Third, PS extract directly interferes with the infectivity of HIV-1 particles before they interact with the host cell and thus has virucidal activity. This is in contrast to entry inhibitors that bind to molecules on the surface of host cells (e.g. SALPs, Maraviroc and Griffithsin [47]) or that interfere with entry processes that take place after the virus has bound to the host cell (T20).
Interestingly, PS extract was also reported to display virucidal activity against herpesvirus [16], but not against influenza virus [15]. This indicates that PS inhibits HIV-1 and herpesviruses by mechanisms different from those involved in inhibition of influenzavirus and suggests that PS extract contains multiple antiviral compounds with different activities.
In addition to the novel mode-of-action, PS extract has several other characteristics that highlight it as an attractive candidate for HIV-1 preventive and therapeutic strategies. We demonstrate that anti-HIV activity of PS extract is exerted by the concerted action of a unique combination of polyphenolic antiviral ingredients. Conceivably, this may reduce the risk of the emergence of drugresistant viruses, compared to treatment with single-molecule drugs.
Furthermore, safety of PS extract was demonstrated in at least 18 clinical trials, involving treatment of children and adults with upper respiratory tract diseases with EPsH7630 over several weeks [12,48]. We show that enrichment of polyphenols by adsorption to polyvinylpyrrolidone further reduces cytotoxicity for PBMCs, improving the therapeutic index of PS-mediated antiviral activity. Clinical safety of PS extracts and the compelling evidence for in vitro anti-HIV-1 activity presented in our study encourages future testing of PS extract and PS-derived polyphenol-enriched formulations in HIV-1 infected individuals.
Because PS extract inhibits HIV-1 entry by a novel mode-ofaction, it may complement the activity of current microbicide candidates and thus improve prevention of HIV-1 transmission. For anti-HIV-1 therapy, PS extract may supplement first-line treatment regimens, which so far do not include entry inhibitors. Furthermore, PS extract may help decrease virus burden during interruption of conventional drug treatment. Another remarkable feature of the anti-HIV-1 activity of PS extract is its robustness against boiling, drying and extreme pH conditions. Thus we show that PS-extract can be easily produced from PS roots and is amenable to long-term storage at room temperature without loss of anti-HIV-1 activity. This indicates the possibility to use PSpolyphenolic drugs as anti-HIV-1 therapy in resource limited settings. Of note, PS is a medicinal plant habituated in South Africa. Hence, it is highly likely that there will be a high level of compliance towards a PS-based antiviral therapeutic regimen in this region, which is strongly affected by the AIDS pandemic.

Conclusions
Here we demonstrate that an aqueous extract from roots of Pelargonium sidoides plants contains robust and potent anti-HIV-1 activity. PS extract prevents HIV-1 particles from attaching to host cells and displays a novel mode-of-action different from other HIV-1 entry inhibitors. Chemical analysis indicates that anti-HIV-1 activity is mediated by the concerted action of multiple polyphenolic compounds which can be separated from components of the extract with higher cytotoxicity by adsorption to polyvinylpyrrolidone.
Based on its potent anti-HIV activity and its extensively studied safety profile, we conclude that PS extract represents a promising lead candidate for the development of a herbal medicine for HIV-1 treatment. The novel mode of anti-HIV activity suggests that PS extract may be useful for complementation of other anti-HIV-1 agents in the therapy of HIV-infected individuals and in protection against HIV-1 infection. Finally, our study demonstrates the amenability of PS extract for experimental analyses, suggesting the use of PS extract as a research tool to advance the development of new anti-HIV agents that protect potential HIV-1 target cells from virus exposure in humans.

Supporting Information
Materials and Methods S1 Assays for testing of inhibition of HIV-1 infection. Quantification of HIV-1 DNA-and RNA-levels. Chromatographic separation of PS extracts. Ultrahigh resolution mass spectrometry. (DOCX) Figure S1 Inhibition of HIV-1 infection by different preparations of PS root extracts, including the commercial herbal medicine EPsH 7630. PS extract from plants were prepared from dry or fresh PS roots as described in the main text (Materials and Methods). For analysis of anti-HIV activity of the commercial herbal medicine Umckaloabo/EPsH7630 (purchased from a local pharmacist), aqueous samples were prepared by removing ethanol in the commercial formulation by evaporation in an Eppendorf Vacuum Concentrator and restoration of the original sample volume with ddH 2 0. Anti-HIV-1 activity was evaluated in LC5-RIC cultures exposed to HIV1 LAI . Each extract concentration was tested in triplicate. Fluorescent signal intensities of treated cultures were normalized to those of untreated cultures assayed in the same plate (100% infection). Mean values (columns) and standard deviation of the mean are indicated for each extract dilution. Medical School, Worcester, MA, USA) for the plasmids pNL4.3DEnv and pSVIIIenvJRFL and Dr. Olaf Kutsch (University of Alabama, Birmingham, Ala, USA) for the CCR5-expression plasmid in the pMSCVpuro backbone. We thank Beatrice Hahn, University of Pennsylvania School of Medicine for permission to use infectious molecular clones of CH077 and STC0r1 clinical isolates and Frank Kirchhoff, University of Ulm, for providing the appropriate plasmids. pNL(AD8) and pSG3DEnv were obtained through the NIH AIDS Reagent Program, Division of AIDS, NIAID, NIH from Dr. Eric O. Freed and Drs. John C. Kappes and Xiaoyun Wu, respectively. We thank Ulrike Protzer for continuous support and encouragement and Ingrid Huelsmeier and Ute Finkel for expert technical assistance.