The Low-Cost Compound Lignosulfonic Acid (LA) Exhibits Broad-Spectrum Anti-HIV and Anti-HSV Activity and Has Potential for Microbicidal Applications

Objectives Lignosulfonic acid (LA), a low-cost lignin-derived polyanionic macromolecule, was extensively studied for its anti-HIV and anti-HSV activity in various cellular assays, its mechanism of viral inhibition and safety profile as potential microbicide. Results LA demonstrated potent inhibitory activity of HIV replication against a wide range of R5 and X4 HIV strains and prevented the uptake of HIV by bystander CD4+ T cells from persistently infected T cells in vitro (IC50: 0.07 – 0.34 μM). LA also inhibited HSV-2 replication in vitro in different cell types (IC50: 0.42 – 1.1 μM) and in rodents in vivo. Furthermore, LA neutralized the HIV-1 and HSV-2 DC-SIGN-mediated viral transfer to CD4+ T cells (IC50: ∼1 μM). In addition, dual HIV-1/HSV-2 infection in T cells was potently blocked by LA (IC50: 0.71 μM). No antiviral activity was observed against the non-enveloped viruses Coxsackie type B4 and Reovirus type 1. LA is defined as a HIV entry inhibitor since it interfered with gp120 binding to the cell surface of T cells. Pretreatment of PBMCs with LA neither increased expression levels of cellular activation markers (CD69, CD25 and HLA-DR), nor enhanced HIV-1 replication. Furthermore, we found that LA had non-antagonistic effects with acyclovir, PRO2000 or LabyA1 (combination index (CI): 0.46 – 1.03) in its anti-HSV-2 activity and synergized with tenofovir (CI: 0.59) in its anti-HIV-1 activity. To identify mechanisms of LA resistance, we generated in vitro a mutant HIV-1 NL4.3LAresistant virus, which acquired seven mutations in the HIV-1 envelope glycoproteins: S160N, V170N, Q280H and R389T in gp120 and K77Q, N113D and H132Y in gp41. Additionally, HIV-1 NL4.3LAresistant virus showed cross-resistance with feglymycin, enfuvirtide, PRO2000 and mAb b12, four well-described HIV binding/fusion inhibitors. Importantly, LA did not affect the growth of vaginal Lactobacilli strains. Conclusion Overall, these data highlight LA as a potential and unique low-cost microbicide displaying broad anti-HIV and anti-HSV activity.


Introduction
According to UNAIDS' latest results, about 2.1 million new human immunodeficiency virus (HIV) infections still occurred worldwide in 2013 [1]. Multiple studies indicate the importance of the interaction between genital herpes simplex type 2 (HSV-2) infections and HIV-1 on the sexual transmission in women [2][3][4][5][6]. The association of HSV-2 with significantly higher amounts of HIV-1 in plasma and genital secretions suggests that antiviral treatment of solely HSV-2 with nucleoside analogues (e.g. acyclovir) could result in a reduced replication rate of HIV-1. Although condom use is still the best way to protect men and women against sexually transmitted pathogens such as HIV and HSV-2, it would be of great benefit for women to develop self-administrating topical microbicides (e.g. vaginal/rectal gels, intravaginal ring systems, suppositories, pills) containing one or more antiviral agents with an exquisite activity against both HSV-2 and HIV-1.
At present, the HIV-1 nucleotide reverse transcriptase inhibitor (NtRTI) tenofovir (Viread) is the most promising microbicidal compound evaluated in clinical trials so far [7]. Topically applied gel-formulated tenofovir has been shown to reduce the sexual transmission of HIV-1 significantly by 39% overall and surprisingly also of HSV-2 by 51% [8]. However, the observed inhibitory activities of tenofovir on HSV-2 replication by targeting the viral DNA polymerase was only achieved at higher drug levels [9]. Acyclovir (Zovirax) is the gold standard drug for treatment of HSV infections and belongs to a group of synthetic drugs called nucleoside analogs [10]. The compound specifically inhibits the herpes DNA polymerase and has little effect on the host cell DNA polymerase. However, studies proved that long-term administration of acyclovir in immunocompromised patients could result in drug-resistant HSV strains [11]. Lisco et al. [12] demonstrated that acyclovir can even act as an HIV-1 reverse transcriptase inhibitor in herpes virus-infected cells, while McMahon et al. [13] found a consistent inhibition of HIV-1 in the absence of herpes viruses. The observed activity of tenofovir and the inconsistent findings of acyclovir indicate the need for antiviral agents targeting both viruses with equal potency. Recent insights highlight also the emergence of acyclovir-resistant HSV strains in immunocompetent individuals treated for herpetic keratitis or encephalitis [11].
Entry inhibitors may even have a better profile as potential microbicide candidates as they prevent infection of the target cells already in the (vaginal/rectal) lumen. Therefore, we focus on the low-cost molecule lignosulfonic acid (LA), which belongs to the family of lignin-derived macromolecules, byproducts formed during the conversion of woodpulp into paper [14]. Previously published reports demonstrated that LA has some very interesting biological properties such as long time usage as an animal feed additive due to anti-pepsin activity and protective effects against gastric ulcer development [15,16]. Preliminary activity of different water-soluble lignins against certain HIV-1 isolates was reported previously [17,18].
Here, we report an extensive evaluation of the consistent broad-spectrum anti-HIV and anti-HSV activity of a low molecular weight variant of LA (mw: *8000 g/mol) in various HIV and HSV target cell lines and in vivo in a mouse model. We also demonstrate its excellent safety profile at the cellular level and at the level of vaginal Lactobacilli microbiota. Hereby highlighting its potential use for topical microbicidal applications.
Peripheral blood mononuclear cells (PBMCs) were isolated out of buffy coats from healthy donors, derived from the blood transfusion center (Red Cross, Belgium) by density centrifugation. The cells were then cultured in RPMI-1640 medium supplemented with 10% FCS and 1% L-glutamine. PBMCs were stimulated with 2 μg/ml phytohemagglutinin (PHA) for 3 days at 37°C, before further usage in HIV infection assays. The preparation of monocyte/macrophages (MDM) was described previously [24].
Monocyte-derived dendritic cells (MDDCs) were generated from freshly isolated PBMCs (cfr. supra). PBMCs were gently rotated at 4°C to form monocyte aggregates. After sedimentation of the mononuclear cell aggregates, the pellet was resuspended in RPMI medium and incubated for 90 min. at 5% CO 2 , 37°C to allow for monocyte adherence. After the incubation period, non-adherent cells were carefully removed by washing three times with PBS. Adherent cells were cultured in RPMI culture medium supplemented with 25 ng/ml IL-4 and 50 ng/ml GM-CSF (Peprotech, London, United Kingdom). Four days later, culture medium was changed with fresh medium containing the same cytokine composition. After 6 days, IL-4 and GM-CSF had induced differentiation of monocytes into immature MDDCs as shown by phenotypical analysis by fluorescently labeled anti-CD1a, anti-CD14, anti-CD80, anti-CD86, anti-DC-SIGN and anti-HLA-DR antibodies (all from BD Biosciences, Erembodegem, Belgium) (S1 Fig).
The human embryonic lung fibroblasts (HEL cells) were obtained from the ATCC and cultured in MEM containing 10% FCS, 1% L-glutamine and 0.3% sodium bicarbonate (Invitrogen).
The HSV-1 KOS and HSV-2 G were used as reference laboratory HSV strains. The HSV-1 wild-type (RV-174), HSV-1 thymidine kinase-deficient (TK -) (RV-179, RV-117), HSV-2 wildtype (RV-124, RV-24) and HSV-2 TK -(RV-129, BA19026589) clinical isolates were derived from virus-infected individuals in Belgium. They were obtained as part of a translational research program granted by the Belgian Ministry of Health as part of the National Cancer Plan for the diagnosis of drug resistance in herpesviruses. The viruses were obtained and used as approved according to the rules of Belgian equivalent of IRB (Institutional review board) (Departement Leefmilieu, Natuur en Energie, protocol SBB219 2011/0011, and the Biosafety Committee at the University of Leuven).

HIV-1 replication assays in MT-4-, TZM-bl-and primary cells
The anti-HIV assays in MT-4 cells and PBMCs have been described previously [29]. MT-4 cells (1x10 6 cells/ml; 50 μl) were pre-incubated for 30 min. at 37°C with the test compounds in 96-well plates (Falcon, BD Biosciences). The laboratory HIV strains (HE, NL4.3, IIIB and ROD) were added according to the 50% tissue culture infectious dose (TCID 50 ) of the viral stock. Cytopathicity was scored microscopically in MT-4 cells 5 days post-infection and IC 50 s were calculated spectrophotometrically using MTS/PES. For the latter viability assay, the 'Cell-Titer96 Aqueous One Solution Proliferation Assay' (Promega, Leiden, The Netherlands) kit was used.
The firefly luciferase and Escherichia coli β-galactosidase expressing CD4 + , CXCR4 + , CCR5 + TZM-bl cells (50 μl; 2x10 5 cells/ml) were resuspended in cell culture medium supplemented with 15 μg/ml diethylaminoethyl-dextran (DEAE-Dextran; Sigma-Aldrich, Diegem, Belgium) and pre-incubated for 30 min. at 37°C in 96-well plates with in cell culture medium diluted test compounds (100 μl). Next, the laboratory HIV-1 strain R5 BaL was added (50 μl) according to the TCID 50 of the viral stock. Two days post-infection, viral replication was measured by luminescence. Steadylite plus reagent (Perkin Elmer, Zaventem, Belgium) was mixed with lyophilized substrate according to manufacturer's guidelines. Supernatant (120 μl) was removed and 75 μl Steadylite plus substrate solution was added to the 96-well plates. Next, the plates were incubated in dark for 10 min. in a closed plate shaker (PHMP, Grant, Shepreth, Cambridgeshire, UK). Finally, cell lysis was scored microscopically and 100 μl was transferred to white lumitrac 96-well plates (Greiner Bio-One, Frickenhausen, Germany) to measure the relative luminescence units (RLUs) using the SpectraMax L microplate reader and Softmax Pro software (Molecular Devices, Sunnyvale, CA, USA) with an integration time of 0.6 sec. and dark adapt of 5 min.
MDM were seeded in 48-well plates in 1 ml medium. After removal of 800 μl of medium, 250 μl of various concentrations of LA was given in triplicate. After 30 min. at 37°C, 1000 pg/ well of HIV-1 BaL was added. Three weeks post-infection, supernatant was collected and viral replication measured using HIV-1 p24 Ag ELISA.

HIV-1 time-of-drug addition experiments
The methodology of time-of-drug addition studies was described previously [30]. Briefly, MT-4 cells (1x10 6 cells/ml) were infected with HIV-1 NL4.3. The test compounds were added in a range from 0-5 h post-infection. Afterwards, HIV-1 replication was determined by p24 HIV-1 Ag ELISA.

Virus-inactivation assay in MT-4 cells
Fifty microliter of product solution (LA low, PRO 2000, HHA, aldrithiol) at a starting concentration of 100 μg/ml was incubated with 50 μl of HIV-1 NL4.3 virus stock (2,000,000 pg/ml) in Nunc tubes (Sigma Aldrich). After 1 hour of incubation at room temperature, all virus samples (with or without product) were diluted to 2,000 pg/ml, 1,000 pg/ml and 200 pg/ml. Subsequently, 100 μl aliquots of the dilutions were added to MT-4 cells (100 μl, 50,000 cells/well) in a 96-well plate. Experiments were carried out in quintuplicate. The MT-4 cells were exposed to concentrations (dilutions) of product that were well below their IC 50 values. Degree of infection was assessed with the cellular viability MTS/PES assay after 5 days of incubation. IC 50 values were determined based on the reference values obtained with the untreated positive control.

Anti-HSV replication assays in epithelial HEL cells
The antiviral assays of HSV replication in HEL cells was described earlier [31]. Shortly, confluent cell cultures in 96-well plates were incubated with various concentrations of LA and simultaneously infected with 100 TCID 50 of laboratory HSV strains or clinical isolates for 3 days at 37°C before viral cytopathogenic effect (CPE) was measured microscopically. The anti-HSV activity is expressed as the IC 50 (concentration required to reduce virus induced CPE by 50%).

HSV-2 time-of-drug addition studies
The time-of-drug addition experiments with HSV-2 G were performed identically as the viral replication assays in HEL cells, but the compounds were added separately at the time of infection (0 h) and 2 h post-infection at 1 fixed concentration: LabyA1 9.6 μM; LA 0.5 μM and acyclovir 10 μM. After 3 days, HSV-2 induced CPE was scored microscopically.

MDDCs HSV-2 infection assays
MDDCs were seeded in 48-well plates (12x10 5 cells/200 μl/well) and pre-incubated with various concentrations of LA (250 μl/well) for 30 min. at 5% CO 2 , 37°C. After the incubation period, 50 μl of 1/40 diluted HSV-2 G was added to the wells, whether or not in the presence of compound. After 5 days of incubation, cells were harvested and processed for further flow cytometric side-scatter analysis. Viral titers of the samples were determined in human embryonic lung fibroblasts [9].

HIV-1 cell-cell cocultivation assay
This method was described in detail previously [20]. In brief, various concentrations of LA (100 μl) were added in a 96-well plate along with CD4 + SupT1 T cells (1x10 5 cells/50 μl). The persistently HIV-1 IIIB infected HUT-78 cells (HUT-78/IIIB) were washed to remove free viral particles, resuspended at a density of 1x10 5 cells/50 μl and immediately mixed with the SupT1 T cells for 24 h at 37°C. The next day, giant cell formation was scored microscopically and afterwards, flow cytometry was used to calculate the IC 50 s. The target SupT1 T cell population was stained with PE-conjugated anti-CD28 (CD28-PE, BD) diluted in PBS/FCS2% for 30 min. at 4°C. After two wash steps, the cells were fixed with a 1% paraformaldehyde solution. Acquisition and analysis occurred on a FACSArray flow cytometer (BD) using Windows-based FAC-SArray System. Aspecific binding was determined using SimulTest control (IgGγ1-FITC/ IgGγ2a-PE; BD).

HSV-2 transmission and subsequent replication assay
MDDCs (1x10 6 cells) were pre-incubated with high amounts of HSV-2 G (170 PFU) for 3 h at 5% CO 2 and 37°C. Afterwards, the cells were extensively washed with PBS/FCS2% and resuspended at a density of 1x10 5 cells/50 μl in RPMI culture medium. LA (100 μl) was diluted in RPMI culture medium and pre-incubated for 30 min. with CD4 + MT-4 cells (1x10 5 cells/50 μl) in a 96-well plate before being added to the same amount of HSV-2 exposed MDDCs. Five days after cocultivation, giant cell formation was scored microscopically and dying MT-4-infected cells were detected by flow cytometric side-scatter analyses.

In vivo anti-HSV-2 activity of LA in infected mice
The Ethical Committee for Animal Care and Use of the University of Leuven approved all animal work (P221/2014). Animals' guidelines and policies were in accordance with the Belgian Royal Decree of November 14 th , 1993 concerning the protection of laboratory animals and the EU Directive 86-609-EEC for the protection of vertebrate animals used for scientific purposes.
Female adult nu-nu mice were treated topically with 50 μl of 2.5% LA, 5% acyclovir, 1% tenofovir (all dissolved in DMSO) or placebo (just DMSO) twice daily for 5 days, starting one day prior to infection. On the day of infection, the skin of the mice was scarificated lumbosacrally and subsequently exposed to a HSV-2 G suspension (500 PFU/50 μl). Development of lesions and mortality were recorded over 25 days. Primary lesions develop at the site of inoculation and the severity of the infection was given a score based on the number and size of the lesions [9,32]. Initially the lesions are very small ("small early lesions") and increase in size and number over time ("advanced lesions"). Swab samples were taken from the lesion sites at day 6 and 7 post-infection by firmly rubbing the lesions. The swabs were subsequently immersed in 3 ml PBS and the viral titers of the samples were determined in human embryonic lung fibroblasts [9]. GraphPad Prism 5 was used for developing the figures and performing Mann Whitney U test (GraphPad Software Inc, La Jolla, CA, USA).

Lactobacillus growth assay
The Lactobacillus strains used in this study are listed in S1 Table. All strains were routinely grown non-shaking in MRS (de Man-Rogosa-Sharpe) medium (BD Difco) at 37°C [33]. The capacity of different Lactobacillus strains to grow in the presence of various concentrations of LA was performed as described previously [34]. Briefly, Lactobacilli grown overnight to stationary phase (OD 600nm *1.8-2.0) were used as inocula for the assay. The assay was performed in 96-well plates. Sterile wells were filled with 200 μl MRS medium. About 0.5x10 7 colony forming units (CFUs) (1/100 of an overnight culture) were added and various concentrations of LA were administered. The plates were incubated without shaking for 24 h at 37°C before the OD 600nm was measured. As a negative control, each strain was grown in compound free medium. Each of the Lactobacillus strains and components were examined in 4 replicates and each experiment was independently performed at least 3 times.

Cellular activation marker expression
The expression of the cellular activation markers CD69 (early), CD25 (late) and HLA-DR (very late) were measured on freshly isolated PBMCs after 3 days of incubation at 37°C with varying concentrations of LA, PRO2000, tenofovir and PHA. Briefly, the cells were washed with PBS/FCS2% and incubated with PerCP-conjugated anti-CD4 mAb (BD Biosciences) together with PE-conjugated anti-CD25 mAb (BD Biosciences), anti-CD69 mAb (Biolegend) or anti-HLA-DR mAb (BD Biosciences) for 30 min. at 4°C. In parallel, the cells were stained with SimulTest control IgGγ1-FITC/IgGγ2a-PE (BD) for aspecific background staining. After washing, the cells were fixed with 1% paraformaldehyde solution and analyzed with the FACS-Calibur and Cell Quest software (BD).

Susceptibility of PBMCs for HIV-1 infection
Freshly isolated PBMCs were cultured in the presence of LA, PHA, tenofovir, PRO2000 and acyclovir for 24 h. The next day, the cells were collected and extensively washed in cell culture medium, before being seeded at a density of 5x10 5 cells/450 μl in a 48-well plate in the presence of 2 ng/ml IL-2. Afterwards, 50 μl of HIV-1 R5 BaL (TCID 50 depending on the viral stock) was given and supernatant was collected 7 days post-infection to measure HIV-1 replication by p24 HIV-1 Ag ELISA.

LA/antiviral drug combination assays
The methodology of analysis of two-drug combinations was described earlier [29,35]. Shortly, IC 50 s of each inhibitor solely were determined in C8166 cells for HSV-2 G and TZM-bl cells for HIV-1 R5 BaL. Afterwards, viral replication was determined using MTS/PES method in the presence of LA in combination with acyclovir, PRO2000 and LabyA1 for HSV-2. The anti-HIV activitiy of LA in combination with maraviroc, tenofovir and PRO2000 was determined using the luminiscence assay. Combination indices were calculated using CalcuSyn software (Bio-Soft, Cambridge, UK) based on the median-effect principle of Chou and Talalay [36].

Selection of HIV-1 NL4.3 LAresistant virus and genotyping of the HIV-1 env region
HIV-1 NL4.3 was added to MT-4 cells in 24-well plates in the presence of increasing concentrations of LA. Every 5 days, viral replication was scored microscopically and compound concentration was increased when full CPE was observed. After 55 passages, HIV-1 NL4.3 LAresistant virus was collected and viral RNA was extracted from cell culture supernatants using QIAamp Viral RNA minikit (Qiagen, Hilden, Germany). The genotyping of the env gene was determined as described earlier [37].

Results
LA has a broad-spectrum anti-HIV activity in vitro with low cytotoxicity First, we investigated the cytotoxicity of LA in various cell types and as shown in Fig 1A, the 50% cytotoxic concentrations (CC 50 s) were all higher than 62.5 μM (or >500 μg/ml).

Inhibitory activity of LA on HIV-1-associated syncytium formation
HIV infection and transmission is not only characterized by free viral particles, but also by virus-infected cells (e.g. T cells) [38]. Therefore, the antiviral activity of LA was determined using the giant cell (syncytium) assay between persistently HIV-infected T cells (HUT-78/IIIB) and non-infected CD4 + target T cells (SupT1 cells). The SupT1 T cells express high amounts of CD28, while HUT-78/IIIB cells do not (data not shown) such that this marker can be used to determine the percentage of SupT1 T cell protection, indicative for the percentage inhibition of syncytium formation. We found that LA neutralized this process very efficiently with an IC 50 of 0.23 ± 0.02 μM (1.8 ± 0.16 μg/ml) (Fig 1D).

LA is an HIV-1 entry inhibitor by interfering with viral envelope gp120 binding
To investigate the mechanism of viral inhibition of LA, time-of-drug addition experiments were performed (Fig 2A). The polyanionic compound dextran sulfate (DS5000), a compound known to target HIV-1 gp120 and the NtRTI tenofovir were included as reference compounds. Inhibition of viral infection by LA was completely abrogated 1 h post-infection, while tenofovir kept his antiviral activity up to 4 h (Fig 2A). As can be seen from Fig 2A, LA has a comparable inhibitory profile as dextran sulfate and this finding suggests that LA acts as a HIV entry inhibitor. To provide additional support that LA inhibits HIV-1 entry by interacting with the viral envelope, we performed a virus inactivation assay. As shown in S2 Table, LA had a mild effect on virus inactivation but less compared to the control compounds HHA and aldrithiol.
Based on the time-of-drug addition experiments, we investigated if LA interfered with the initial gp120 binding to the cell surface. Using flow cytometry, virus binding was detected with the anti-gp120 mAb 9205 recognizing the tip of the V3-loop and the 2G12 mAb, which binds to high-mannose type N-glycans that lie around the C4-V4 region on gp120. Similar to FGM (p = 0.0318) and PRO2000 (p = 0.0041), LA inhibited significantly (p = 0.0264) the binding of HIV-1 NL4.3 to the cell surface of SupT1 T cells as detected by the 9205 mAb. Likewise, 2G12 mAb binding was reduced in the presence of LA and this effect was dose-dependent (S2 Fig). As expected the CXCR4 antagonist AMD3100 (p = 0.9097) had no activity in the virus binding assay (Fig 2B and S2 Fig).
When we pre-incubated the cells with the above-mentioned entry inhibitors, washed the cells extensively and added the same amount of virus, neither LA (p = 0.1695), FGM (p = 0.9296), PRO2000 (p = 0.1344) nor AMD3100 (p = 0.3228) was able to significantly inhibit Dual Anti-HIV and Anti-HSV Activity of LA the binding of HIV-1 NL4.3 to CD4. In summary, these results show that LA binds to the envelope glycoproteins of HIV and presumably not to the cell surface (Fig 2C).

LA inhibits the DC-SIGN-mediated route of viral transmission
In the next set of experiments, we investigated the binding of LA to DC-SIGN, an important attachment receptor involved in sexual transmission of HIV-1 to CD4 + T cells [39]. At the highest concentrations tested (12.5 μM), it appeared that LA was not able to bind to the DC--SIGN receptor, since no dose-dependent inhibition of anti-DC-SIGN mAb was observed ( Fig  3A).
Secondly, we pre-incubated high amounts of the R5/X4 HIV-1 strain HE with LA, PRO2000 or HHA prior to exposure to Raji.DC-SIGN + cells. As shown in Fig 3B, the polyanionic compounds LA and PRO2000 did not inhibit the capture of HIV-1 to DC-SIGN, while the  (Fig 3C, first panel). After 48 h of coculture, the infected CD4 + C8166 target T cells lost CD4 expression and died due to HIV-1 infection (second panel). LA inhibited this process dose-dependently (third and fourth panel) with a mean IC 50 of 1.09 ± 0.025 μM (8.7 ± 0.2 μg/ml), comparable to the polyanionic compound PRO2000 (IC 50 : 1.2 μM; data not shown). The mean IC 50 -value obtained with the HIV-1 p-24 Ag ELISA was 0.8 ± 0.03 μM (6.2 ± 0.26 μg/ml), which was in the same range as with MDDCs 0.2 ± 0.02 μM (1.6 ± 0.16 μg/ml) (data not shown). The class of CBAs (e.g. HHA, griffithsin) were previously Dual Anti-HIV and Anti-HSV Activity of LA described to be potent inhibitors of this process [20,40,41], as no viral transmission was observed in the presence of 0.79 μM griffithsin (data not shown).

Potent broad-spectrum anti-HSV activity of LA
First, we evaluated the broad-spectrum anti-HSV activity of LA using the CPE reduction assay in HEL fibroblasts against various wild-type and acyclovir resistant (thymidine kinase negative [TK -]) HSV-1 and HSV-2 strains. LA had a consistent and broad-spectrum anti-HSV activity with an IC 50 ranging between 0.42 and 0.62 μM (3.4 and 5.0 μg/ml), comparable with acyclovir (IC 50 : 0.13-0.34 μM) (Fig 4A), and remained equally active against a variety of clinical mutant TK -HSV-1 and HSV-2 strains. These mutations resulted in a more than 200-fold decrease in antiviral activity of acyclovir ( Fig 4A). Furthermore, inhibitory activity of LA was also found against Canine herpes virus in Madin-Darby Canin Kidney (MDCK) cells with an IC 50 of 0.53 ± 0.03 μM (4.2 ± 0.24 μg/ml) (data not shown).
To investigate the mechanism of action of LA on HSV-2 infection, we performed once more time-of-drug addition experiments in HEL cells. As shown in Fig 4B, when the compounds were added simultaneously with the virus (0 h), LA as well as the reference compounds LabyA1 (a HSV-2 entry inhibitor) and acyclovir (a DNA polymerase inhibitor) blocked virus infection. When given 2 h post-infection, solely acyclovir was able to inhibit viral replication, while LA and LabyA1 lost most of their antiviral activity (Fig 4B), the latter in agreement with a previous report [31]. These data clearly indicate that LA acts as a HSV entry inhibitor.

Anti-HSV-2 activity of LA in vivo
In the next set of experiments, we evaluated the anti-HSV-2 activity of LA in vivo. Female nunu mice were lumbosacrally scarificated and subsequently inoculated with HSV-2. Each group was treated topically twice daily for 5 days, starting 1 day prior to infection. At days 6 and 7 post-infection, swab samples were taken from the site of inoculation of animals that presented advanced lesions or small early lesion and viral titers of the samples were determined. The highest viral titers were observed in the placebo-arm. Treatment with a solution containing 2.5% LA (p = 0.0159) or 1% tenofovir (p = 0.0195), both dissolved in DMSO, showed significant lower viral titers compared to the placebo after 7 days (Fig 6). Comparable results were found on day 6 post-infection (data not shown). No viral titers were determined in the acyclovir-treated group, as no lesions were observed. Based on these results we can conclude, firstly, that acyclovir completely suppresses HSV-2 replication and secondly that LA and tenofovir have comparable antiviral effects in vivo (Fig 6).
The potent HIV-1 inhibitor AMD3100 lacked anti-HSV-2 activity (IC 50 >1200 nM), while HIV-1 NL4.3 was dose-dependently inhibited with an IC 50 of 18.0 ± 2.7 nM (Fig 7B). The MTS/PES method revealed no inhibitory activity of AMD3100 in the dual HIV-1/HSV-2 infection model. However when viral replication of HIV-1 and HSV-2 was measured by flow cytometry using the above-mentioned antibodies, the analysis clearly demonstrated that AMD3100 inhibited solely HIV-1 replication (IC 50 = 23.4 ± 5.4 nM), as the IC 50 for HSV-2 was >1200 nM. On the other hand, acyclovir potently inhibited HSV-2 G replication (IC 50 = 0.85 ± 0.15 μM) in T cells, while lacking anti-HIV-1 NL4.3 activity in single infection conditions. In the co-infection model, no anti-HIV/HSV activity could be detected for acyclovir (IC 50 >444 μM; Fig 7C) using the MTS/PES method. However, flow cytometric data indicated that acyclovir was still able to inhibit HSV-2 replication in the latter model with an IC 50 of 1.3 ± 0.9 μM, while having no anti-HIV activity (IC 50 >444 μM).

Lack of stimulatory effects of LA on HIV-1 target cells
To be a potential microbicidal candidate, LA may not activate HIV susceptible target cells. Therefore, LA was thoroughly investigated for its potential cell activating properties. Here, we used the mitogenic lectin PHA as positive control. Freshly isolated PBMCs were exposed to various concentrations of LA, tenofovir or PHA and after 72 h, the expression of the activation markers CD69 (early), CD25 (late) and HLA-DR (very late) on CD4 + T cells was determined (C) MDDCs were generated from buffy coat and exposed to HSV-2 G for 2h. After several washing steps, virus exposed MDDCs were cocultivated with uninfected CD4 + MT-4 cells for 5 days. Viral-induced cytopathicity was scored light microscopically and supernatant was retained for further analysis of viral content. Bars represent the percentage of viral replication after 5 days of co-cultivation, relative to the positive untreated control. Mean ± SEM of 3 independent experiments is shown. Dual Anti-HIV and Anti-HSV Activity of LA the amount of CD4 + CD69 + cells (33.0 ± 1.8%; p<0.0001) (Fig 8A, left). Compared to untreated conditions (15.7 ± 1.8%) LA had no impact on the number of CD4 + CD25 + expressing T cells (18.7 ± 2.2%; p = 0.3251), whereas PHA dramatically increased the amount of CD4 + CD25 + cells (39.8 ± 2.6%; p<0.0001) (Fig 8A, middle). In addition, LA did not affect the expression of the very late activation marker HLA-DR. The number of CD4 + HLA-DR + cells after LA treatment was almost identical to that of the negative control (5.4 ± 1.1% and 3.9 ± 0.6%, p = 0.2557; respectively). A significant 5-fold increase (p = 0.0005) in the amount of CD4 +-HLA-DR + T cells was observed with PHA (20.2 ± 3.2%; Fig 8A, right).
In the next set of experiments, we determined whether LA treatment of PBMCs for 24 h influenced their susceptibility to HIV-1 infection. Freshly isolated PBMCs from 3 different donors were pretreated with LA, tenofovir, PRO2000 or acyclovir. Once more the mitogenic lectin PHA was included as positive control. After 24 h, the cells were extensively washed and infected with the HIV-1 R5 strain BaL in the absence of compounds. Seven days post-infection, HIV-1 p24 Ag production was determined in the supernatant. The p24 Ag production of untreated cells (PC) was 1806 pg/ml (Fig 8B). Pretreatment with LA did not enhance the cellular susceptibility to HIV-1 considering its p24 Ag concentration of 2152 pg/ml, which was comparable with another polyanionic compound PRO2000 (1947 pg/ml) and the well-known antiviral agents tenofovir (1579 pg/ml) and acyclovir (1321 pg/ml) (Fig 8B). In contrast,

No growth inhibitory effects of LA on the vaginal microbiota
The vaginal (and also gut) microbiota is very important for the health of the host and may preferentially not be affected by potential microbicidal agents that would offer protection against sexually transmitted diseases such as HIV and HSV [46]. Therefore, we investigated in detail the effects of LA on the growth capacity of the normal vaginal microbiota, which is composed mainly out of Lactobacillus spp. All of the tested Lactobacillus strains showed the ability to grow in the presence of concentrations of LA as high as 31 μM (or 250 μg/ml) (Fig 9).

In general synergistic effects of LA in combination assays
An effective microbicide will presumably consist out of at least two different antiviral drugs. Therefore, we investigated if LA acts synergistically with acyclovir, PRO2000 or LabyA1 on Dual Anti-HIV and Anti-HSV Activity of LA HSV-2 G replication in C8166 cells. As shown in Fig 10A, the IC 50 s of LA decreased 10-fold in combination with acyclovir (IC 50 : 0.12 ± 0.02 μM (0.96 ± 0.16 μg/ml); p = 0.0089), 4-fold with PRO2000 (IC 50 : 0.46 ± 0.05 μM (3.7 ± 0.4 μg/ml); p = 0.0165) or 4-fold with LabyA1 (IC 50 : 0.29 ± 0.01 μM (2.3 ± 0.08 μg/ml); p = 0.0183). For each combination with LA, we observed a 1.3-to 2-fold dose reduction, which was only found to be significant when LA was combined with LabyA1 (p = 0.0027). CIs were determined to investigate the potential synergistic/additive/antagonistic effects, and as overviewed in Fig 10B, a synergistic interaction was observed with PRO2000 and acyclovir, with the highest levels of synergy in the LA/acyclovir combination. When LA was combined with LabyA1, an additive interaction was observed at the CI 50level, but increased synergistic effects were seen with increasing doses (CI 75 and CI 90 ).
CI analysis revealed antagonistic to additive effects in combination with PRO2000, while synergy was seen at each inhibitory level with tenofovir ( Fig 10D). LA/maraviroc combinations showed additivity at CI 50 -levels and increasing synergistic effects with increasing doses.

Inhibitory effects of LA independent on viral input levels
The amount of infectious virions at the time of transmission is not exactly known. Here, we investigated if LA is able to inhibit HIV-1 infection and replication at increased viral inputs (Table 1). These data clearly indicate that LA kept consistently its anti-HIV activity, with only a 1.8-fold decrease in its IC 50 , even at a 50-fold higher HIV-1 NL4.3 concentration. The other polyanionic compound PRO2000 showed a 4.5-fold decrease. Enfuvirtide (T20), used as reference compound was completely inactive at the highest concentrations tested. In addition, a 3.5-fold increase in IC 50 for LA was observed with increasing concentrations of HIV-2 ROD, while T20 lacked all anti-HIV-2 activity (data not shown).

LAresistant virus
We generated in vitro an NL4.3 LAresistant HIV-1 strain to gain more insight in the mechanism of action of LA. Wild-type NL4.3 was cultured in MT-4 cells in the presence of increasing concentrations of LA. As a control, wild-type NL4.3 virus was subcultivated in parallel in the absence of compound. The HIV-1 NL4.3 LAresistant virus acquired four mutations throughout gp120: S160N in the V2 loop, V170N in the C2 region, Q280H in the V3 loop and R389T in the C5 region. Two additional mixed mutations were found at R118K/R and A310A/V. In gp41, 3 mutations were observed: K77Q adjacent to HR1 domain, N113D, which belongs to glycosylation sequon "NXT" and H132Y inside HR2 domain. The HIV-1 NL4.3 LAresistant virus showed a 10.5-fold decrease in sensitivity to LA (Fig 11A).

Discussion
In this study, we investigated the antiviral activity and microbicidal properties of the novel low-cost sulfonated polyanionic compound, lignosulfonic acid (LA). LA demonstrated a potent and consistent broad-spectrum anti-HIV activity in the lower μM-range with low, if any, cytotoxicity in replication and cocultivation assays (Fig 1). These findings are in accordance with recently published data [18], where Qui and colleagues demonstrated the binding of a sulfonated lignin to the V3 loop of gp120 [18]. This loop plays a critical role in the entry process of HIV by interacting with the chemokine receptors CCR5 and/or CXCR4 [47]. This observation could explain its broad-spectrum anti-HIV activity against both R5 and X4 viruses. However, in the past, many concerns have been raised as X4 and R5 HIV-1 strains differ substantially in their sensitivities to polyanionic compounds, due to differences in positive charges [48]. As indicated for PRO2000 [49], LA inhibited equally well X4 and R5 viruses.
HIV-1 binding experiments demonstrated that LA, just like the included reference compounds FGM [24] and PRO2000 [50], significantly inhibited the initial interaction of free HIV-1 virions to the cell surface of target CD4 + T cells (Fig 2B). These results correspond to those obtained with other water-soluble lignins using rgp120 and sCD4 [18]. We hypothesize that this reflects inhibition by LA of the interaction between gp120 and cellular CD4. Nevertheless, we also need to consider the possibility that LA compromises virion binding to other T cell surface molecules such as heparan sulfate proteoglycans [51]. As a washing step after pre-exposure of the CD4 + T cells to compounds (i.e., LA, PRO2000, FGM) did not prevent virus binding, the interaction between LA and the cell surface appears to be rather weak (Fig 2C). These findings led us to believe that LA mainly binds to the HIV-1 envelope glycoproteins and not to the cell surface. Nevertheless, we (data not shown) and others observed a dose dependent binding inhibition of anti-CD4 mAb binding to the CD4 receptor in the presence of LA variants [18]. This could represent an additional mode of interference with gp120-CD4 interaction. However, this anti-CD4 mAb binding inhibition was only observed in the high LA concentration range (100-200 μg/ml) [18].
Epitope mapping experiments using the neutralizing mAbs 2G12, targeting the N-linked glycans on gp120; b12, recognizing the CD4 binding site on gp120; and 447-52D, binding to a part of the V3 loop on gp120; indicated that a certain sulfonated lignin significantly inhibited dose-dependently the antibody binding to the CD4 binding site and the V3 loop on gp120 [18]. These findings suggest that LA interacts with multiple sites on gp120. In addition, we generated an in vitro LA-resistant HIV-1 strain. The observed mutations throughout gp120 gave indeed another indication that LA interacts with different regions on gp120. We found that LA is able to interfere with the CD4/gp120 binding as the V170N mutation occurred in a sequence of the C2 region, which is part of the CD4-induced epitope [52]. The R389T substitution in the C5 region is situated adjacent to the V4 loop and was described previously as another important sequence involved in CD4 and mAb b12 binding [52,53]. The mutation Q280H on the tip of the V3 loop explains the interaction of LA with the V3 loop and this mutation was also observed with previously reported polyanionic compounds [54]. The mutations around the CD4 binding motifs and in the V3 loop could explain the observed cross-resistance with PRO2000 and FGM (Fig 11), two compounds described to inhibit CD4/gp120 binding [24,49,50]. The moderate cross-resistance with b12 mAb, which targets the CD4-binding site on gp120, could be explained by the observed mutation at one of the b12 mAb contact sites and a reported inhibition of b12 mAb by a sulfonated lignin for binding to gp120 by Qui et al. [18]. The lack of 2G12 mAb binding inhibition by this lignin on gp120 [18] also corresponds with our findings, as no high-type mannose N-linked glycans were deleted on gp120. On that account, the mannose-specific CBA HHA kept its full antiviral activity (Fig 11) in our crossresistance studies. Finally, the K77Q (adjacent to HR1) and H132Y (in domain HR2) mutations, both in gp41, reduced the activity of T20, the FDA-approved gp41 fusion inhibitor. Although the cross-resistance of a LA-resistant strain to the only FDA-approved fusion inhibitor might seem cumbersome, the use of T20 in daily practice is drastically reduced due to the fact that it needs to be injected twice daily and because of the continuous emergence of T20 resistant mutations [55][56][57].
Previous studies highlighted the clinical importance of genital HSV-2 infections as a co-factor for increased HIV transmission in women [2,6,58]. We showed for the first time that LA possessed a comparable antiviral activity (lower μM-range) against wild-type and acyclovirresistant HSV-1 and HSV-2 strains (Fig 4A). These observations are very important regarding the occurrence of acyclovir-resistant strains not only in immunocompromised patients, but also in immunocompetent individuals [11,59]. Another important observation is that LA could also inhibit HSV-2 infection in MDDCs, one of the key cells of the innate immune system involved for T cell responses. Additionally, MDDCs could also contribute to viral spread in the mucosa [60]. HSV-2 infection results in the expression and secretion of retinoic acid, which induces the expression of α4ß7 integrin on CD4 + T cells to make them more vulnerable for HIV-1 infection [44]. Recently, Stefanidou et al. demonstrated that HSV-2 infection prevents MDDCs maturation, induce apoptosis and triggers the release of pro-inflammatory cytokines, which could also increase the rate of HIV-1 infection [61]. Here, we also showed for the first time that LA had no effect on the DC-SIGN (or Dendritic cell-specific ICAM-3 grabbing non-integrin) receptor (Fig 3). DC-SIGN acts as a pathogen recognition receptor for viruses and bacteria [62]. Some pathogens, such as HIV and HSV, are captured by DC-SIGN and misuse its function [63] so that they are subsequently transmitted to uninfected target cells [40,45]. We could demonstrate that both viruses are dose-dependently inhibited by LA in DC-SIGN-related cellular transmission pathways (Figs 3 and 5).
As LA is a highly negatively charged compound, it cannot cross the cellular membrane and based on the HSV-2 time-of-drug addition experiments (Fig 4B) we believe that LA binds to one or more envelope glycoproteins of HSV. These in vitro broad-spectrum antiviral activity data indicate that LA acts as a dual HIV/HSV entry inhibitor (Figs 2 and 4) with equal potency against both viruses. Another indication that LA targets the envelope proteins is the lack of antiviral activity of LA against non-enveloped viruses such as the Picornaviridae Coxsackie type B4 and the Reoviridae Reovirus type 1 (>100 μg/ml or 12.5 μM; data not shown). These findings are in accordance with those described for other polyanions [64].
Acyclovir has been described to possess anti-HIV activity by acting as an reverse transcriptase inhibitor in herpes virus negative cells [13]. Other research groups found such inhibitory activity only when certain herpes viruses were already present and pre-established an infection in CD4 + T cells [12]. The lack of activity of acyclovir against HIV in our experimental co--infection model could be explained by the fact that we infected the CD4 + T cells with both viruses simultaneously. However, these inconsistent data of acyclovir on HIV inhibition, rendering it not an ideal molecule for microbicidal applications, unless used in combination with other anti-HIV agents. Therefore, we investigated LA in combination with other antiviral inhibitors. Additive to moderate synergistic effects were observed against HSV-2 infections with PRO2000, acyclovir and the recently described potent lantibiotic HIV/HSV peptide inhibitor LabyA1 [31]. Additivity to synergy was observed with maraviroc and tenofovir against HIV-1 replication, while antagonistic to additive effects were observed with PRO2000. This discrepancy between synergy and antagonism with the LA/PRO2000 combination, could be explained by the differences in envelope glycoproteins of HIV and HSV and the cross-resistance profiles of both polyanions to HIV (Fig 11). To date, tenofovir, applied as a 1% vaginal gel, still is the only drug that significantly inhibited HIV-1 transmission by 39% overall and of HSV-2 by 51% and this even at suboptimal concentrations [8]. The potent inhibitory activity against HIV and HSV-2, together with the observed synergy (and additivity) with tenofovir and other anti-HIV and anti-HSV drugs, makes us believe that LA is a strong microbicide candidate. Especially, as Agarwal and colleagues demonstrated also the feasibility of conjugating chemically nucleoside reverse transcriptase inhibitors with cellulose sulfate, displaying antiviral and contraceptive properties [65].
In addition, cellulose sulfate, which failed throughout clinical microbicide trials, and polystyrene sulfonate combine antiviral activity against HIV-1 and HSV-2 with contraceptive activity [66]. Tollner et al., [67,68] demonstrated that LA also exhibits contraceptive properties as in vitro fertilization of macaque oocytes was blocked when sperm was treated before or after capacitation. Importantly, LA (Fig 9), as well as polystyrene sulfonate, did not harm the vaginal microbiota, but the anti-HIV activity of polystyrene sulfonate was not so potent, indicating superior microbicidal properties for LA [69]. As a vaginal or rectal microbicide may not be toxic nor stimulate HIV target cells, we showed that LA did not activate PBMCs, nor increased their susceptibility for HIV-1 infection as our control agents tenofovir, PRO2000 or acyclovir (Fig 7). In contrast, application and removal of cellulose sulfate and carrageenan, resulted in an enhanced subsequent HIV-1 infection in vitro [70]. Further studies pointed out that cellulose sulfate altered the epithelial architecture by affecting junctional proteins (e.g. zona occludens 1 (ZO-A1)) [71,72]. Interestingly, Qui et al., mentioned that a water-soluble sulfonated lignin had no effects on the epithelial integrity on different epithelial monolayers [18]. Previously described in vitro and in vivo data demonstrated a good safety profile of PRO2000 [73,74]. Our in vivo experiments in mice indicated that LA did not show superiority over acyclovir in suppressing HSV-2 infection, however LA demonstrated a significant antiviral effect comparable with tenofovir [9]. We could state, for the first time, that LA demonstrated an anti-HSV-2 activity in vivo, evaluating its effect on lesion viral titers (Fig 6). However, the more moderate anti-herpetic activity in vivo, compared to its potency in vitro could be attributed to the administration route or its formulation. Previous studies in mice, using vaginally applied PRO2000 in solution and a gel-formulation, demonstrated protective effects to genital HSV-2 infection [75]. Further optimizations of in vivo studies are planned in the near future using a gel-containing LA preparation.
In conclusion, these in vitro and in vivo results of inhibition of HIV and HSV transmission and infection by LA, together with its high scale availability and safety profile, make it a very promising microbicidal candidate. combined with or without compounds for 2 hours at RT. Thereafter, cells were extensively washed and gp120 binding was evaluated in all the virus treated conditions with the antihuman 2G12 mAb + RaH-IgG-FITC. The bars represent the percentages of anti-gp120 binding relative to the positive control (d). Each value represents the mean ± SEM of 3 independent experiments. Ã p<0.05, ÃÃ p<0.01, ÃÃÃ p<0.005, ÃÃÃ p<0.001 compared to the nontreated control, according to the one-way Anova and Dunnett's multi comparison post-hoc test. (DOCX) S1